olive5 commited on
Commit
ee6e8a0
·
verified ·
1 Parent(s): 529661c

Upload folder using huggingface_hub

Browse files
Files changed (36) hide show
  1. 3B1B_LinearAlgebra/01 - Vectors | Chapter 1, Essence of linear algebra.en.vtt +475 -0
  2. 3B1B_LinearAlgebra/01 - Vectors | Chapter 1, Essence of linear algebra.mp4 +3 -0
  3. 3B1B_LinearAlgebra/02 - Linear combinations, span, and basis vectors | Chapter 2, Essence of linear algebra.en.vtt +439 -0
  4. 3B1B_LinearAlgebra/02 - Linear combinations, span, and basis vectors | Chapter 2, Essence of linear algebra.mp4 +3 -0
  5. 3B1B_LinearAlgebra/03 - Linear transformations and matrices | Chapter 3, Essence of linear algebra.en.vtt +484 -0
  6. 3B1B_LinearAlgebra/03 - Linear transformations and matrices | Chapter 3, Essence of linear algebra.mp4 +3 -0
  7. 3B1B_LinearAlgebra/04 - Matrix multiplication as composition | Chapter 4, Essence of linear algebra.en.vtt +454 -0
  8. 3B1B_LinearAlgebra/04 - Matrix multiplication as composition | Chapter 4, Essence of linear algebra.mp4 +3 -0
  9. 3B1B_LinearAlgebra/05 - Three-dimensional linear transformations | Chapter 5, Essence of linear algebra.en.vtt +190 -0
  10. 3B1B_LinearAlgebra/05 - Three-dimensional linear transformations | Chapter 5, Essence of linear algebra.mp4 +3 -0
  11. 3B1B_LinearAlgebra/06 - The determinant | Chapter 6, Essence of linear algebra.en.vtt +454 -0
  12. 3B1B_LinearAlgebra/06 - The determinant | Chapter 6, Essence of linear algebra.mp4 +3 -0
  13. 3B1B_LinearAlgebra/07 - Inverse matrices, column space and null space | Chapter 7, Essence of linear algebra.en.vtt +532 -0
  14. 3B1B_LinearAlgebra/07 - Inverse matrices, column space and null space | Chapter 7, Essence of linear algebra.mp4 +3 -0
  15. 3B1B_LinearAlgebra/08 - Nonsquare matrices as transformations between dimensions | Chapter 8, Essence of linear algebra.en-orig.vtt +944 -0
  16. 3B1B_LinearAlgebra/08 - Nonsquare matrices as transformations between dimensions | Chapter 8, Essence of linear algebra.en.vtt +944 -0
  17. 3B1B_LinearAlgebra/08 - Nonsquare matrices as transformations between dimensions | Chapter 8, Essence of linear algebra.mp4 +3 -0
  18. 3B1B_LinearAlgebra/09 - Dot products and duality | Chapter 9, Essence of linear algebra.en-orig.vtt +0 -0
  19. 3B1B_LinearAlgebra/09 - Dot products and duality | Chapter 9, Essence of linear algebra.en.vtt +0 -0
  20. 3B1B_LinearAlgebra/09 - Dot products and duality | Chapter 9, Essence of linear algebra.mp4 +3 -0
  21. 3B1B_LinearAlgebra/10 - Cross products | Chapter 10, Essence of linear algebra.en-orig.vtt +1896 -0
  22. 3B1B_LinearAlgebra/10 - Cross products | Chapter 10, Essence of linear algebra.en.vtt +1896 -0
  23. 3B1B_LinearAlgebra/10 - Cross products | Chapter 10, Essence of linear algebra.mp4 +3 -0
  24. 3B1B_LinearAlgebra/11 - Cross products in the light of linear transformations | Chapter 11, Essence of linear algebra.en.vtt +532 -0
  25. 3B1B_LinearAlgebra/11 - Cross products in the light of linear transformations | Chapter 11, Essence of linear algebra.mp4 +3 -0
  26. 3B1B_LinearAlgebra/12 - Cramer's rule, explained geometrically | Chapter 12, Essence of linear algebra.en.vtt +562 -0
  27. 3B1B_LinearAlgebra/12 - Cramer's rule, explained geometrically | Chapter 12, Essence of linear algebra.mp4 +3 -0
  28. 3B1B_LinearAlgebra/13 - Change of basis | Chapter 13, Essence of linear algebra.en.vtt +520 -0
  29. 3B1B_LinearAlgebra/13 - Change of basis | Chapter 13, Essence of linear algebra.mp4 +3 -0
  30. 3B1B_LinearAlgebra/14 - Eigenvectors and eigenvalues | Chapter 14, Essence of linear algebra.en.vtt +694 -0
  31. 3B1B_LinearAlgebra/14 - Eigenvectors and eigenvalues | Chapter 14, Essence of linear algebra.mp4 +3 -0
  32. 3B1B_LinearAlgebra/15 - A quick trick for computing eigenvalues | Chapter 15, Essence of linear algebra.en.vtt +619 -0
  33. 3B1B_LinearAlgebra/15 - A quick trick for computing eigenvalues | Chapter 15, Essence of linear algebra.mp4 +3 -0
  34. 3B1B_LinearAlgebra/16 - Abstract vector spaces | Chapter 16, Essence of linear algebra.en.vtt +670 -0
  35. 3B1B_LinearAlgebra/16 - Abstract vector spaces | Chapter 16, Essence of linear algebra.mp4 +3 -0
  36. 3B1B_LinearAlgebra/ids.tsv +16 -0
3B1B_LinearAlgebra/01 - Vectors | Chapter 1, Essence of linear algebra.en.vtt ADDED
@@ -0,0 +1,475 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:00.000 --> 00:00:10.920
6
+ [Translated by Grant Sanderson. Submit corrections at criblate.com]
7
+
8
+ 00:00:10.920 --> 00:00:15.220
9
+ The fundamental, root-of-it-all building block for linear algebra is the vector.
10
+
11
+ 00:00:15.720 --> 00:00:19.840
12
+ So it's worth making sure that we're all on the same page about what exactly a vector is.
13
+
14
+ 00:00:20.380 --> 00:00:24.231
15
+ You see, broadly speaking, there are three distinct but related ideas about vectors,
16
+
17
+ 00:00:24.231 --> 00:00:26.478
18
+ which I'll call the physics student perspective,
19
+
20
+ 00:00:26.478 --> 00:00:30.100
21
+ the computer science student perspective, and the mathematician's perspective.
22
+
23
+ 00:00:30.880 --> 00:00:34.400
24
+ The physics student perspective is that vectors are arrows pointing in space.
25
+
26
+ 00:00:34.940 --> 00:00:38.338
27
+ What defines a given vector is its length and the direction it's pointing,
28
+
29
+ 00:00:38.338 --> 00:00:41.691
30
+ but as long as those two facts are the same, you can move it all around,
31
+
32
+ 00:00:41.691 --> 00:00:43.160
33
+ and it's still the same vector.
34
+
35
+ 00:00:44.040 --> 00:00:46.493
36
+ Vectors that live in the flat plane are two-dimensional,
37
+
38
+ 00:00:46.493 --> 00:00:50.040
39
+ and those sitting in broader space that you and I live in are three-dimensional.
40
+
41
+ 00:00:51.720 --> 00:00:55.640
42
+ The computer science perspective is that vectors are ordered lists of numbers.
43
+
44
+ 00:00:55.640 --> 00:00:59.276
45
+ For example, let's say you were doing some analytics about house prices,
46
+
47
+ 00:00:59.276 --> 00:01:02.760
48
+ and the only features you cared about were square footage and price.
49
+
50
+ 00:01:03.020 --> 00:01:05.398
51
+ You might model each house with a pair of numbers,
52
+
53
+ 00:01:05.398 --> 00:01:08.680
54
+ the first indicating square footage and the second indicating price.
55
+
56
+ 00:01:09.320 --> 00:01:11.040
57
+ Notice the order matters here.
58
+
59
+ 00:01:12.400 --> 00:01:15.861
60
+ In the lingo, you'd be modeling houses as two-dimensional vectors,
61
+
62
+ 00:01:15.861 --> 00:01:19.688
63
+ where in this context, vector is pretty much just a fancy word for list,
64
+
65
+ 00:01:19.688 --> 00:01:24.040
66
+ and what makes it two-dimensional is the fact that the length of that list is two.
67
+
68
+ 00:01:25.640 --> 00:01:29.301
69
+ The mathematician, on the other hand, seeks to generalize both these views,
70
+
71
+ 00:01:29.301 --> 00:01:33.646
72
+ basically saying that a vector can be anything where there's a sensible notion of adding
73
+
74
+ 00:01:33.646 --> 00:01:36.086
75
+ two vectors and multiplying a vector by a number,
76
+
77
+ 00:01:36.086 --> 00:01:38.820
78
+ operations that I'll talk about later on in this video.
79
+
80
+ 00:01:39.580 --> 00:01:43.760
81
+ The details of this view are rather abstract, and I actually think it's healthy to ignore
82
+
83
+ 00:01:43.760 --> 00:01:47.940
84
+ it until the last video of this series, favoring a more concrete setting in the interim.
85
+
86
+ 00:01:48.400 --> 00:01:51.421
87
+ But the reason I bring it up here is that it hints at the fact
88
+
89
+ 00:01:51.421 --> 00:01:54.150
90
+ that the ideas of vector addition and multiplication by
91
+
92
+ 00:01:54.150 --> 00:01:57.220
93
+ numbers will play an important role throughout linear algebra.
94
+
95
+ 00:01:58.000 --> 00:02:01.067
96
+ But before I talk about those operations, let's just settle in on
97
+
98
+ 00:02:01.067 --> 00:02:04.040
99
+ a specific thought to have in mind when I say the word vector.
100
+
101
+ 00:02:04.740 --> 00:02:07.499
102
+ Given the geometric focus that I'm shooting for here,
103
+
104
+ 00:02:07.499 --> 00:02:10.206
105
+ whenever I introduce a new topic involving vectors,
106
+
107
+ 00:02:10.206 --> 00:02:13.330
108
+ I want you to first think about an arrow, and specifically,
109
+
110
+ 00:02:13.330 --> 00:02:16.974
111
+ think about that arrow inside a coordinate system, like the xy-plane,
112
+
113
+ 00:02:16.974 --> 00:02:18.900
114
+ with its tail sitting at the origin.
115
+
116
+ 00:02:19.680 --> 00:02:22.508
117
+ This is a little bit different from the physics student perspective,
118
+
119
+ 00:02:22.508 --> 00:02:24.920
120
+ where vectors can freely sit anywhere they want in space.
121
+
122
+ 00:02:25.420 --> 00:02:27.844
123
+ In linear algebra, it's almost always the case
124
+
125
+ 00:02:27.844 --> 00:02:30.320
126
+ that your vector will be rooted at the origin.
127
+
128
+ 00:02:30.940 --> 00:02:34.558
129
+ Then, once you understand a new concept in the context of arrows in space,
130
+
131
+ 00:02:34.558 --> 00:02:37.589
132
+ we'll translate it over to the list of numbers point of view,
133
+
134
+ 00:02:37.589 --> 00:02:40.620
135
+ which we can do by considering the coordinates of the vector.
136
+
137
+ 00:02:41.440 --> 00:02:45.156
138
+ Now, while I'm sure that many of you are already familiar with this coordinate system,
139
+
140
+ 00:02:45.156 --> 00:02:48.612
141
+ it's worth walking through explicitly, since this is where all of the important
142
+
143
+ 00:02:48.612 --> 00:02:51.680
144
+ back and forth happens between the two perspectives of linear algebra.
145
+
146
+ 00:02:52.740 --> 00:02:56.758
147
+ Focusing our attention on two dimensions for the moment, you have a horizontal line,
148
+
149
+ 00:02:56.758 --> 00:02:59.580
150
+ called the x-axis, and a vertical line, called the y-axis.
151
+
152
+ 00:03:00.260 --> 00:03:02.364
153
+ The place where they intersect is called the origin,
154
+
155
+ 00:03:02.364 --> 00:03:05.520
156
+ which you should think of as the center of space and the root of all vectors.
157
+
158
+ 00:03:06.380 --> 00:03:08.672
159
+ After choosing an arbitrary length to represent one,
160
+
161
+ 00:03:08.672 --> 00:03:11.360
162
+ you make tick marks on each axis to represent this distance.
163
+
164
+ 00:03:12.320 --> 00:03:14.612
165
+ When I want to convey the idea of 2D space as a whole,
166
+
167
+ 00:03:14.612 --> 00:03:16.691
168
+ which you'll see comes up a lot in these videos,
169
+
170
+ 00:03:16.691 --> 00:03:19.408
171
+ I'll extend these tick marks to make grid lines, but right now,
172
+
173
+ 00:03:19.408 --> 00:03:21.360
174
+ they'll actually get a little bit in the way.
175
+
176
+ 00:03:22.000 --> 00:03:26.402
177
+ The coordinates of a vector is a pair of numbers that basically gives instructions
178
+
179
+ 00:03:26.402 --> 00:03:30.160
180
+ for how to get from the tail of that vector at the origin to its tip.
181
+
182
+ 00:03:30.880 --> 00:03:33.954
183
+ The first number tells you how far to walk along the x-axis,
184
+
185
+ 00:03:33.954 --> 00:03:36.310
186
+ positive numbers indicating rightward motion,
187
+
188
+ 00:03:36.310 --> 00:03:40.459
189
+ negative numbers indicating leftward motion, and the second number tells you how
190
+
191
+ 00:03:40.459 --> 00:03:42.867
192
+ far to walk parallel to the y-axis after that,
193
+
194
+ 00:03:42.867 --> 00:03:47.170
195
+ positive numbers indicating upward motion, and negative numbers indicating downward
196
+
197
+ 00:03:47.170 --> 00:03:47.580
198
+ motion.
199
+
200
+ 00:03:48.140 --> 00:03:51.143
201
+ To distinguish vectors from points, the convention is to write
202
+
203
+ 00:03:51.143 --> 00:03:54.340
204
+ this pair of numbers vertically with square brackets around them.
205
+
206
+ 00:03:56.340 --> 00:03:59.602
207
+ Every pair of numbers gives you one and only one vector,
208
+
209
+ 00:03:59.602 --> 00:04:03.680
210
+ and every vector is associated with one and only one pair of numbers.
211
+
212
+ 00:04:04.640 --> 00:04:05.500
213
+ What about in three dimensions?
214
+
215
+ 00:04:06.200 --> 00:04:08.912
216
+ Well, you add a third axis, called the z-axis,
217
+
218
+ 00:04:08.912 --> 00:04:12.862
219
+ which is perpendicular to both the x and y-axes, and in this case,
220
+
221
+ 00:04:12.862 --> 00:04:16.340
222
+ each vector is associated with ordered triplet of numbers.
223
+
224
+ 00:04:16.860 --> 00:04:19.847
225
+ The first tells you how far to move along the x-axis,
226
+
227
+ 00:04:19.847 --> 00:04:23.284
228
+ the second tells you how far to move parallel to the y-axis,
229
+
230
+ 00:04:23.284 --> 00:04:27.680
231
+ and the third one tells you how far to then move parallel to this new z-axis.
232
+
233
+ 00:04:28.400 --> 00:04:31.815
234
+ Every triplet of numbers gives you one unique vector in space,
235
+
236
+ 00:04:31.815 --> 00:04:35.560
237
+ and every vector in space gives you exactly one triplet of numbers.
238
+
239
+ 00:04:36.900 --> 00:04:40.100
240
+ All right, so back to vector addition and multiplication by numbers.
241
+
242
+ 00:04:40.460 --> 00:04:44.780
243
+ After all, every topic in linear algebra is going to center around these two operations.
244
+
245
+ 00:04:45.440 --> 00:04:47.640
246
+ Luckily, each one's pretty straightforward to define.
247
+
248
+ 00:04:48.480 --> 00:04:51.520
249
+ Let's say we have two vectors, one pointing up and a little to the right,
250
+
251
+ 00:04:51.520 --> 00:04:53.560
252
+ and the other one pointing right and down a bit.
253
+
254
+ 00:04:53.960 --> 00:04:56.820
255
+ To add these two vectors, move the second one so
256
+
257
+ 00:04:56.820 --> 00:04:59.680
258
+ that its tail sits at the tip of the first one.
259
+
260
+ 00:05:00.300 --> 00:05:04.392
261
+ Then, if you draw a new vector from the tail of the first one to
262
+
263
+ 00:05:04.392 --> 00:05:08.740
264
+ where the tip of the second one sits, that new vector is their sum.
265
+
266
+ 00:05:12.080 --> 00:05:15.520
267
+ This definition of addition, by the way, is pretty much the only time
268
+
269
+ 00:05:15.520 --> 00:05:18.860
270
+ in linear algebra where we let vectors stray away from the origin.
271
+
272
+ 00:05:19.720 --> 00:05:21.480
273
+ Now, why is this a reasonable thing to do?
274
+
275
+ 00:05:21.740 --> 00:05:24.020
276
+ Why this definition of addition and not some other one?
277
+
278
+ 00:05:25.520 --> 00:05:29.945
279
+ Well, the way I like to think about it is that each vector represents a certain movement,
280
+
281
+ 00:05:29.945 --> 00:05:32.680
282
+ a step with a certain distance and direction in space.
283
+
284
+ 00:05:33.980 --> 00:05:37.661
285
+ If you take a step along the first vector, then take a step in the direction
286
+
287
+ 00:05:37.661 --> 00:05:41.148
288
+ and distance described by the second vector, the overall effect is just
289
+
290
+ 00:05:41.148 --> 00:05:44.780
291
+ the same as if you moved along the sum of those two vectors to start with.
292
+
293
+ 00:05:45.260 --> 00:05:47.382
294
+ You could think about this as an extension of how
295
+
296
+ 00:05:47.382 --> 00:05:49.460
297
+ we think about adding numbers on a number line.
298
+
299
+ 00:05:50.180 --> 00:05:53.472
300
+ One way that we teach kids to think about this, say with 2 plus 5,
301
+
302
+ 00:05:53.472 --> 00:05:57.960
303
+ is to think of moving two steps to the right followed by another five steps to the right.
304
+
305
+ 00:05:57.960 --> 00:06:01.720
306
+ The overall effect is the same as if you just took seven steps to the right.
307
+
308
+ 00:06:02.660 --> 00:06:05.480
309
+ In fact, let's see how vector addition looks numerically.
310
+
311
+ 00:06:06.020 --> 00:06:11.629
312
+ The first vector here has coordinates 1, 2, and the second one has coordinates 3,
313
+
314
+ 00:06:11.629 --> 00:06:12.460
315
+ negative 1.
316
+
317
+ 00:06:14.360 --> 00:06:17.293
318
+ When you take the vector sum using this tip-to-tail method,
319
+
320
+ 00:06:17.293 --> 00:06:21.420
321
+ you can think of a four-step path from the origin to the tip of the second vector.
322
+
323
+ 00:06:21.840 --> 00:06:25.620
324
+ Walk 1 to the right, then 2 up, then 3 to the right, then 1 down.
325
+
326
+ 00:06:26.920 --> 00:06:31.283
327
+ Reorganizing these steps so that you first do all of the rightward motion,
328
+
329
+ 00:06:31.283 --> 00:06:35.114
330
+ then do all the vertical motion, you can read it as saying first
331
+
332
+ 00:06:35.114 --> 00:06:38.180
333
+ move 1 plus 3 to the right, then move 2 minus 1 up.
334
+
335
+ 00:06:40.080 --> 00:06:44.920
336
+ So the new vector has coordinates 1 plus 3 and 2 plus negative 1.
337
+
338
+ 00:06:45.600 --> 00:06:49.066
339
+ In general, vector addition in this list of numbers conception
340
+
341
+ 00:06:49.066 --> 00:06:52.700
342
+ looks like matching up their terms and adding each one together.
343
+
344
+ 00:06:54.640 --> 00:06:58.360
345
+ The other fundamental vector operation is multiplication by a number.
346
+
347
+ 00:06:58.860 --> 00:07:01.380
348
+ Now this is best understood just by looking at a few examples.
349
+
350
+ 00:07:01.840 --> 00:07:04.941
351
+ If you take the number 2 and multiply it by a given vector,
352
+
353
+ 00:07:04.941 --> 00:07:09.620
354
+ it means you stretch out that vector so that it's two times as long as when you started.
355
+
356
+ 00:07:10.500 --> 00:07:13.012
357
+ If you multiply that vector by, say, one-third,
358
+
359
+ 00:07:13.012 --> 00:07:16.860
360
+ it means you squish it down so that it's one-third the original length.
361
+
362
+ 00:07:17.640 --> 00:07:21.258
363
+ When you multiply it by a negative number, like negative 1.8,
364
+
365
+ 00:07:21.258 --> 00:07:26.300
366
+ then the vector first gets flipped around, then stretched out by that factor of 1.8.
367
+
368
+ 00:07:27.360 --> 00:07:31.934
369
+ This process of stretching or squishing or sometimes reversing the direction of
370
+
371
+ 00:07:31.934 --> 00:07:36.682
372
+ a vector is called scaling, and whenever you catch a number like two or one-third
373
+
374
+ 00:07:36.682 --> 00:07:41.140
375
+ or negative 1.8 acting like this, scaling some vector, you call it a scalar.
376
+
377
+ 00:07:41.940 --> 00:07:46.201
378
+ In fact, throughout linear algebra, one of the main things that numbers do is scale
379
+
380
+ 00:07:46.201 --> 00:07:50.769
381
+ vectors, so it's common to use the word scalar pretty much interchangeably with the word
382
+
383
+ 00:07:50.769 --> 00:07:51.180
384
+ number.
385
+
386
+ 00:07:52.020 --> 00:07:55.483
387
+ Numerically, stretching out a vector by a factor of, say, 2,
388
+
389
+ 00:07:55.483 --> 00:07:59.580
390
+ corresponds with multiplying each of its components by that factor, 2.
391
+
392
+ 00:08:00.300 --> 00:08:03.062
393
+ So in the conception of vectors as lists of numbers,
394
+
395
+ 00:08:03.062 --> 00:08:07.046
396
+ multiplying a given vector by a scalar means multiplying each one of those
397
+
398
+ 00:08:07.046 --> 00:08:08.480
399
+ components by that scalar.
400
+
401
+ 00:08:10.220 --> 00:08:14.594
402
+ You'll see in the following videos what I mean when I say linear algebra topics tend to
403
+
404
+ 00:08:14.594 --> 00:08:17.058
405
+ revolve around these two fundamental operations,
406
+
407
+ 00:08:17.058 --> 00:08:19.220
408
+ vector addition and scalar multiplication.
409
+
410
+ 00:08:19.980 --> 00:08:22.830
411
+ And I'll talk more in the last video about how and why the
412
+
413
+ 00:08:22.830 --> 00:08:25.287
414
+ mathematician thinks only about these operations,
415
+
416
+ 00:08:25.287 --> 00:08:29.120
417
+ independent and abstracted away from however you choose to represent vectors.
418
+
419
+ 00:08:29.800 --> 00:08:33.821
420
+ In truth, it doesn't matter whether you think about vectors as fundamentally being arrows
421
+
422
+ 00:08:33.821 --> 00:08:37.210
423
+ in space, like I'm suggesting you do, that happen to have a nice numerical
424
+
425
+ 00:08:37.210 --> 00:08:40.825
426
+ representation, or fundamentally as lists of numbers that happen to have a nice
427
+
428
+ 00:08:40.825 --> 00:08:42.000
429
+ geometric interpretation.
430
+
431
+ 00:08:42.520 --> 00:08:45.976
432
+ The usefulness of linear algebra has less to do with either one of these
433
+
434
+ 00:08:45.976 --> 00:08:49.720
435
+ views than it does with the ability to translate back and forth between them.
436
+
437
+ 00:08:50.140 --> 00:08:53.538
438
+ It gives the data analyst a nice way to conceptualize many lists
439
+
440
+ 00:08:53.538 --> 00:08:56.988
441
+ of numbers in a visual way, which can seriously clarify patterns
442
+
443
+ 00:08:56.988 --> 00:09:00.280
444
+ in data and give a global view of what certain operations do.
445
+
446
+ 00:09:00.820 --> 00:09:04.155
447
+ And on the flip side, it gives people like physicists and computer
448
+
449
+ 00:09:04.155 --> 00:09:07.742
450
+ graphics programmers a language to describe space and the manipulation
451
+
452
+ 00:09:07.742 --> 00:09:11.380
453
+ of space using numbers that can be crunched and run through a computer.
454
+
455
+ 00:09:12.300 --> 00:09:15.871
456
+ When I do math-y animations, for example, I start by thinking about what's
457
+
458
+ 00:09:15.871 --> 00:09:20.068
459
+ actually going on in space, and then get the computer to represent things numerically,
460
+
461
+ 00:09:20.068 --> 00:09:23.060
462
+ thereby figuring out where to place the pixels on the screen.
463
+
464
+ 00:09:23.480 --> 00:09:26.580
465
+ And doing that usually relies on a lot of linear algebra understanding.
466
+
467
+ 00:09:27.840 --> 00:09:38.959
468
+ So there are your vector basics, and in the next video I'll start getting into some
469
+
470
+ 00:09:38.959 --> 00:09:49.944
471
+ pretty neat concepts surrounding vectors like span, bases, and linear dependence.
472
+
473
+ 00:09:49.944 --> 00:09:51.820
474
+ See you then!
475
+
3B1B_LinearAlgebra/01 - Vectors | Chapter 1, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4ac7bc175415303e03fa71a9283c4a60737cf169153d3f2d73f1de3614078754
3
+ size 18772234
3B1B_LinearAlgebra/02 - Linear combinations, span, and basis vectors | Chapter 2, Essence of linear algebra.en.vtt ADDED
@@ -0,0 +1,439 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:11.880 --> 00:00:16.191
6
+ In the last video, along with the ideas of vector addition and scalar multiplication,
7
+
8
+ 00:00:16.191 --> 00:00:19.952
9
+ I described vector coordinates, where there's this back and forth between,
10
+
11
+ 00:00:19.952 --> 00:00:22.860
12
+ for example, pairs of numbers and two-dimensional vectors.
13
+
14
+ 00:00:23.800 --> 00:00:27.036
15
+ Now, I imagine the vector coordinates were already familiar to a lot of you,
16
+
17
+ 00:00:27.036 --> 00:00:30.314
18
+ but there's another kind of interesting way to think about these coordinates,
19
+
20
+ 00:00:30.314 --> 00:00:32.080
21
+ which is pretty central to linear algebra.
22
+
23
+ 00:00:32.840 --> 00:00:36.547
24
+ When you have a pair of numbers that's meant to describe a vector,
25
+
26
+ 00:00:36.547 --> 00:00:40.698
27
+ like 3, negative 2, I want you to think about each coordinate as a scalar,
28
+
29
+ 00:00:40.698 --> 00:00:44.240
30
+ meaning, think about how each one stretches or squishes vectors.
31
+
32
+ 00:00:45.140 --> 00:00:48.841
33
+ In the xy coordinate system, there are two very special vectors,
34
+
35
+ 00:00:48.841 --> 00:00:52.712
36
+ the one pointing to the right with length 1, commonly called i-hat,
37
+
38
+ 00:00:52.712 --> 00:00:57.666
39
+ or the unit vector in the x direction, and the one pointing straight up with length 1,
40
+
41
+ 00:00:57.666 --> 00:01:01.140
42
+ commonly called j-hat, or the unit vector in the y direction.
43
+
44
+ 00:01:02.440 --> 00:01:06.772
45
+ Now, think of the x coordinate of our vector as a scalar that scales i-hat,
46
+
47
+ 00:01:06.772 --> 00:01:11.560
48
+ stretching it by a factor of 3, and the y coordinate as a scalar that scales j-hat,
49
+
50
+ 00:01:11.560 --> 00:01:14.240
51
+ flipping it and stretching it by a factor of 2.
52
+
53
+ 00:01:14.880 --> 00:01:17.820
54
+ In this sense, the vector that these coordinates
55
+
56
+ 00:01:17.820 --> 00:01:20.340
57
+ describe is the sum of two scaled vectors.
58
+
59
+ 00:01:20.340 --> 00:01:25.560
60
+ That's a surprisingly important concept, this idea of adding together two scaled vectors.
61
+
62
+ 00:01:27.320 --> 00:01:30.440
63
+ Those two vectors, i-hat and j-hat, have a special name, by the way.
64
+
65
+ 00:01:30.900 --> 00:01:33.560
66
+ Together, they're called the basis of a coordinate system.
67
+
68
+ 00:01:34.240 --> 00:01:38.300
69
+ What this means, basically, is that when you think about coordinates as scalars,
70
+
71
+ 00:01:38.300 --> 00:01:41.660
72
+ the basis vectors are what those scalars actually, you know, scale.
73
+
74
+ 00:01:42.320 --> 00:01:45.660
75
+ There's also a more technical definition, but I'll get to that later.
76
+
77
+ 00:01:47.180 --> 00:01:51.352
78
+ By framing our coordinate system in terms of these two special basis vectors,
79
+
80
+ 00:01:51.352 --> 00:01:53.920
81
+ it raises a pretty interesting and subtle point.
82
+
83
+ 00:01:54.460 --> 00:01:57.371
84
+ We could have chosen different basis vectors and
85
+
86
+ 00:01:57.371 --> 00:02:00.520
87
+ gotten a completely reasonable new coordinate system.
88
+
89
+ 00:02:01.100 --> 00:02:03.763
90
+ For example, take some vector pointing up and to the right,
91
+
92
+ 00:02:03.763 --> 00:02:06.960
93
+ along with some other vector pointing down and to the right in some way.
94
+
95
+ 00:02:07.620 --> 00:02:12.437
96
+ Take a moment to think about all the different vectors that you can get by choosing two
97
+
98
+ 00:02:12.437 --> 00:02:17.200
99
+ scalars, using each one to scale one of the vectors, then adding together what you get.
100
+
101
+ 00:02:17.920 --> 00:02:21.500
102
+ Which two-dimensional vectors can you reach by altering the choices of scalars?
103
+
104
+ 00:02:24.580 --> 00:02:28.168
105
+ The answer is that you can reach every possible two-dimensional vector,
106
+
107
+ 00:02:28.168 --> 00:02:30.660
108
+ and I think it's a good puzzle to contemplate why.
109
+
110
+ 00:02:32.320 --> 00:02:36.822
111
+ A new pair of basis vectors like this still gives us a valid way to go back and forth
112
+
113
+ 00:02:36.822 --> 00:02:39.649
114
+ between pairs of numbers and two-dimensional vectors,
115
+
116
+ 00:02:39.649 --> 00:02:44.099
117
+ but the association is definitely different from the one that you get using the more
118
+
119
+ 00:02:44.099 --> 00:02:45.880
120
+ standard basis of i-hat and j-hat.
121
+
122
+ 00:02:46.460 --> 00:02:49.325
123
+ This is something I'll go into much more detail on later,
124
+
125
+ 00:02:49.325 --> 00:02:52.882
126
+ describing the exact relationship between different coordinate systems,
127
+
128
+ 00:02:52.882 --> 00:02:56.587
129
+ but for right now, I just want you to appreciate the fact that any time we
130
+
131
+ 00:02:56.587 --> 00:03:00.391
132
+ describe vectors numerically, it depends on an implicit choice of what basis
133
+
134
+ 00:03:00.391 --> 00:03:01.380
135
+ vectors we're using.
136
+
137
+ 00:03:02.360 --> 00:03:05.972
138
+ So any time that you're scaling two vectors and adding them like this,
139
+
140
+ 00:03:05.972 --> 00:03:08.720
141
+ it's called a linear combination of those two vectors.
142
+
143
+ 00:03:11.120 --> 00:03:12.660
144
+ Where does this word linear come from?
145
+
146
+ 00:03:12.840 --> 00:03:14.400
147
+ Why does this have anything to do with lines?
148
+
149
+ 00:03:14.940 --> 00:03:18.449
150
+ Well, this isn't the etymology, but one way I like to think about it
151
+
152
+ 00:03:18.449 --> 00:03:22.873
153
+ is that if you fix one of those scalars and let the other one change its value freely,
154
+
155
+ 00:03:22.873 --> 00:03:25.620
156
+ the tip of the resulting vector draws a straight line.
157
+
158
+ 00:03:29.160 --> 00:03:32.511
159
+ Now, if you let both scalars range freely and consider every possible
160
+
161
+ 00:03:32.511 --> 00:03:35.480
162
+ vector that you can get, there are two things that can happen.
163
+
164
+ 00:03:36.240 --> 00:03:40.120
165
+ For most pairs of vectors, you'll be able to reach every possible point in the plane.
166
+
167
+ 00:03:40.600 --> 00:03:42.940
168
+ Every two-dimensional vector is within your grasp.
169
+
170
+ 00:03:43.560 --> 00:03:47.582
171
+ However, in the unlucky case where your two original vectors happen to line up,
172
+
173
+ 00:03:47.582 --> 00:03:52.008
174
+ the tip of the resulting vector is limited to just this single line passing through the
175
+
176
+ 00:03:52.008 --> 00:03:52.360
177
+ origin.
178
+
179
+ 00:03:53.980 --> 00:03:56.120
180
+ Actually, technically there's a third possibility too.
181
+
182
+ 00:03:56.480 --> 00:04:00.160
183
+ Both your vectors could be zero, in which case you'd just be stuck at the origin.
184
+
185
+ 00:04:01.400 --> 00:04:02.380
186
+ Here's some more terminology.
187
+
188
+ 00:04:02.840 --> 00:04:07.171
189
+ The set of all possible vectors that you can reach with a linear combination
190
+
191
+ 00:04:07.171 --> 00:04:10.940
192
+ of a given pair of vectors is called the span of those two vectors.
193
+
194
+ 00:04:14.680 --> 00:04:18.283
195
+ So restating what we just saw in this lingo, the span of most
196
+
197
+ 00:04:18.283 --> 00:04:22.409
198
+ pairs of 2D vectors is all vectors of 2D space, but when they line up,
199
+
200
+ 00:04:22.409 --> 00:04:25.780
201
+ their span is all vectors whose tip sit on a certain line.
202
+
203
+ 00:04:27.160 --> 00:04:29.280
204
+ Remember how I said that linear algebra revolves
205
+
206
+ 00:04:29.280 --> 00:04:31.400
207
+ around vector addition and scalar multiplication?
208
+
209
+ 00:04:31.960 --> 00:04:35.410
210
+ Well, the span of two vectors is basically a way of asking what
211
+
212
+ 00:04:35.410 --> 00:04:40.155
213
+ are all the possible vectors you can reach using only these two fundamental operations,
214
+
215
+ 00:04:40.155 --> 00:04:42.420
216
+ vector addition and scalar multiplication.
217
+
218
+ 00:04:43.620 --> 00:04:47.220
219
+ This is a good time to talk about how people commonly think about vectors as points.
220
+
221
+ 00:04:47.940 --> 00:04:51.931
222
+ It gets really crowded to think about a whole collection of vectors sitting on a line,
223
+
224
+ 00:04:51.931 --> 00:04:55.556
225
+ and more crowded still to think about all two-dimensional vectors all at once,
226
+
227
+ 00:04:55.556 --> 00:04:56.520
228
+ filling up the plane.
229
+
230
+ 00:04:57.220 --> 00:05:00.123
231
+ So when dealing with collections of vectors like this,
232
+
233
+ 00:05:00.123 --> 00:05:03.397
234
+ it's common to represent each one with just a point in space,
235
+
236
+ 00:05:03.397 --> 00:05:06.195
237
+ the point at the tip of that vector where, as usual,
238
+
239
+ 00:05:06.195 --> 00:05:09.680
240
+ I want you thinking about that vector with its tail on the origin.
241
+
242
+ 00:05:10.580 --> 00:05:14.067
243
+ That way, if you want to think about every possible vector whose
244
+
245
+ 00:05:14.067 --> 00:05:17.340
246
+ tip sits on a certain line, just think about the line itself.
247
+
248
+ 00:05:19.980 --> 00:05:24.237
249
+ Likewise, to think about all possible two-dimensional vectors all at once,
250
+
251
+ 00:05:24.237 --> 00:05:27.360
252
+ conceptualize each one as the point where its tip sits.
253
+
254
+ 00:05:27.360 --> 00:05:30.790
255
+ So in effect, what you'll be thinking about is the infinite flat
256
+
257
+ 00:05:30.790 --> 00:05:34.380
258
+ sheet of two-dimensional space itself, leaving the arrows out of it.
259
+
260
+ 00:05:36.140 --> 00:05:39.740
261
+ In general, if you're thinking about a vector on its own, think of it as an arrow.
262
+
263
+ 00:05:40.160 --> 00:05:42.397
264
+ And if you're dealing with a collection of vectors,
265
+
266
+ 00:05:42.397 --> 00:05:44.420
267
+ it's convenient to think of them all as points.
268
+
269
+ 00:05:45.240 --> 00:05:48.687
270
+ So for our span example, the span of most pairs of vectors ends
271
+
272
+ 00:05:48.687 --> 00:05:51.920
273
+ up being the entire infinite sheet of two-dimensional space.
274
+
275
+ 00:05:52.180 --> 00:05:54.880
276
+ But if they line up, their span is just a line.
277
+
278
+ 00:05:58.200 --> 00:06:00.659
279
+ The idea of span gets a lot more interesting if we
280
+
281
+ 00:06:00.659 --> 00:06:03.360
282
+ start thinking about vectors in three-dimensional space.
283
+
284
+ 00:06:04.080 --> 00:06:07.250
285
+ For example, if you take two vectors in 3D space that are not
286
+
287
+ 00:06:07.250 --> 00:06:10.780
288
+ pointing in the same direction, what does it mean to take their span?
289
+
290
+ 00:06:13.340 --> 00:06:17.346
291
+ Well, their span is the collection of all possible linear combinations
292
+
293
+ 00:06:17.346 --> 00:06:21.295
294
+ of those two vectors, meaning all possible vectors you get by scaling
295
+
296
+ 00:06:21.295 --> 00:06:25.020
297
+ each of the two of them in some way and then adding them together.
298
+
299
+ 00:06:25.780 --> 00:06:28.957
300
+ You can kind of imagine turning two different knobs to change
301
+
302
+ 00:06:28.957 --> 00:06:31.469
303
+ the two scalars defining the linear combination,
304
+
305
+ 00:06:31.469 --> 00:06:35.160
306
+ adding the scaled vectors and following the tip of the resulting vector.
307
+
308
+ 00:06:36.040 --> 00:06:38.430
309
+ That tip will trace out some kind of flat sheet
310
+
311
+ 00:06:38.430 --> 00:06:41.120
312
+ cutting through the origin of three-dimensional space.
313
+
314
+ 00:06:41.940 --> 00:06:44.560
315
+ This flat sheet is the span of the two vectors.
316
+
317
+ 00:06:45.120 --> 00:06:48.101
318
+ Or more precisely, the set of all possible vectors whose
319
+
320
+ 00:06:48.101 --> 00:06:51.240
321
+ tips sit on that flat sheet is the span of your two vectors.
322
+
323
+ 00:06:51.880 --> 00:06:53.360
324
+ Isn't that a beautiful mental image?
325
+
326
+ 00:06:54.480 --> 00:06:56.946
327
+ So, what happens if we add a third vector and
328
+
329
+ 00:06:56.946 --> 00:06:59.360
330
+ consider the span of all three of those guys?
331
+
332
+ 00:07:00.700 --> 00:07:03.004
333
+ A linear combination of three vectors is defined
334
+
335
+ 00:07:03.004 --> 00:07:04.980
336
+ pretty much the same way as it is for two.
337
+
338
+ 00:07:05.380 --> 00:07:09.130
339
+ You'll choose three different scalars, scale each of those vectors,
340
+
341
+ 00:07:09.130 --> 00:07:10.840
342
+ and then add them all together.
343
+
344
+ 00:07:15.980 --> 00:07:20.960
345
+ And again, the span of these vectors is the set of all possible linear combinations.
346
+
347
+ 00:07:24.320 --> 00:07:25.960
348
+ Two different things could happen here.
349
+
350
+ 00:07:26.320 --> 00:07:30.055
351
+ If your third vector happens to be sitting on the span of the first two,
352
+
353
+ 00:07:30.055 --> 00:07:31.540
354
+ then the span doesn't change.
355
+
356
+ 00:07:31.820 --> 00:07:33.940
357
+ You're sort of trapped on that same flat sheet.
358
+
359
+ 00:07:34.500 --> 00:07:37.785
360
+ In other words, adding a scaled version of that third vector to the
361
+
362
+ 00:07:37.785 --> 00:07:41.120
363
+ linear combination doesn't really give you access to any new vectors.
364
+
365
+ 00:07:42.720 --> 00:07:44.988
366
+ But if you just randomly choose a third vector,
367
+
368
+ 00:07:44.988 --> 00:07:48.060
369
+ it's almost certainly not sitting on the span of those first two.
370
+
371
+ 00:07:48.700 --> 00:07:51.259
372
+ Then, since it's pointing in a separate direction,
373
+
374
+ 00:07:51.259 --> 00:07:54.320
375
+ it unlocks access to every possible three-dimensional vector.
376
+
377
+ 00:07:55.520 --> 00:07:59.889
378
+ One way I like to think about this is that as you scale that new third vector,
379
+
380
+ 00:07:59.889 --> 00:08:04.480
381
+ it moves around that span sheet of the first two, sweeping it through all of space.
382
+
383
+ 00:08:05.900 --> 00:08:10.162
384
+ Another way to think about it is that you're making full use of the three freely changing
385
+
386
+ 00:08:10.162 --> 00:08:14.140
387
+ scalars that you have at your disposal to access the full three dimensions of space.
388
+
389
+ 00:08:16.640 --> 00:08:21.065
390
+ Now, in the case where the third vector was already sitting on the span of the first two,
391
+
392
+ 00:08:21.065 --> 00:08:23.475
393
+ or the case where two vectors happen to line up,
394
+
395
+ 00:08:23.475 --> 00:08:27.605
396
+ we want some terminology to describe the fact that at least one of these vectors is
397
+
398
+ 00:08:27.605 --> 00:08:29.720
399
+ redundant, not adding anything to our span.
400
+
401
+ 00:08:30.820 --> 00:08:35.144
402
+ Whenever this happens, where you have multiple vectors and you could remove one without
403
+
404
+ 00:08:35.144 --> 00:08:39.419
405
+ reducing the span, the relevant terminology is to say that they are linearly dependent.
406
+
407
+ 00:08:40.380 --> 00:08:44.578
408
+ Another way of phrasing that would be to say that one of the vectors can be expressed
409
+
410
+ 00:08:44.578 --> 00:08:48.680
411
+ as a linear combination of the others, since it's already in the span of the others.
412
+
413
+ 00:08:52.980 --> 00:08:57.424
414
+ On the other hand, if each vector really does add another dimension to the span,
415
+
416
+ 00:08:57.424 --> 00:08:59.620
417
+ they're said to be linearly independent.
418
+
419
+ 00:09:06.340 --> 00:09:09.376
420
+ So with all of that terminology, and hopefully with some good mental
421
+
422
+ 00:09:09.376 --> 00:09:12.280
423
+ images to go with it, let me leave you with a puzzle before we go.
424
+
425
+ 00:09:12.280 --> 00:09:16.048
426
+ The technical definition of a basis of a space is a
427
+
428
+ 00:09:16.048 --> 00:09:20.180
429
+ set of linearly independent vectors that span that space.
430
+
431
+ 00:09:22.040 --> 00:09:26.925
432
+ Now, given how I described a basis earlier, and given your current understanding of the
433
+
434
+ 00:09:26.925 --> 00:09:31.700
435
+ words span and linearly independent, think about why this definition would make sense.
436
+
437
+ 00:09:33.880 --> 00:09:37.880
438
+ In the next video, I'll get into matrices in transforming space. See you then!
439
+
3B1B_LinearAlgebra/02 - Linear combinations, span, and basis vectors | Chapter 2, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:61e2647750deb9ab4f5d1537ac553c00e06815d17fedc9923f62cffd6457bdeb
3
+ size 27243339
3B1B_LinearAlgebra/03 - Linear transformations and matrices | Chapter 3, Essence of linear algebra.en.vtt ADDED
@@ -0,0 +1,484 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:00.000 --> 00:00:12.040
6
+ [Submit subtitle corrections at criblate.com]
7
+
8
+ 00:00:12.040 --> 00:00:12.920
9
+ Hey everyone!
10
+
11
+ 00:00:13.320 --> 00:00:16.232
12
+ If I had to choose just one topic that makes all of the others in
13
+
14
+ 00:00:16.232 --> 00:00:19.189
15
+ linear algebra start to click, and which too often goes unlearned
16
+
17
+ 00:00:19.189 --> 00:00:22.280
18
+ the first time a student takes linear algebra, it would be this one.
19
+
20
+ 00:00:22.700 --> 00:00:26.200
21
+ The idea of a linear transformation and its relation to matrices.
22
+
23
+ 00:00:26.950 --> 00:00:30.935
24
+ For this video, I'm just going to focus on what these transformations look like in the
25
+
26
+ 00:00:30.935 --> 00:00:35.060
27
+ case of two dimensions, and how they relate to the idea of matrix-vector multiplication.
28
+
29
+ 00:00:35.880 --> 00:00:38.717
30
+ In particular, I want to show you a way to think about
31
+
32
+ 00:00:38.717 --> 00:00:42.080
33
+ matrix-vector multiplication that doesn't rely on memorization.
34
+
35
+ 00:00:43.160 --> 00:00:46.580
36
+ To start, let's just parse this term, linear transformation.
37
+
38
+ 00:00:47.420 --> 00:00:49.880
39
+ Transformation is essentially a fancy word for function.
40
+
41
+ 00:00:50.260 --> 00:00:53.980
42
+ It's something that takes in inputs and spits out an output for each one.
43
+
44
+ 00:00:53.980 --> 00:00:56.364
45
+ Specifically, in the context of linear algebra,
46
+
47
+ 00:00:56.364 --> 00:00:59.812
48
+ we like to think about transformations that take in some vector and
49
+
50
+ 00:00:59.812 --> 00:01:01.080
51
+ spit out another vector.
52
+
53
+ 00:01:02.500 --> 00:01:06.380
54
+ So why use the word transformation instead of function if they mean the same thing?
55
+
56
+ 00:01:07.120 --> 00:01:11.340
57
+ Well, it's to be suggestive of a certain way to visualize this input-output relation.
58
+
59
+ 00:01:11.860 --> 00:01:15.800
60
+ You see, a great way to understand functions of vectors is to use movement.
61
+
62
+ 00:01:16.780 --> 00:01:20.914
63
+ If a transformation takes some input vector to some output vector,
64
+
65
+ 00:01:20.914 --> 00:01:24.860
66
+ we imagine that input vector moving over to the output vector.
67
+
68
+ 00:01:25.680 --> 00:01:28.442
69
+ Then to understand the transformation as a whole,
70
+
71
+ 00:01:28.442 --> 00:01:32.445
72
+ we might imagine watching every possible input vector move over to its
73
+
74
+ 00:01:32.445 --> 00:01:34.080
75
+ corresponding output vector.
76
+
77
+ 00:01:34.980 --> 00:01:38.119
78
+ It gets really crowded to think about all of the vectors all at once,
79
+
80
+ 00:01:38.119 --> 00:01:39.120
81
+ each one as an arrow.
82
+
83
+ 00:01:39.500 --> 00:01:43.570
84
+ So as I mentioned last video, a nice trick is to conceptualize each vector
85
+
86
+ 00:01:43.570 --> 00:01:47.420
87
+ not as an arrow, but as a single point, the point where its tip sits.
88
+
89
+ 00:01:48.030 --> 00:01:52.211
90
+ That way, to think about a transformation taking every possible input vector to
91
+
92
+ 00:01:52.211 --> 00:01:56.340
93
+ some output vector, we watch every point in space moving to some other point.
94
+
95
+ 00:01:57.220 --> 00:01:59.617
96
+ In the case of transformations in two dimensions,
97
+
98
+ 00:01:59.617 --> 00:02:02.747
99
+ to get a better feel for the whole shape of the transformation,
100
+
101
+ 00:02:02.747 --> 00:02:05.780
102
+ I like to do this with all of the points on an infinite grid.
103
+
104
+ 00:02:06.560 --> 00:02:09.403
105
+ I also sometimes like to keep a copy of the grid in the background,
106
+
107
+ 00:02:09.403 --> 00:02:12.840
108
+ just to help keep track of where everything ends up relative to where it starts.
109
+
110
+ 00:02:14.460 --> 00:02:19.238
111
+ The effect for various transformations moving around all of the points in space is,
112
+
113
+ 00:02:19.238 --> 00:02:21.080
114
+ you've got to admit, beautiful.
115
+
116
+ 00:02:21.880 --> 00:02:24.640
117
+ It gives the feeling of squishing and morphing space itself.
118
+
119
+ 00:02:25.600 --> 00:02:29.920
120
+ As you can imagine though, arbitrary transformations can look pretty complicated.
121
+
122
+ 00:02:30.380 --> 00:02:34.630
123
+ But luckily, linear algebra limits itself to a special type of transformation,
124
+
125
+ 00:02:34.630 --> 00:02:38.280
126
+ ones that are easier to understand, called linear transformations.
127
+
128
+ 00:02:39.120 --> 00:02:43.060
129
+ Visually speaking, a transformation is linear if it has two properties.
130
+
131
+ 00:02:43.700 --> 00:02:46.901
132
+ All lines must remain lines without getting curved,
133
+
134
+ 00:02:46.901 --> 00:02:49.600
135
+ and the origin must remain fixed in place.
136
+
137
+ 00:02:50.620 --> 00:02:53.968
138
+ For example, this right here would not be a linear transformation,
139
+
140
+ 00:02:53.968 --> 00:02:55.540
141
+ since the lines get all curvy.
142
+
143
+ 00:02:56.100 --> 00:02:59.003
144
+ And this one right here, although it keeps the lines straight,
145
+
146
+ 00:02:59.003 --> 00:03:01.860
147
+ is not a linear transformation, because it moves the origin.
148
+
149
+ 00:03:02.680 --> 00:03:05.919
150
+ This one here fixes the origin, and it might look like it keeps lines straight,
151
+
152
+ 00:03:05.919 --> 00:03:09.240
153
+ but that's just because I'm only showing the horizontal and vertical grid lines.
154
+
155
+ 00:03:09.540 --> 00:03:12.362
156
+ When you see what it does to a diagonal line, it becomes clear
157
+
158
+ 00:03:12.362 --> 00:03:15.320
159
+ that it's not at all linear, since it turns that line all curvy.
160
+
161
+ 00:03:16.760 --> 00:03:19.605
162
+ In general, you should think of linear transformations
163
+
164
+ 00:03:19.605 --> 00:03:22.240
165
+ as keeping grid lines parallel and evenly spaced.
166
+
167
+ 00:03:23.400 --> 00:03:27.540
168
+ Some linear transformations are simple to think about, like rotations about the origin.
169
+
170
+ 00:03:28.120 --> 00:03:30.600
171
+ Others are a little trickier to describe with words.
172
+
173
+ 00:03:32.040 --> 00:03:35.480
174
+ So, how do you think you could describe these transformations numerically?
175
+
176
+ 00:03:35.480 --> 00:03:39.604
177
+ If you were, say, programming some animations to make a video teaching the topic,
178
+
179
+ 00:03:39.604 --> 00:03:43.524
180
+ what formula do you give the computer so that if you give it the coordinates
181
+
182
+ 00:03:43.524 --> 00:03:47.240
183
+ of a vector, it can give you the coordinates of where that vector lands?
184
+
185
+ 00:03:48.480 --> 00:03:52.511
186
+ It turns out that you only need to record where the two basis vectors,
187
+
188
+ 00:03:52.511 --> 00:03:56.600
189
+ i-hat and j-hat, each land, and everything else will follow from that.
190
+
191
+ 00:03:57.500 --> 00:04:01.669
192
+ For example, consider the vector v with coordinates (-1, 2),
193
+
194
+ 00:04:01.669 --> 00:04:05.700
195
+ meaning that it equals −1 times i-hat plus 2 times j-hat.
196
+
197
+ 00:04:08.680 --> 00:04:12.919
198
+ If we play some transformation and follow where all three of these vectors go,
199
+
200
+ 00:04:12.919 --> 00:04:17.593
201
+ the property that grid lines remain parallel and evenly spaced has a really important
202
+
203
+ 00:04:17.593 --> 00:04:18.300
204
+ consequence.
205
+
206
+ 00:04:19.100 --> 00:04:22.250
207
+ The place where v lands will be -1 times the vector where
208
+
209
+ 00:04:22.250 --> 00:04:25.400
210
+ i-hat landed plus 2 times the vector where j-hat landed.
211
+
212
+ 00:04:25.980 --> 00:04:30.280
213
+ In other words, it started off as a certain linear combination of i-hat and j-hat,
214
+
215
+ 00:04:30.280 --> 00:04:34.580
216
+ and it ends up as that same linear combination of where those two vectors landed.
217
+
218
+ 00:04:35.620 --> 00:04:40.920
219
+ This means you can deduce where v must go based only on where i-hat and j-hat each land.
220
+
221
+ 00:04:41.580 --> 00:04:44.540
222
+ This is why I like keeping a copy of the original grid in the background.
223
+
224
+ 00:04:45.080 --> 00:04:49.727
225
+ For the transformation shown here, we can read off that i-hat lands on the
226
+
227
+ 00:04:49.727 --> 00:04:54.940
228
+ coordinates (1, -2), and j-hat lands on the x-axis over at the coordinates (3, 0).
229
+
230
+ 00:04:55.540 --> 00:05:00.764
231
+ This means that the vector represented by -1 i-hat plus 2 times j-hat
232
+
233
+ 00:05:00.764 --> 00:05:06.140
234
+ ends up at -1 times the vector (1, -2) plus 2 times the vector (3, 0).
235
+
236
+ 00:05:07.100 --> 00:05:11.680
237
+ Adding that all together, you can deduce that it has to land on the vector (5, 2).
238
+
239
+ 00:05:14.260 --> 00:05:17.240
240
+ This is a good point to pause and ponder, because it's pretty important.
241
+
242
+ 00:05:18.520 --> 00:05:21.799
243
+ Now, given that I'm actually showing you the full transformation,
244
+
245
+ 00:05:21.799 --> 00:05:25.280
246
+ you could have just looked to see that v has the coordinates (5, 2).
247
+
248
+ 00:05:25.760 --> 00:05:29.519
249
+ But the cool part here is that this gives us a technique to deduce
250
+
251
+ 00:05:29.519 --> 00:05:33.279
252
+ where any vectors land so long as we have a record of where i-hat
253
+
254
+ 00:05:33.279 --> 00:05:37.380
255
+ and j-hat each land without needing to watch the transformation itself.
256
+
257
+ 00:05:38.600 --> 00:05:42.462
258
+ Write the vector with more general coordinates, x and y,
259
+
260
+ 00:05:42.462 --> 00:05:47.083
261
+ and it will land on x times the vector where i-hat lands, (1, -2),
262
+
263
+ 00:05:47.083 --> 00:05:50.600
264
+ plus y times the vector where j-hat lands, (3, 0).
265
+
266
+ 00:05:51.860 --> 00:05:58.100
267
+ Carrying out that sum, you see that it lands at (1x + 3y, -2x + 0y).
268
+
269
+ 00:05:58.740 --> 00:06:03.580
270
+ I give you any vector, and you can tell me where that vector lands using this formula.
271
+
272
+ 00:06:04.860 --> 00:06:08.981
273
+ What all of this is saying is that a two-dimensional linear transformation
274
+
275
+ 00:06:08.981 --> 00:06:11.543
276
+ is completely described by just four numbers,
277
+
278
+ 00:06:11.543 --> 00:06:16.500
279
+ the two coordinates for where i-hat lands and the two coordinates for where j-hat lands.
280
+
281
+ 00:06:17.080 --> 00:06:17.640
282
+ Isn't that cool?
283
+
284
+ 00:06:18.380 --> 00:06:23.765
285
+ It's common to package these coordinates into a 2x2 grid of numbers called a 2x2 matrix,
286
+
287
+ 00:06:23.765 --> 00:06:27.621
288
+ where you can interpret the columns as the two special vectors
289
+
290
+ 00:06:27.621 --> 00:06:29.640
291
+ where i-hat and j-hat each land.
292
+
293
+ 00:06:30.380 --> 00:06:35.479
294
+ If you're given a 2x2 matrix describing a linear transformation and some specific vector,
295
+
296
+ 00:06:35.479 --> 00:06:39.662
297
+ and you want to know where that linear transformation takes that vector,
298
+
299
+ 00:06:39.662 --> 00:06:44.189
300
+ you can take the coordinates of the vector, multiply them by the corresponding
301
+
302
+ 00:06:44.189 --> 00:06:47.340
303
+ columns of the matrix, then add together what you get.
304
+
305
+ 00:06:48.180 --> 00:06:52.720
306
+ This corresponds with the idea of adding the scaled versions of our new basis vectors.
307
+
308
+ 00:06:54.720 --> 00:06:57.795
309
+ Let's see what this looks like in the most general case,
310
+
311
+ 00:06:57.795 --> 00:07:00.540
312
+ where your matrix has entries "a", "b", "c", "d".
313
+
314
+ 00:07:01.100 --> 00:07:03.670
315
+ And remember, this matrix is just a way of packaging the
316
+
317
+ 00:07:03.670 --> 00:07:06.240
318
+ information needed to describe a linear transformation.
319
+
320
+ 00:07:06.240 --> 00:07:09.256
321
+ Always remember to interpret that first column, (a, c),
322
+
323
+ 00:07:09.256 --> 00:07:13.259
324
+ as the place where the first basis vector lands, and that second column,
325
+
326
+ 00:07:13.259 --> 00:07:16.440
327
+ (b, d), as the place where the second basis vector lands.
328
+
329
+ 00:07:17.500 --> 00:07:21.000
330
+ When we apply this transformation to some vector (x, y), what do you get?
331
+
332
+ 00:07:22.060 --> 00:07:26.980
333
+ Well, it'll be x times (a, c) plus y times (b, d).
334
+
335
+ 00:07:28.060 --> 00:07:33.300
336
+ Putting this together, you get a vector (ax + by, cx + dy).
337
+
338
+ 00:07:33.980 --> 00:07:37.115
339
+ You could even define this as matrix vector multiplication,
340
+
341
+ 00:07:37.115 --> 00:07:40.940
342
+ when you put the matrix on the left of the vector like it's a function.
343
+
344
+ 00:07:41.660 --> 00:07:44.076
345
+ Then, you could make high schoolers memorize this without
346
+
347
+ 00:07:44.076 --> 00:07:46.620
348
+ showing them the crucial part that makes it feel intuitive.
349
+
350
+ 00:07:48.300 --> 00:07:51.280
351
+ But, isn't it more fun to think about these columns as the
352
+
353
+ 00:07:51.280 --> 00:07:54.517
354
+ transformed versions of your basis vectors, and to think about
355
+
356
+ 00:07:54.517 --> 00:07:57.960
357
+ the result as the appropriate linear combination of those vectors?
358
+
359
+ 00:08:00.720 --> 00:08:03.780
360
+ Let's practice describing a few linear transformations with matrices.
361
+
362
+ 00:08:04.580 --> 00:08:09.204
363
+ For example, if we rotate all of space 90 degrees counterclockwise,
364
+
365
+ 00:08:09.204 --> 00:08:12.240
366
+ then i-hat lands on the coordinates (0, 1).
367
+
368
+ 00:08:13.980 --> 00:08:17.180
369
+ And j-hat lands on the coordinates (-1, 0).
370
+
371
+ 00:08:17.980 --> 00:08:21.960
372
+ So the matrix we end up with has columns (0, 1), (-1, 0).
373
+
374
+ 00:08:22.880 --> 00:08:26.576
375
+ To figure out what happens to any vector after a 90-degree rotation,
376
+
377
+ 00:08:26.576 --> 00:08:29.620
378
+ you could just multiply its coordinates by this matrix.
379
+
380
+ 00:08:31.560 --> 00:08:34.300
381
+ Here's a fun transformation with a special name, called a shear.
382
+
383
+ 00:08:35.000 --> 00:08:39.160
384
+ In it, i-hat remains fixed, so the first column of the matrix is (1, 0).
385
+
386
+ 00:08:39.600 --> 00:08:42.481
387
+ But j-hat moves over to the coordinates (1, 1),
388
+
389
+ 00:08:42.481 --> 00:08:45.300
390
+ which become the second column of the matrix.
391
+
392
+ 00:08:45.300 --> 00:08:49.902
393
+ And at the risk of being redundant here, figuring out how a shear transforms
394
+
395
+ 00:08:49.902 --> 00:08:54.080
396
+ a given vector comes down to multiplying this matrix by that vector.
397
+
398
+ 00:08:55.760 --> 00:08:59.468
399
+ Let's say we want to go the other way around, starting with a matrix,
400
+
401
+ 00:08:59.468 --> 00:09:03.875
402
+ say with columns (1, 2) and (3, 1), and we want to deduce what its transformation
403
+
404
+ 00:09:03.875 --> 00:09:04.520
405
+ looks like.
406
+
407
+ 00:09:04.960 --> 00:09:07.440
408
+ Pause and take a moment to see if you can imagine it.
409
+
410
+ 00:09:08.420 --> 00:09:15.100
411
+ One way to do this is to first move i-hat to (1, 2), then move j-hat to (3, 1).
412
+
413
+ 00:09:15.100 --> 00:09:17.551
414
+ Always moving the rest of space in such a way
415
+
416
+ 00:09:17.551 --> 00:09:20.220
417
+ that keeps gridlines parallel and evenly spaced.
418
+
419
+ 00:09:21.680 --> 00:09:26.168
420
+ If the vectors that i-hat and j-hat land on are linearly dependent, which,
421
+
422
+ 00:09:26.168 --> 00:09:31.019
423
+ if you recall from last video, means that one is a scaled version of the other,
424
+
425
+ 00:09:31.019 --> 00:09:36.174
426
+ it means that the linear transformation squishes all of 2D space onto the line where
427
+
428
+ 00:09:36.174 --> 00:09:41.268
429
+ those two vectors sit, also known as the one-dimensional span of those two linearly
430
+
431
+ 00:09:41.268 --> 00:09:42.420
432
+ dependent vectors.
433
+
434
+ 00:09:44.420 --> 00:09:48.851
435
+ To sum up, linear transformations are a way to move around space such that
436
+
437
+ 00:09:48.851 --> 00:09:53.940
438
+ gridlines remain parallel and evenly spaced, and such that the origin remains fixed.
439
+
440
+ 00:09:54.540 --> 00:09:58.941
441
+ Delightfully, these transformations can be described using only a handful of numbers,
442
+
443
+ 00:09:58.941 --> 00:10:01.530
444
+ the coordinates of where each basis vector lands.
445
+
446
+ 00:10:02.760 --> 00:10:06.176
447
+ Matrices give us a language to describe these transformations,
448
+
449
+ 00:10:06.176 --> 00:10:08.765
450
+ where the columns represent those coordinates,
451
+
452
+ 00:10:08.765 --> 00:10:12.511
453
+ and matrix-vector multiplication is just a way to compute what that
454
+
455
+ 00:10:12.511 --> 00:10:14.660
456
+ transformation does to a given vector.
457
+
458
+ 00:10:15.360 --> 00:10:18.753
459
+ The important takeaway here is that every time you see a matrix,
460
+
461
+ 00:10:18.753 --> 00:10:21.880
462
+ you can interpret it as a certain transformation of space.
463
+
464
+ 00:10:22.580 --> 00:10:24.779
465
+ Once you really digest this idea, you're in a
466
+
467
+ 00:10:24.779 --> 00:10:27.320
468
+ great position to understand linear algebra deeply.
469
+
470
+ 00:10:27.660 --> 00:10:32.230
471
+ Almost all of the topics coming up, from matrix multiplication to determinants,
472
+
473
+ 00:10:32.230 --> 00:10:36.626
474
+ change of basis, eigenvalues, all of these will become easier to understand
475
+
476
+ 00:10:36.626 --> 00:10:40.560
477
+ once you start thinking about matrices as transformations of space.
478
+
479
+ 00:10:41.300 --> 00:10:43.740
480
+ Most immediately, in the next video, I'll be talking
481
+
482
+ 00:10:43.740 --> 00:10:46.320
483
+ about multiplying two matrices together. See you then!
484
+
3B1B_LinearAlgebra/03 - Linear transformations and matrices | Chapter 3, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:08411463bef7ab6d54d145c11bdb53817b794d6cc917115c185556cec574d7f0
3
+ size 46616392
3B1B_LinearAlgebra/04 - Matrix multiplication as composition | Chapter 4, Essence of linear algebra.en.vtt ADDED
@@ -0,0 +1,454 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:00.000 --> 00:00:10.940
6
+ [Submit subtitle corrections at criblate.com]
7
+
8
+ 00:00:10.940 --> 00:00:13.674
9
+ Hey everyone, where we last left off, I showed what linear
10
+
11
+ 00:00:13.674 --> 00:00:16.880
12
+ transformations look like and how to represent them using matrices.
13
+
14
+ 00:00:18.320 --> 00:00:21.110
15
+ This is worth a quick recap because it's just really important,
16
+
17
+ 00:00:21.110 --> 00:00:23.634
18
+ but of course if this feels like more than just a recap,
19
+
20
+ 00:00:23.634 --> 00:00:25.140
21
+ go back and watch the full video.
22
+
23
+ 00:00:25.780 --> 00:00:29.683
24
+ Technically speaking, linear transformations are functions with vectors as
25
+
26
+ 00:00:29.683 --> 00:00:33.427
27
+ inputs and vectors as outputs, but I showed last time how we can think
28
+
29
+ 00:00:33.427 --> 00:00:37.119
30
+ about them visually as smooshing around space in such a way that grid
31
+
32
+ 00:00:37.119 --> 00:00:41.180
33
+ lines stay parallel and evenly spaced, and so that the origin remains fixed.
34
+
35
+ 00:00:41.820 --> 00:00:46.580
36
+ The key takeaway was that a linear transformation is completely determined by where it
37
+
38
+ 00:00:46.580 --> 00:00:51.340
39
+ takes the basis vectors of the space, which for two dimensions means i-hat and j-hat.
40
+
41
+ 00:00:51.340 --> 00:00:54.431
42
+ This is because any other vector could be described
43
+
44
+ 00:00:54.431 --> 00:00:57.340
45
+ as a linear combination of those basis vectors.
46
+
47
+ 00:00:57.940 --> 00:01:02.340
48
+ A vector with coordinates (x, y) is x times i-hat plus y times j-hat.
49
+
50
+ 00:01:03.460 --> 00:01:06.515
51
+ After going through the transformation, this property that grid
52
+
53
+ 00:01:06.515 --> 00:01:09.860
54
+ lines remain parallel and evenly spaced has a wonderful consequence.
55
+
56
+ 00:01:10.500 --> 00:01:14.057
57
+ The place where your vector lands will be x times the transformed
58
+
59
+ 00:01:14.057 --> 00:01:17.560
60
+ version of i-hat plus y times the transformed version of j-hat.
61
+
62
+ 00:01:18.240 --> 00:01:23.277
63
+ This means if you keep a record of the coordinates where i-hat lands and the coordinates
64
+
65
+ 00:01:23.277 --> 00:01:27.226
66
+ where j-hat lands, you can compute that a vector which starts at (x,
67
+
68
+ 00:01:27.226 --> 00:01:32.148
69
+ y) must land on x times the new coordinates of i-hat plus y times the new coordinates
70
+
71
+ 00:01:32.148 --> 00:01:32.720
72
+ of j-hat.
73
+
74
+ 00:01:33.560 --> 00:01:37.363
75
+ The convention is to record the coordinates of where i-hat and j-hat
76
+
77
+ 00:01:37.363 --> 00:01:41.278
78
+ land as the columns of a matrix, and to define this sum of the scaled
79
+
80
+ 00:01:41.278 --> 00:01:45.360
81
+ versions of those columns by x and y to be matrix-vector multiplication.
82
+
83
+ 00:01:46.050 --> 00:01:50.072
84
+ In this way, a matrix represents a specific linear transformation,
85
+
86
+ 00:01:50.072 --> 00:01:53.363
87
+ and multiplying a matrix by a vector is what it means
88
+
89
+ 00:01:53.363 --> 00:01:57.080
90
+ computationally to apply that transformation to that vector.
91
+
92
+ 00:01:58.800 --> 00:02:00.880
93
+ Alright, recap over, on to the new stuff.
94
+
95
+ 00:02:01.600 --> 00:02:04.202
96
+ Oftentimes, you find yourself wanting to describe the
97
+
98
+ 00:02:04.202 --> 00:02:07.000
99
+ effects of applying one transformation and then another.
100
+
101
+ 00:02:07.620 --> 00:02:11.076
102
+ For example, maybe you want to describe what happens when you first
103
+
104
+ 00:02:11.076 --> 00:02:14.480
105
+ rotate the plane 90 degrees counterclockwise, then apply a shear.
106
+
107
+ 00:02:15.260 --> 00:02:17.726
108
+ The overall effect here, from start to finish,
109
+
110
+ 00:02:17.726 --> 00:02:21.800
111
+ is another linear transformation, distinct from the rotation and the shear.
112
+
113
+ 00:02:22.280 --> 00:02:25.066
114
+ This new linear transformation is commonly called the
115
+
116
+ 00:02:25.066 --> 00:02:28.220
117
+ composition of the two separate transformations we applied.
118
+
119
+ 00:02:28.920 --> 00:02:32.066
120
+ And like any linear transformation, it can be described
121
+
122
+ 00:02:32.066 --> 00:02:35.440
123
+ with a matrix all of its own by following i-hat and j-hat.
124
+
125
+ 00:02:36.020 --> 00:02:41.032
126
+ In this example, the ultimate landing spot for i-hat after both transformations is (1,
127
+
128
+ 00:02:41.032 --> 00:02:44.120
129
+ 1), so let's make that the first column of a matrix.
130
+
131
+ 00:02:44.960 --> 00:02:48.474
132
+ Likewise, j-hat ultimately ends up at the location (-1,
133
+
134
+ 00:02:48.474 --> 00:02:51.860
135
+ 0), so we make that the second column of the matrix.
136
+
137
+ 00:02:52.680 --> 00:02:57.664
138
+ This new matrix captures the overall effect of applying a rotation then a shear,
139
+
140
+ 00:02:57.664 --> 00:03:01.340
141
+ but as one single action, rather than two successive ones.
142
+
143
+ 00:03:03.040 --> 00:03:04.880
144
+ Here's one way to think about that new matrix.
145
+
146
+ 00:03:05.420 --> 00:03:09.900
147
+ If you were to take some vector and pump it through the rotation, then the shear,
148
+
149
+ 00:03:09.900 --> 00:03:14.546
150
+ the long way to compute where it ends up is to first multiply it on the left by the
151
+
152
+ 00:03:14.546 --> 00:03:19.358
153
+ rotation matrix, then take whatever you get and multiply that on the left by the shear
154
+
155
+ 00:03:19.358 --> 00:03:19.800
156
+ matrix.
157
+
158
+ 00:03:20.460 --> 00:03:23.202
159
+ This is, numerically speaking, what it means to
160
+
161
+ 00:03:23.202 --> 00:03:26.060
162
+ apply a rotation then a shear to a given vector.
163
+
164
+ 00:03:26.800 --> 00:03:31.635
165
+ But whatever you get should be the same as just applying this new composition matrix that
166
+
167
+ 00:03:31.635 --> 00:03:35.330
168
+ we just found by that same vector, no matter what vector you chose,
169
+
170
+ 00:03:35.330 --> 00:03:39.948
171
+ since this new matrix is supposed to capture the same overall effect as the rotation
172
+
173
+ 00:03:39.948 --> 00:03:40.980
174
+ then shear action.
175
+
176
+ 00:03:42.480 --> 00:03:45.809
177
+ Based on how things are written down here, I think it's reasonable to
178
+
179
+ 00:03:45.809 --> 00:03:49.380
180
+ call this new matrix the product of the original two matrices, don't you?
181
+
182
+ 00:03:50.420 --> 00:03:53.932
183
+ We can think about how to compute that product more generally in just a moment,
184
+
185
+ 00:03:53.932 --> 00:03:56.600
186
+ but it's way too easy to get lost in the forest of numbers.
187
+
188
+ 00:03:56.600 --> 00:04:00.440
189
+ Always remember that multiplying two matrices like this has the
190
+
191
+ 00:04:00.440 --> 00:04:04.280
192
+ geometric meaning of applying one transformation then another.
193
+
194
+ 00:04:05.860 --> 00:04:09.660
195
+ One thing that's kind of weird here is that this has us reading from right to left.
196
+
197
+ 00:04:10.040 --> 00:04:13.403
198
+ You first apply the transformation represented by the matrix on the right,
199
+
200
+ 00:04:13.403 --> 00:04:16.720
201
+ then you apply the transformation represented by the matrix on the left.
202
+
203
+ 00:04:17.400 --> 00:04:21.478
204
+ This stems from function notation, since we write functions on the left of variables,
205
+
206
+ 00:04:21.478 --> 00:04:25.460
207
+ so every time you compose two functions, you always have to read it right to left.
208
+
209
+ 00:04:25.920 --> 00:04:28.980
210
+ Good news for the Hebrew readers, bad news for the rest of us.
211
+
212
+ 00:04:29.880 --> 00:04:31.100
213
+ Let's look at another example.
214
+
215
+ 00:04:31.760 --> 00:04:36.860
216
+ Take the matrix with columns (1, 1) and (-2, 0), whose transformation looks like this.
217
+
218
+ 00:04:37.980 --> 00:04:39.060
219
+ And let's call it M1.
220
+
221
+ 00:04:40.100 --> 00:04:43.115
222
+ Next, take the matrix with columns (0, 1) and (2,
223
+
224
+ 00:04:43.115 --> 00:04:45.700
225
+ 0), whose transformation looks like this.
226
+
227
+ 00:04:47.520 --> 00:04:49.240
228
+ And let's call that guy M2.
229
+
230
+ 00:04:49.920 --> 00:04:54.120
231
+ The total effect of applying M1 then M2 gives us a new transformation,
232
+
233
+ 00:04:54.120 --> 00:04:55.680
234
+ so let's find its matrix.
235
+
236
+ 00:04:56.280 --> 00:05:00.409
237
+ But this time, let's see if we can do it without watching the animations,
238
+
239
+ 00:05:00.409 --> 00:05:03.860
240
+ and instead just using the numerical entries in each matrix.
241
+
242
+ 00:05:04.740 --> 00:05:07.140
243
+ First, we need to figure out where i-hat goes.
244
+
245
+ 00:05:08.040 --> 00:05:12.352
246
+ After applying M1, the new coordinates of i-hat, by definition,
247
+
248
+ 00:05:12.352 --> 00:05:15.980
249
+ are given by that first column of M1, namely (1, 1).
250
+
251
+ 00:05:16.780 --> 00:05:23.500
252
+ To see what happens after applying M2, multiply the matrix for M2 by that vector, (1, 1).
253
+
254
+ 00:05:25.300 --> 00:05:29.880
255
+ Working it out, the way I described last video, you'll get the vector (2, 1).
256
+
257
+ 00:05:30.700 --> 00:05:33.100
258
+ This will be the first column of the composition matrix.
259
+
260
+ 00:05:34.520 --> 00:05:40.275
261
+ Likewise, to follow j-hat, the second column of M1 tells us that it first lands on (-2,
262
+
263
+ 00:05:40.275 --> 00:05:40.540
264
+ 0).
265
+
266
+ 00:05:42.700 --> 00:05:49.190
267
+ Then, when we apply M2 to that vector, you can work out the matrix-vector product
268
+
269
+ 00:05:49.190 --> 00:05:55.200
270
+ to get (0, -2), which becomes the second column of our composition matrix.
271
+
272
+ 00:05:56.640 --> 00:06:00.780
273
+ Let me talk through that same process again, but this time I'll show variable entries
274
+
275
+ 00:06:00.780 --> 00:06:04.920
276
+ in each matrix, just to show that the same line of reasoning works for any matrices.
277
+
278
+ 00:06:05.540 --> 00:06:08.216
279
+ This is more symbol-heavy and will require some more room,
280
+
281
+ 00:06:08.216 --> 00:06:12.091
282
+ but it should be pretty satisfying for anyone who has previously been taught matrix
283
+
284
+ 00:06:12.091 --> 00:06:13.660
285
+ multiplication the more rote way.
286
+
287
+ 00:06:14.460 --> 00:06:17.736
288
+ To follow where i-hat goes, start by looking at the first column of
289
+
290
+ 00:06:17.736 --> 00:06:21.060
291
+ the matrix on the right, since this is where i-hat initially lands.
292
+
293
+ 00:06:22.000 --> 00:06:26.124
294
+ Multiplying that column by the matrix on the left is how you can tell where the
295
+
296
+ 00:06:26.124 --> 00:06:30.300
297
+ intermediate version of i-hat ends up after applying the second transformation.
298
+
299
+ 00:06:31.620 --> 00:06:34.939
300
+ So the first column of the composition matrix will always equal
301
+
302
+ 00:06:34.939 --> 00:06:38.100
303
+ the left matrix times the first column of the right matrix.
304
+
305
+ 00:06:42.160 --> 00:06:47.140
306
+ Likewise, j-hat will always initially land on the second column of the right matrix.
307
+
308
+ 00:06:48.940 --> 00:06:53.541
309
+ So multiplying the left matrix by this second column will give its final location,
310
+
311
+ 00:06:53.541 --> 00:06:57.020
312
+ and hence that's the second column of the composition matrix.
313
+
314
+ 00:07:00.620 --> 00:07:04.754
315
+ Notice there's a lot of symbols here, and it's common to be taught this formula as
316
+
317
+ 00:07:04.754 --> 00:07:09.040
318
+ something to memorize, along with a certain algorithmic process to help remember it.
319
+
320
+ 00:07:09.160 --> 00:07:12.087
321
+ But I really do think that before memorizing that process,
322
+
323
+ 00:07:12.087 --> 00:07:15.014
324
+ you should get in the habit of thinking about what matrix
325
+
326
+ 00:07:15.014 --> 00:07:18.900
327
+ multiplication really represents, applying one transformation after another.
328
+
329
+ 00:07:19.620 --> 00:07:22.842
330
+ Trust me, this will give you a much better conceptual framework that
331
+
332
+ 00:07:22.842 --> 00:07:26.300
333
+ makes the properties of matrix multiplication much easier to understand.
334
+
335
+ 00:07:27.060 --> 00:07:28.360
336
+ For example, here's a question.
337
+
338
+ 00:07:28.880 --> 00:07:32.840
339
+ Does it matter what order we put the two matrices in when we multiply them?
340
+
341
+ 00:07:33.620 --> 00:07:37.000
342
+ Well, let's think through a simple example, like the one from earlier.
343
+
344
+ 00:07:37.640 --> 00:07:41.402
345
+ Take a shear, which fixes i-hat and smooshes j-hat over to the right,
346
+
347
+ 00:07:41.402 --> 00:07:42.820
348
+ and a 90 degree rotation.
349
+
350
+ 00:07:43.600 --> 00:07:48.732
351
+ If you first do the shear, then rotate, we can see that i-hat ends up at (0,
352
+
353
+ 00:07:48.732 --> 00:07:50.960
354
+ 1) and j-hat ends up at (-1, 1).
355
+
356
+ 00:07:51.320 --> 00:07:53.060
357
+ Both are generally pointing close together.
358
+
359
+ 00:07:53.860 --> 00:07:58.593
360
+ If you first rotate, then do the shear, i-hat ends up over at (1, 1),
361
+
362
+ 00:07:58.593 --> 00:08:04.491
363
+ and j-hat is off in a different direction at (-1, 0), and they're pointing, you know,
364
+
365
+ 00:08:04.491 --> 00:08:05.520
366
+ farther apart.
367
+
368
+ 00:08:06.380 --> 00:08:10.660
369
+ The overall effect here is clearly different, so evidently, order totally does matter.
370
+
371
+ 00:08:12.200 --> 00:08:14.494
372
+ Notice, by thinking in terms of transformations,
373
+
374
+ 00:08:14.494 --> 00:08:17.840
375
+ that's the kind of thing that you can do in your head by visualizing.
376
+
377
+ 00:08:18.220 --> 00:08:19.900
378
+ No matrix multiplication necessary.
379
+
380
+ 00:08:21.480 --> 00:08:25.248
381
+ I remember when I first took linear algebra, there was this one homework
382
+
383
+ 00:08:25.248 --> 00:08:29.120
384
+ problem that asked us to prove that matrix multiplication is associative.
385
+
386
+ 00:08:29.560 --> 00:08:32.810
387
+ This means that if you have three matrices, A, B, and C,
388
+
389
+ 00:08:32.810 --> 00:08:37.918
390
+ and you multiply them all together, it shouldn't matter if you first compute A times B,
391
+
392
+ 00:08:37.918 --> 00:08:41.806
393
+ then multiply the result by C, or if you first multiply B times C,
394
+
395
+ 00:08:41.806 --> 00:08:44.360
396
+ then multiply that result by A on the left.
397
+
398
+ 00:08:44.940 --> 00:08:47.400
399
+ In other words, it doesn't matter where you put the parentheses.
400
+
401
+ 00:08:48.380 --> 00:08:52.205
402
+ Now, if you try to work through this numerically, like I did back then,
403
+
404
+ 00:08:52.205 --> 00:08:55.760
405
+ it's horrible, just horrible, and unenlightening for that matter.
406
+
407
+ 00:08:55.760 --> 00:08:59.299
408
+ But when you think about matrix multiplication as applying one
409
+
410
+ 00:08:59.299 --> 00:09:02.780
411
+ transformation after another, this property is just trivial.
412
+
413
+ 00:09:03.300 --> 00:09:04.000
414
+ Can you see why?
415
+
416
+ 00:09:04.860 --> 00:09:08.655
417
+ What it's saying is that if you first apply C, then B,
418
+
419
+ 00:09:08.655 --> 00:09:12.380
420
+ then A, it's the same as applying C, then B, then A.
421
+
422
+ 00:09:12.820 --> 00:09:14.380
423
+ I mean, there's nothing to prove.
424
+
425
+ 00:09:14.540 --> 00:09:18.660
426
+ You're just applying the same three things one after the other, all in the same order.
427
+
428
+ 00:09:19.460 --> 00:09:21.540
429
+ This might feel like cheating, but it's not.
430
+
431
+ 00:09:21.540 --> 00:09:25.809
432
+ This is an honest-to-goodness proof that matrix multiplication is associative,
433
+
434
+ 00:09:25.809 --> 00:09:30.680
435
+ and even better than that, it's a good explanation for why that property should be true.
436
+
437
+ 00:09:31.560 --> 00:09:34.507
438
+ I really do encourage you to play around more with this idea,
439
+
440
+ 00:09:34.507 --> 00:09:38.082
441
+ imagining two different transformations, thinking about what happens when
442
+
443
+ 00:09:38.082 --> 00:09:42.140
444
+ you apply one after the other, and then working out the matrix product numerically.
445
+
446
+ 00:09:42.600 --> 00:09:46.440
447
+ Trust me, this is the kind of playtime that really makes the idea sink in.
448
+
449
+ 00:09:47.200 --> 00:09:49.667
450
+ In the next video, I'll start talking about extending
451
+
452
+ 00:09:49.667 --> 00:09:52.180
453
+ these ideas beyond just two dimensions. See you then!
454
+
3B1B_LinearAlgebra/04 - Matrix multiplication as composition | Chapter 4, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a4c2ea0d2e3d5e6435ac3cdedcb09ee21d83a9e6e264577ca3cbeea3e35f6330
3
+ size 40565913
3B1B_LinearAlgebra/05 - Three-dimensional linear transformations | Chapter 5, Essence of linear algebra.en.vtt ADDED
@@ -0,0 +1,190 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:00.000 --> 00:00:13.460
6
+ [Submit subtitle corrections at criblate.com]
7
+
8
+ 00:00:13.460 --> 00:00:16.416
9
+ Hey folks, I've got a relatively quick video for you today,
10
+
11
+ 00:00:16.416 --> 00:00:18.520
12
+ just sort of a footnote between chapters.
13
+
14
+ 00:00:19.060 --> 00:00:22.561
15
+ In the last two videos I talked about linear transformations and matrices,
16
+
17
+ 00:00:22.561 --> 00:00:25.636
18
+ but I only showed the specific case of transformations that take
19
+
20
+ 00:00:25.636 --> 00:00:28.380
21
+ two-dimensional vectors to other two-dimensional vectors.
22
+
23
+ 00:00:28.920 --> 00:00:32.140
24
+ In general throughout the series we'll work mainly in two dimensions,
25
+
26
+ 00:00:32.140 --> 00:00:36.060
27
+ mostly because it's easier to actually see on the screen and wrap your mind around.
28
+
29
+ 00:00:36.500 --> 00:00:40.243
30
+ But more importantly than that, once you get all the core ideas in two dimensions,
31
+
32
+ 00:00:40.243 --> 00:00:42.800
33
+ they carry over pretty seamlessly to higher dimensions.
34
+
35
+ 00:00:43.800 --> 00:00:47.163
36
+ Nevertheless, it's good to peek our heads outside of flatland now and then to,
37
+
38
+ 00:00:47.163 --> 00:00:51.000
39
+ you know, see what it means to apply these ideas in more than just those two dimensions.
40
+
41
+ 00:00:52.340 --> 00:00:55.814
42
+ For example, consider a linear transformation with three-dimensional
43
+
44
+ 00:00:55.814 --> 00:00:58.880
45
+ vectors as inputs and three-dimensional vectors as outputs.
46
+
47
+ 00:01:00.160 --> 00:01:04.989
48
+ We can visualize this by smooshing around all the points in three-dimensional space,
49
+
50
+ 00:01:04.989 --> 00:01:09.301
51
+ as represented by a grid, in such a way that keeps the grid lines parallel
52
+
53
+ 00:01:09.301 --> 00:01:12.520
54
+ and evenly spaced, and which fixes the origin in place.
55
+
56
+ 00:01:13.460 --> 00:01:17.954
57
+ And just as with two dimensions, every point of space that we see moving around is
58
+
59
+ 00:01:17.954 --> 00:01:21.461
60
+ really just a proxy for a vector who has its tip at that point,
61
+
62
+ 00:01:21.461 --> 00:01:25.900
63
+ and what we're really doing is thinking about input vectors moving over to their
64
+
65
+ 00:01:25.900 --> 00:01:27.160
66
+ corresponding outputs.
67
+
68
+ 00:01:27.900 --> 00:01:30.876
69
+ And just as with two dimensions, one of these transformations
70
+
71
+ 00:01:30.876 --> 00:01:33.560
72
+ is completely described by where the basis vectors go.
73
+
74
+ 00:01:34.160 --> 00:01:37.680
75
+ But now, there are three standard basis vectors that we typically use.
76
+
77
+ 00:01:38.000 --> 00:01:42.451
78
+ The unit vector in the x direction, i-hat, the unit vector in the y direction,
79
+
80
+ 00:01:42.451 --> 00:01:46.560
81
+ j-hat, and a new guy, the unit vector in the z direction, called k-hat.
82
+
83
+ 00:01:47.140 --> 00:01:51.682
84
+ In fact, I think it's easier to think about these transformations by only following those
85
+
86
+ 00:01:51.682 --> 00:01:56.020
87
+ basis vectors, since the full 3D grid representing all points can get kind of messy.
88
+
89
+ 00:01:56.920 --> 00:01:59.762
90
+ By leaving a copy of the original axes in the background,
91
+
92
+ 00:01:59.762 --> 00:02:04.000
93
+ we can think about the coordinates of where each of these three basis vectors lands.
94
+
95
+ 00:02:05.820 --> 00:02:10.460
96
+ Record the coordinates of these three vectors as the columns of a 3x3 matrix.
97
+
98
+ 00:02:11.260 --> 00:02:16.160
99
+ This gives a matrix that completely describes the transformation using only nine numbers.
100
+
101
+ 00:02:17.200 --> 00:02:20.545
102
+ As a simple example, consider the transformation
103
+
104
+ 00:02:20.545 --> 00:02:23.960
105
+ that rotates space 90 degrees around the y-axis.
106
+
107
+ 00:02:24.860 --> 00:02:30.100
108
+ So that would mean that it takes i-hat to the coordinates (0, 0, -1) on the z-axis.
109
+
110
+ 00:02:30.820 --> 00:02:34.280
111
+ It doesn't move j-hat, so it stays at the coordinates (0, 1, 0).
112
+
113
+ 00:02:34.880 --> 00:02:38.840
114
+ And then k-hat moves over to the x-axis at (1, 0, 0).
115
+
116
+ 00:02:40.990 --> 00:02:43.848
117
+ Those three sets of coordinates become the columns of
118
+
119
+ 00:02:43.848 --> 00:02:46.760
120
+ a matrix that describes that rotation transformation.
121
+
122
+ 00:02:49.760 --> 00:02:52.625
123
+ To see where a vector with coordinates (x, y, z) lands,
124
+
125
+ 00:02:52.625 --> 00:02:56.220
126
+ the reasoning is almost identical to what it was for two dimensions.
127
+
128
+ 00:02:56.940 --> 00:03:00.490
129
+ Each of those coordinates can be thought of as instructions for how to
130
+
131
+ 00:03:00.490 --> 00:03:04.040
132
+ scale each basis vector so that they add together to get your vector.
133
+
134
+ 00:03:06.260 --> 00:03:08.917
135
+ And the important part, just like the 2D case,
136
+
137
+ 00:03:08.917 --> 00:03:14.000
138
+ is that this scaling and adding process works both before and after the transformation.
139
+
140
+ 00:03:16.380 --> 00:03:20.326
141
+ So to see where your vector lands, you multiply those coordinates by the
142
+
143
+ 00:03:20.326 --> 00:03:24.820
144
+ corresponding columns of the matrix, and then you add together the three results.
145
+
146
+ 00:03:29.600 --> 00:03:31.500
147
+ Multiplying two matrices is also similar.
148
+
149
+ 00:03:32.020 --> 00:03:35.522
150
+ Whenever you see two 3x3 matrices getting multiplied together,
151
+
152
+ 00:03:35.522 --> 00:03:39.984
153
+ you should imagine first applying the transformation encoded by the right one,
154
+
155
+ 00:03:39.984 --> 00:03:43.260
156
+ then applying the transformation encoded by the left one.
157
+
158
+ 00:03:44.060 --> 00:03:48.246
159
+ It turns out that 3D matrix multiplication is actually pretty important for fields
160
+
161
+ 00:03:48.246 --> 00:03:52.584
162
+ like computer graphics and robotics, since things like rotations in three dimensions
163
+
164
+ 00:03:52.584 --> 00:03:56.821
165
+ can be pretty hard to describe, but they're easier to wrap your mind around if you
166
+
167
+ 00:03:56.821 --> 00:04:01.160
168
+ can break them down as the composition of separate, easier-to-think-about rotations.
169
+
170
+ 00:04:04.360 --> 00:04:07.034
171
+ Performing this matrix multiplication numerically is,
172
+
173
+ 00:04:07.034 --> 00:04:09.860
174
+ once again, pretty similar to the two-dimensional case.
175
+
176
+ 00:04:10.480 --> 00:04:14.674
177
+ In fact, a good way to test your understanding of the last video would be to try to
178
+
179
+ 00:04:14.674 --> 00:04:18.615
180
+ reason through what specifically this matrix multiplication should look like,
181
+
182
+ 00:04:18.615 --> 00:04:22.506
183
+ thinking closely about how it relates to the idea of applying two successive
184
+
185
+ 00:04:22.506 --> 00:04:23.820
186
+ transformations in space.
187
+
188
+ 00:04:32.140 --> 00:04:34.500
189
+ In the next video, I'll start getting into the determinant.
190
+
3B1B_LinearAlgebra/05 - Three-dimensional linear transformations | Chapter 5, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4cec650b6e1fe9ebebbba802339150b1bb8e9d2f9c9832ac7e16e123d0799df2
3
+ size 38590230
3B1B_LinearAlgebra/06 - The determinant | Chapter 6, Essence of linear algebra.en.vtt ADDED
@@ -0,0 +1,454 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:00.000 --> 00:00:11.980
6
+ [Submit subtitle corrections at criblate.com]
7
+
8
+ 00:00:11.980 --> 00:00:13.000
9
+ Hello, hello again.
10
+
11
+ 00:00:13.520 --> 00:00:16.486
12
+ So moving forward, I'll be assuming that you have a visual understanding
13
+
14
+ 00:00:16.486 --> 00:00:19.328
15
+ of linear transformations and how they're represented with matrices,
16
+
17
+ 00:00:19.328 --> 00:00:21.840
18
+ the way that I've been talking about in the last few videos.
19
+
20
+ 00:00:22.660 --> 00:00:25.743
21
+ If you think about a couple of these linear transformations,
22
+
23
+ 00:00:25.743 --> 00:00:28.878
24
+ you might notice how some of them seem to stretch space out,
25
+
26
+ 00:00:28.878 --> 00:00:30.420
27
+ while others squish it on in.
28
+
29
+ 00:00:31.140 --> 00:00:34.905
30
+ One thing that turns out to be pretty useful for understanding one of these
31
+
32
+ 00:00:34.905 --> 00:00:38.920
33
+ transformations is to measure exactly how much it stretches or squishes things.
34
+
35
+ 00:00:39.520 --> 00:00:42.607
36
+ More specifically, to measure the factor by which
37
+
38
+ 00:00:42.607 --> 00:00:45.820
39
+ the area of a given region increases or decreases.
40
+
41
+ 00:00:47.180 --> 00:00:50.880
42
+ For example, look at the matrix with columns (3, 0) and (0, 2).
43
+
44
+ 00:00:51.320 --> 00:00:56.180
45
+ It scales i-hat by a factor of 3 and scales j-hat by a factor of 2.
46
+
47
+ 00:00:56.700 --> 00:01:02.141
48
+ Now, if we focus our attention on the 1 by 1 square whose bottom sits on i-hat and whose
49
+
50
+ 00:01:02.141 --> 00:01:07.520
51
+ left side sits on j-hat, after the transformation, this turns into a 2 by 3 rectangle.
52
+
53
+ 00:01:08.380 --> 00:01:12.579
54
+ Since this region started out with area 1 and ended up with area 6,
55
+
56
+ 00:01:12.579 --> 00:01:17.280
57
+ we can say the linear transformation has scaled its area by a factor of 6.
58
+
59
+ 00:01:18.180 --> 00:01:22.109
60
+ Compare that to a shear, whose matrix has columns (1, 0) and (1,
61
+
62
+ 00:01:22.109 --> 00:01:26.100
63
+ 1), meaning i-hat stays in place and j-hat moves over to (1, 1).
64
+
65
+ 00:01:27.000 --> 00:01:30.831
66
+ That same unit square determined by i-hat and j-hat gets slanted and
67
+
68
+ 00:01:30.831 --> 00:01:35.112
69
+ turned into a parallelogram, but the area of that parallelogram is still 1,
70
+
71
+ 00:01:35.112 --> 00:01:38.380
72
+ since its base and height each continue to have length 1.
73
+
74
+ 00:01:39.180 --> 00:01:41.851
75
+ So even though this transformation smushes things about,
76
+
77
+ 00:01:41.851 --> 00:01:45.620
78
+ it seems to leave areas unchanged, at least in the case of that 1 unit square.
79
+
80
+ 00:01:46.820 --> 00:01:51.616
81
+ Actually though, if you know how much the area of that one single unit square changes,
82
+
83
+ 00:01:51.616 --> 00:01:55.520
84
+ it can tell you how the area of any possible region in space changes.
85
+
86
+ 00:01:56.300 --> 00:01:59.967
87
+ For starters, notice that whatever happens to one square in the grid
88
+
89
+ 00:01:59.967 --> 00:02:03.580
90
+ has to happen to any other square in the grid, no matter the size.
91
+
92
+ 00:02:04.340 --> 00:02:08.040
93
+ This follows from the fact that grid lines remain parallel and evenly spaced.
94
+
95
+ 00:02:08.760 --> 00:02:13.514
96
+ Then, any shape that's not a grid square can be approximated by grid squares pretty well,
97
+
98
+ 00:02:13.514 --> 00:02:17.520
99
+ with arbitrarily good approximations if you use small enough grid squares.
100
+
101
+ 00:02:17.520 --> 00:02:22.488
102
+ So, since the areas of all those tiny grid squares are being scaled by some single
103
+
104
+ 00:02:22.488 --> 00:02:27.820
105
+ amount, the area of the blob as a whole will also be scaled by that same single amount.
106
+
107
+ 00:02:28.900 --> 00:02:33.265
108
+ This very special scaling factor, the factor by which a linear transformation
109
+
110
+ 00:02:33.265 --> 00:02:37.120
111
+ changes any area, is called the determinant of that transformation.
112
+
113
+ 00:02:39.120 --> 00:02:42.284
114
+ I'll show how to compute the determinant of a transformation using
115
+
116
+ 00:02:42.284 --> 00:02:45.927
117
+ its matrix later on in this video, but understanding what it represents is,
118
+
119
+ 00:02:45.927 --> 00:02:48.420
120
+ trust me, much more important than the computation.
121
+
122
+ 00:02:49.580 --> 00:02:53.395
123
+ For example, the determinant of a transformation would be 3 if that
124
+
125
+ 00:02:53.395 --> 00:02:57.040
126
+ transformation increases the area of a region by a factor of 3.
127
+ 41
128
+ 00:02:58,180 --> 00:02:57,622
129
+ The determinant of a transformation would be ½
130
+ 42
131
+ 00:02:57,622 --> 00:02:57,040
132
+ if it squishes down all areas by a factor of ½.
133
+
134
+ 00:02:58.180 --> 00:03:03.356
135
+ The determinant of a transformation would be 1 half if it squishes down all
136
+
137
+ 00:03:03.356 --> 00:03:08.531
138
+ areas by a factor of 1 half. And the determinant of a 2D transformation is
139
+
140
+ 00:03:08.531 --> 00:03:13.500
141
+ 0 if it squishes all of space onto a line, or even onto a single point.
142
+
143
+ 00:03:14.000 --> 00:03:16.760
144
+ Since then, the area of any region would become zero.
145
+
146
+ 00:03:17.620 --> 00:03:19.600
147
+ That last example will prove to be pretty important.
148
+
149
+ 00:03:20.020 --> 00:03:23.212
150
+ It means that checking if the determinant of a given matrix is zero
151
+
152
+ 00:03:23.212 --> 00:03:26.214
153
+ will give a way of computing whether or not the transformation
154
+
155
+ 00:03:26.214 --> 00:03:29.740
156
+ associated with that matrix squishes everything into a smaller dimension.
157
+
158
+ 00:03:30.520 --> 00:03:34.233
159
+ You'll see in the next few videos why this is even a useful thing to think about,
160
+
161
+ 00:03:34.233 --> 00:03:37.258
162
+ but for now, I just want to lay down all of the visual intuition,
163
+
164
+ 00:03:37.258 --> 00:03:40.100
165
+ which, in and of itself, is a beautiful thing to think about.
166
+
167
+ 00:03:42.120 --> 00:03:45.560
168
+ Okay, I need to confess that what I've said so far is not quite right.
169
+
170
+ 00:03:45.880 --> 00:03:49.280
171
+ The full concept of the determinant allows for negative values.
172
+
173
+ 00:03:49.720 --> 00:03:53.480
174
+ But what would the idea of scaling an area by a negative amount even mean?
175
+
176
+ 00:03:54.940 --> 00:03:56.960
177
+ This has to do with the idea of orientation.
178
+
179
+ 00:03:57.800 --> 00:04:02.680
180
+ For example, notice how this transformation gives the sensation of flipping space over.
181
+
182
+ 00:04:03.240 --> 00:04:05.839
183
+ If you were thinking of 2D space as a sheet of paper,
184
+
185
+ 00:04:05.839 --> 00:04:09.860
186
+ a transformation like that one seems to turn over that sheet onto the other side.
187
+
188
+ 00:04:10.640 --> 00:04:15.040
189
+ Any transformations that do this are said to invert the orientation of space.
190
+
191
+ 00:04:15.840 --> 00:04:18.600
192
+ Another way to think about it is in terms of i-hat and j-hat.
193
+
194
+ 00:04:19.160 --> 00:04:23.060
195
+ Notice that in their starting positions, j-hat is to the left of i-hat.
196
+
197
+ 00:04:23.620 --> 00:04:27.494
198
+ If, after a transformation, j-hat is now on the right of i-hat,
199
+
200
+ 00:04:27.494 --> 00:04:30.200
201
+ the orientation of space has been inverted.
202
+
203
+ 00:04:32.120 --> 00:04:35.108
204
+ Whenever this happens, whenever the orientation of space is inverted,
205
+
206
+ 00:04:35.108 --> 00:04:36.580
207
+ the determinant will be negative.
208
+
209
+ 00:04:37.460 --> 00:04:39.604
210
+ The absolute value of the determinant, though,
211
+
212
+ 00:04:39.604 --> 00:04:42.400
213
+ still tells you the factor by which areas have been scaled.
214
+
215
+ 00:04:43.020 --> 00:04:45.935
216
+ For example, the matrix with columns (1, 1) and (2,
217
+
218
+ 00:04:45.935 --> 00:04:50.680
219
+ -1) encodes a transformation that has determinant, I'll just tell you, negative 3.
220
+
221
+ 00:04:51.460 --> 00:04:56.280
222
+ And what this means is that space gets flipped over and areas are scaled by a factor of 3.
223
+
224
+ 00:04:57.780 --> 00:05:00.878
225
+ So why would this idea of a negative area scaling factor
226
+
227
+ 00:05:00.878 --> 00:05:03.700
228
+ be a natural way to describe orientation flipping?
229
+
230
+ 00:05:04.260 --> 00:05:07.172
231
+ Think about the series of transformations you get by
232
+
233
+ 00:05:07.172 --> 00:05:10.140
234
+ slowly letting i-hat get closer and closer to j-hat.
235
+
236
+ 00:05:10.720 --> 00:05:15.096
237
+ As i-hat gets closer, all of the areas in space are getting squished more and more,
238
+
239
+ 00:05:15.096 --> 00:05:17.100
240
+ meaning the determinant approaches 0.
241
+
242
+ 00:05:17.820 --> 00:05:21.640
243
+ Once i-hat lines up perfectly with j-hat, the determinant is 0.
244
+
245
+ 00:05:22.440 --> 00:05:24.781
246
+ Then, if i-hat continues the way that it was going,
247
+
248
+ 00:05:24.781 --> 00:05:28.270
249
+ doesn't it kind of feel natural for the determinant to keep decreasing into
250
+
251
+ 00:05:28.270 --> 00:05:29.280
252
+ the negative numbers?
253
+
254
+ 00:05:30.680 --> 00:05:33.560
255
+ So that's the understanding of determinants in two dimensions.
256
+
257
+ 00:05:33.560 --> 00:05:35.940
258
+ What do you think it should mean for three dimensions?
259
+
260
+ 00:05:36.920 --> 00:05:40.107
261
+ It also tells you how much a transformation scales things,
262
+
263
+ 00:05:40.107 --> 00:05:43.240
264
+ but this time, it tells you how much volumes get scaled.
265
+
266
+ 00:05:45.340 --> 00:05:49.894
267
+ Just as in two dimensions, where this is easiest to think about by focusing on
268
+
269
+ 00:05:49.894 --> 00:05:54.273
270
+ one particular square with an area 1 and watching only what happens to it,
271
+
272
+ 00:05:54.273 --> 00:05:58.769
273
+ in three dimensions, it helps to focus your attention on the specific 1 by 1
274
+
275
+ 00:05:58.769 --> 00:06:03.440
276
+ by 1 cube whose edges are resting on the basis vectors, i-hat, j-hat and k-hat.
277
+
278
+ 00:06:04.320 --> 00:06:09.300
279
+ After the transformation, that cube might get warped into some kind of slanty slanty cube.
280
+
281
+ 00:06:10.340 --> 00:06:13.207
282
+ This shape, by the way, has the best name ever, parallelipiped,
283
+
284
+ 00:06:13.207 --> 00:06:16.712
285
+ a name that's made even more delightful when your professor has a nice thick
286
+
287
+ 00:06:16.712 --> 00:06:17.440
288
+ Russian accent.
289
+
290
+ 00:06:18.520 --> 00:06:21.020
291
+ Since this cube starts out with a volume of 1,
292
+
293
+ 00:06:21.020 --> 00:06:24.716
294
+ and the determinant gives the factor by which any volume is scaled,
295
+
296
+ 00:06:24.716 --> 00:06:28.412
297
+ you can think of the determinant simply as being the volume of that
298
+
299
+ 00:06:28.412 --> 00:06:30.640
300
+ parallelipiped that the cube turns into.
301
+
302
+ 00:06:32.380 --> 00:06:37.469
303
+ A determinant of 0 would mean that all of space is squished onto something with 0 volume,
304
+
305
+ 00:06:37.469 --> 00:06:42.500
306
+ meaning either a flat plane, a line, or, in the most extreme case, onto a single point.
307
+
308
+ 00:06:43.760 --> 00:06:46.363
309
+ Those of you who watched chapter 2 will recognize this as
310
+
311
+ 00:06:46.363 --> 00:06:49.240
312
+ meaning that the columns of the matrix are linearly dependent.
313
+
314
+ 00:06:49.760 --> 00:06:50.420
315
+ Can you see why?
316
+
317
+ 00:06:54.920 --> 00:06:56.640
318
+ What about negative determinants?
319
+
320
+ 00:06:56.780 --> 00:06:58.100
321
+ What should that mean for three dimensions?
322
+
323
+ 00:06:58.780 --> 00:07:02.680
324
+ One way to describe orientation in 3D is with the right hand rule.
325
+
326
+ 00:07:03.300 --> 00:07:06.437
327
+ Point the forefinger of your right hand in the direction of i-hat,
328
+
329
+ 00:07:06.437 --> 00:07:09.100
330
+ stick out your middle finger in the direction of j-hat,
331
+
332
+ 00:07:09.100 --> 00:07:12.760
333
+ and notice how when you point your thumb up, it's in the direction of k-hat.
334
+
335
+ 00:07:14.880 --> 00:07:17.567
336
+ If you can still do that after the transformation,
337
+
338
+ 00:07:17.567 --> 00:07:20.900
339
+ orientation has not changed, and the determinant is positive.
340
+
341
+ 00:07:21.540 --> 00:07:25.332
342
+ Otherwise, if after the transformation it only makes sense to do that with
343
+
344
+ 00:07:25.332 --> 00:07:29.380
345
+ your left hand, orientation has been flipped, and the determinant is negative.
346
+
347
+ 00:07:31.900 --> 00:07:34.993
348
+ So, if you haven't seen it before, you're probably wondering by now,
349
+
350
+ 00:07:34.993 --> 00:07:37.040
351
+ how do you actually compute the determinant?
352
+
353
+ 00:07:37.560 --> 00:07:44.420
354
+ For a 2x2 matrix with entries a, b, c, d, the formula is a times d minus b times c.
355
+
356
+ 00:07:45.740 --> 00:07:48.500
357
+ Here's part of an intuition for where this formula comes from.
358
+
359
+ 00:07:48.880 --> 00:07:51.780
360
+ Let's say that the terms b and c both happened to be 0.
361
+
362
+ 00:07:51.780 --> 00:07:56.502
363
+ Then, the term a tells you how much i-hat is stretched in the x direction,
364
+
365
+ 00:07:56.502 --> 00:08:01.160
366
+ and the term d tells you how much j-hat is stretched in the y direction.
367
+
368
+ 00:08:02.760 --> 00:08:06.148
369
+ So, since those other terms are 0, it should make sense that a
370
+
371
+ 00:08:06.148 --> 00:08:10.628
372
+ times d gives the area of the rectangle that our favorite unit square turns into,
373
+
374
+ 00:08:10.628 --> 00:08:13.360
375
+ kind of like the 3, 0, 0, 2 example from earlier.
376
+
377
+ 00:08:15.360 --> 00:08:18.525
378
+ Even if only one of b or c are 0, you'll have
379
+
380
+ 00:08:18.525 --> 00:08:21.760
381
+ a parallelogram with a base a and a height d.
382
+
383
+ 00:08:21.780 --> 00:08:24.500
384
+ So, the area should still be a times d.
385
+
386
+ 00:08:25.460 --> 00:08:28.361
387
+ Loosely speaking, if both b and c are non-zero,
388
+
389
+ 00:08:28.361 --> 00:08:33.238
390
+ then that b times c term tells you how much this parallelogram is stretched or
391
+
392
+ 00:08:33.238 --> 00:08:35.460
393
+ squished in the diagonal direction.
394
+
395
+ 00:08:36.660 --> 00:08:40.176
396
+ For those of you hungry for a more precise description of this b times c term,
397
+
398
+ 00:08:40.176 --> 00:08:42.880
399
+ here's a helpful diagram if you'd like to pause and ponder.
400
+
401
+ 00:08:43.980 --> 00:08:48.188
402
+ Now, if you feel like computing determinants by hand is something that you need to know,
403
+
404
+ 00:08:48.188 --> 00:08:51.200
405
+ the only way to get it down is to just practice it with a few.
406
+
407
+ 00:08:51.200 --> 00:08:53.343
408
+ There's really not that much I can say or animate
409
+
410
+ 00:08:53.343 --> 00:08:55.180
411
+ that's going to drill in the computation.
412
+
413
+ 00:08:56.120 --> 00:08:58.640
414
+ This is all triply true for three-dimensional determinants.
415
+
416
+ 00:08:59.040 --> 00:09:02.299
417
+ There is a formula, and if you feel like that's something you need to know,
418
+
419
+ 00:09:02.299 --> 00:09:04.689
420
+ you should practice with a few matrices, or, you know,
421
+
422
+ 00:09:04.689 --> 00:09:06.340
423
+ go watch Sal Khan work through a few.
424
+
425
+ 00:09:07.240 --> 00:09:10.344
426
+ Honestly, though, I don't think that those computations fall within
427
+
428
+ 00:09:10.344 --> 00:09:13.078
429
+ the essence of linear algebra, but I definitely think that
430
+
431
+ 00:09:13.078 --> 00:09:16.460
432
+ understanding what the determinant represents falls within that essence.
433
+
434
+ 00:09:18.060 --> 00:09:20.640
435
+ Here's kind of a fun question to think about before the next video.
436
+
437
+ 00:09:20.640 --> 00:09:25.390
438
+ If you multiply two matrices together, the determinant of the resulting matrix
439
+
440
+ 00:09:25.390 --> 00:09:30.080
441
+ is the same as the product of the determinants of the original two matrices.
442
+
443
+ 00:09:31.100 --> 00:09:34.629
444
+ If you tried to justify this with numbers, it would take a really long time,
445
+
446
+ 00:09:34.629 --> 00:09:37.880
447
+ but see if you can explain why this makes sense in just one sentence.
448
+
449
+ 00:09:42.000 --> 00:09:46.663
450
+ Next up, I'll be relating the idea of linear transformations covered so far to one of
451
+
452
+ 00:09:46.663 --> 00:09:51.600
453
+ the areas where linear algebra is most useful, linear systems of equations. See you then!
454
+
3B1B_LinearAlgebra/06 - The determinant | Chapter 6, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:19cb359ce0ee1b724a0c3367bdcfc71b43524d8284d55a873b45547e7650d987
3
+ size 59458776
3B1B_LinearAlgebra/07 - Inverse matrices, column space and null space | Chapter 7, Essence of linear algebra.en.vtt ADDED
@@ -0,0 +1,532 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:10.240 --> 00:00:14.958
6
+ As you can probably tell by now, the bulk of this series is on understanding matrix
7
+
8
+ 00:00:14.958 --> 00:00:19.340
9
+ and vector operations through that more visual lens of linear transformations.
10
+
11
+ 00:00:19.980 --> 00:00:24.348
12
+ This video is no exception, describing the concepts of inverse matrices,
13
+
14
+ 00:00:24.348 --> 00:00:27.520
15
+ column space, rank, and null space through that lens.
16
+
17
+ 00:00:27.520 --> 00:00:30.902
18
+ A forewarning though, I'm not going to talk about the methods for actually
19
+
20
+ 00:00:30.902 --> 00:00:34.240
21
+ computing these things, and some would argue that that's pretty important.
22
+
23
+ 00:00:34.840 --> 00:00:39.353
24
+ There are a lot of very good resources for learning those methods outside this series,
25
+
26
+ 00:00:39.353 --> 00:00:42.000
27
+ keywords Gaussian elimination and row echelon form.
28
+
29
+ 00:00:42.540 --> 00:00:46.340
30
+ I think most of the value that I actually have to add here is on the intuition half.
31
+
32
+ 00:00:46.900 --> 00:00:50.480
33
+ Plus, in practice, we usually get software to compute this stuff for us anyway.
34
+
35
+ 00:00:51.500 --> 00:00:53.920
36
+ First, a few words on the usefulness of linear algebra.
37
+
38
+ 00:00:54.300 --> 00:00:57.845
39
+ By now, you already have a hint for how it's used in describing the manipulation
40
+
41
+ 00:00:57.845 --> 00:01:01.040
42
+ of space, which is useful for things like computer graphics and robotics.
43
+
44
+ 00:01:01.500 --> 00:01:04.256
45
+ But one of the main reasons that linear algebra is more
46
+
47
+ 00:01:04.256 --> 00:01:07.259
48
+ broadly applicable and required for just about any technical
49
+
50
+ 00:01:07.259 --> 00:01:10.460
51
+ discipline is that it lets us solve certain systems of equations.
52
+
53
+ 00:01:11.380 --> 00:01:14.766
54
+ When I say system of equations, I mean you have a list of variables,
55
+
56
+ 00:01:14.766 --> 00:01:17.760
57
+ things you don't know, and a list of equations relating them.
58
+
59
+ 00:01:18.340 --> 00:01:21.600
60
+ In a lot of situations, those equations can get very complicated.
61
+
62
+ 00:01:22.120 --> 00:01:25.300
63
+ But, if you're lucky, they might take on a certain special form.
64
+
65
+ 00:01:26.440 --> 00:01:29.920
66
+ Within each equation, the only thing happening to each variable is
67
+
68
+ 00:01:29.920 --> 00:01:33.399
69
+ that it's scaled by some constant, and the only thing happening to
70
+
71
+ 00:01:33.399 --> 00:01:36.880
72
+ each of those scaled variables is that they're added to each other.
73
+
74
+ 00:01:37.540 --> 00:01:41.394
75
+ So no exponents or fancy functions or multiplying two variables together,
76
+
77
+ 00:01:41.394 --> 00:01:42.280
78
+ things like that.
79
+
80
+ 00:01:43.420 --> 00:01:47.612
81
+ The typical way to organize this sort of special system of equations is to
82
+
83
+ 00:01:47.612 --> 00:01:52.140
84
+ throw all the variables on the left and put any lingering constants on the right.
85
+
86
+ 00:01:53.600 --> 00:01:56.231
87
+ It's also nice to vertically line up the common variables,
88
+
89
+ 00:01:56.231 --> 00:01:59.174
90
+ and to do that, you might need to throw in some zero coefficients
91
+
92
+ 00:01:59.174 --> 00:02:01.940
93
+ whenever the variable doesn't show up in one of the equations.
94
+
95
+ 00:02:04.540 --> 00:02:07.240
96
+ This is called a linear system of equations.
97
+
98
+ 00:02:08.100 --> 00:02:11.180
99
+ You might notice that this looks a lot like matrix-vector multiplication.
100
+
101
+ 00:02:11.820 --> 00:02:15.629
102
+ In fact, you can package all of the equations together into a single
103
+
104
+ 00:02:15.629 --> 00:02:19.161
105
+ vector equation where you have the matrix containing all of the
106
+
107
+ 00:02:19.161 --> 00:02:22.915
108
+ constant coefficients and a vector containing all of the variables,
109
+
110
+ 00:02:22.915 --> 00:02:26.780
111
+ and their matrix-vector product equals some different constant vector.
112
+
113
+ 00:02:28.640 --> 00:02:33.060
114
+ Let's name that constant matrix A, denote the vector holding the variables
115
+
116
+ 00:02:33.060 --> 00:02:37.480
117
+ with a bold-faced X, and call the constant vector on the right-hand side V.
118
+
119
+ 00:02:38.860 --> 00:02:41.030
120
+ This is more than just a notational trick to get
121
+
122
+ 00:02:41.030 --> 00:02:42.980
123
+ our system of equations written on one line.
124
+
125
+ 00:02:43.340 --> 00:02:46.780
126
+ It sheds light on a pretty cool geometric interpretation for the problem.
127
+
128
+ 00:02:47.620 --> 00:02:51.134
129
+ The matrix A corresponds with some linear transformation,
130
+
131
+ 00:02:51.134 --> 00:02:55.132
132
+ so solving Ax equals V means we're looking for a vector X, which,
133
+
134
+ 00:02:55.132 --> 00:02:57.920
135
+ after applying the transformation, lands on V.
136
+
137
+ 00:02:59.940 --> 00:03:01.780
138
+ Think about what's happening here for a moment.
139
+
140
+ 00:03:02.060 --> 00:03:05.299
141
+ You can hold in your head this really complicated idea of multiple
142
+
143
+ 00:03:05.299 --> 00:03:08.538
144
+ variables all intermingling with each other just by thinking about
145
+
146
+ 00:03:08.538 --> 00:03:12.600
147
+ squishing and morphing space and trying to figure out which vector lands on another.
148
+
149
+ 00:03:13.160 --> 00:03:13.760
150
+ Cool, right?
151
+
152
+ 00:03:14.600 --> 00:03:18.680
153
+ To start simple, let's say you have a system with two equations and two unknowns.
154
+
155
+ 00:03:19.000 --> 00:03:23.960
156
+ This means the matrix A is a 2x2 matrix, and V and X are each two-dimensional vectors.
157
+
158
+ 00:03:25.600 --> 00:03:29.620
159
+ Now, how we think about the solutions to this equation depends on whether the
160
+
161
+ 00:03:29.620 --> 00:03:33.693
162
+ transformation associated with A squishes all of space into a lower dimension,
163
+
164
+ 00:03:33.693 --> 00:03:38.023
165
+ like a line or a point, or if it leaves everything spanning the full two dimensions
166
+
167
+ 00:03:38.023 --> 00:03:38.900
168
+ where it started.
169
+
170
+ 00:03:40.320 --> 00:03:44.180
171
+ In the language of the last video, we subdivide into the cases where
172
+
173
+ 00:03:44.180 --> 00:03:48.040
174
+ A has zero determinant and the case where A has non-zero determinant.
175
+
176
+ 00:03:51.300 --> 00:03:54.845
177
+ Let's start with the most likely case, where the determinant is non-zero,
178
+
179
+ 00:03:54.845 --> 00:03:57.720
180
+ meaning space does not get squished into a zero-area region.
181
+
182
+ 00:03:58.600 --> 00:04:02.793
183
+ In this case, there will always be one and only one vector that lands on V,
184
+
185
+ 00:04:02.793 --> 00:04:06.160
186
+ and you can find it by playing the transformation in reverse.
187
+
188
+ 00:04:06.700 --> 00:04:10.141
189
+ Following where V goes as we rewind the tape like this,
190
+
191
+ 00:04:10.141 --> 00:04:13.460
192
+ you'll find the vector x such that A times x equals V.
193
+
194
+ 00:04:15.400 --> 00:04:19.930
195
+ When you play the transformation in reverse, it actually corresponds to a separate
196
+
197
+ 00:04:19.930 --> 00:04:24.680
198
+ linear transformation, commonly called the inverse of A, denoted A to the negative one.
199
+
200
+ 00:04:25.360 --> 00:04:29.031
201
+ For example, if A was a counterclockwise rotation by 90 degrees,
202
+
203
+ 00:04:29.031 --> 00:04:32.760
204
+ then the inverse of A would be a clockwise rotation by 90 degrees.
205
+
206
+ 00:04:34.320 --> 00:04:38.019
207
+ If A was a rightward shear that pushes j-hat one unit to the right,
208
+
209
+ 00:04:38.019 --> 00:04:42.480
210
+ the inverse of A would be a leftward shear that pushes j-hat one unit to the left.
211
+
212
+ 00:04:44.100 --> 00:04:48.633
213
+ In general, A inverse is the unique transformation with the property that if you first
214
+
215
+ 00:04:48.633 --> 00:04:51.708
216
+ apply A, then follow it with the transformation A inverse,
217
+
218
+ 00:04:51.708 --> 00:04:53.480
219
+ you end up back where you started.
220
+
221
+ 00:04:54.540 --> 00:04:58.638
222
+ Applying one transformation after another is captured algebraically with
223
+
224
+ 00:04:58.638 --> 00:05:02.961
225
+ matrix multiplication, so the core property of this transformation A inverse
226
+
227
+ 00:05:02.961 --> 00:05:07.340
228
+ is that A inverse times A equals the matrix that corresponds to doing nothing.
229
+
230
+ 00:05:08.200 --> 00:05:11.320
231
+ The transformation that does nothing is called the identity transformation.
232
+
233
+ 00:05:11.780 --> 00:05:18.080
234
+ It leaves i-hat and j-hat each where they are, unmoved, so its columns are 1,0 and 0,1.
235
+
236
+ 00:05:19.980 --> 00:05:23.960
237
+ Once you find this inverse, which in practice you'd do with a computer,
238
+
239
+ 00:05:23.960 --> 00:05:27.720
240
+ you can solve your equation by multiplying this inverse matrix by v.
241
+
242
+ 00:05:29.960 --> 00:05:33.258
243
+ And again, what this means geometrically is that you're
244
+
245
+ 00:05:33.258 --> 00:05:36.440
246
+ playing the transformation in reverse and following v.
247
+
248
+ 00:05:40.200 --> 00:05:44.316
249
+ This non-zero determinant case, which for a random choice of matrix is by far the
250
+
251
+ 00:05:44.316 --> 00:05:48.383
252
+ most likely one, corresponds with the idea that if you have two unknowns and two
253
+
254
+ 00:05:48.383 --> 00:05:52.400
255
+ equations, it's almost certainly the case that there's a single unique solution.
256
+
257
+ 00:05:53.680 --> 00:05:56.184
258
+ This idea also makes sense in higher dimensions,
259
+
260
+ 00:05:56.184 --> 00:05:59.200
261
+ when the number of equations equals the number of unknowns.
262
+
263
+ 00:05:59.380 --> 00:06:04.045
264
+ Again, the system of equations can be translated to the geometric
265
+
266
+ 00:06:04.045 --> 00:06:09.064
267
+ interpretation where you have some transformation A and some vector v,
268
+
269
+ 00:06:09.064 --> 00:06:12.740
270
+ and you're looking for the vector x that lands on v.
271
+
272
+ 00:06:15.740 --> 00:06:20.535
273
+ As long as the transformation A doesn't squish all of space into a lower dimension,
274
+
275
+ 00:06:20.535 --> 00:06:25.559
276
+ meaning its determinant is non-zero, there will be an inverse transformation A inverse,
277
+
278
+ 00:06:25.559 --> 00:06:29.270
279
+ with the property that if you first do A, then you do A inverse,
280
+
281
+ 00:06:29.270 --> 00:06:31.040
282
+ it's the same as doing nothing.
283
+
284
+ 00:06:33.540 --> 00:06:36.574
285
+ And to solve your equation, you just have to multiply
286
+
287
+ 00:06:36.574 --> 00:06:39.440
288
+ that reverse transformation matrix by the vector v.
289
+
290
+ 00:06:43.500 --> 00:06:47.671
291
+ But when the determinant is zero, and the transformation associated with the
292
+
293
+ 00:06:47.671 --> 00:06:52.060
294
+ system of equations squishes space into a smaller dimension, there is no inverse.
295
+
296
+ 00:06:52.480 --> 00:06:55.460
297
+ You cannot unsquish a line to turn it into a plane.
298
+
299
+ 00:06:55.980 --> 00:06:58.060
300
+ At least that's not something that a function can do.
301
+
302
+ 00:06:58.360 --> 00:07:02.980
303
+ That would require transforming each individual vector into a whole line full of vectors.
304
+
305
+ 00:07:03.740 --> 00:07:06.740
306
+ But functions can only take a single input to a single output.
307
+
308
+ 00:07:08.400 --> 00:07:11.208
309
+ Similarly, for three equations and three unknowns,
310
+
311
+ 00:07:11.208 --> 00:07:14.568
312
+ there will be no inverse if the corresponding transformation
313
+
314
+ 00:07:14.568 --> 00:07:19.140
315
+ squishes 3D space onto the plane, or even if it squishes it onto a line or a point.
316
+
317
+ 00:07:19.920 --> 00:07:22.200
318
+ Those all correspond to a determinant of zero,
319
+
320
+ 00:07:22.200 --> 00:07:25.160
321
+ since any region is squished into something with zero volume.
322
+
323
+ 00:07:26.700 --> 00:07:30.640
324
+ It's still possible that a solution exists even when there is no inverse.
325
+
326
+ 00:07:30.720 --> 00:07:35.020
327
+ It's just that when your transformation squishes space onto, say, a line,
328
+
329
+ 00:07:35.020 --> 00:07:39.380
330
+ you have to be lucky enough that the vector v lives somewhere on that line.
331
+
332
+ 00:07:43.300 --> 00:07:45.976
333
+ You might notice that some of these zero determinant
334
+
335
+ 00:07:45.976 --> 00:07:48.300
336
+ cases feel a lot more restrictive than others.
337
+
338
+ 00:07:48.840 --> 00:07:52.621
339
+ Given a 3x3 matrix, for example, it seems a lot harder for a solution
340
+
341
+ 00:07:52.621 --> 00:07:56.566
342
+ to exist when it squishes space onto a line compared to when it squishes
343
+
344
+ 00:07:56.566 --> 00:08:00.240
345
+ things onto a plane, even though both of those are zero determinant.
346
+
347
+ 00:08:02.600 --> 00:08:06.100
348
+ We have some language that's a bit more specific than just saying zero determinant.
349
+
350
+ 00:08:06.520 --> 00:08:09.231
351
+ When the output of a transformation is a line,
352
+
353
+ 00:08:09.231 --> 00:08:13.500
354
+ meaning it's one-dimensional, we say the transformation has a rank of one.
355
+
356
+ 00:08:15.140 --> 00:08:18.073
357
+ If all the vectors land on some two-dimensional plane,
358
+
359
+ 00:08:18.073 --> 00:08:20.420
360
+ we say the transformation has a rank of two.
361
+
362
+ 00:08:22.920 --> 00:08:27.480
363
+ So the word rank means the number of dimensions in the output of a transformation.
364
+
365
+ 00:08:28.400 --> 00:08:32.720
366
+ For instance, in the case of 2x2 matrices, rank two is the best that it can be.
367
+
368
+ 00:08:33.080 --> 00:08:37.292
369
+ It means the basis vectors continue to span the full two dimensions of space,
370
+
371
+ 00:08:37.292 --> 00:08:39.020
372
+ and the determinant is not zero.
373
+
374
+ 00:08:39.419 --> 00:08:42.639
375
+ But for 3x3 matrices, rank two means that we've collapsed,
376
+
377
+ 00:08:42.639 --> 00:08:46.460
378
+ but not as much as they would have collapsed for a rank one situation.
379
+
380
+ 00:08:47.240 --> 00:08:52.076
381
+ If a 3D transformation has a non-zero determinant and its output fills all of 3D space,
382
+
383
+ 00:08:52.076 --> 00:08:53.340
384
+ it has a rank of three.
385
+
386
+ 00:08:54.520 --> 00:08:58.620
387
+ This set of all possible outputs for your matrix, whether it's a line,
388
+
389
+ 00:08:58.620 --> 00:09:02.720
390
+ a plane, 3D space, whatever, is called the column space of your matrix.
391
+
392
+ 00:09:04.140 --> 00:09:06.280
393
+ You can probably guess where that name comes from.
394
+
395
+ 00:09:06.560 --> 00:09:10.784
396
+ The columns of your matrix tell you where the basis vectors land,
397
+
398
+ 00:09:10.784 --> 00:09:15.840
399
+ and the span of those transformed basis vectors gives you all possible outputs.
400
+
401
+ 00:09:16.360 --> 00:09:21.140
402
+ In other words, the column space is the span of the columns of your matrix.
403
+
404
+ 00:09:23.300 --> 00:09:26.147
405
+ So a more precise definition of rank would be that
406
+
407
+ 00:09:26.147 --> 00:09:28.940
408
+ it's the number of dimensions in the column space.
409
+
410
+ 00:09:29.940 --> 00:09:34.490
411
+ When this rank is as high as it can be, meaning it equals the number of columns,
412
+
413
+ 00:09:34.490 --> 00:09:36.120
414
+ we call the matrix full rank.
415
+
416
+ 00:09:38.540 --> 00:09:42.272
417
+ Notice the zero vector will always be included in the column space,
418
+
419
+ 00:09:42.272 --> 00:09:45.840
420
+ since linear transformations must keep the origin fixed in place.
421
+
422
+ 00:09:46.900 --> 00:09:49.353
423
+ For a full rank transformation, the only vector
424
+
425
+ 00:09:49.353 --> 00:09:51.960
426
+ that lands at the origin is the zero vector itself.
427
+
428
+ 00:09:52.460 --> 00:09:56.107
429
+ But for matrices that aren't full rank, which squish to a smaller dimension,
430
+
431
+ 00:09:56.107 --> 00:09:58.760
432
+ you can have a whole bunch of vectors that land on zero.
433
+
434
+ 00:10:01.640 --> 00:10:05.128
435
+ If a 2D transformation squishes space onto a line, for example,
436
+
437
+ 00:10:05.128 --> 00:10:09.707
438
+ there is a separate line in a different direction full of vectors that get squished
439
+
440
+ 00:10:09.707 --> 00:10:10.580
441
+ onto the origin.
442
+
443
+ 00:10:11.780 --> 00:10:14.491
444
+ If a 3D transformation squishes space onto a plane,
445
+
446
+ 00:10:14.491 --> 00:10:17.620
447
+ there's also a full line of vectors that land on the origin.
448
+
449
+ 00:10:20.520 --> 00:10:23.740
450
+ If a 3D transformation squishes all of space onto a line,
451
+
452
+ 00:10:23.740 --> 00:10:27.460
453
+ then there's a whole plane full of vectors that land on the origin.
454
+
455
+ 00:10:32.800 --> 00:10:37.045
456
+ This set of vectors that lands on the origin is called the null space,
457
+
458
+ 00:10:37.045 --> 00:10:38.780
459
+ or the kernel of your matrix.
460
+
461
+ 00:10:39.360 --> 00:10:41.795
462
+ It's the space of all vectors that become null,
463
+
464
+ 00:10:41.795 --> 00:10:44.180
465
+ in the sense that they land on the zero vector.
466
+
467
+ 00:10:45.680 --> 00:10:49.946
468
+ In terms of the linear system of equations, when v happens to be the zero vector,
469
+
470
+ 00:10:49.946 --> 00:10:53.640
471
+ the null space gives you all of the possible solutions to the equation.
472
+
473
+ 00:10:56.420 --> 00:10:58.886
474
+ So that's a very high level overview of how to
475
+
476
+ 00:10:58.886 --> 00:11:01.720
477
+ think about linear systems of equations geometrically.
478
+
479
+ 00:11:02.300 --> 00:11:05.999
480
+ Each system has some kind of linear transformation associated with it,
481
+
482
+ 00:11:05.999 --> 00:11:10.114
483
+ and when that transformation has an inverse, you can use that inverse to solve
484
+
485
+ 00:11:10.114 --> 00:11:10.740
486
+ your system.
487
+
488
+ 00:11:12.280 --> 00:11:17.240
489
+ Otherwise, the idea of column space lets us understand when a solution even exists,
490
+
491
+ 00:11:17.240 --> 00:11:20.841
492
+ and the idea of a null space helps us to understand what the
493
+
494
+ 00:11:20.841 --> 00:11:23.440
495
+ set of all possible solutions can look like.
496
+
497
+ 00:11:24.980 --> 00:11:27.397
498
+ Again, there's a lot that I haven't covered here,
499
+
500
+ 00:11:27.397 --> 00:11:29.380
501
+ most notably how to compute these things.
502
+
503
+ 00:11:29.800 --> 00:11:32.304
504
+ I also had to limit my scope to examples where the
505
+
506
+ 00:11:32.304 --> 00:11:34.760
507
+ number of equations equals the number of unknowns.
508
+
509
+ 00:11:34.880 --> 00:11:37.557
510
+ But the goal here is not to try to teach everything,
511
+
512
+ 00:11:37.557 --> 00:11:41.094
513
+ it's that you come away with a strong intuition for inverse matrices,
514
+
515
+ 00:11:41.094 --> 00:11:44.731
516
+ column space, and null space, and that those intuitions make any future
517
+
518
+ 00:11:44.731 --> 00:11:46.500
519
+ learning that you do more fruitful.
520
+
521
+ 00:11:47.660 --> 00:11:51.880
522
+ Next video, by popular request, will be a brief footnote about non-square matrices.
523
+
524
+ 00:11:51.880 --> 00:11:54.677
525
+ Then after that, I'm going to give you my take on dot products,
526
+
527
+ 00:11:54.677 --> 00:11:58.348
528
+ and something pretty cool that happens when you view them under the light of linear
529
+
530
+ 00:11:58.348 --> 00:11:59.660
531
+ transformations. See you then!
532
+
3B1B_LinearAlgebra/07 - Inverse matrices, column space and null space | Chapter 7, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2d2c8a4bcf8217e5fe98569978ebdac9547ab1cfcfb6580b7e2b46ce0948ede6
3
+ size 79158162
3B1B_LinearAlgebra/08 - Nonsquare matrices as transformations between dimensions | Chapter 8, Essence of linear algebra.en-orig.vtt ADDED
@@ -0,0 +1,944 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:11.240 --> 00:00:12.750 align:start position:0%
6
+
7
+ Hey<00:00:11.320><c> everyone,</c><00:00:11.960><c> I've</c><00:00:12.080><c> got</c><00:00:12.200><c> another</c><00:00:12.480><c> quick</c>
8
+
9
+ 00:00:12.750 --> 00:00:12.760 align:start position:0%
10
+ Hey everyone, I've got another quick
11
+
12
+
13
+ 00:00:12.760 --> 00:00:15.190 align:start position:0%
14
+ Hey everyone, I've got another quick
15
+ footnote<00:00:13.120><c> for</c><00:00:13.240><c> you</c><00:00:13.360><c> between</c><00:00:13.640><c> chapters</c><00:00:14.040><c> today.</c>
16
+
17
+ 00:00:15.190 --> 00:00:15.200 align:start position:0%
18
+ footnote for you between chapters today.
19
+
20
+
21
+ 00:00:15.200 --> 00:00:16.030 align:start position:0%
22
+ footnote for you between chapters today.
23
+ When<00:00:15.280><c> I've</c><00:00:15.360><c> talked</c><00:00:15.600><c> about</c><00:00:15.800><c> linear</c>
24
+
25
+ 00:00:16.030 --> 00:00:16.040 align:start position:0%
26
+ When I've talked about linear
27
+
28
+
29
+ 00:00:16.040 --> 00:00:17.830 align:start position:0%
30
+ When I've talked about linear
31
+ transformation<00:00:16.680><c> so</c><00:00:16.840><c> far,</c><00:00:17.320><c> I've</c><00:00:17.440><c> only</c><00:00:17.640><c> really</c>
32
+
33
+ 00:00:17.830 --> 00:00:17.840 align:start position:0%
34
+ transformation so far, I've only really
35
+
36
+
37
+ 00:00:17.840 --> 00:00:19.550 align:start position:0%
38
+ transformation so far, I've only really
39
+ talked<00:00:18.120><c> about</c><00:00:18.320><c> transformations</c><00:00:19.000><c> from</c><00:00:19.240><c> 2D</c>
40
+
41
+ 00:00:19.550 --> 00:00:19.560 align:start position:0%
42
+ talked about transformations from 2D
43
+
44
+
45
+ 00:00:19.560 --> 00:00:21.830 align:start position:0%
46
+ talked about transformations from 2D
47
+ vectors<00:00:19.960><c> to</c><00:00:20.120><c> other</c><00:00:20.360><c> 2D</c><00:00:20.600><c> vectors</c><00:00:21.320><c> represented</c>
48
+
49
+ 00:00:21.830 --> 00:00:21.840 align:start position:0%
50
+ vectors to other 2D vectors represented
51
+
52
+
53
+ 00:00:21.840 --> 00:00:24.590 align:start position:0%
54
+ vectors to other 2D vectors represented
55
+ with<00:00:22.160><c> 2</c><00:00:22.280><c> by</c><00:00:22.440><c> 2</c><00:00:22.600><c> matrices</c><00:00:23.520><c> or</c><00:00:23.640><c> from</c><00:00:23.800><c> 3D</c><00:00:24.160><c> vectors</c>
56
+
57
+ 00:00:24.590 --> 00:00:24.600 align:start position:0%
58
+ with 2 by 2 matrices or from 3D vectors
59
+
60
+
61
+ 00:00:24.600 --> 00:00:26.990 align:start position:0%
62
+ with 2 by 2 matrices or from 3D vectors
63
+ to<00:00:24.720><c> other</c><00:00:24.920><c> 3D</c><00:00:25.280><c> vectors</c><00:00:26.120><c> represented</c><00:00:26.640><c> with</c><00:00:26.800><c> 3</c>
64
+
65
+ 00:00:26.990 --> 00:00:27.000 align:start position:0%
66
+ to other 3D vectors represented with 3
67
+
68
+
69
+ 00:00:27.000 --> 00:00:28.990 align:start position:0%
70
+ to other 3D vectors represented with 3
71
+ by<00:00:27.160><c> 3</c><00:00:27.400><c> matrices.</c>
72
+
73
+ 00:00:28.990 --> 00:00:29.000 align:start position:0%
74
+ by 3 matrices.
75
+
76
+
77
+ 00:00:29.000 --> 00:00:30.630 align:start position:0%
78
+ by 3 matrices.
79
+ But<00:00:29.200><c> several</c><00:00:29.480><c> commenters</c><00:00:29.960><c> have</c><00:00:30.080><c> asked</c><00:00:30.320><c> about</c>
80
+
81
+ 00:00:30.630 --> 00:00:30.640 align:start position:0%
82
+ But several commenters have asked about
83
+
84
+
85
+ 00:00:30.640 --> 00:00:32.350 align:start position:0%
86
+ But several commenters have asked about
87
+ non-square<00:00:31.200><c> matrices,</c><00:00:31.920><c> so</c><00:00:32.040><c> I</c><00:00:32.080><c> thought</c><00:00:32.279><c> I'd</c>
88
+
89
+ 00:00:32.350 --> 00:00:32.360 align:start position:0%
90
+ non-square matrices, so I thought I'd
91
+
92
+
93
+ 00:00:32.360 --> 00:00:33.670 align:start position:0%
94
+ non-square matrices, so I thought I'd
95
+ take<00:00:32.520><c> a</c><00:00:32.599><c> moment</c><00:00:32.840><c> to</c><00:00:32.920><c> just</c><00:00:33.160><c> show</c><00:00:33.360><c> what</c><00:00:33.480><c> those</c>
96
+
97
+ 00:00:33.670 --> 00:00:33.680 align:start position:0%
98
+ take a moment to just show what those
99
+
100
+
101
+ 00:00:33.680 --> 00:00:35.270 align:start position:0%
102
+ take a moment to just show what those
103
+ mean<00:00:33.840><c> geometrically.</c>
104
+
105
+ 00:00:35.270 --> 00:00:35.280 align:start position:0%
106
+ mean geometrically.
107
+
108
+
109
+ 00:00:35.280 --> 00:00:37.030 align:start position:0%
110
+ mean geometrically.
111
+ By<00:00:35.400><c> now</c><00:00:35.600><c> in</c><00:00:35.680><c> the</c><00:00:35.760><c> series,</c><00:00:36.320><c> you</c><00:00:36.440><c> actually</c><00:00:36.800><c> have</c>
112
+
113
+ 00:00:37.030 --> 00:00:37.040 align:start position:0%
114
+ By now in the series, you actually have
115
+
116
+
117
+ 00:00:37.040 --> 00:00:38.590 align:start position:0%
118
+ By now in the series, you actually have
119
+ most<00:00:37.280><c> of</c><00:00:37.360><c> the</c><00:00:37.440><c> background</c><00:00:37.880><c> you</c><00:00:37.960><c> need</c><00:00:38.200><c> to</c><00:00:38.320><c> start</c>
120
+
121
+ 00:00:38.590 --> 00:00:38.600 align:start position:0%
122
+ most of the background you need to start
123
+
124
+
125
+ 00:00:38.600 --> 00:00:40.110 align:start position:0%
126
+ most of the background you need to start
127
+ pondering<00:00:39.040><c> a</c><00:00:39.080><c> question</c><00:00:39.480><c> like</c><00:00:39.640><c> this</c><00:00:39.840><c> on</c><00:00:39.960><c> your</c>
128
+
129
+ 00:00:40.110 --> 00:00:40.120 align:start position:0%
130
+ pondering a question like this on your
131
+
132
+
133
+ 00:00:40.120 --> 00:00:41.630 align:start position:0%
134
+ pondering a question like this on your
135
+ own,<00:00:40.640><c> but</c><00:00:40.760><c> I'll</c><00:00:40.880><c> start</c><00:00:41.080><c> talking</c><00:00:41.360><c> through</c><00:00:41.520><c> it</c>
136
+
137
+ 00:00:41.630 --> 00:00:41.640 align:start position:0%
138
+ own, but I'll start talking through it
139
+
140
+
141
+ 00:00:41.640 --> 00:00:44.030 align:start position:0%
142
+ own, but I'll start talking through it
143
+ just<00:00:41.800><c> to</c><00:00:41.840><c> give</c><00:00:42.000><c> a</c><00:00:42.040><c> little</c><00:00:42.280><c> mental</c><00:00:42.640><c> momentum.</c>
144
+
145
+ 00:00:44.030 --> 00:00:44.040 align:start position:0%
146
+ just to give a little mental momentum.
147
+
148
+
149
+ 00:00:44.040 --> 00:00:45.630 align:start position:0%
150
+ just to give a little mental momentum.
151
+ It's<00:00:44.200><c> perfectly</c><00:00:44.640><c> reasonable</c><00:00:45.080><c> to</c><00:00:45.200><c> talk</c><00:00:45.400><c> about</c>
152
+
153
+ 00:00:45.630 --> 00:00:45.640 align:start position:0%
154
+ It's perfectly reasonable to talk about
155
+
156
+
157
+ 00:00:45.640 --> 00:00:48.510 align:start position:0%
158
+ It's perfectly reasonable to talk about
159
+ transformations<00:00:46.520><c> between</c><00:00:47.160><c> dimensions,</c><00:00:48.320><c> such</c>
160
+
161
+ 00:00:48.510 --> 00:00:48.520 align:start position:0%
162
+ transformations between dimensions, such
163
+
164
+
165
+ 00:00:48.520 --> 00:00:50.750 align:start position:0%
166
+ transformations between dimensions, such
167
+ as<00:00:48.640><c> one</c><00:00:48.800><c> that</c><00:00:48.920><c> takes</c><00:00:49.200><c> 2D</c><00:00:49.600><c> vectors</c><00:00:50.240><c> to</c><00:00:50.360><c> 3D</c>
168
+
169
+ 00:00:50.750 --> 00:00:50.760 align:start position:0%
170
+ as one that takes 2D vectors to 3D
171
+
172
+
173
+ 00:00:50.760 --> 00:00:53.270 align:start position:0%
174
+ as one that takes 2D vectors to 3D
175
+ vectors.<00:00:51.800><c> Again,</c><00:00:52.360><c> what</c><00:00:52.520><c> makes</c><00:00:52.800><c> one</c><00:00:52.960><c> of</c><00:00:53.040><c> these</c>
176
+
177
+ 00:00:53.270 --> 00:00:53.280 align:start position:0%
178
+ vectors. Again, what makes one of these
179
+
180
+
181
+ 00:00:53.280 --> 00:00:55.190 align:start position:0%
182
+ vectors. Again, what makes one of these
183
+ linear<00:00:54.040><c> is</c><00:00:54.200><c> that</c><00:00:54.320><c> grid</c><00:00:54.600><c> lines</c><00:00:54.880><c> remain</c>
184
+
185
+ 00:00:55.190 --> 00:00:55.200 align:start position:0%
186
+ linear is that grid lines remain
187
+
188
+
189
+ 00:00:55.200 --> 00:00:57.430 align:start position:0%
190
+ linear is that grid lines remain
191
+ parallel<00:00:55.840><c> and</c><00:00:56.040><c> evenly</c><00:00:56.400><c> spaced</c><00:00:57.080><c> and</c><00:00:57.200><c> that</c><00:00:57.320><c> the</c>
192
+
193
+ 00:00:57.430 --> 00:00:57.440 align:start position:0%
194
+ parallel and evenly spaced and that the
195
+
196
+
197
+ 00:00:57.440 --> 00:00:59.990 align:start position:0%
198
+ parallel and evenly spaced and that the
199
+ origin<00:00:58.040><c> maps</c><00:00:58.320><c> to</c><00:00:58.440><c> the</c><00:00:58.600><c> origin.</c>
200
+
201
+ 00:00:59.990 --> 00:01:00.000 align:start position:0%
202
+ origin maps to the origin.
203
+
204
+
205
+ 00:01:00.000 --> 00:01:01.510 align:start position:0%
206
+ origin maps to the origin.
207
+ What<00:01:00.120><c> I</c><00:01:00.160><c> have</c><00:01:00.360><c> pictured</c><00:01:00.640><c> here</c><00:01:00.920><c> is</c><00:01:01.040><c> the</c><00:01:01.160><c> input</c>
208
+
209
+ 00:01:01.510 --> 00:01:01.520 align:start position:0%
210
+ What I have pictured here is the input
211
+
212
+
213
+ 00:01:01.520 --> 00:01:03.190 align:start position:0%
214
+ What I have pictured here is the input
215
+ space<00:01:01.800><c> on</c><00:01:01.880><c> the</c><00:01:02.000><c> left,</c><00:01:02.400><c> which</c><00:01:02.560><c> is</c><00:01:02.640><c> just</c><00:01:02.920><c> 2D</c>
216
+
217
+ 00:01:03.190 --> 00:01:03.200 align:start position:0%
218
+ space on the left, which is just 2D
219
+
220
+
221
+ 00:01:03.200 --> 00:01:04.829 align:start position:0%
222
+ space on the left, which is just 2D
223
+ space,<00:01:03.920><c> and</c><00:01:04.040><c> the</c><00:01:04.239><c> output</c><00:01:04.640><c> of</c><00:01:04.720><c> the</c>
224
+
225
+ 00:01:04.829 --> 00:01:04.839 align:start position:0%
226
+ space, and the output of the
227
+
228
+
229
+ 00:01:04.839 --> 00:01:06.950 align:start position:0%
230
+ space, and the output of the
231
+ transformation<00:01:05.560><c> shown</c><00:01:05.800><c> on</c><00:01:05.880><c> the</c><00:01:06.000><c> right.</c>
232
+
233
+ 00:01:06.950 --> 00:01:06.960 align:start position:0%
234
+ transformation shown on the right.
235
+
236
+
237
+ 00:01:06.960 --> 00:01:08.430 align:start position:0%
238
+ transformation shown on the right.
239
+ The<00:01:07.040><c> reason</c><00:01:07.320><c> I'm</c><00:01:07.400><c> not</c><00:01:07.600><c> showing</c><00:01:07.840><c> the</c><00:01:08.000><c> inputs</c>
240
+
241
+ 00:01:08.430 --> 00:01:08.440 align:start position:0%
242
+ The reason I'm not showing the inputs
243
+
244
+
245
+ 00:01:08.440 --> 00:01:10.270 align:start position:0%
246
+ The reason I'm not showing the inputs
247
+ move<00:01:08.760><c> over</c><00:01:09.000><c> to</c><00:01:09.120><c> the</c><00:01:09.280><c> outputs</c><00:01:09.680><c> like</c><00:01:09.840><c> I</c><00:01:09.920><c> usually</c>
248
+
249
+ 00:01:10.270 --> 00:01:10.280 align:start position:0%
250
+ move over to the outputs like I usually
251
+
252
+
253
+ 00:01:10.280 --> 00:01:12.830 align:start position:0%
254
+ move over to the outputs like I usually
255
+ do<00:01:10.800><c> is</c><00:01:10.880><c> not</c><00:01:11.080><c> just</c><00:01:11.360><c> animation</c><00:01:11.840><c> laziness.</c><00:01:12.680><c> It's</c>
256
+
257
+ 00:01:12.830 --> 00:01:12.840 align:start position:0%
258
+ do is not just animation laziness. It's
259
+
260
+
261
+ 00:01:12.840 --> 00:01:15.150 align:start position:0%
262
+ do is not just animation laziness. It's
263
+ worth<00:01:13.040><c> emphasizing</c><00:01:13.640><c> that</c><00:01:13.800><c> 2D</c><00:01:14.120><c> vector</c><00:01:14.440><c> inputs</c>
264
+
265
+ 00:01:15.150 --> 00:01:15.160 align:start position:0%
266
+ worth emphasizing that 2D vector inputs
267
+
268
+
269
+ 00:01:15.160 --> 00:01:17.190 align:start position:0%
270
+ worth emphasizing that 2D vector inputs
271
+ are<00:01:15.320><c> very</c><00:01:15.640><c> different</c><00:01:15.960><c> animals</c><00:01:16.440><c> from</c><00:01:16.600><c> these</c><00:01:16.840><c> 3D</c>
272
+
273
+ 00:01:17.190 --> 00:01:17.200 align:start position:0%
274
+ are very different animals from these 3D
275
+
276
+
277
+ 00:01:17.200 --> 00:01:19.230 align:start position:0%
278
+ are very different animals from these 3D
279
+ vector<00:01:17.520><c> outputs</c><00:01:18.200><c> living</c><00:01:18.520><c> in</c><00:01:18.640><c> a</c><00:01:18.680><c> completely</c>
280
+
281
+ 00:01:19.230 --> 00:01:19.240 align:start position:0%
282
+ vector outputs living in a completely
283
+
284
+
285
+ 00:01:19.240 --> 00:01:21.710 align:start position:0%
286
+ vector outputs living in a completely
287
+ separate<00:01:19.720><c> unconnected</c><00:01:20.400><c> space.</c>
288
+
289
+ 00:01:21.710 --> 00:01:21.720 align:start position:0%
290
+ separate unconnected space.
291
+
292
+
293
+ 00:01:21.720 --> 00:01:23.190 align:start position:0%
294
+ separate unconnected space.
295
+ Encoding<00:01:22.160><c> one</c><00:01:22.320><c> of</c><00:01:22.360><c> these</c><00:01:22.520><c> transformations</c>
296
+
297
+ 00:01:23.190 --> 00:01:23.200 align:start position:0%
298
+ Encoding one of these transformations
299
+
300
+
301
+ 00:01:23.200 --> 00:01:25.430 align:start position:0%
302
+ Encoding one of these transformations
303
+ with<00:01:23.360><c> a</c><00:01:23.400><c> matrix</c><00:01:24.160><c> is</c><00:01:24.560><c> really</c><00:01:24.840><c> just</c><00:01:25.080><c> the</c><00:01:25.160><c> same</c>
304
+
305
+ 00:01:25.430 --> 00:01:25.440 align:start position:0%
306
+ with a matrix is really just the same
307
+
308
+
309
+ 00:01:25.440 --> 00:01:27.350 align:start position:0%
310
+ with a matrix is really just the same
311
+ thing<00:01:25.640><c> as</c><00:01:25.760><c> what</c><00:01:25.880><c> we've</c><00:01:26.040><c> done</c><00:01:26.240><c> before.</c><00:01:27.240><c> You</c>
312
+
313
+ 00:01:27.350 --> 00:01:27.360 align:start position:0%
314
+ thing as what we've done before. You
315
+
316
+
317
+ 00:01:27.360 --> 00:01:29.430 align:start position:0%
318
+ thing as what we've done before. You
319
+ look<00:01:27.520><c> at</c><00:01:27.640><c> where</c><00:01:27.840><c> each</c><00:01:28.080><c> basis</c><00:01:28.480><c> vector</c><00:01:28.760><c> lands</c>
320
+
321
+ 00:01:29.430 --> 00:01:29.440 align:start position:0%
322
+ look at where each basis vector lands
323
+
324
+
325
+ 00:01:29.440 --> 00:01:31.150 align:start position:0%
326
+ look at where each basis vector lands
327
+ and<00:01:29.560><c> write</c><00:01:29.760><c> the</c><00:01:29.920><c> coordinates</c><00:01:30.560><c> of</c><00:01:30.640><c> the</c><00:01:30.760><c> landing</c>
328
+
329
+ 00:01:31.150 --> 00:01:31.160 align:start position:0%
330
+ and write the coordinates of the landing
331
+
332
+
333
+ 00:01:31.160 --> 00:01:33.750 align:start position:0%
334
+ and write the coordinates of the landing
335
+ spots<00:01:31.720><c> as</c><00:01:31.880><c> the</c><00:01:32.000><c> columns</c><00:01:32.440><c> of</c><00:01:32.560><c> a</c><00:01:32.600><c> matrix.</c><00:01:33.680><c> For</c>
336
+
337
+ 00:01:33.750 --> 00:01:33.760 align:start position:0%
338
+ spots as the columns of a matrix. For
339
+
340
+
341
+ 00:01:33.760 --> 00:01:35.830 align:start position:0%
342
+ spots as the columns of a matrix. For
343
+ example,<00:01:34.200><c> what</c><00:01:34.360><c> you're</c><00:01:34.480><c> looking</c><00:01:34.760><c> at</c><00:01:34.880><c> here</c><00:01:35.680><c> is</c>
344
+
345
+ 00:01:35.830 --> 00:01:35.840 align:start position:0%
346
+ example, what you're looking at here is
347
+
348
+
349
+ 00:01:35.840 --> 00:01:37.910 align:start position:0%
350
+ example, what you're looking at here is
351
+ an<00:01:36.000><c> output</c><00:01:36.320><c> of</c><00:01:36.400><c> a</c><00:01:36.440><c> transformation</c><00:01:37.440><c> that</c><00:01:37.560><c> takes</c>
352
+
353
+ 00:01:37.910 --> 00:01:37.920 align:start position:0%
354
+ an output of a transformation that takes
355
+
356
+
357
+ 00:01:37.920 --> 00:01:42.350 align:start position:0%
358
+ an output of a transformation that takes
359
+ I<00:01:38.120><c> hat</c><00:01:38.600><c> to</c><00:01:38.680><c> the</c><00:01:38.800><c> coordinates</c><00:01:39.520><c> 2</c><00:01:40.000><c> -1</c><00:01:40.760><c> -2</c><00:01:41.920><c> and</c><00:01:42.120><c> J</c>
360
+
361
+ 00:01:42.350 --> 00:01:42.360 align:start position:0%
362
+ I hat to the coordinates 2 -1 -2 and J
363
+
364
+
365
+ 00:01:42.360 --> 00:01:47.870 align:start position:0%
366
+ I hat to the coordinates 2 -1 -2 and J
367
+ hat<00:01:42.720><c> to</c><00:01:42.800><c> the</c><00:01:42.920><c> coordinates</c><00:01:43.560><c> 0</c><00:01:44.200><c> 1</c><00:01:44.760><c> 1.</c>
368
+
369
+ 00:01:47.870 --> 00:01:47.880 align:start position:0%
370
+
371
+
372
+
373
+ 00:01:47.880 --> 00:01:49.790 align:start position:0%
374
+
375
+ Notice,<00:01:48.320><c> this</c><00:01:48.560><c> means</c><00:01:48.840><c> the</c><00:01:48.960><c> matrix</c><00:01:49.400><c> encoding</c>
376
+
377
+ 00:01:49.790 --> 00:01:49.800 align:start position:0%
378
+ Notice, this means the matrix encoding
379
+
380
+
381
+ 00:01:49.800 --> 00:01:52.190 align:start position:0%
382
+ Notice, this means the matrix encoding
383
+ our<00:01:49.880><c> transformation</c><00:01:51.000><c> has</c><00:01:51.200><c> three</c><00:01:51.520><c> rows</c><00:01:52.000><c> and</c>
384
+
385
+ 00:01:52.190 --> 00:01:52.200 align:start position:0%
386
+ our transformation has three rows and
387
+
388
+
389
+ 00:01:52.200 --> 00:01:54.150 align:start position:0%
390
+ our transformation has three rows and
391
+ two<00:01:52.360><c> columns,</c><00:01:53.240><c> which</c><00:01:53.560><c> to</c><00:01:53.680><c> use</c><00:01:53.800><c> standard</c>
392
+
393
+ 00:01:54.150 --> 00:01:54.160 align:start position:0%
394
+ two columns, which to use standard
395
+
396
+
397
+ 00:01:54.160 --> 00:01:57.750 align:start position:0%
398
+ two columns, which to use standard
399
+ terminology<00:01:55.080><c> makes</c><00:01:55.280><c> it</c><00:01:55.400><c> a</c><00:01:55.480><c> 3</c><00:01:55.800><c> by</c><00:01:55.960><c> 2</c><00:01:56.200><c> matrix.</c>
400
+
401
+ 00:01:57.750 --> 00:01:57.760 align:start position:0%
402
+ terminology makes it a 3 by 2 matrix.
403
+
404
+
405
+ 00:01:57.760 --> 00:01:59.630 align:start position:0%
406
+ terminology makes it a 3 by 2 matrix.
407
+ In<00:01:57.920><c> the</c><00:01:58.000><c> language</c><00:01:58.400><c> of</c><00:01:58.480><c> last</c><00:01:58.720><c> video,</c><00:01:59.480><c> the</c>
408
+
409
+ 00:01:59.630 --> 00:01:59.640 align:start position:0%
410
+ In the language of last video, the
411
+
412
+
413
+ 00:01:59.640 --> 00:02:01.910 align:start position:0%
414
+ In the language of last video, the
415
+ column<00:02:00.240><c> space</c><00:02:00.680><c> of</c><00:02:00.760><c> this</c><00:02:00.960><c> matrix,</c><00:02:01.560><c> the</c><00:02:01.680><c> place</c>
416
+
417
+ 00:02:01.910 --> 00:02:01.920 align:start position:0%
418
+ column space of this matrix, the place
419
+
420
+
421
+ 00:02:01.920 --> 00:02:04.110 align:start position:0%
422
+ column space of this matrix, the place
423
+ where<00:02:02.040><c> all</c><00:02:02.200><c> the</c><00:02:02.280><c> vectors</c><00:02:02.680><c> land,</c><00:02:03.360><c> is</c><00:02:03.520><c> a</c><00:02:03.600><c> 2D</c>
424
+
425
+ 00:02:04.110 --> 00:02:04.120 align:start position:0%
426
+ where all the vectors land, is a 2D
427
+
428
+
429
+ 00:02:04.120 --> 00:02:06.310 align:start position:0%
430
+ where all the vectors land, is a 2D
431
+ plane<00:02:04.520><c> slicing</c><00:02:05.000><c> through</c><00:02:05.200><c> the</c><00:02:05.360><c> origin</c><00:02:05.840><c> of</c><00:02:05.960><c> 3D</c>
432
+
433
+ 00:02:06.310 --> 00:02:06.320 align:start position:0%
434
+ plane slicing through the origin of 3D
435
+
436
+
437
+ 00:02:06.320 --> 00:02:07.350 align:start position:0%
438
+ plane slicing through the origin of 3D
439
+ space.
440
+
441
+ 00:02:07.350 --> 00:02:07.360 align:start position:0%
442
+ space.
443
+
444
+
445
+ 00:02:07.360 --> 00:02:09.669 align:start position:0%
446
+ space.
447
+ But<00:02:07.480><c> the</c><00:02:07.560><c> matrix</c><00:02:07.920><c> is</c><00:02:08.039><c> still</c><00:02:08.399><c> full</c><00:02:08.800><c> rank</c><00:02:09.399><c> since</c>
448
+
449
+ 00:02:09.669 --> 00:02:09.679 align:start position:0%
450
+ But the matrix is still full rank since
451
+
452
+
453
+ 00:02:09.679 --> 00:02:11.550 align:start position:0%
454
+ But the matrix is still full rank since
455
+ the<00:02:09.800><c> number</c><00:02:10.080><c> of</c><00:02:10.200><c> dimensions</c><00:02:10.800><c> in</c><00:02:11.000><c> this</c><00:02:11.200><c> column</c>
456
+
457
+ 00:02:11.550 --> 00:02:11.560 align:start position:0%
458
+ the number of dimensions in this column
459
+
460
+
461
+ 00:02:11.560 --> 00:02:13.430 align:start position:0%
462
+ the number of dimensions in this column
463
+ space<00:02:12.240><c> is</c><00:02:12.400><c> the</c><00:02:12.480><c> same</c><00:02:12.880><c> as</c><00:02:13.000><c> the</c><00:02:13.120><c> number</c><00:02:13.360><c> of</c>
464
+
465
+ 00:02:13.430 --> 00:02:13.440 align:start position:0%
466
+ space is the same as the number of
467
+
468
+
469
+ 00:02:13.440 --> 00:02:17.390 align:start position:0%
470
+ space is the same as the number of
471
+ dimensions<00:02:14.200><c> of</c><00:02:14.360><c> the</c><00:02:14.520><c> input</c><00:02:14.960><c> space.</c>
472
+
473
+ 00:02:17.390 --> 00:02:17.400 align:start position:0%
474
+ dimensions of the input space.
475
+
476
+
477
+ 00:02:17.400 --> 00:02:19.350 align:start position:0%
478
+ dimensions of the input space.
479
+ So,<00:02:17.600><c> if</c><00:02:17.760><c> you</c><00:02:17.880><c> see</c><00:02:18.040><c> a</c><00:02:18.120><c> 3</c><00:02:18.200><c> by</c><00:02:18.480><c> 2</c><00:02:18.640><c> matrix</c><00:02:19.080><c> out</c><00:02:19.240><c> in</c>
480
+
481
+ 00:02:19.350 --> 00:02:19.360 align:start position:0%
482
+ So, if you see a 3 by 2 matrix out in
483
+
484
+
485
+ 00:02:19.360 --> 00:02:21.150 align:start position:0%
486
+ So, if you see a 3 by 2 matrix out in
487
+ the<00:02:19.440><c> wild,</c><00:02:20.120><c> you</c><00:02:20.240><c> can</c><00:02:20.400><c> know</c><00:02:20.560><c> that</c><00:02:20.720><c> it</c><00:02:20.840><c> has</c><00:02:21.000><c> the</c>
488
+
489
+ 00:02:21.150 --> 00:02:21.160 align:start position:0%
490
+ the wild, you can know that it has the
491
+
492
+
493
+ 00:02:21.160 --> 00:02:23.470 align:start position:0%
494
+ the wild, you can know that it has the
495
+ geometric<00:02:21.720><c> interpretation</c><00:02:22.760><c> of</c><00:02:22.920><c> mapping</c><00:02:23.320><c> two</c>
496
+
497
+ 00:02:23.470 --> 00:02:23.480 align:start position:0%
498
+ geometric interpretation of mapping two
499
+
500
+
501
+ 00:02:23.480 --> 00:02:26.030 align:start position:0%
502
+ geometric interpretation of mapping two
503
+ dimensions<00:02:24.320><c> to</c><00:02:24.440><c> three</c><00:02:24.640><c> dimensions</c><00:02:25.640><c> since</c><00:02:25.920><c> the</c>
504
+
505
+ 00:02:26.030 --> 00:02:26.040 align:start position:0%
506
+ dimensions to three dimensions since the
507
+
508
+
509
+ 00:02:26.040 --> 00:02:27.750 align:start position:0%
510
+ dimensions to three dimensions since the
511
+ two<00:02:26.240><c> columns</c><00:02:26.720><c> indicate</c><00:02:27.080><c> that</c><00:02:27.200><c> the</c><00:02:27.360><c> input</c>
512
+
513
+ 00:02:27.750 --> 00:02:27.760 align:start position:0%
514
+ two columns indicate that the input
515
+
516
+
517
+ 00:02:27.760 --> 00:02:30.030 align:start position:0%
518
+ two columns indicate that the input
519
+ space<00:02:28.160><c> has</c><00:02:28.400><c> two</c><00:02:28.600><c> basis</c><00:02:28.960><c> vectors</c><00:02:29.840><c> and</c><00:02:29.960><c> the</c>
520
+
521
+ 00:02:30.030 --> 00:02:30.040 align:start position:0%
522
+ space has two basis vectors and the
523
+
524
+
525
+ 00:02:30.040 --> 00:02:32.230 align:start position:0%
526
+ space has two basis vectors and the
527
+ three<00:02:30.320><c> rows</c><00:02:31.000><c> indicate</c><00:02:31.440><c> that</c><00:02:31.560><c> the</c><00:02:31.720><c> landing</c>
528
+
529
+ 00:02:32.230 --> 00:02:32.240 align:start position:0%
530
+ three rows indicate that the landing
531
+
532
+
533
+ 00:02:32.240 --> 00:02:34.750 align:start position:0%
534
+ three rows indicate that the landing
535
+ spots<00:02:32.680><c> for</c><00:02:32.840><c> each</c><00:02:33.080><c> of</c><00:02:33.160><c> those</c><00:02:33.400><c> basis</c><00:02:33.760><c> vectors</c><00:02:34.600><c> is</c>
536
+
537
+ 00:02:34.750 --> 00:02:34.760 align:start position:0%
538
+ spots for each of those basis vectors is
539
+
540
+
541
+ 00:02:34.760 --> 00:02:36.230 align:start position:0%
542
+ spots for each of those basis vectors is
543
+ described<00:02:35.280><c> with</c><00:02:35.480><c> three</c><00:02:35.840><c> separate</c>
544
+
545
+ 00:02:36.230 --> 00:02:36.240 align:start position:0%
546
+ described with three separate
547
+
548
+
549
+ 00:02:36.240 --> 00:02:37.830 align:start position:0%
550
+ described with three separate
551
+ coordinates.
552
+
553
+ 00:02:37.830 --> 00:02:37.840 align:start position:0%
554
+ coordinates.
555
+
556
+
557
+ 00:02:37.840 --> 00:02:40.110 align:start position:0%
558
+ coordinates.
559
+ Likewise,<00:02:38.400><c> if</c><00:02:38.560><c> you</c><00:02:38.640><c> see</c><00:02:38.840><c> a</c><00:02:38.960><c> 2</c><00:02:39.200><c> by</c><00:02:39.360><c> 3</c><00:02:39.640><c> matrix</c>
560
+
561
+ 00:02:40.110 --> 00:02:40.120 align:start position:0%
562
+ Likewise, if you see a 2 by 3 matrix
563
+
564
+
565
+ 00:02:40.120 --> 00:02:42.190 align:start position:0%
566
+ Likewise, if you see a 2 by 3 matrix
567
+ with<00:02:40.400><c> two</c><00:02:40.560><c> rows</c><00:02:40.840><c> and</c><00:02:40.960><c> three</c><00:02:41.160><c> columns,</c><00:02:42.040><c> what</c><00:02:42.160><c> do</c>
568
+
569
+ 00:02:42.190 --> 00:02:42.200 align:start position:0%
570
+ with two rows and three columns, what do
571
+
572
+
573
+ 00:02:42.200 --> 00:02:43.630 align:start position:0%
574
+ with two rows and three columns, what do
575
+ you<00:02:42.280><c> think</c><00:02:42.440><c> that</c><00:02:42.640><c> means?</c>
576
+
577
+ 00:02:43.630 --> 00:02:43.640 align:start position:0%
578
+ you think that means?
579
+
580
+
581
+ 00:02:43.640 --> 00:02:45.630 align:start position:0%
582
+ you think that means?
583
+ Well,<00:02:44.240><c> the</c><00:02:44.360><c> three</c><00:02:44.600><c> columns</c><00:02:45.120><c> indicate</c><00:02:45.520><c> that</c>
584
+
585
+ 00:02:45.630 --> 00:02:45.640 align:start position:0%
586
+ Well, the three columns indicate that
587
+
588
+
589
+ 00:02:45.640 --> 00:02:46.950 align:start position:0%
590
+ Well, the three columns indicate that
591
+ you're<00:02:45.760><c> starting</c><00:02:46.120><c> in</c><00:02:46.240><c> a</c><00:02:46.280><c> space</c><00:02:46.640><c> that</c><00:02:46.800><c> has</c>
592
+
593
+ 00:02:46.950 --> 00:02:46.960 align:start position:0%
594
+ you're starting in a space that has
595
+
596
+
597
+ 00:02:46.960 --> 00:02:48.870 align:start position:0%
598
+ you're starting in a space that has
599
+ three<00:02:47.240><c> basis</c><00:02:47.600><c> vectors,</c><00:02:48.320><c> so</c><00:02:48.440><c> we're</c><00:02:48.520><c> starting</c>
600
+
601
+ 00:02:48.870 --> 00:02:48.880 align:start position:0%
602
+ three basis vectors, so we're starting
603
+
604
+
605
+ 00:02:48.880 --> 00:02:51.190 align:start position:0%
606
+ three basis vectors, so we're starting
607
+ in<00:02:48.960><c> three</c><00:02:49.120><c> dimensions.</c><00:02:50.280><c> And</c><00:02:50.400><c> the</c><00:02:50.480><c> two</c><00:02:50.720><c> rows</c>
608
+
609
+ 00:02:51.190 --> 00:02:51.200 align:start position:0%
610
+ in three dimensions. And the two rows
611
+
612
+
613
+ 00:02:51.200 --> 00:02:53.110 align:start position:0%
614
+ in three dimensions. And the two rows
615
+ indicate<00:02:51.640><c> that</c><00:02:51.760><c> the</c><00:02:51.880><c> landing</c><00:02:52.360><c> spot</c><00:02:52.720><c> for</c><00:02:52.920><c> each</c>
616
+
617
+ 00:02:53.110 --> 00:02:53.120 align:start position:0%
618
+ indicate that the landing spot for each
619
+
620
+
621
+ 00:02:53.120 --> 00:02:55.110 align:start position:0%
622
+ indicate that the landing spot for each
623
+ of<00:02:53.200><c> those</c><00:02:53.440><c> three</c><00:02:53.680><c> basis</c><00:02:54.080><c> vectors</c><00:02:54.960><c> is</c>
624
+
625
+ 00:02:55.110 --> 00:02:55.120 align:start position:0%
626
+ of those three basis vectors is
627
+
628
+
629
+ 00:02:55.120 --> 00:02:57.590 align:start position:0%
630
+ of those three basis vectors is
631
+ described<00:02:55.680><c> with</c><00:02:55.920><c> only</c><00:02:56.200><c> two</c><00:02:56.480><c> coordinates,</c><00:02:57.480><c> so</c>
632
+
633
+ 00:02:57.590 --> 00:02:57.600 align:start position:0%
634
+ described with only two coordinates, so
635
+
636
+
637
+ 00:02:57.600 --> 00:03:00.430 align:start position:0%
638
+ described with only two coordinates, so
639
+ they<00:02:57.720><c> must</c><00:02:57.920><c> be</c><00:02:58.040><c> landing</c><00:02:58.640><c> in</c><00:02:58.840><c> two</c><00:02:59.040><c> dimensions.</c>
640
+
641
+ 00:03:00.430 --> 00:03:00.440 align:start position:0%
642
+ they must be landing in two dimensions.
643
+
644
+
645
+ 00:03:00.440 --> 00:03:02.910 align:start position:0%
646
+ they must be landing in two dimensions.
647
+ So,<00:03:00.560><c> it's</c><00:03:00.680><c> a</c><00:03:00.720><c> transformation</c><00:03:01.480><c> from</c><00:03:01.840><c> 3D</c><00:03:02.200><c> space</c>
648
+
649
+ 00:03:02.910 --> 00:03:02.920 align:start position:0%
650
+ So, it's a transformation from 3D space
651
+
652
+
653
+ 00:03:02.920 --> 00:03:05.390 align:start position:0%
654
+ So, it's a transformation from 3D space
655
+ onto<00:03:03.360><c> the</c><00:03:03.480><c> 2D</c><00:03:03.840><c> plane.</c>
656
+
657
+ 00:03:05.390 --> 00:03:05.400 align:start position:0%
658
+ onto the 2D plane.
659
+
660
+
661
+ 00:03:05.400 --> 00:03:07.190 align:start position:0%
662
+ onto the 2D plane.
663
+ A<00:03:05.480><c> transformation</c><00:03:06.080><c> that</c><00:03:06.200><c> should</c><00:03:06.360><c> feel</c><00:03:06.640><c> very</c>
664
+
665
+ 00:03:07.190 --> 00:03:07.200 align:start position:0%
666
+ A transformation that should feel very
667
+
668
+
669
+ 00:03:07.200 --> 00:03:08.790 align:start position:0%
670
+ A transformation that should feel very
671
+ uncomfortable<00:03:07.920><c> if</c><00:03:08.040><c> you</c><00:03:08.160><c> imagine</c><00:03:08.520><c> going</c>
672
+
673
+ 00:03:08.790 --> 00:03:08.800 align:start position:0%
674
+ uncomfortable if you imagine going
675
+
676
+
677
+ 00:03:08.800 --> 00:03:13.870 align:start position:0%
678
+ uncomfortable if you imagine going
679
+ through<00:03:09.040><c> it.</c>
680
+
681
+ 00:03:13.870 --> 00:03:13.880 align:start position:0%
682
+
683
+
684
+
685
+ 00:03:13.880 --> 00:03:15.150 align:start position:0%
686
+
687
+ You<00:03:14.000><c> could</c><00:03:14.160><c> also</c><00:03:14.360><c> have</c><00:03:14.480><c> a</c><00:03:14.520><c> transformation</c>
688
+
689
+ 00:03:15.150 --> 00:03:15.160 align:start position:0%
690
+ You could also have a transformation
691
+
692
+
693
+ 00:03:15.160 --> 00:03:17.790 align:start position:0%
694
+ You could also have a transformation
695
+ from<00:03:15.440><c> two</c><00:03:15.640><c> dimensions</c><00:03:16.200><c> to</c><00:03:16.360><c> one</c><00:03:16.560><c> dimension.</c>
696
+
697
+ 00:03:17.790 --> 00:03:17.800 align:start position:0%
698
+ from two dimensions to one dimension.
699
+
700
+
701
+ 00:03:17.800 --> 00:03:19.430 align:start position:0%
702
+ from two dimensions to one dimension.
703
+ One-dimensional<00:03:18.360><c> space</c><00:03:18.680><c> is</c><00:03:18.920><c> really</c><00:03:19.120><c> just</c><00:03:19.360><c> the</c>
704
+
705
+ 00:03:19.430 --> 00:03:19.440 align:start position:0%
706
+ One-dimensional space is really just the
707
+
708
+
709
+ 00:03:19.440 --> 00:03:21.270 align:start position:0%
710
+ One-dimensional space is really just the
711
+ number<00:03:19.720><c> line,</c><00:03:20.320><c> so</c><00:03:20.440><c> a</c><00:03:20.480><c> transformation</c><00:03:21.120><c> like</c>
712
+
713
+ 00:03:21.270 --> 00:03:21.280 align:start position:0%
714
+ number line, so a transformation like
715
+
716
+
717
+ 00:03:21.280 --> 00:03:24.310 align:start position:0%
718
+ number line, so a transformation like
719
+ this<00:03:21.800><c> takes</c><00:03:22.160><c> in</c><00:03:22.480><c> 2D</c><00:03:22.880><c> vectors</c><00:03:23.640><c> and</c><00:03:23.800><c> spits</c><00:03:24.160><c> out</c>
720
+
721
+ 00:03:24.310 --> 00:03:24.320 align:start position:0%
722
+ this takes in 2D vectors and spits out
723
+
724
+
725
+ 00:03:24.320 --> 00:03:25.710 align:start position:0%
726
+ this takes in 2D vectors and spits out
727
+ numbers.
728
+
729
+ 00:03:25.710 --> 00:03:25.720 align:start position:0%
730
+ numbers.
731
+
732
+
733
+ 00:03:25.720 --> 00:03:27.390 align:start position:0%
734
+ numbers.
735
+ Thinking<00:03:26.000><c> about</c><00:03:26.440><c> grid</c><00:03:26.720><c> lines</c><00:03:27.000><c> remaining</c>
736
+
737
+ 00:03:27.390 --> 00:03:27.400 align:start position:0%
738
+ Thinking about grid lines remaining
739
+
740
+
741
+ 00:03:27.400 --> 00:03:29.150 align:start position:0%
742
+ Thinking about grid lines remaining
743
+ parallel<00:03:27.840><c> and</c><00:03:27.960><c> evenly</c><00:03:28.320><c> spaced</c><00:03:28.680><c> is</c><00:03:28.920><c> a</c><00:03:28.960><c> little</c>
744
+
745
+ 00:03:29.150 --> 00:03:29.160 align:start position:0%
746
+ parallel and evenly spaced is a little
747
+
748
+
749
+ 00:03:29.160 --> 00:03:30.550 align:start position:0%
750
+ parallel and evenly spaced is a little
751
+ bit<00:03:29.320><c> messy</c><00:03:29.840><c> due</c><00:03:29.960><c> to</c><00:03:30.040><c> all</c><00:03:30.240><c> of</c><00:03:30.320><c> the</c>
752
+
753
+ 00:03:30.550 --> 00:03:30.560 align:start position:0%
754
+ bit messy due to all of the
755
+
756
+
757
+ 00:03:30.560 --> 00:03:32.470 align:start position:0%
758
+ bit messy due to all of the
759
+ squishification<00:03:31.360><c> happening</c><00:03:31.720><c> here.</c><00:03:32.320><c> So,</c><00:03:32.400><c> in</c>
760
+
761
+ 00:03:32.470 --> 00:03:32.480 align:start position:0%
762
+ squishification happening here. So, in
763
+
764
+
765
+ 00:03:32.480 --> 00:03:34.550 align:start position:0%
766
+ squishification happening here. So, in
767
+ this<00:03:32.640><c> case,</c><00:03:33.160><c> the</c><00:03:33.320><c> visual</c><00:03:33.800><c> understanding</c><00:03:34.440><c> for</c>
768
+
769
+ 00:03:34.550 --> 00:03:34.560 align:start position:0%
770
+ this case, the visual understanding for
771
+
772
+
773
+ 00:03:34.560 --> 00:03:36.790 align:start position:0%
774
+ this case, the visual understanding for
775
+ what<00:03:34.760><c> linearity</c><00:03:35.360><c> means</c><00:03:36.120><c> is</c><00:03:36.240><c> that</c><00:03:36.400><c> if</c><00:03:36.520><c> you</c><00:03:36.640><c> have</c>
776
+
777
+ 00:03:36.790 --> 00:03:36.800 align:start position:0%
778
+ what linearity means is that if you have
779
+
780
+
781
+ 00:03:36.800 --> 00:03:39.310 align:start position:0%
782
+ what linearity means is that if you have
783
+ a<00:03:36.880><c> line</c><00:03:37.320><c> of</c><00:03:37.480><c> evenly</c><00:03:37.880><c> spaced</c><00:03:38.240><c> dots,</c><00:03:39.000><c> it</c><00:03:39.120><c> would</c>
784
+
785
+ 00:03:39.310 --> 00:03:39.320 align:start position:0%
786
+ a line of evenly spaced dots, it would
787
+
788
+
789
+ 00:03:39.320 --> 00:03:41.350 align:start position:0%
790
+ a line of evenly spaced dots, it would
791
+ remain<00:03:39.800><c> evenly</c><00:03:40.240><c> spaced</c><00:03:40.680><c> once</c><00:03:40.880><c> they're</c><00:03:41.040><c> mapped</c>
792
+
793
+ 00:03:41.350 --> 00:03:41.360 align:start position:0%
794
+ remain evenly spaced once they're mapped
795
+
796
+
797
+ 00:03:41.360 --> 00:03:43.670 align:start position:0%
798
+ remain evenly spaced once they're mapped
799
+ onto<00:03:41.560><c> the</c><00:03:41.680><c> number</c><00:03:41.920><c> line.</c>
800
+
801
+ 00:03:43.670 --> 00:03:43.680 align:start position:0%
802
+ onto the number line.
803
+
804
+
805
+ 00:03:43.680 --> 00:03:45.510 align:start position:0%
806
+ onto the number line.
807
+ One<00:03:43.800><c> of</c><00:03:43.840><c> these</c><00:03:44.040><c> transformations</c><00:03:44.960><c> is</c><00:03:45.120><c> encoded</c>
808
+
809
+ 00:03:45.510 --> 00:03:45.520 align:start position:0%
810
+ One of these transformations is encoded
811
+
812
+
813
+ 00:03:45.520 --> 00:03:48.150 align:start position:0%
814
+ One of these transformations is encoded
815
+ with<00:03:45.720><c> a</c><00:03:45.920><c> 1</c><00:03:46.240><c> by</c><00:03:46.400><c> 2</c><00:03:46.600><c> matrix,</c><00:03:47.360><c> each</c><00:03:47.560><c> of</c><00:03:47.640><c> whose</c><00:03:47.920><c> two</c>
816
+
817
+ 00:03:48.150 --> 00:03:48.160 align:start position:0%
818
+ with a 1 by 2 matrix, each of whose two
819
+
820
+
821
+ 00:03:48.160 --> 00:03:51.070 align:start position:0%
822
+ with a 1 by 2 matrix, each of whose two
823
+ columns<00:03:48.840><c> has</c><00:03:49.120><c> just</c><00:03:49.360><c> a</c><00:03:49.440><c> single</c><00:03:49.760><c> entry.</c><00:03:50.720><c> The</c><00:03:50.840><c> two</c>
824
+
825
+ 00:03:51.070 --> 00:03:51.080 align:start position:0%
826
+ columns has just a single entry. The two
827
+
828
+
829
+ 00:03:51.080 --> 00:03:52.510 align:start position:0%
830
+ columns has just a single entry. The two
831
+ columns<00:03:51.520><c> represent</c><00:03:51.920><c> where</c><00:03:52.040><c> the</c><00:03:52.120><c> basis</c>
832
+
833
+ 00:03:52.510 --> 00:03:52.520 align:start position:0%
834
+ columns represent where the basis
835
+
836
+
837
+ 00:03:52.520 --> 00:03:54.310 align:start position:0%
838
+ columns represent where the basis
839
+ vectors<00:03:52.880><c> land</c><00:03:53.520><c> and</c><00:03:53.760><c> each</c><00:03:53.960><c> one</c><00:03:54.120><c> of</c><00:03:54.160><c> those</c>
840
+
841
+ 00:03:54.310 --> 00:03:54.320 align:start position:0%
842
+ vectors land and each one of those
843
+
844
+
845
+ 00:03:54.320 --> 00:03:56.430 align:start position:0%
846
+ vectors land and each one of those
847
+ columns<00:03:54.720><c> requires</c><00:03:55.080><c> just</c><00:03:55.440><c> one</c><00:03:55.640><c> number,</c><00:03:56.320><c> the</c>
848
+
849
+ 00:03:56.430 --> 00:03:56.440 align:start position:0%
850
+ columns requires just one number, the
851
+
852
+
853
+ 00:03:56.440 --> 00:03:59.190 align:start position:0%
854
+ columns requires just one number, the
855
+ number<00:03:56.720><c> that</c><00:03:56.880><c> that</c><00:03:57.040><c> basis</c><00:03:57.400><c> vector</c><00:03:57.640><c> landed</c><00:03:58.000><c> on.</c>
856
+
857
+ 00:03:59.190 --> 00:03:59.200 align:start position:0%
858
+ number that that basis vector landed on.
859
+
860
+
861
+ 00:03:59.200 --> 00:04:00.630 align:start position:0%
862
+ number that that basis vector landed on.
863
+ This<00:03:59.360><c> is</c><00:03:59.480><c> actually</c><00:03:59.800><c> a</c><00:03:59.880><c> surprisingly</c>
864
+
865
+ 00:04:00.630 --> 00:04:00.640 align:start position:0%
866
+ This is actually a surprisingly
867
+
868
+
869
+ 00:04:00.640 --> 00:04:02.510 align:start position:0%
870
+ This is actually a surprisingly
871
+ meaningful<00:04:01.160><c> type</c><00:04:01.400><c> of</c><00:04:01.520><c> transformation</c><00:04:02.360><c> with</c>
872
+
873
+ 00:04:02.510 --> 00:04:02.520 align:start position:0%
874
+ meaningful type of transformation with
875
+
876
+
877
+ 00:04:02.520 --> 00:04:04.270 align:start position:0%
878
+ meaningful type of transformation with
879
+ close<00:04:02.760><c> ties</c><00:04:03.000><c> to</c><00:04:03.120><c> the</c><00:04:03.240><c> dot</c><00:04:03.480><c> product</c><00:04:04.120><c> and</c><00:04:04.240><c> I'll</c>
880
+
881
+ 00:04:04.270 --> 00:04:04.280 align:start position:0%
882
+ close ties to the dot product and I'll
883
+
884
+
885
+ 00:04:04.280 --> 00:04:06.270 align:start position:0%
886
+ close ties to the dot product and I'll
887
+ be<00:04:04.400><c> talking</c><00:04:04.680><c> about</c><00:04:04.920><c> that</c><00:04:05.040><c> next</c><00:04:05.240><c> video.</c>
888
+
889
+ 00:04:06.270 --> 00:04:06.280 align:start position:0%
890
+ be talking about that next video.
891
+
892
+
893
+ 00:04:06.280 --> 00:04:07.910 align:start position:0%
894
+ be talking about that next video.
895
+ Until<00:04:06.520><c> then,</c><00:04:07.040><c> I</c><00:04:07.120><c> encourage</c><00:04:07.480><c> you</c><00:04:07.600><c> to</c><00:04:07.720><c> play</c>
896
+
897
+ 00:04:07.910 --> 00:04:07.920 align:start position:0%
898
+ Until then, I encourage you to play
899
+
900
+
901
+ 00:04:07.920 --> 00:04:09.630 align:start position:0%
902
+ Until then, I encourage you to play
903
+ around<00:04:08.240><c> with</c><00:04:08.360><c> this</c><00:04:08.560><c> idea</c><00:04:08.880><c> on</c><00:04:09.000><c> your</c><00:04:09.160><c> own</c>
904
+
905
+ 00:04:09.630 --> 00:04:09.640 align:start position:0%
906
+ around with this idea on your own
907
+
908
+
909
+ 00:04:09.640 --> 00:04:10.990 align:start position:0%
910
+ around with this idea on your own
911
+ contemplating<00:04:10.240><c> the</c><00:04:10.320><c> meanings</c><00:04:10.640><c> of</c><00:04:10.760><c> things</c>
912
+
913
+ 00:04:10.990 --> 00:04:11.000 align:start position:0%
914
+ contemplating the meanings of things
915
+
916
+
917
+ 00:04:11.000 --> 00:04:12.830 align:start position:0%
918
+ contemplating the meanings of things
919
+ like<00:04:11.240><c> matrix</c><00:04:11.600><c> multiplication</c><00:04:12.360><c> and</c><00:04:12.560><c> linear</c>
920
+
921
+ 00:04:12.830 --> 00:04:12.840 align:start position:0%
922
+ like matrix multiplication and linear
923
+
924
+
925
+ 00:04:12.840 --> 00:04:14.990 align:start position:0%
926
+ like matrix multiplication and linear
927
+ systems<00:04:13.200><c> of</c><00:04:13.320><c> equations</c><00:04:14.200><c> in</c><00:04:14.320><c> the</c><00:04:14.400><c> context</c><00:04:14.880><c> of</c>
928
+
929
+ 00:04:14.990 --> 00:04:15.000 align:start position:0%
930
+ systems of equations in the context of
931
+
932
+
933
+ 00:04:15.000 --> 00:04:16.949 align:start position:0%
934
+ systems of equations in the context of
935
+ transformations<00:04:16.040><c> between</c><00:04:16.600><c> different</c>
936
+
937
+ 00:04:16.949 --> 00:04:16.959 align:start position:0%
938
+ transformations between different
939
+
940
+
941
+ 00:04:16.959 --> 00:04:20.560 align:start position:0%
942
+ transformations between different
943
+ dimensions.<00:04:18.079><c> Have</c><00:04:18.239><c> fun.</c>
944
+
3B1B_LinearAlgebra/08 - Nonsquare matrices as transformations between dimensions | Chapter 8, Essence of linear algebra.en.vtt ADDED
@@ -0,0 +1,944 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:11.240 --> 00:00:12.750 align:start position:0%
6
+
7
+ Hey<00:00:11.320><c> everyone,</c><00:00:11.960><c> I've</c><00:00:12.080><c> got</c><00:00:12.200><c> another</c><00:00:12.480><c> quick</c>
8
+
9
+ 00:00:12.750 --> 00:00:12.760 align:start position:0%
10
+ Hey everyone, I've got another quick
11
+
12
+
13
+ 00:00:12.760 --> 00:00:15.190 align:start position:0%
14
+ Hey everyone, I've got another quick
15
+ footnote<00:00:13.120><c> for</c><00:00:13.240><c> you</c><00:00:13.360><c> between</c><00:00:13.640><c> chapters</c><00:00:14.040><c> today.</c>
16
+
17
+ 00:00:15.190 --> 00:00:15.200 align:start position:0%
18
+ footnote for you between chapters today.
19
+
20
+
21
+ 00:00:15.200 --> 00:00:16.030 align:start position:0%
22
+ footnote for you between chapters today.
23
+ When<00:00:15.280><c> I've</c><00:00:15.360><c> talked</c><00:00:15.600><c> about</c><00:00:15.800><c> linear</c>
24
+
25
+ 00:00:16.030 --> 00:00:16.040 align:start position:0%
26
+ When I've talked about linear
27
+
28
+
29
+ 00:00:16.040 --> 00:00:17.830 align:start position:0%
30
+ When I've talked about linear
31
+ transformation<00:00:16.680><c> so</c><00:00:16.840><c> far,</c><00:00:17.320><c> I've</c><00:00:17.440><c> only</c><00:00:17.640><c> really</c>
32
+
33
+ 00:00:17.830 --> 00:00:17.840 align:start position:0%
34
+ transformation so far, I've only really
35
+
36
+
37
+ 00:00:17.840 --> 00:00:19.550 align:start position:0%
38
+ transformation so far, I've only really
39
+ talked<00:00:18.120><c> about</c><00:00:18.320><c> transformations</c><00:00:19.000><c> from</c><00:00:19.240><c> 2D</c>
40
+
41
+ 00:00:19.550 --> 00:00:19.560 align:start position:0%
42
+ talked about transformations from 2D
43
+
44
+
45
+ 00:00:19.560 --> 00:00:21.830 align:start position:0%
46
+ talked about transformations from 2D
47
+ vectors<00:00:19.960><c> to</c><00:00:20.120><c> other</c><00:00:20.360><c> 2D</c><00:00:20.600><c> vectors</c><00:00:21.320><c> represented</c>
48
+
49
+ 00:00:21.830 --> 00:00:21.840 align:start position:0%
50
+ vectors to other 2D vectors represented
51
+
52
+
53
+ 00:00:21.840 --> 00:00:24.590 align:start position:0%
54
+ vectors to other 2D vectors represented
55
+ with<00:00:22.160><c> 2</c><00:00:22.280><c> by</c><00:00:22.440><c> 2</c><00:00:22.600><c> matrices</c><00:00:23.520><c> or</c><00:00:23.640><c> from</c><00:00:23.800><c> 3D</c><00:00:24.160><c> vectors</c>
56
+
57
+ 00:00:24.590 --> 00:00:24.600 align:start position:0%
58
+ with 2 by 2 matrices or from 3D vectors
59
+
60
+
61
+ 00:00:24.600 --> 00:00:26.990 align:start position:0%
62
+ with 2 by 2 matrices or from 3D vectors
63
+ to<00:00:24.720><c> other</c><00:00:24.920><c> 3D</c><00:00:25.280><c> vectors</c><00:00:26.120><c> represented</c><00:00:26.640><c> with</c><00:00:26.800><c> 3</c>
64
+
65
+ 00:00:26.990 --> 00:00:27.000 align:start position:0%
66
+ to other 3D vectors represented with 3
67
+
68
+
69
+ 00:00:27.000 --> 00:00:28.990 align:start position:0%
70
+ to other 3D vectors represented with 3
71
+ by<00:00:27.160><c> 3</c><00:00:27.400><c> matrices.</c>
72
+
73
+ 00:00:28.990 --> 00:00:29.000 align:start position:0%
74
+ by 3 matrices.
75
+
76
+
77
+ 00:00:29.000 --> 00:00:30.630 align:start position:0%
78
+ by 3 matrices.
79
+ But<00:00:29.200><c> several</c><00:00:29.480><c> commenters</c><00:00:29.960><c> have</c><00:00:30.080><c> asked</c><00:00:30.320><c> about</c>
80
+
81
+ 00:00:30.630 --> 00:00:30.640 align:start position:0%
82
+ But several commenters have asked about
83
+
84
+
85
+ 00:00:30.640 --> 00:00:32.350 align:start position:0%
86
+ But several commenters have asked about
87
+ non-square<00:00:31.200><c> matrices,</c><00:00:31.920><c> so</c><00:00:32.040><c> I</c><00:00:32.080><c> thought</c><00:00:32.279><c> I'd</c>
88
+
89
+ 00:00:32.350 --> 00:00:32.360 align:start position:0%
90
+ non-square matrices, so I thought I'd
91
+
92
+
93
+ 00:00:32.360 --> 00:00:33.670 align:start position:0%
94
+ non-square matrices, so I thought I'd
95
+ take<00:00:32.520><c> a</c><00:00:32.599><c> moment</c><00:00:32.840><c> to</c><00:00:32.920><c> just</c><00:00:33.160><c> show</c><00:00:33.360><c> what</c><00:00:33.480><c> those</c>
96
+
97
+ 00:00:33.670 --> 00:00:33.680 align:start position:0%
98
+ take a moment to just show what those
99
+
100
+
101
+ 00:00:33.680 --> 00:00:35.270 align:start position:0%
102
+ take a moment to just show what those
103
+ mean<00:00:33.840><c> geometrically.</c>
104
+
105
+ 00:00:35.270 --> 00:00:35.280 align:start position:0%
106
+ mean geometrically.
107
+
108
+
109
+ 00:00:35.280 --> 00:00:37.030 align:start position:0%
110
+ mean geometrically.
111
+ By<00:00:35.400><c> now</c><00:00:35.600><c> in</c><00:00:35.680><c> the</c><00:00:35.760><c> series,</c><00:00:36.320><c> you</c><00:00:36.440><c> actually</c><00:00:36.800><c> have</c>
112
+
113
+ 00:00:37.030 --> 00:00:37.040 align:start position:0%
114
+ By now in the series, you actually have
115
+
116
+
117
+ 00:00:37.040 --> 00:00:38.590 align:start position:0%
118
+ By now in the series, you actually have
119
+ most<00:00:37.280><c> of</c><00:00:37.360><c> the</c><00:00:37.440><c> background</c><00:00:37.880><c> you</c><00:00:37.960><c> need</c><00:00:38.200><c> to</c><00:00:38.320><c> start</c>
120
+
121
+ 00:00:38.590 --> 00:00:38.600 align:start position:0%
122
+ most of the background you need to start
123
+
124
+
125
+ 00:00:38.600 --> 00:00:40.110 align:start position:0%
126
+ most of the background you need to start
127
+ pondering<00:00:39.040><c> a</c><00:00:39.080><c> question</c><00:00:39.480><c> like</c><00:00:39.640><c> this</c><00:00:39.840><c> on</c><00:00:39.960><c> your</c>
128
+
129
+ 00:00:40.110 --> 00:00:40.120 align:start position:0%
130
+ pondering a question like this on your
131
+
132
+
133
+ 00:00:40.120 --> 00:00:41.630 align:start position:0%
134
+ pondering a question like this on your
135
+ own,<00:00:40.640><c> but</c><00:00:40.760><c> I'll</c><00:00:40.880><c> start</c><00:00:41.080><c> talking</c><00:00:41.360><c> through</c><00:00:41.520><c> it</c>
136
+
137
+ 00:00:41.630 --> 00:00:41.640 align:start position:0%
138
+ own, but I'll start talking through it
139
+
140
+
141
+ 00:00:41.640 --> 00:00:44.030 align:start position:0%
142
+ own, but I'll start talking through it
143
+ just<00:00:41.800><c> to</c><00:00:41.840><c> give</c><00:00:42.000><c> a</c><00:00:42.040><c> little</c><00:00:42.280><c> mental</c><00:00:42.640><c> momentum.</c>
144
+
145
+ 00:00:44.030 --> 00:00:44.040 align:start position:0%
146
+ just to give a little mental momentum.
147
+
148
+
149
+ 00:00:44.040 --> 00:00:45.630 align:start position:0%
150
+ just to give a little mental momentum.
151
+ It's<00:00:44.200><c> perfectly</c><00:00:44.640><c> reasonable</c><00:00:45.080><c> to</c><00:00:45.200><c> talk</c><00:00:45.400><c> about</c>
152
+
153
+ 00:00:45.630 --> 00:00:45.640 align:start position:0%
154
+ It's perfectly reasonable to talk about
155
+
156
+
157
+ 00:00:45.640 --> 00:00:48.510 align:start position:0%
158
+ It's perfectly reasonable to talk about
159
+ transformations<00:00:46.520><c> between</c><00:00:47.160><c> dimensions,</c><00:00:48.320><c> such</c>
160
+
161
+ 00:00:48.510 --> 00:00:48.520 align:start position:0%
162
+ transformations between dimensions, such
163
+
164
+
165
+ 00:00:48.520 --> 00:00:50.750 align:start position:0%
166
+ transformations between dimensions, such
167
+ as<00:00:48.640><c> one</c><00:00:48.800><c> that</c><00:00:48.920><c> takes</c><00:00:49.200><c> 2D</c><00:00:49.600><c> vectors</c><00:00:50.240><c> to</c><00:00:50.360><c> 3D</c>
168
+
169
+ 00:00:50.750 --> 00:00:50.760 align:start position:0%
170
+ as one that takes 2D vectors to 3D
171
+
172
+
173
+ 00:00:50.760 --> 00:00:53.270 align:start position:0%
174
+ as one that takes 2D vectors to 3D
175
+ vectors.<00:00:51.800><c> Again,</c><00:00:52.360><c> what</c><00:00:52.520><c> makes</c><00:00:52.800><c> one</c><00:00:52.960><c> of</c><00:00:53.040><c> these</c>
176
+
177
+ 00:00:53.270 --> 00:00:53.280 align:start position:0%
178
+ vectors. Again, what makes one of these
179
+
180
+
181
+ 00:00:53.280 --> 00:00:55.190 align:start position:0%
182
+ vectors. Again, what makes one of these
183
+ linear<00:00:54.040><c> is</c><00:00:54.200><c> that</c><00:00:54.320><c> grid</c><00:00:54.600><c> lines</c><00:00:54.880><c> remain</c>
184
+
185
+ 00:00:55.190 --> 00:00:55.200 align:start position:0%
186
+ linear is that grid lines remain
187
+
188
+
189
+ 00:00:55.200 --> 00:00:57.430 align:start position:0%
190
+ linear is that grid lines remain
191
+ parallel<00:00:55.840><c> and</c><00:00:56.040><c> evenly</c><00:00:56.400><c> spaced</c><00:00:57.080><c> and</c><00:00:57.200><c> that</c><00:00:57.320><c> the</c>
192
+
193
+ 00:00:57.430 --> 00:00:57.440 align:start position:0%
194
+ parallel and evenly spaced and that the
195
+
196
+
197
+ 00:00:57.440 --> 00:00:59.990 align:start position:0%
198
+ parallel and evenly spaced and that the
199
+ origin<00:00:58.040><c> maps</c><00:00:58.320><c> to</c><00:00:58.440><c> the</c><00:00:58.600><c> origin.</c>
200
+
201
+ 00:00:59.990 --> 00:01:00.000 align:start position:0%
202
+ origin maps to the origin.
203
+
204
+
205
+ 00:01:00.000 --> 00:01:01.510 align:start position:0%
206
+ origin maps to the origin.
207
+ What<00:01:00.120><c> I</c><00:01:00.160><c> have</c><00:01:00.360><c> pictured</c><00:01:00.640><c> here</c><00:01:00.920><c> is</c><00:01:01.040><c> the</c><00:01:01.160><c> input</c>
208
+
209
+ 00:01:01.510 --> 00:01:01.520 align:start position:0%
210
+ What I have pictured here is the input
211
+
212
+
213
+ 00:01:01.520 --> 00:01:03.190 align:start position:0%
214
+ What I have pictured here is the input
215
+ space<00:01:01.800><c> on</c><00:01:01.880><c> the</c><00:01:02.000><c> left,</c><00:01:02.400><c> which</c><00:01:02.560><c> is</c><00:01:02.640><c> just</c><00:01:02.920><c> 2D</c>
216
+
217
+ 00:01:03.190 --> 00:01:03.200 align:start position:0%
218
+ space on the left, which is just 2D
219
+
220
+
221
+ 00:01:03.200 --> 00:01:04.829 align:start position:0%
222
+ space on the left, which is just 2D
223
+ space,<00:01:03.920><c> and</c><00:01:04.040><c> the</c><00:01:04.239><c> output</c><00:01:04.640><c> of</c><00:01:04.720><c> the</c>
224
+
225
+ 00:01:04.829 --> 00:01:04.839 align:start position:0%
226
+ space, and the output of the
227
+
228
+
229
+ 00:01:04.839 --> 00:01:06.950 align:start position:0%
230
+ space, and the output of the
231
+ transformation<00:01:05.560><c> shown</c><00:01:05.800><c> on</c><00:01:05.880><c> the</c><00:01:06.000><c> right.</c>
232
+
233
+ 00:01:06.950 --> 00:01:06.960 align:start position:0%
234
+ transformation shown on the right.
235
+
236
+
237
+ 00:01:06.960 --> 00:01:08.430 align:start position:0%
238
+ transformation shown on the right.
239
+ The<00:01:07.040><c> reason</c><00:01:07.320><c> I'm</c><00:01:07.400><c> not</c><00:01:07.600><c> showing</c><00:01:07.840><c> the</c><00:01:08.000><c> inputs</c>
240
+
241
+ 00:01:08.430 --> 00:01:08.440 align:start position:0%
242
+ The reason I'm not showing the inputs
243
+
244
+
245
+ 00:01:08.440 --> 00:01:10.270 align:start position:0%
246
+ The reason I'm not showing the inputs
247
+ move<00:01:08.760><c> over</c><00:01:09.000><c> to</c><00:01:09.120><c> the</c><00:01:09.280><c> outputs</c><00:01:09.680><c> like</c><00:01:09.840><c> I</c><00:01:09.920><c> usually</c>
248
+
249
+ 00:01:10.270 --> 00:01:10.280 align:start position:0%
250
+ move over to the outputs like I usually
251
+
252
+
253
+ 00:01:10.280 --> 00:01:12.830 align:start position:0%
254
+ move over to the outputs like I usually
255
+ do<00:01:10.800><c> is</c><00:01:10.880><c> not</c><00:01:11.080><c> just</c><00:01:11.360><c> animation</c><00:01:11.840><c> laziness.</c><00:01:12.680><c> It's</c>
256
+
257
+ 00:01:12.830 --> 00:01:12.840 align:start position:0%
258
+ do is not just animation laziness. It's
259
+
260
+
261
+ 00:01:12.840 --> 00:01:15.150 align:start position:0%
262
+ do is not just animation laziness. It's
263
+ worth<00:01:13.040><c> emphasizing</c><00:01:13.640><c> that</c><00:01:13.800><c> 2D</c><00:01:14.120><c> vector</c><00:01:14.440><c> inputs</c>
264
+
265
+ 00:01:15.150 --> 00:01:15.160 align:start position:0%
266
+ worth emphasizing that 2D vector inputs
267
+
268
+
269
+ 00:01:15.160 --> 00:01:17.190 align:start position:0%
270
+ worth emphasizing that 2D vector inputs
271
+ are<00:01:15.320><c> very</c><00:01:15.640><c> different</c><00:01:15.960><c> animals</c><00:01:16.440><c> from</c><00:01:16.600><c> these</c><00:01:16.840><c> 3D</c>
272
+
273
+ 00:01:17.190 --> 00:01:17.200 align:start position:0%
274
+ are very different animals from these 3D
275
+
276
+
277
+ 00:01:17.200 --> 00:01:19.230 align:start position:0%
278
+ are very different animals from these 3D
279
+ vector<00:01:17.520><c> outputs</c><00:01:18.200><c> living</c><00:01:18.520><c> in</c><00:01:18.640><c> a</c><00:01:18.680><c> completely</c>
280
+
281
+ 00:01:19.230 --> 00:01:19.240 align:start position:0%
282
+ vector outputs living in a completely
283
+
284
+
285
+ 00:01:19.240 --> 00:01:21.710 align:start position:0%
286
+ vector outputs living in a completely
287
+ separate<00:01:19.720><c> unconnected</c><00:01:20.400><c> space.</c>
288
+
289
+ 00:01:21.710 --> 00:01:21.720 align:start position:0%
290
+ separate unconnected space.
291
+
292
+
293
+ 00:01:21.720 --> 00:01:23.190 align:start position:0%
294
+ separate unconnected space.
295
+ Encoding<00:01:22.160><c> one</c><00:01:22.320><c> of</c><00:01:22.360><c> these</c><00:01:22.520><c> transformations</c>
296
+
297
+ 00:01:23.190 --> 00:01:23.200 align:start position:0%
298
+ Encoding one of these transformations
299
+
300
+
301
+ 00:01:23.200 --> 00:01:25.430 align:start position:0%
302
+ Encoding one of these transformations
303
+ with<00:01:23.360><c> a</c><00:01:23.400><c> matrix</c><00:01:24.160><c> is</c><00:01:24.560><c> really</c><00:01:24.840><c> just</c><00:01:25.080><c> the</c><00:01:25.160><c> same</c>
304
+
305
+ 00:01:25.430 --> 00:01:25.440 align:start position:0%
306
+ with a matrix is really just the same
307
+
308
+
309
+ 00:01:25.440 --> 00:01:27.350 align:start position:0%
310
+ with a matrix is really just the same
311
+ thing<00:01:25.640><c> as</c><00:01:25.760><c> what</c><00:01:25.880><c> we've</c><00:01:26.040><c> done</c><00:01:26.240><c> before.</c><00:01:27.240><c> You</c>
312
+
313
+ 00:01:27.350 --> 00:01:27.360 align:start position:0%
314
+ thing as what we've done before. You
315
+
316
+
317
+ 00:01:27.360 --> 00:01:29.430 align:start position:0%
318
+ thing as what we've done before. You
319
+ look<00:01:27.520><c> at</c><00:01:27.640><c> where</c><00:01:27.840><c> each</c><00:01:28.080><c> basis</c><00:01:28.480><c> vector</c><00:01:28.760><c> lands</c>
320
+
321
+ 00:01:29.430 --> 00:01:29.440 align:start position:0%
322
+ look at where each basis vector lands
323
+
324
+
325
+ 00:01:29.440 --> 00:01:31.150 align:start position:0%
326
+ look at where each basis vector lands
327
+ and<00:01:29.560><c> write</c><00:01:29.760><c> the</c><00:01:29.920><c> coordinates</c><00:01:30.560><c> of</c><00:01:30.640><c> the</c><00:01:30.760><c> landing</c>
328
+
329
+ 00:01:31.150 --> 00:01:31.160 align:start position:0%
330
+ and write the coordinates of the landing
331
+
332
+
333
+ 00:01:31.160 --> 00:01:33.750 align:start position:0%
334
+ and write the coordinates of the landing
335
+ spots<00:01:31.720><c> as</c><00:01:31.880><c> the</c><00:01:32.000><c> columns</c><00:01:32.440><c> of</c><00:01:32.560><c> a</c><00:01:32.600><c> matrix.</c><00:01:33.680><c> For</c>
336
+
337
+ 00:01:33.750 --> 00:01:33.760 align:start position:0%
338
+ spots as the columns of a matrix. For
339
+
340
+
341
+ 00:01:33.760 --> 00:01:35.830 align:start position:0%
342
+ spots as the columns of a matrix. For
343
+ example,<00:01:34.200><c> what</c><00:01:34.360><c> you're</c><00:01:34.480><c> looking</c><00:01:34.760><c> at</c><00:01:34.880><c> here</c><00:01:35.680><c> is</c>
344
+
345
+ 00:01:35.830 --> 00:01:35.840 align:start position:0%
346
+ example, what you're looking at here is
347
+
348
+
349
+ 00:01:35.840 --> 00:01:37.910 align:start position:0%
350
+ example, what you're looking at here is
351
+ an<00:01:36.000><c> output</c><00:01:36.320><c> of</c><00:01:36.400><c> a</c><00:01:36.440><c> transformation</c><00:01:37.440><c> that</c><00:01:37.560><c> takes</c>
352
+
353
+ 00:01:37.910 --> 00:01:37.920 align:start position:0%
354
+ an output of a transformation that takes
355
+
356
+
357
+ 00:01:37.920 --> 00:01:42.350 align:start position:0%
358
+ an output of a transformation that takes
359
+ I<00:01:38.120><c> hat</c><00:01:38.600><c> to</c><00:01:38.680><c> the</c><00:01:38.800><c> coordinates</c><00:01:39.520><c> 2</c><00:01:40.000><c> -1</c><00:01:40.760><c> -2</c><00:01:41.920><c> and</c><00:01:42.120><c> J</c>
360
+
361
+ 00:01:42.350 --> 00:01:42.360 align:start position:0%
362
+ I hat to the coordinates 2 -1 -2 and J
363
+
364
+
365
+ 00:01:42.360 --> 00:01:47.870 align:start position:0%
366
+ I hat to the coordinates 2 -1 -2 and J
367
+ hat<00:01:42.720><c> to</c><00:01:42.800><c> the</c><00:01:42.920><c> coordinates</c><00:01:43.560><c> 0</c><00:01:44.200><c> 1</c><00:01:44.760><c> 1.</c>
368
+
369
+ 00:01:47.870 --> 00:01:47.880 align:start position:0%
370
+
371
+
372
+
373
+ 00:01:47.880 --> 00:01:49.790 align:start position:0%
374
+
375
+ Notice,<00:01:48.320><c> this</c><00:01:48.560><c> means</c><00:01:48.840><c> the</c><00:01:48.960><c> matrix</c><00:01:49.400><c> encoding</c>
376
+
377
+ 00:01:49.790 --> 00:01:49.800 align:start position:0%
378
+ Notice, this means the matrix encoding
379
+
380
+
381
+ 00:01:49.800 --> 00:01:52.190 align:start position:0%
382
+ Notice, this means the matrix encoding
383
+ our<00:01:49.880><c> transformation</c><00:01:51.000><c> has</c><00:01:51.200><c> three</c><00:01:51.520><c> rows</c><00:01:52.000><c> and</c>
384
+
385
+ 00:01:52.190 --> 00:01:52.200 align:start position:0%
386
+ our transformation has three rows and
387
+
388
+
389
+ 00:01:52.200 --> 00:01:54.150 align:start position:0%
390
+ our transformation has three rows and
391
+ two<00:01:52.360><c> columns,</c><00:01:53.240><c> which</c><00:01:53.560><c> to</c><00:01:53.680><c> use</c><00:01:53.800><c> standard</c>
392
+
393
+ 00:01:54.150 --> 00:01:54.160 align:start position:0%
394
+ two columns, which to use standard
395
+
396
+
397
+ 00:01:54.160 --> 00:01:57.750 align:start position:0%
398
+ two columns, which to use standard
399
+ terminology<00:01:55.080><c> makes</c><00:01:55.280><c> it</c><00:01:55.400><c> a</c><00:01:55.480><c> 3</c><00:01:55.800><c> by</c><00:01:55.960><c> 2</c><00:01:56.200><c> matrix.</c>
400
+
401
+ 00:01:57.750 --> 00:01:57.760 align:start position:0%
402
+ terminology makes it a 3 by 2 matrix.
403
+
404
+
405
+ 00:01:57.760 --> 00:01:59.630 align:start position:0%
406
+ terminology makes it a 3 by 2 matrix.
407
+ In<00:01:57.920><c> the</c><00:01:58.000><c> language</c><00:01:58.400><c> of</c><00:01:58.480><c> last</c><00:01:58.720><c> video,</c><00:01:59.480><c> the</c>
408
+
409
+ 00:01:59.630 --> 00:01:59.640 align:start position:0%
410
+ In the language of last video, the
411
+
412
+
413
+ 00:01:59.640 --> 00:02:01.910 align:start position:0%
414
+ In the language of last video, the
415
+ column<00:02:00.240><c> space</c><00:02:00.680><c> of</c><00:02:00.760><c> this</c><00:02:00.960><c> matrix,</c><00:02:01.560><c> the</c><00:02:01.680><c> place</c>
416
+
417
+ 00:02:01.910 --> 00:02:01.920 align:start position:0%
418
+ column space of this matrix, the place
419
+
420
+
421
+ 00:02:01.920 --> 00:02:04.110 align:start position:0%
422
+ column space of this matrix, the place
423
+ where<00:02:02.040><c> all</c><00:02:02.200><c> the</c><00:02:02.280><c> vectors</c><00:02:02.680><c> land,</c><00:02:03.360><c> is</c><00:02:03.520><c> a</c><00:02:03.600><c> 2D</c>
424
+
425
+ 00:02:04.110 --> 00:02:04.120 align:start position:0%
426
+ where all the vectors land, is a 2D
427
+
428
+
429
+ 00:02:04.120 --> 00:02:06.310 align:start position:0%
430
+ where all the vectors land, is a 2D
431
+ plane<00:02:04.520><c> slicing</c><00:02:05.000><c> through</c><00:02:05.200><c> the</c><00:02:05.360><c> origin</c><00:02:05.840><c> of</c><00:02:05.960><c> 3D</c>
432
+
433
+ 00:02:06.310 --> 00:02:06.320 align:start position:0%
434
+ plane slicing through the origin of 3D
435
+
436
+
437
+ 00:02:06.320 --> 00:02:07.350 align:start position:0%
438
+ plane slicing through the origin of 3D
439
+ space.
440
+
441
+ 00:02:07.350 --> 00:02:07.360 align:start position:0%
442
+ space.
443
+
444
+
445
+ 00:02:07.360 --> 00:02:09.669 align:start position:0%
446
+ space.
447
+ But<00:02:07.480><c> the</c><00:02:07.560><c> matrix</c><00:02:07.920><c> is</c><00:02:08.039><c> still</c><00:02:08.399><c> full</c><00:02:08.800><c> rank</c><00:02:09.399><c> since</c>
448
+
449
+ 00:02:09.669 --> 00:02:09.679 align:start position:0%
450
+ But the matrix is still full rank since
451
+
452
+
453
+ 00:02:09.679 --> 00:02:11.550 align:start position:0%
454
+ But the matrix is still full rank since
455
+ the<00:02:09.800><c> number</c><00:02:10.080><c> of</c><00:02:10.200><c> dimensions</c><00:02:10.800><c> in</c><00:02:11.000><c> this</c><00:02:11.200><c> column</c>
456
+
457
+ 00:02:11.550 --> 00:02:11.560 align:start position:0%
458
+ the number of dimensions in this column
459
+
460
+
461
+ 00:02:11.560 --> 00:02:13.430 align:start position:0%
462
+ the number of dimensions in this column
463
+ space<00:02:12.240><c> is</c><00:02:12.400><c> the</c><00:02:12.480><c> same</c><00:02:12.880><c> as</c><00:02:13.000><c> the</c><00:02:13.120><c> number</c><00:02:13.360><c> of</c>
464
+
465
+ 00:02:13.430 --> 00:02:13.440 align:start position:0%
466
+ space is the same as the number of
467
+
468
+
469
+ 00:02:13.440 --> 00:02:17.390 align:start position:0%
470
+ space is the same as the number of
471
+ dimensions<00:02:14.200><c> of</c><00:02:14.360><c> the</c><00:02:14.520><c> input</c><00:02:14.960><c> space.</c>
472
+
473
+ 00:02:17.390 --> 00:02:17.400 align:start position:0%
474
+ dimensions of the input space.
475
+
476
+
477
+ 00:02:17.400 --> 00:02:19.350 align:start position:0%
478
+ dimensions of the input space.
479
+ So,<00:02:17.600><c> if</c><00:02:17.760><c> you</c><00:02:17.880><c> see</c><00:02:18.040><c> a</c><00:02:18.120><c> 3</c><00:02:18.200><c> by</c><00:02:18.480><c> 2</c><00:02:18.640><c> matrix</c><00:02:19.080><c> out</c><00:02:19.240><c> in</c>
480
+
481
+ 00:02:19.350 --> 00:02:19.360 align:start position:0%
482
+ So, if you see a 3 by 2 matrix out in
483
+
484
+
485
+ 00:02:19.360 --> 00:02:21.150 align:start position:0%
486
+ So, if you see a 3 by 2 matrix out in
487
+ the<00:02:19.440><c> wild,</c><00:02:20.120><c> you</c><00:02:20.240><c> can</c><00:02:20.400><c> know</c><00:02:20.560><c> that</c><00:02:20.720><c> it</c><00:02:20.840><c> has</c><00:02:21.000><c> the</c>
488
+
489
+ 00:02:21.150 --> 00:02:21.160 align:start position:0%
490
+ the wild, you can know that it has the
491
+
492
+
493
+ 00:02:21.160 --> 00:02:23.470 align:start position:0%
494
+ the wild, you can know that it has the
495
+ geometric<00:02:21.720><c> interpretation</c><00:02:22.760><c> of</c><00:02:22.920><c> mapping</c><00:02:23.320><c> two</c>
496
+
497
+ 00:02:23.470 --> 00:02:23.480 align:start position:0%
498
+ geometric interpretation of mapping two
499
+
500
+
501
+ 00:02:23.480 --> 00:02:26.030 align:start position:0%
502
+ geometric interpretation of mapping two
503
+ dimensions<00:02:24.320><c> to</c><00:02:24.440><c> three</c><00:02:24.640><c> dimensions</c><00:02:25.640><c> since</c><00:02:25.920><c> the</c>
504
+
505
+ 00:02:26.030 --> 00:02:26.040 align:start position:0%
506
+ dimensions to three dimensions since the
507
+
508
+
509
+ 00:02:26.040 --> 00:02:27.750 align:start position:0%
510
+ dimensions to three dimensions since the
511
+ two<00:02:26.240><c> columns</c><00:02:26.720><c> indicate</c><00:02:27.080><c> that</c><00:02:27.200><c> the</c><00:02:27.360><c> input</c>
512
+
513
+ 00:02:27.750 --> 00:02:27.760 align:start position:0%
514
+ two columns indicate that the input
515
+
516
+
517
+ 00:02:27.760 --> 00:02:30.030 align:start position:0%
518
+ two columns indicate that the input
519
+ space<00:02:28.160><c> has</c><00:02:28.400><c> two</c><00:02:28.600><c> basis</c><00:02:28.960><c> vectors</c><00:02:29.840><c> and</c><00:02:29.960><c> the</c>
520
+
521
+ 00:02:30.030 --> 00:02:30.040 align:start position:0%
522
+ space has two basis vectors and the
523
+
524
+
525
+ 00:02:30.040 --> 00:02:32.230 align:start position:0%
526
+ space has two basis vectors and the
527
+ three<00:02:30.320><c> rows</c><00:02:31.000><c> indicate</c><00:02:31.440><c> that</c><00:02:31.560><c> the</c><00:02:31.720><c> landing</c>
528
+
529
+ 00:02:32.230 --> 00:02:32.240 align:start position:0%
530
+ three rows indicate that the landing
531
+
532
+
533
+ 00:02:32.240 --> 00:02:34.750 align:start position:0%
534
+ three rows indicate that the landing
535
+ spots<00:02:32.680><c> for</c><00:02:32.840><c> each</c><00:02:33.080><c> of</c><00:02:33.160><c> those</c><00:02:33.400><c> basis</c><00:02:33.760><c> vectors</c><00:02:34.600><c> is</c>
536
+
537
+ 00:02:34.750 --> 00:02:34.760 align:start position:0%
538
+ spots for each of those basis vectors is
539
+
540
+
541
+ 00:02:34.760 --> 00:02:36.230 align:start position:0%
542
+ spots for each of those basis vectors is
543
+ described<00:02:35.280><c> with</c><00:02:35.480><c> three</c><00:02:35.840><c> separate</c>
544
+
545
+ 00:02:36.230 --> 00:02:36.240 align:start position:0%
546
+ described with three separate
547
+
548
+
549
+ 00:02:36.240 --> 00:02:37.830 align:start position:0%
550
+ described with three separate
551
+ coordinates.
552
+
553
+ 00:02:37.830 --> 00:02:37.840 align:start position:0%
554
+ coordinates.
555
+
556
+
557
+ 00:02:37.840 --> 00:02:40.110 align:start position:0%
558
+ coordinates.
559
+ Likewise,<00:02:38.400><c> if</c><00:02:38.560><c> you</c><00:02:38.640><c> see</c><00:02:38.840><c> a</c><00:02:38.960><c> 2</c><00:02:39.200><c> by</c><00:02:39.360><c> 3</c><00:02:39.640><c> matrix</c>
560
+
561
+ 00:02:40.110 --> 00:02:40.120 align:start position:0%
562
+ Likewise, if you see a 2 by 3 matrix
563
+
564
+
565
+ 00:02:40.120 --> 00:02:42.190 align:start position:0%
566
+ Likewise, if you see a 2 by 3 matrix
567
+ with<00:02:40.400><c> two</c><00:02:40.560><c> rows</c><00:02:40.840><c> and</c><00:02:40.960><c> three</c><00:02:41.160><c> columns,</c><00:02:42.040><c> what</c><00:02:42.160><c> do</c>
568
+
569
+ 00:02:42.190 --> 00:02:42.200 align:start position:0%
570
+ with two rows and three columns, what do
571
+
572
+
573
+ 00:02:42.200 --> 00:02:43.630 align:start position:0%
574
+ with two rows and three columns, what do
575
+ you<00:02:42.280><c> think</c><00:02:42.440><c> that</c><00:02:42.640><c> means?</c>
576
+
577
+ 00:02:43.630 --> 00:02:43.640 align:start position:0%
578
+ you think that means?
579
+
580
+
581
+ 00:02:43.640 --> 00:02:45.630 align:start position:0%
582
+ you think that means?
583
+ Well,<00:02:44.240><c> the</c><00:02:44.360><c> three</c><00:02:44.600><c> columns</c><00:02:45.120><c> indicate</c><00:02:45.520><c> that</c>
584
+
585
+ 00:02:45.630 --> 00:02:45.640 align:start position:0%
586
+ Well, the three columns indicate that
587
+
588
+
589
+ 00:02:45.640 --> 00:02:46.950 align:start position:0%
590
+ Well, the three columns indicate that
591
+ you're<00:02:45.760><c> starting</c><00:02:46.120><c> in</c><00:02:46.240><c> a</c><00:02:46.280><c> space</c><00:02:46.640><c> that</c><00:02:46.800><c> has</c>
592
+
593
+ 00:02:46.950 --> 00:02:46.960 align:start position:0%
594
+ you're starting in a space that has
595
+
596
+
597
+ 00:02:46.960 --> 00:02:48.870 align:start position:0%
598
+ you're starting in a space that has
599
+ three<00:02:47.240><c> basis</c><00:02:47.600><c> vectors,</c><00:02:48.320><c> so</c><00:02:48.440><c> we're</c><00:02:48.520><c> starting</c>
600
+
601
+ 00:02:48.870 --> 00:02:48.880 align:start position:0%
602
+ three basis vectors, so we're starting
603
+
604
+
605
+ 00:02:48.880 --> 00:02:51.190 align:start position:0%
606
+ three basis vectors, so we're starting
607
+ in<00:02:48.960><c> three</c><00:02:49.120><c> dimensions.</c><00:02:50.280><c> And</c><00:02:50.400><c> the</c><00:02:50.480><c> two</c><00:02:50.720><c> rows</c>
608
+
609
+ 00:02:51.190 --> 00:02:51.200 align:start position:0%
610
+ in three dimensions. And the two rows
611
+
612
+
613
+ 00:02:51.200 --> 00:02:53.110 align:start position:0%
614
+ in three dimensions. And the two rows
615
+ indicate<00:02:51.640><c> that</c><00:02:51.760><c> the</c><00:02:51.880><c> landing</c><00:02:52.360><c> spot</c><00:02:52.720><c> for</c><00:02:52.920><c> each</c>
616
+
617
+ 00:02:53.110 --> 00:02:53.120 align:start position:0%
618
+ indicate that the landing spot for each
619
+
620
+
621
+ 00:02:53.120 --> 00:02:55.110 align:start position:0%
622
+ indicate that the landing spot for each
623
+ of<00:02:53.200><c> those</c><00:02:53.440><c> three</c><00:02:53.680><c> basis</c><00:02:54.080><c> vectors</c><00:02:54.960><c> is</c>
624
+
625
+ 00:02:55.110 --> 00:02:55.120 align:start position:0%
626
+ of those three basis vectors is
627
+
628
+
629
+ 00:02:55.120 --> 00:02:57.590 align:start position:0%
630
+ of those three basis vectors is
631
+ described<00:02:55.680><c> with</c><00:02:55.920><c> only</c><00:02:56.200><c> two</c><00:02:56.480><c> coordinates,</c><00:02:57.480><c> so</c>
632
+
633
+ 00:02:57.590 --> 00:02:57.600 align:start position:0%
634
+ described with only two coordinates, so
635
+
636
+
637
+ 00:02:57.600 --> 00:03:00.430 align:start position:0%
638
+ described with only two coordinates, so
639
+ they<00:02:57.720><c> must</c><00:02:57.920><c> be</c><00:02:58.040><c> landing</c><00:02:58.640><c> in</c><00:02:58.840><c> two</c><00:02:59.040><c> dimensions.</c>
640
+
641
+ 00:03:00.430 --> 00:03:00.440 align:start position:0%
642
+ they must be landing in two dimensions.
643
+
644
+
645
+ 00:03:00.440 --> 00:03:02.910 align:start position:0%
646
+ they must be landing in two dimensions.
647
+ So,<00:03:00.560><c> it's</c><00:03:00.680><c> a</c><00:03:00.720><c> transformation</c><00:03:01.480><c> from</c><00:03:01.840><c> 3D</c><00:03:02.200><c> space</c>
648
+
649
+ 00:03:02.910 --> 00:03:02.920 align:start position:0%
650
+ So, it's a transformation from 3D space
651
+
652
+
653
+ 00:03:02.920 --> 00:03:05.390 align:start position:0%
654
+ So, it's a transformation from 3D space
655
+ onto<00:03:03.360><c> the</c><00:03:03.480><c> 2D</c><00:03:03.840><c> plane.</c>
656
+
657
+ 00:03:05.390 --> 00:03:05.400 align:start position:0%
658
+ onto the 2D plane.
659
+
660
+
661
+ 00:03:05.400 --> 00:03:07.190 align:start position:0%
662
+ onto the 2D plane.
663
+ A<00:03:05.480><c> transformation</c><00:03:06.080><c> that</c><00:03:06.200><c> should</c><00:03:06.360><c> feel</c><00:03:06.640><c> very</c>
664
+
665
+ 00:03:07.190 --> 00:03:07.200 align:start position:0%
666
+ A transformation that should feel very
667
+
668
+
669
+ 00:03:07.200 --> 00:03:08.790 align:start position:0%
670
+ A transformation that should feel very
671
+ uncomfortable<00:03:07.920><c> if</c><00:03:08.040><c> you</c><00:03:08.160><c> imagine</c><00:03:08.520><c> going</c>
672
+
673
+ 00:03:08.790 --> 00:03:08.800 align:start position:0%
674
+ uncomfortable if you imagine going
675
+
676
+
677
+ 00:03:08.800 --> 00:03:13.870 align:start position:0%
678
+ uncomfortable if you imagine going
679
+ through<00:03:09.040><c> it.</c>
680
+
681
+ 00:03:13.870 --> 00:03:13.880 align:start position:0%
682
+
683
+
684
+
685
+ 00:03:13.880 --> 00:03:15.150 align:start position:0%
686
+
687
+ You<00:03:14.000><c> could</c><00:03:14.160><c> also</c><00:03:14.360><c> have</c><00:03:14.480><c> a</c><00:03:14.520><c> transformation</c>
688
+
689
+ 00:03:15.150 --> 00:03:15.160 align:start position:0%
690
+ You could also have a transformation
691
+
692
+
693
+ 00:03:15.160 --> 00:03:17.790 align:start position:0%
694
+ You could also have a transformation
695
+ from<00:03:15.440><c> two</c><00:03:15.640><c> dimensions</c><00:03:16.200><c> to</c><00:03:16.360><c> one</c><00:03:16.560><c> dimension.</c>
696
+
697
+ 00:03:17.790 --> 00:03:17.800 align:start position:0%
698
+ from two dimensions to one dimension.
699
+
700
+
701
+ 00:03:17.800 --> 00:03:19.430 align:start position:0%
702
+ from two dimensions to one dimension.
703
+ One-dimensional<00:03:18.360><c> space</c><00:03:18.680><c> is</c><00:03:18.920><c> really</c><00:03:19.120><c> just</c><00:03:19.360><c> the</c>
704
+
705
+ 00:03:19.430 --> 00:03:19.440 align:start position:0%
706
+ One-dimensional space is really just the
707
+
708
+
709
+ 00:03:19.440 --> 00:03:21.270 align:start position:0%
710
+ One-dimensional space is really just the
711
+ number<00:03:19.720><c> line,</c><00:03:20.320><c> so</c><00:03:20.440><c> a</c><00:03:20.480><c> transformation</c><00:03:21.120><c> like</c>
712
+
713
+ 00:03:21.270 --> 00:03:21.280 align:start position:0%
714
+ number line, so a transformation like
715
+
716
+
717
+ 00:03:21.280 --> 00:03:24.310 align:start position:0%
718
+ number line, so a transformation like
719
+ this<00:03:21.800><c> takes</c><00:03:22.160><c> in</c><00:03:22.480><c> 2D</c><00:03:22.880><c> vectors</c><00:03:23.640><c> and</c><00:03:23.800><c> spits</c><00:03:24.160><c> out</c>
720
+
721
+ 00:03:24.310 --> 00:03:24.320 align:start position:0%
722
+ this takes in 2D vectors and spits out
723
+
724
+
725
+ 00:03:24.320 --> 00:03:25.710 align:start position:0%
726
+ this takes in 2D vectors and spits out
727
+ numbers.
728
+
729
+ 00:03:25.710 --> 00:03:25.720 align:start position:0%
730
+ numbers.
731
+
732
+
733
+ 00:03:25.720 --> 00:03:27.390 align:start position:0%
734
+ numbers.
735
+ Thinking<00:03:26.000><c> about</c><00:03:26.440><c> grid</c><00:03:26.720><c> lines</c><00:03:27.000><c> remaining</c>
736
+
737
+ 00:03:27.390 --> 00:03:27.400 align:start position:0%
738
+ Thinking about grid lines remaining
739
+
740
+
741
+ 00:03:27.400 --> 00:03:29.150 align:start position:0%
742
+ Thinking about grid lines remaining
743
+ parallel<00:03:27.840><c> and</c><00:03:27.960><c> evenly</c><00:03:28.320><c> spaced</c><00:03:28.680><c> is</c><00:03:28.920><c> a</c><00:03:28.960><c> little</c>
744
+
745
+ 00:03:29.150 --> 00:03:29.160 align:start position:0%
746
+ parallel and evenly spaced is a little
747
+
748
+
749
+ 00:03:29.160 --> 00:03:30.550 align:start position:0%
750
+ parallel and evenly spaced is a little
751
+ bit<00:03:29.320><c> messy</c><00:03:29.840><c> due</c><00:03:29.960><c> to</c><00:03:30.040><c> all</c><00:03:30.240><c> of</c><00:03:30.320><c> the</c>
752
+
753
+ 00:03:30.550 --> 00:03:30.560 align:start position:0%
754
+ bit messy due to all of the
755
+
756
+
757
+ 00:03:30.560 --> 00:03:32.470 align:start position:0%
758
+ bit messy due to all of the
759
+ squishification<00:03:31.360><c> happening</c><00:03:31.720><c> here.</c><00:03:32.320><c> So,</c><00:03:32.400><c> in</c>
760
+
761
+ 00:03:32.470 --> 00:03:32.480 align:start position:0%
762
+ squishification happening here. So, in
763
+
764
+
765
+ 00:03:32.480 --> 00:03:34.550 align:start position:0%
766
+ squishification happening here. So, in
767
+ this<00:03:32.640><c> case,</c><00:03:33.160><c> the</c><00:03:33.320><c> visual</c><00:03:33.800><c> understanding</c><00:03:34.440><c> for</c>
768
+
769
+ 00:03:34.550 --> 00:03:34.560 align:start position:0%
770
+ this case, the visual understanding for
771
+
772
+
773
+ 00:03:34.560 --> 00:03:36.790 align:start position:0%
774
+ this case, the visual understanding for
775
+ what<00:03:34.760><c> linearity</c><00:03:35.360><c> means</c><00:03:36.120><c> is</c><00:03:36.240><c> that</c><00:03:36.400><c> if</c><00:03:36.520><c> you</c><00:03:36.640><c> have</c>
776
+
777
+ 00:03:36.790 --> 00:03:36.800 align:start position:0%
778
+ what linearity means is that if you have
779
+
780
+
781
+ 00:03:36.800 --> 00:03:39.310 align:start position:0%
782
+ what linearity means is that if you have
783
+ a<00:03:36.880><c> line</c><00:03:37.320><c> of</c><00:03:37.480><c> evenly</c><00:03:37.880><c> spaced</c><00:03:38.240><c> dots,</c><00:03:39.000><c> it</c><00:03:39.120><c> would</c>
784
+
785
+ 00:03:39.310 --> 00:03:39.320 align:start position:0%
786
+ a line of evenly spaced dots, it would
787
+
788
+
789
+ 00:03:39.320 --> 00:03:41.350 align:start position:0%
790
+ a line of evenly spaced dots, it would
791
+ remain<00:03:39.800><c> evenly</c><00:03:40.240><c> spaced</c><00:03:40.680><c> once</c><00:03:40.880><c> they're</c><00:03:41.040><c> mapped</c>
792
+
793
+ 00:03:41.350 --> 00:03:41.360 align:start position:0%
794
+ remain evenly spaced once they're mapped
795
+
796
+
797
+ 00:03:41.360 --> 00:03:43.670 align:start position:0%
798
+ remain evenly spaced once they're mapped
799
+ onto<00:03:41.560><c> the</c><00:03:41.680><c> number</c><00:03:41.920><c> line.</c>
800
+
801
+ 00:03:43.670 --> 00:03:43.680 align:start position:0%
802
+ onto the number line.
803
+
804
+
805
+ 00:03:43.680 --> 00:03:45.510 align:start position:0%
806
+ onto the number line.
807
+ One<00:03:43.800><c> of</c><00:03:43.840><c> these</c><00:03:44.040><c> transformations</c><00:03:44.960><c> is</c><00:03:45.120><c> encoded</c>
808
+
809
+ 00:03:45.510 --> 00:03:45.520 align:start position:0%
810
+ One of these transformations is encoded
811
+
812
+
813
+ 00:03:45.520 --> 00:03:48.150 align:start position:0%
814
+ One of these transformations is encoded
815
+ with<00:03:45.720><c> a</c><00:03:45.920><c> 1</c><00:03:46.240><c> by</c><00:03:46.400><c> 2</c><00:03:46.600><c> matrix,</c><00:03:47.360><c> each</c><00:03:47.560><c> of</c><00:03:47.640><c> whose</c><00:03:47.920><c> two</c>
816
+
817
+ 00:03:48.150 --> 00:03:48.160 align:start position:0%
818
+ with a 1 by 2 matrix, each of whose two
819
+
820
+
821
+ 00:03:48.160 --> 00:03:51.070 align:start position:0%
822
+ with a 1 by 2 matrix, each of whose two
823
+ columns<00:03:48.840><c> has</c><00:03:49.120><c> just</c><00:03:49.360><c> a</c><00:03:49.440><c> single</c><00:03:49.760><c> entry.</c><00:03:50.720><c> The</c><00:03:50.840><c> two</c>
824
+
825
+ 00:03:51.070 --> 00:03:51.080 align:start position:0%
826
+ columns has just a single entry. The two
827
+
828
+
829
+ 00:03:51.080 --> 00:03:52.510 align:start position:0%
830
+ columns has just a single entry. The two
831
+ columns<00:03:51.520><c> represent</c><00:03:51.920><c> where</c><00:03:52.040><c> the</c><00:03:52.120><c> basis</c>
832
+
833
+ 00:03:52.510 --> 00:03:52.520 align:start position:0%
834
+ columns represent where the basis
835
+
836
+
837
+ 00:03:52.520 --> 00:03:54.310 align:start position:0%
838
+ columns represent where the basis
839
+ vectors<00:03:52.880><c> land</c><00:03:53.520><c> and</c><00:03:53.760><c> each</c><00:03:53.960><c> one</c><00:03:54.120><c> of</c><00:03:54.160><c> those</c>
840
+
841
+ 00:03:54.310 --> 00:03:54.320 align:start position:0%
842
+ vectors land and each one of those
843
+
844
+
845
+ 00:03:54.320 --> 00:03:56.430 align:start position:0%
846
+ vectors land and each one of those
847
+ columns<00:03:54.720><c> requires</c><00:03:55.080><c> just</c><00:03:55.440><c> one</c><00:03:55.640><c> number,</c><00:03:56.320><c> the</c>
848
+
849
+ 00:03:56.430 --> 00:03:56.440 align:start position:0%
850
+ columns requires just one number, the
851
+
852
+
853
+ 00:03:56.440 --> 00:03:59.190 align:start position:0%
854
+ columns requires just one number, the
855
+ number<00:03:56.720><c> that</c><00:03:56.880><c> that</c><00:03:57.040><c> basis</c><00:03:57.400><c> vector</c><00:03:57.640><c> landed</c><00:03:58.000><c> on.</c>
856
+
857
+ 00:03:59.190 --> 00:03:59.200 align:start position:0%
858
+ number that that basis vector landed on.
859
+
860
+
861
+ 00:03:59.200 --> 00:04:00.630 align:start position:0%
862
+ number that that basis vector landed on.
863
+ This<00:03:59.360><c> is</c><00:03:59.480><c> actually</c><00:03:59.800><c> a</c><00:03:59.880><c> surprisingly</c>
864
+
865
+ 00:04:00.630 --> 00:04:00.640 align:start position:0%
866
+ This is actually a surprisingly
867
+
868
+
869
+ 00:04:00.640 --> 00:04:02.510 align:start position:0%
870
+ This is actually a surprisingly
871
+ meaningful<00:04:01.160><c> type</c><00:04:01.400><c> of</c><00:04:01.520><c> transformation</c><00:04:02.360><c> with</c>
872
+
873
+ 00:04:02.510 --> 00:04:02.520 align:start position:0%
874
+ meaningful type of transformation with
875
+
876
+
877
+ 00:04:02.520 --> 00:04:04.270 align:start position:0%
878
+ meaningful type of transformation with
879
+ close<00:04:02.760><c> ties</c><00:04:03.000><c> to</c><00:04:03.120><c> the</c><00:04:03.240><c> dot</c><00:04:03.480><c> product</c><00:04:04.120><c> and</c><00:04:04.240><c> I'll</c>
880
+
881
+ 00:04:04.270 --> 00:04:04.280 align:start position:0%
882
+ close ties to the dot product and I'll
883
+
884
+
885
+ 00:04:04.280 --> 00:04:06.270 align:start position:0%
886
+ close ties to the dot product and I'll
887
+ be<00:04:04.400><c> talking</c><00:04:04.680><c> about</c><00:04:04.920><c> that</c><00:04:05.040><c> next</c><00:04:05.240><c> video.</c>
888
+
889
+ 00:04:06.270 --> 00:04:06.280 align:start position:0%
890
+ be talking about that next video.
891
+
892
+
893
+ 00:04:06.280 --> 00:04:07.910 align:start position:0%
894
+ be talking about that next video.
895
+ Until<00:04:06.520><c> then,</c><00:04:07.040><c> I</c><00:04:07.120><c> encourage</c><00:04:07.480><c> you</c><00:04:07.600><c> to</c><00:04:07.720><c> play</c>
896
+
897
+ 00:04:07.910 --> 00:04:07.920 align:start position:0%
898
+ Until then, I encourage you to play
899
+
900
+
901
+ 00:04:07.920 --> 00:04:09.630 align:start position:0%
902
+ Until then, I encourage you to play
903
+ around<00:04:08.240><c> with</c><00:04:08.360><c> this</c><00:04:08.560><c> idea</c><00:04:08.880><c> on</c><00:04:09.000><c> your</c><00:04:09.160><c> own</c>
904
+
905
+ 00:04:09.630 --> 00:04:09.640 align:start position:0%
906
+ around with this idea on your own
907
+
908
+
909
+ 00:04:09.640 --> 00:04:10.990 align:start position:0%
910
+ around with this idea on your own
911
+ contemplating<00:04:10.240><c> the</c><00:04:10.320><c> meanings</c><00:04:10.640><c> of</c><00:04:10.760><c> things</c>
912
+
913
+ 00:04:10.990 --> 00:04:11.000 align:start position:0%
914
+ contemplating the meanings of things
915
+
916
+
917
+ 00:04:11.000 --> 00:04:12.830 align:start position:0%
918
+ contemplating the meanings of things
919
+ like<00:04:11.240><c> matrix</c><00:04:11.600><c> multiplication</c><00:04:12.360><c> and</c><00:04:12.560><c> linear</c>
920
+
921
+ 00:04:12.830 --> 00:04:12.840 align:start position:0%
922
+ like matrix multiplication and linear
923
+
924
+
925
+ 00:04:12.840 --> 00:04:14.990 align:start position:0%
926
+ like matrix multiplication and linear
927
+ systems<00:04:13.200><c> of</c><00:04:13.320><c> equations</c><00:04:14.200><c> in</c><00:04:14.320><c> the</c><00:04:14.400><c> context</c><00:04:14.880><c> of</c>
928
+
929
+ 00:04:14.990 --> 00:04:15.000 align:start position:0%
930
+ systems of equations in the context of
931
+
932
+
933
+ 00:04:15.000 --> 00:04:16.949 align:start position:0%
934
+ systems of equations in the context of
935
+ transformations<00:04:16.040><c> between</c><00:04:16.600><c> different</c>
936
+
937
+ 00:04:16.949 --> 00:04:16.959 align:start position:0%
938
+ transformations between different
939
+
940
+
941
+ 00:04:16.959 --> 00:04:20.560 align:start position:0%
942
+ transformations between different
943
+ dimensions.<00:04:18.079><c> Have</c><00:04:18.239><c> fun.</c>
944
+
3B1B_LinearAlgebra/08 - Nonsquare matrices as transformations between dimensions | Chapter 8, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:afb1de60bf7bd159dad4f5fd079d875b88f8b2a87abfc88697363930f769f960
3
+ size 25399372
3B1B_LinearAlgebra/09 - Dot products and duality | Chapter 9, Essence of linear algebra.en-orig.vtt ADDED
The diff for this file is too large to render. See raw diff
 
3B1B_LinearAlgebra/09 - Dot products and duality | Chapter 9, Essence of linear algebra.en.vtt ADDED
The diff for this file is too large to render. See raw diff
 
3B1B_LinearAlgebra/09 - Dot products and duality | Chapter 9, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a1480c887903b68f34baafc6bbdf685b03b03852c6b1e0e42f2eedc59356cf5a
3
+ size 31315321
3B1B_LinearAlgebra/10 - Cross products | Chapter 10, Essence of linear algebra.en-orig.vtt ADDED
@@ -0,0 +1,1896 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:09.200 --> 00:00:10.830 align:start position:0%
6
+
7
+ Last<00:00:09.440><c> video,</c><00:00:09.880><c> I</c><00:00:09.960><c> talked</c><00:00:10.200><c> about</c><00:00:10.440><c> the</c><00:00:10.560><c> dot</c>
8
+
9
+ 00:00:10.830 --> 00:00:10.840 align:start position:0%
10
+ Last video, I talked about the dot
11
+
12
+
13
+ 00:00:10.840 --> 00:00:12.630 align:start position:0%
14
+ Last video, I talked about the dot
15
+ product,<00:00:11.560><c> showing</c><00:00:11.880><c> both</c><00:00:12.120><c> the</c><00:00:12.200><c> standard</c>
16
+
17
+ 00:00:12.630 --> 00:00:12.640 align:start position:0%
18
+ product, showing both the standard
19
+
20
+
21
+ 00:00:12.640 --> 00:00:14.630 align:start position:0%
22
+ product, showing both the standard
23
+ introduction<00:00:13.160><c> to</c><00:00:13.240><c> the</c><00:00:13.320><c> topic,</c><00:00:14.080><c> as</c><00:00:14.280><c> well</c><00:00:14.440><c> as</c><00:00:14.560><c> a</c>
24
+
25
+ 00:00:14.630 --> 00:00:14.640 align:start position:0%
26
+ introduction to the topic, as well as a
27
+
28
+
29
+ 00:00:14.640 --> 00:00:16.510 align:start position:0%
30
+ introduction to the topic, as well as a
31
+ deeper<00:00:15.040><c> view</c><00:00:15.240><c> of</c><00:00:15.400><c> how</c><00:00:15.520><c> it</c><00:00:15.640><c> relates</c><00:00:16.000><c> to</c><00:00:16.160><c> linear</c>
32
+
33
+ 00:00:16.510 --> 00:00:16.520 align:start position:0%
34
+ deeper view of how it relates to linear
35
+
36
+
37
+ 00:00:16.520 --> 00:00:18.750 align:start position:0%
38
+ deeper view of how it relates to linear
39
+ transformations.
40
+
41
+ 00:00:18.750 --> 00:00:18.760 align:start position:0%
42
+ transformations.
43
+
44
+
45
+ 00:00:18.760 --> 00:00:20.070 align:start position:0%
46
+ transformations.
47
+ I'd<00:00:18.880><c> like</c><00:00:19.040><c> to</c><00:00:19.120><c> do</c><00:00:19.240><c> the</c><00:00:19.320><c> same</c><00:00:19.560><c> thing</c><00:00:19.720><c> for</c><00:00:19.840><c> cross</c>
48
+
49
+ 00:00:20.070 --> 00:00:20.080 align:start position:0%
50
+ I'd like to do the same thing for cross
51
+
52
+
53
+ 00:00:20.080 --> 00:00:21.750 align:start position:0%
54
+ I'd like to do the same thing for cross
55
+ products,<00:00:20.560><c> which</c><00:00:20.880><c> also</c><00:00:21.160><c> have</c><00:00:21.320><c> a</c><00:00:21.360><c> standard</c>
56
+
57
+ 00:00:21.750 --> 00:00:21.760 align:start position:0%
58
+ products, which also have a standard
59
+
60
+
61
+ 00:00:21.760 --> 00:00:23.670 align:start position:0%
62
+ products, which also have a standard
63
+ introduction<00:00:22.680><c> along</c><00:00:22.960><c> with</c><00:00:23.160><c> a</c><00:00:23.280><c> deeper</c>
64
+
65
+ 00:00:23.670 --> 00:00:23.680 align:start position:0%
66
+ introduction along with a deeper
67
+
68
+
69
+ 00:00:23.680 --> 00:00:25.230 align:start position:0%
70
+ introduction along with a deeper
71
+ understanding<00:00:24.360><c> in</c><00:00:24.520><c> the</c><00:00:24.600><c> light</c><00:00:24.840><c> of</c><00:00:24.960><c> linear</c>
72
+
73
+ 00:00:25.230 --> 00:00:25.240 align:start position:0%
74
+ understanding in the light of linear
75
+
76
+
77
+ 00:00:25.240 --> 00:00:27.150 align:start position:0%
78
+ understanding in the light of linear
79
+ transformations,<00:00:26.400><c> but</c><00:00:26.560><c> this</c><00:00:26.680><c> time</c><00:00:27.040><c> I'm</c>
80
+
81
+ 00:00:27.150 --> 00:00:27.160 align:start position:0%
82
+ transformations, but this time I'm
83
+
84
+
85
+ 00:00:27.160 --> 00:00:29.670 align:start position:0%
86
+ transformations, but this time I'm
87
+ dividing<00:00:27.560><c> it</c><00:00:27.680><c> into</c><00:00:27.920><c> two</c><00:00:28.080><c> separate</c><00:00:28.400><c> videos.</c>
88
+
89
+ 00:00:29.670 --> 00:00:29.680 align:start position:0%
90
+ dividing it into two separate videos.
91
+
92
+
93
+ 00:00:29.680 --> 00:00:31.350 align:start position:0%
94
+ dividing it into two separate videos.
95
+ Here,<00:00:30.120><c> I'll</c><00:00:30.240><c> try</c><00:00:30.400><c> to</c><00:00:30.520><c> hit</c><00:00:30.640><c> the</c><00:00:30.760><c> main</c><00:00:31.040><c> points</c>
96
+
97
+ 00:00:31.350 --> 00:00:31.360 align:start position:0%
98
+ Here, I'll try to hit the main points
99
+
100
+
101
+ 00:00:31.360 --> 00:00:32.950 align:start position:0%
102
+ Here, I'll try to hit the main points
103
+ that<00:00:31.520><c> students</c><00:00:31.880><c> are</c><00:00:31.960><c> usually</c><00:00:32.279><c> shown</c><00:00:32.640><c> about</c>
104
+
105
+ 00:00:32.950 --> 00:00:32.960 align:start position:0%
106
+ that students are usually shown about
107
+
108
+
109
+ 00:00:32.960 --> 00:00:34.590 align:start position:0%
110
+ that students are usually shown about
111
+ the<00:00:33.040><c> cross</c><00:00:33.320><c> product,</c><00:00:34.040><c> and</c><00:00:34.160><c> in</c><00:00:34.240><c> the</c><00:00:34.320><c> next</c>
112
+
113
+ 00:00:34.590 --> 00:00:34.600 align:start position:0%
114
+ the cross product, and in the next
115
+
116
+
117
+ 00:00:34.600 --> 00:00:36.750 align:start position:0%
118
+ the cross product, and in the next
119
+ video,<00:00:35.360><c> I'll</c><00:00:35.440><c> be</c><00:00:35.560><c> showing</c><00:00:35.840><c> a</c><00:00:35.880><c> view</c><00:00:36.120><c> which</c><00:00:36.360><c> is</c>
120
+
121
+ 00:00:36.750 --> 00:00:36.760 align:start position:0%
122
+ video, I'll be showing a view which is
123
+
124
+
125
+ 00:00:36.760 --> 00:00:38.830 align:start position:0%
126
+ video, I'll be showing a view which is
127
+ less<00:00:37.080><c> commonly</c><00:00:37.640><c> taught,</c><00:00:38.160><c> but</c><00:00:38.520><c> really</c>
128
+
129
+ 00:00:38.830 --> 00:00:38.840 align:start position:0%
130
+ less commonly taught, but really
131
+
132
+
133
+ 00:00:38.840 --> 00:00:40.830 align:start position:0%
134
+ less commonly taught, but really
135
+ satisfying<00:00:39.480><c> when</c><00:00:39.600><c> you</c><00:00:39.720><c> learn</c><00:00:39.960><c> it.</c><00:00:40.720><c> We'll</c>
136
+
137
+ 00:00:40.830 --> 00:00:40.840 align:start position:0%
138
+ satisfying when you learn it. We'll
139
+
140
+
141
+ 00:00:40.840 --> 00:00:42.870 align:start position:0%
142
+ satisfying when you learn it. We'll
143
+ start<00:00:41.080><c> in</c><00:00:41.200><c> two</c><00:00:41.320><c> dimensions.</c><00:00:42.280><c> If</c><00:00:42.400><c> you</c><00:00:42.480><c> have</c><00:00:42.720><c> two</c>
144
+
145
+ 00:00:42.870 --> 00:00:42.880 align:start position:0%
146
+ start in two dimensions. If you have two
147
+
148
+
149
+ 00:00:42.880 --> 00:00:45.430 align:start position:0%
150
+ start in two dimensions. If you have two
151
+ vectors,<00:00:43.560><c> V</c><00:00:43.920><c> and</c><00:00:44.080><c> W,</c><00:00:44.920><c> think</c><00:00:45.080><c> about</c><00:00:45.320><c> the</c>
152
+
153
+ 00:00:45.430 --> 00:00:45.440 align:start position:0%
154
+ vectors, V and W, think about the
155
+
156
+
157
+ 00:00:45.440 --> 00:00:47.990 align:start position:0%
158
+ vectors, V and W, think about the
159
+ parallelogram<00:00:46.360><c> that</c><00:00:46.520><c> they</c><00:00:46.640><c> span</c><00:00:47.040><c> out.</c><00:00:47.880><c> What</c><00:00:47.960><c> I</c>
160
+
161
+ 00:00:47.990 --> 00:00:48.000 align:start position:0%
162
+ parallelogram that they span out. What I
163
+
164
+
165
+ 00:00:48.000 --> 00:00:50.070 align:start position:0%
166
+ parallelogram that they span out. What I
167
+ mean<00:00:48.200><c> by</c><00:00:48.320><c> that</c><00:00:48.880><c> is</c><00:00:49.000><c> that</c><00:00:49.120><c> if</c><00:00:49.240><c> you</c><00:00:49.360><c> take</c><00:00:49.560><c> a</c><00:00:49.600><c> copy</c>
168
+
169
+ 00:00:50.070 --> 00:00:50.080 align:start position:0%
170
+ mean by that is that if you take a copy
171
+
172
+
173
+ 00:00:50.080 --> 00:00:53.270 align:start position:0%
174
+ mean by that is that if you take a copy
175
+ of<00:00:50.200><c> V</c><00:00:50.840><c> and</c><00:00:51.040><c> move</c><00:00:51.320><c> its</c><00:00:51.520><c> tail</c><00:00:51.880><c> to</c><00:00:52.000><c> the</c><00:00:52.120><c> tip</c><00:00:52.320><c> of</c><00:00:52.440><c> W,</c>
176
+
177
+ 00:00:53.270 --> 00:00:53.280 align:start position:0%
178
+ of V and move its tail to the tip of W,
179
+
180
+
181
+ 00:00:53.280 --> 00:00:55.030 align:start position:0%
182
+ of V and move its tail to the tip of W,
183
+ and<00:00:53.400><c> you</c><00:00:53.480><c> take</c><00:00:53.640><c> a</c><00:00:53.720><c> copy</c><00:00:54.000><c> of</c><00:00:54.120><c> W</c><00:00:54.560><c> and</c><00:00:54.680><c> move</c><00:00:54.840><c> its</c>
184
+
185
+ 00:00:55.030 --> 00:00:55.040 align:start position:0%
186
+ and you take a copy of W and move its
187
+
188
+
189
+ 00:00:55.040 --> 00:00:57.510 align:start position:0%
190
+ and you take a copy of W and move its
191
+ tail<00:00:55.360><c> to</c><00:00:55.480><c> the</c><00:00:55.560><c> tip</c><00:00:55.760><c> of</c><00:00:55.880><c> V,</c><00:00:56.800><c> the</c><00:00:56.880><c> four</c><00:00:57.120><c> vectors</c>
192
+
193
+ 00:00:57.510 --> 00:00:57.520 align:start position:0%
194
+ tail to the tip of V, the four vectors
195
+
196
+
197
+ 00:00:57.520 --> 00:00:59.630 align:start position:0%
198
+ tail to the tip of V, the four vectors
199
+ now<00:00:57.720><c> on</c><00:00:57.800><c> the</c><00:00:57.880><c> screen</c><00:00:58.640><c> enclose</c><00:00:59.240><c> a</c><00:00:59.320><c> certain</c>
200
+
201
+ 00:00:59.630 --> 00:00:59.640 align:start position:0%
202
+ now on the screen enclose a certain
203
+
204
+
205
+ 00:00:59.640 --> 00:01:01.910 align:start position:0%
206
+ now on the screen enclose a certain
207
+ parallelogram.
208
+
209
+ 00:01:01.910 --> 00:01:01.920 align:start position:0%
210
+ parallelogram.
211
+
212
+
213
+ 00:01:01.920 --> 00:01:04.350 align:start position:0%
214
+ parallelogram.
215
+ The<00:01:02.080><c> cross</c><00:01:02.480><c> product</c><00:01:02.880><c> of</c><00:01:03.000><c> V</c><00:01:03.160><c> and</c><00:01:03.280><c> W,</c><00:01:04.080><c> written</c>
216
+
217
+ 00:01:04.350 --> 00:01:04.360 align:start position:0%
218
+ The cross product of V and W, written
219
+
220
+
221
+ 00:01:04.360 --> 00:01:06.790 align:start position:0%
222
+ The cross product of V and W, written
223
+ with<00:01:04.480><c> the</c><00:01:04.720><c> X-shaped</c><00:01:05.320><c> multiplication</c><00:01:06.000><c> symbol,</c>
224
+
225
+ 00:01:06.790 --> 00:01:06.800 align:start position:0%
226
+ with the X-shaped multiplication symbol,
227
+
228
+
229
+ 00:01:06.800 --> 00:01:10.830 align:start position:0%
230
+ with the X-shaped multiplication symbol,
231
+ is<00:01:06.960><c> the</c><00:01:07.160><c> area</c><00:01:07.680><c> of</c><00:01:07.800><c> this</c><00:01:08.000><c> parallelogram.</c>
232
+
233
+ 00:01:10.830 --> 00:01:10.840 align:start position:0%
234
+ is the area of this parallelogram.
235
+
236
+
237
+ 00:01:10.840 --> 00:01:12.670 align:start position:0%
238
+ is the area of this parallelogram.
239
+ Well,<00:01:11.200><c> almost.</c><00:01:11.760><c> We</c><00:01:11.840><c> also</c><00:01:12.040><c> need</c><00:01:12.200><c> to</c><00:01:12.280><c> consider</c>
240
+
241
+ 00:01:12.670 --> 00:01:12.680 align:start position:0%
242
+ Well, almost. We also need to consider
243
+
244
+
245
+ 00:01:12.680 --> 00:01:15.390 align:start position:0%
246
+ Well, almost. We also need to consider
247
+ orientation.<00:01:13.880><c> Basically,</c><00:01:14.400><c> if</c><00:01:14.760><c> V</c><00:01:15.000><c> is</c><00:01:15.120><c> on</c><00:01:15.240><c> the</c>
248
+
249
+ 00:01:15.390 --> 00:01:15.400 align:start position:0%
250
+ orientation. Basically, if V is on the
251
+
252
+
253
+ 00:01:15.400 --> 00:01:18.630 align:start position:0%
254
+ orientation. Basically, if V is on the
255
+ right<00:01:15.760><c> of</c><00:01:15.880><c> W,</c><00:01:16.640><c> then</c><00:01:16.880><c> V</c><00:01:17.120><c> cross</c><00:01:17.400><c> W</c><00:01:17.840><c> is</c><00:01:18.040><c> positive</c>
256
+
257
+ 00:01:18.630 --> 00:01:18.640 align:start position:0%
258
+ right of W, then V cross W is positive
259
+
260
+
261
+ 00:01:18.640 --> 00:01:19.870 align:start position:0%
262
+ right of W, then V cross W is positive
263
+ and<00:01:18.840><c> equal</c><00:01:19.080><c> to</c><00:01:19.200><c> the</c><00:01:19.320><c> area</c><00:01:19.680><c> of</c><00:01:19.800><c> the</c>
264
+
265
+ 00:01:19.870 --> 00:01:19.880 align:start position:0%
266
+ and equal to the area of the
267
+
268
+
269
+ 00:01:19.880 --> 00:01:21.510 align:start position:0%
270
+ and equal to the area of the
271
+ parallelogram.
272
+
273
+ 00:01:21.510 --> 00:01:21.520 align:start position:0%
274
+ parallelogram.
275
+
276
+
277
+ 00:01:21.520 --> 00:01:23.870 align:start position:0%
278
+ parallelogram.
279
+ But<00:01:21.640><c> if</c><00:01:21.800><c> V</c><00:01:22.040><c> is</c><00:01:22.160><c> on</c><00:01:22.280><c> the</c><00:01:22.400><c> left</c><00:01:22.800><c> of</c><00:01:22.880><c> W,</c><00:01:23.680><c> then</c><00:01:23.800><c> the</c>
280
+
281
+ 00:01:23.870 --> 00:01:23.880 align:start position:0%
282
+ But if V is on the left of W, then the
283
+
284
+
285
+ 00:01:23.880 --> 00:01:25.910 align:start position:0%
286
+ But if V is on the left of W, then the
287
+ cross<00:01:24.120><c> product</c><00:01:24.440><c> is</c><00:01:24.600><c> negative,</c><00:01:25.280><c> namely</c><00:01:25.640><c> the</c>
288
+
289
+ 00:01:25.910 --> 00:01:25.920 align:start position:0%
290
+ cross product is negative, namely the
291
+
292
+
293
+ 00:01:25.920 --> 00:01:28.550 align:start position:0%
294
+ cross product is negative, namely the
295
+ negative<00:01:26.400><c> area</c><00:01:26.800><c> of</c><00:01:26.920><c> that</c><00:01:27.120><c> parallelogram.</c>
296
+
297
+ 00:01:28.550 --> 00:01:28.560 align:start position:0%
298
+ negative area of that parallelogram.
299
+
300
+
301
+ 00:01:28.560 --> 00:01:30.990 align:start position:0%
302
+ negative area of that parallelogram.
303
+ Notice,<00:01:29.040><c> this</c><00:01:29.280><c> means</c><00:01:29.520><c> that</c><00:01:29.800><c> order</c><00:01:30.120><c> matters.</c>
304
+
305
+ 00:01:30.990 --> 00:01:31.000 align:start position:0%
306
+ Notice, this means that order matters.
307
+
308
+
309
+ 00:01:31.000 --> 00:01:33.790 align:start position:0%
310
+ Notice, this means that order matters.
311
+ If<00:01:31.160><c> you</c><00:01:31.320><c> swapped</c><00:01:31.760><c> V</c><00:01:31.920><c> and</c><00:01:32.040><c> W,</c><00:01:32.720><c> instead</c><00:01:33.120><c> taking</c><00:01:33.400><c> W</c>
312
+
313
+ 00:01:33.790 --> 00:01:33.800 align:start position:0%
314
+ If you swapped V and W, instead taking W
315
+
316
+
317
+ 00:01:33.800 --> 00:01:35.830 align:start position:0%
318
+ If you swapped V and W, instead taking W
319
+ cross<00:01:34.080><c> V,</c><00:01:34.760><c> the</c><00:01:34.840><c> cross</c><00:01:35.080><c> product</c><00:01:35.440><c> would</c><00:01:35.560><c> become</c>
320
+
321
+ 00:01:35.830 --> 00:01:35.840 align:start position:0%
322
+ cross V, the cross product would become
323
+
324
+
325
+ 00:01:35.840 --> 00:01:38.950 align:start position:0%
326
+ cross V, the cross product would become
327
+ the<00:01:35.960><c> negative</c><00:01:36.520><c> of</c><00:01:36.680><c> whatever</c><00:01:37.040><c> it</c><00:01:37.160><c> was</c><00:01:37.360><c> before.</c>
328
+
329
+ 00:01:38.950 --> 00:01:38.960 align:start position:0%
330
+ the negative of whatever it was before.
331
+
332
+
333
+ 00:01:38.960 --> 00:01:40.470 align:start position:0%
334
+ the negative of whatever it was before.
335
+ The<00:01:39.040><c> way</c><00:01:39.240><c> I</c><00:01:39.360><c> always</c><00:01:39.640><c> remember</c><00:01:39.960><c> the</c><00:01:40.080><c> ordering</c>
336
+
337
+ 00:01:40.470 --> 00:01:40.480 align:start position:0%
338
+ The way I always remember the ordering
339
+
340
+
341
+ 00:01:40.480 --> 00:01:42.110 align:start position:0%
342
+ The way I always remember the ordering
343
+ here<00:01:41.040><c> is</c><00:01:41.160><c> that</c><00:01:41.360><c> when</c><00:01:41.480><c> you</c><00:01:41.560><c> take</c><00:01:41.760><c> the</c><00:01:41.880><c> cross</c>
344
+
345
+ 00:01:42.110 --> 00:01:42.120 align:start position:0%
346
+ here is that when you take the cross
347
+
348
+
349
+ 00:01:42.120 --> 00:01:44.150 align:start position:0%
350
+ here is that when you take the cross
351
+ product<00:01:42.480><c> of</c><00:01:42.600><c> the</c><00:01:42.760><c> two</c><00:01:43.120><c> basis</c><00:01:43.560><c> vectors</c><00:01:44.000><c> in</c>
352
+
353
+ 00:01:44.150 --> 00:01:44.160 align:start position:0%
354
+ product of the two basis vectors in
355
+
356
+
357
+ 00:01:44.160 --> 00:01:46.830 align:start position:0%
358
+ product of the two basis vectors in
359
+ order,<00:01:44.760><c> I</c><00:01:44.920><c> hat</c><00:01:45.240><c> cross</c><00:01:45.520><c> J</c><00:01:45.760><c> hat,</c><00:01:46.360><c> the</c><00:01:46.480><c> result</c>
360
+
361
+ 00:01:46.830 --> 00:01:46.840 align:start position:0%
362
+ order, I hat cross J hat, the result
363
+
364
+
365
+ 00:01:46.840 --> 00:01:49.150 align:start position:0%
366
+ order, I hat cross J hat, the result
367
+ should<00:01:46.960><c> be</c><00:01:47.080><c> positive.</c><00:01:48.160><c> In</c><00:01:48.240><c> fact,</c><00:01:48.760><c> the</c><00:01:48.880><c> order</c>
368
+
369
+ 00:01:49.150 --> 00:01:49.160 align:start position:0%
370
+ should be positive. In fact, the order
371
+
372
+
373
+ 00:01:49.160 --> 00:01:51.230 align:start position:0%
374
+ should be positive. In fact, the order
375
+ of<00:01:49.240><c> your</c><00:01:49.360><c> basis</c><00:01:49.720><c> vectors</c><00:01:50.160><c> is</c><00:01:50.360><c> what</c><00:01:50.520><c> defines</c>
376
+
377
+ 00:01:51.230 --> 00:01:51.240 align:start position:0%
378
+ of your basis vectors is what defines
379
+
380
+
381
+ 00:01:51.240 --> 00:01:52.710 align:start position:0%
382
+ of your basis vectors is what defines
383
+ orientation.
384
+
385
+ 00:01:52.710 --> 00:01:52.720 align:start position:0%
386
+ orientation.
387
+
388
+
389
+ 00:01:52.720 --> 00:01:54.630 align:start position:0%
390
+ orientation.
391
+ So,<00:01:52.960><c> since</c><00:01:53.320><c> I</c><00:01:53.440><c> hat</c><00:01:53.720><c> is</c><00:01:53.840><c> on</c><00:01:53.960><c> the</c><00:01:54.040><c> right</c><00:01:54.320><c> of</c><00:01:54.440><c> J</c>
392
+
393
+ 00:01:54.630 --> 00:01:54.640 align:start position:0%
394
+ So, since I hat is on the right of J
395
+
396
+
397
+ 00:01:54.640 --> 00:01:57.470 align:start position:0%
398
+ So, since I hat is on the right of J
399
+ hat,<00:01:55.240><c> I</c><00:01:55.280><c> remember</c><00:01:55.600><c> that</c><00:01:55.840><c> V</c><00:01:56.040><c> cross</c><00:01:56.320><c> W</c><00:01:57.040><c> has</c><00:01:57.240><c> to</c><00:01:57.320><c> be</c>
400
+
401
+ 00:01:57.470 --> 00:01:57.480 align:start position:0%
402
+ hat, I remember that V cross W has to be
403
+
404
+
405
+ 00:01:57.480 --> 00:01:59.430 align:start position:0%
406
+ hat, I remember that V cross W has to be
407
+ positive<00:01:57.960><c> whenever</c><00:01:58.320><c> V</c><00:01:58.680><c> is</c><00:01:58.800><c> on</c><00:01:58.920><c> the</c><00:01:59.040><c> right</c><00:01:59.320><c> of</c>
408
+
409
+ 00:01:59.430 --> 00:01:59.440 align:start position:0%
410
+ positive whenever V is on the right of
411
+
412
+
413
+ 00:01:59.440 --> 00:02:01.550 align:start position:0%
414
+ positive whenever V is on the right of
415
+ W.
416
+
417
+ 00:02:01.550 --> 00:02:01.560 align:start position:0%
418
+ W.
419
+
420
+
421
+ 00:02:01.560 --> 00:02:03.430 align:start position:0%
422
+ W.
423
+ So,<00:02:01.920><c> for</c><00:02:02.080><c> example,</c><00:02:02.720><c> with</c><00:02:02.840><c> the</c><00:02:02.920><c> vectors</c><00:02:03.200><c> shown</c>
424
+
425
+ 00:02:03.430 --> 00:02:03.440 align:start position:0%
426
+ So, for example, with the vectors shown
427
+
428
+
429
+ 00:02:03.440 --> 00:02:05.430 align:start position:0%
430
+ So, for example, with the vectors shown
431
+ here,<00:02:04.000><c> I'll</c><00:02:04.120><c> just</c><00:02:04.400><c> tell</c><00:02:04.600><c> you</c><00:02:04.760><c> that</c><00:02:04.920><c> the</c><00:02:05.040><c> area</c>
432
+
433
+ 00:02:05.430 --> 00:02:05.440 align:start position:0%
434
+ here, I'll just tell you that the area
435
+
436
+
437
+ 00:02:05.440 --> 00:02:07.830 align:start position:0%
438
+ here, I'll just tell you that the area
439
+ of<00:02:05.560><c> that</c><00:02:05.760><c> parallelogram</c><00:02:06.479><c> is</c><00:02:06.680><c> seven.</c><00:02:07.600><c> And</c>
440
+
441
+ 00:02:07.830 --> 00:02:07.840 align:start position:0%
442
+ of that parallelogram is seven. And
443
+
444
+
445
+ 00:02:07.840 --> 00:02:10.350 align:start position:0%
446
+ of that parallelogram is seven. And
447
+ since<00:02:08.080><c> V</c><00:02:08.399><c> is</c><00:02:08.560><c> on</c><00:02:08.679><c> the</c><00:02:08.800><c> left</c><00:02:09.160><c> of</c><00:02:09.280><c> W,</c><00:02:10.000><c> the</c><00:02:10.119><c> cross</c>
448
+
449
+ 00:02:10.350 --> 00:02:10.360 align:start position:0%
450
+ since V is on the left of W, the cross
451
+
452
+
453
+ 00:02:10.360 --> 00:02:12.390 align:start position:0%
454
+ since V is on the left of W, the cross
455
+ product<00:02:10.720><c> should</c><00:02:10.840><c> be</c><00:02:10.960><c> negative.</c><00:02:11.600><c> So,</c><00:02:12.000><c> V</c><00:02:12.160><c> cross</c>
456
+
457
+ 00:02:12.390 --> 00:02:12.400 align:start position:0%
458
+ product should be negative. So, V cross
459
+
460
+
461
+ 00:02:12.400 --> 00:02:15.830 align:start position:0%
462
+ product should be negative. So, V cross
463
+ W<00:02:12.880><c> is</c><00:02:13.120><c> -7.</c>
464
+
465
+ 00:02:15.830 --> 00:02:15.840 align:start position:0%
466
+ W is -7.
467
+
468
+
469
+ 00:02:15.840 --> 00:02:17.350 align:start position:0%
470
+ W is -7.
471
+ But,<00:02:16.040><c> of</c><00:02:16.120><c> course,</c><00:02:16.640><c> you</c><00:02:16.760><c> want</c><00:02:16.920><c> to</c><00:02:16.959><c> be</c><00:02:17.080><c> able</c><00:02:17.280><c> to</c>
472
+
473
+ 00:02:17.350 --> 00:02:17.360 align:start position:0%
474
+ But, of course, you want to be able to
475
+
476
+
477
+ 00:02:17.360 --> 00:02:19.070 align:start position:0%
478
+ But, of course, you want to be able to
479
+ compute<00:02:17.720><c> this</c><00:02:17.959><c> without</c><00:02:18.320><c> someone</c><00:02:18.640><c> telling</c><00:02:18.959><c> you</c>
480
+
481
+ 00:02:19.070 --> 00:02:19.080 align:start position:0%
482
+ compute this without someone telling you
483
+
484
+
485
+ 00:02:19.080 --> 00:02:20.430 align:start position:0%
486
+ compute this without someone telling you
487
+ the<00:02:19.200><c> area.</c>
488
+
489
+ 00:02:20.430 --> 00:02:20.440 align:start position:0%
490
+ the area.
491
+
492
+
493
+ 00:02:20.440 --> 00:02:22.830 align:start position:0%
494
+ the area.
495
+ This<00:02:20.800><c> is</c><00:02:20.959><c> where</c><00:02:21.120><c> the</c><00:02:21.240><c> determinant</c><00:02:21.840><c> comes</c><00:02:22.120><c> in.</c>
496
+
497
+ 00:02:22.830 --> 00:02:22.840 align:start position:0%
498
+ This is where the determinant comes in.
499
+
500
+
501
+ 00:02:22.840 --> 00:02:24.630 align:start position:0%
502
+ This is where the determinant comes in.
503
+ So,<00:02:23.200><c> if</c><00:02:23.360><c> you</c><00:02:23.480><c> didn't</c><00:02:23.800><c> see</c><00:02:23.959><c> chapter</c><00:02:24.280><c> five</c><00:02:24.520><c> of</c>
504
+
505
+ 00:02:24.630 --> 00:02:24.640 align:start position:0%
506
+ So, if you didn't see chapter five of
507
+
508
+
509
+ 00:02:24.640 --> 00:02:25.949 align:start position:0%
510
+ So, if you didn't see chapter five of
511
+ this<00:02:24.800><c> series</c><00:02:25.280><c> where</c><00:02:25.400><c> I</c><00:02:25.480><c> talk</c><00:02:25.680><c> about</c><00:02:25.880><c> the</c>
512
+
513
+ 00:02:25.949 --> 00:02:25.959 align:start position:0%
514
+ this series where I talk about the
515
+
516
+
517
+ 00:02:25.959 --> 00:02:27.990 align:start position:0%
518
+ this series where I talk about the
519
+ determinant,<00:02:26.840><c> now</c><00:02:27.120><c> would</c><00:02:27.280><c> be</c><00:02:27.440><c> a</c><00:02:27.520><c> really</c><00:02:27.800><c> good</c>
520
+
521
+ 00:02:27.990 --> 00:02:28.000 align:start position:0%
522
+ determinant, now would be a really good
523
+
524
+
525
+ 00:02:28.000 --> 00:02:30.190 align:start position:0%
526
+ determinant, now would be a really good
527
+ time<00:02:28.280><c> to</c><00:02:28.400><c> go</c><00:02:28.560><c> take</c><00:02:28.760><c> a</c><00:02:28.800><c> look.</c><00:02:29.520><c> Even</c><00:02:29.760><c> if</c><00:02:29.880><c> you</c><00:02:30.000><c> did</c>
528
+
529
+ 00:02:30.190 --> 00:02:30.200 align:start position:0%
530
+ time to go take a look. Even if you did
531
+
532
+
533
+ 00:02:30.200 --> 00:02:31.750 align:start position:0%
534
+ time to go take a look. Even if you did
535
+ see<00:02:30.320><c> it,</c><00:02:30.480><c> but</c><00:02:30.600><c> it</c><00:02:30.680><c> was</c><00:02:30.800><c> a</c><00:02:30.840><c> while</c><00:02:31.080><c> ago,</c><00:02:31.640><c> I'd</c>
536
+
537
+ 00:02:31.750 --> 00:02:31.760 align:start position:0%
538
+ see it, but it was a while ago, I'd
539
+
540
+
541
+ 00:02:31.760 --> 00:02:33.310 align:start position:0%
542
+ see it, but it was a while ago, I'd
543
+ recommend<00:02:32.200><c> taking</c><00:02:32.520><c> another</c><00:02:32.800><c> look</c><00:02:33.040><c> just</c><00:02:33.240><c> to</c>
544
+
545
+ 00:02:33.310 --> 00:02:33.320 align:start position:0%
546
+ recommend taking another look just to
547
+
548
+
549
+ 00:02:33.320 --> 00:02:34.670 align:start position:0%
550
+ recommend taking another look just to
551
+ make<00:02:33.480><c> sure</c><00:02:33.600><c> those</c><00:02:33.800><c> ideas</c><00:02:34.120><c> are</c><00:02:34.200><c> fresh</c><00:02:34.440><c> in</c><00:02:34.560><c> your</c>
552
+
553
+ 00:02:34.670 --> 00:02:34.680 align:start position:0%
554
+ make sure those ideas are fresh in your
555
+
556
+
557
+ 00:02:34.680 --> 00:02:37.230 align:start position:0%
558
+ make sure those ideas are fresh in your
559
+ mind.
560
+
561
+ 00:02:37.230 --> 00:02:37.240 align:start position:0%
562
+
563
+
564
+
565
+ 00:02:37.240 --> 00:02:40.070 align:start position:0%
566
+
567
+ For<00:02:37.440><c> the</c><00:02:37.600><c> 2D</c><00:02:37.920><c> cross</c><00:02:38.160><c> product,</c><00:02:38.720><c> V</c><00:02:38.880><c> cross</c><00:02:39.160><c> W,</c>
568
+
569
+ 00:02:40.070 --> 00:02:40.080 align:start position:0%
570
+ For the 2D cross product, V cross W,
571
+
572
+
573
+ 00:02:40.080 --> 00:02:41.870 align:start position:0%
574
+ For the 2D cross product, V cross W,
575
+ what<00:02:40.320><c> you</c><00:02:40.400><c> do</c><00:02:40.760><c> is</c><00:02:40.920><c> you</c><00:02:41.080><c> write</c><00:02:41.320><c> the</c><00:02:41.440><c> coordinates</c>
576
+
577
+ 00:02:41.870 --> 00:02:41.880 align:start position:0%
578
+ what you do is you write the coordinates
579
+
580
+
581
+ 00:02:41.880 --> 00:02:44.350 align:start position:0%
582
+ what you do is you write the coordinates
583
+ of<00:02:42.000><c> V</c><00:02:42.400><c> as</c><00:02:42.560><c> the</c><00:02:42.680><c> first</c><00:02:42.959><c> column</c><00:02:43.240><c> of</c><00:02:43.360><c> a</c><00:02:43.400><c> matrix,</c>
584
+
585
+ 00:02:44.350 --> 00:02:44.360 align:start position:0%
586
+ of V as the first column of a matrix,
587
+
588
+
589
+ 00:02:44.360 --> 00:02:46.030 align:start position:0%
590
+ of V as the first column of a matrix,
591
+ and<00:02:44.480><c> you</c><00:02:44.560><c> take</c><00:02:44.720><c> the</c><00:02:44.800><c> coordinates</c><00:02:45.280><c> of</c><00:02:45.400><c> W</c><00:02:45.920><c> and</c>
592
+
593
+ 00:02:46.030 --> 00:02:46.040 align:start position:0%
594
+ and you take the coordinates of W and
595
+
596
+
597
+ 00:02:46.040 --> 00:02:47.790 align:start position:0%
598
+ and you take the coordinates of W and
599
+ make<00:02:46.240><c> them</c><00:02:46.360><c> the</c><00:02:46.440><c> second</c><00:02:46.800><c> column.</c><00:02:47.560><c> Then,</c><00:02:47.720><c> you</c>
600
+
601
+ 00:02:47.790 --> 00:02:47.800 align:start position:0%
602
+ make them the second column. Then, you
603
+
604
+
605
+ 00:02:47.800 --> 00:02:51.270 align:start position:0%
606
+ make them the second column. Then, you
607
+ just<00:02:48.120><c> compute</c><00:02:48.440><c> the</c><00:02:48.560><c> determinant.</c>
608
+
609
+ 00:02:51.270 --> 00:02:51.280 align:start position:0%
610
+ just compute the determinant.
611
+
612
+
613
+ 00:02:51.280 --> 00:02:53.350 align:start position:0%
614
+ just compute the determinant.
615
+ This<00:02:51.560><c> is</c><00:02:51.680><c> because</c><00:02:52.120><c> a</c><00:02:52.160><c> matrix</c><00:02:52.640><c> whose</c><00:02:52.840><c> columns</c>
616
+
617
+ 00:02:53.350 --> 00:02:53.360 align:start position:0%
618
+ This is because a matrix whose columns
619
+
620
+
621
+ 00:02:53.360 --> 00:02:55.630 align:start position:0%
622
+ This is because a matrix whose columns
623
+ represent<00:02:53.800><c> V</c><00:02:54.000><c> and</c><00:02:54.120><c> W</c><00:02:54.959><c> corresponds</c><00:02:55.440><c> with</c><00:02:55.560><c> a</c>
624
+
625
+ 00:02:55.630 --> 00:02:55.640 align:start position:0%
626
+ represent V and W corresponds with a
627
+
628
+
629
+ 00:02:55.640 --> 00:02:57.910 align:start position:0%
630
+ represent V and W corresponds with a
631
+ linear<00:02:56.080><c> transformation</c><00:02:57.280><c> that</c><00:02:57.440><c> moves</c><00:02:57.840><c> the</c>
632
+
633
+ 00:02:57.910 --> 00:02:57.920 align:start position:0%
634
+ linear transformation that moves the
635
+
636
+
637
+ 00:02:57.920 --> 00:03:00.870 align:start position:0%
638
+ linear transformation that moves the
639
+ basis<00:02:58.320><c> vectors</c><00:02:58.880><c> I</c><00:02:59.000><c> hat</c><00:02:59.280><c> and</c><00:02:59.400><c> J</c><00:02:59.600><c> hat</c><00:03:00.160><c> to</c><00:03:00.280><c> V</c><00:03:00.720><c> and</c>
640
+
641
+ 00:03:00.870 --> 00:03:00.880 align:start position:0%
642
+ basis vectors I hat and J hat to V and
643
+
644
+
645
+ 00:03:00.880 --> 00:03:06.550 align:start position:0%
646
+ basis vectors I hat and J hat to V and
647
+ W.
648
+
649
+ 00:03:06.550 --> 00:03:06.560 align:start position:0%
650
+
651
+
652
+
653
+ 00:03:06.560 --> 00:03:08.470 align:start position:0%
654
+
655
+ The<00:03:06.760><c> determinant</c><00:03:07.360><c> is</c><00:03:07.560><c> all</c><00:03:07.760><c> about</c><00:03:08.040><c> measuring</c>
656
+
657
+ 00:03:08.470 --> 00:03:08.480 align:start position:0%
658
+ The determinant is all about measuring
659
+
660
+
661
+ 00:03:08.480 --> 00:03:10.030 align:start position:0%
662
+ The determinant is all about measuring
663
+ how<00:03:08.680><c> areas</c><00:03:09.160><c> change</c><00:03:09.640><c> due</c><00:03:09.840><c> to</c><00:03:10.000><c> a</c>
664
+
665
+ 00:03:10.030 --> 00:03:10.040 align:start position:0%
666
+ how areas change due to a
667
+
668
+
669
+ 00:03:10.040 --> 00:03:12.310 align:start position:0%
670
+ how areas change due to a
671
+ transformation.<00:03:11.280><c> And</c><00:03:11.440><c> the</c><00:03:11.560><c> prototypical</c>
672
+
673
+ 00:03:12.310 --> 00:03:12.320 align:start position:0%
674
+ transformation. And the prototypical
675
+
676
+
677
+ 00:03:12.320 --> 00:03:14.630 align:start position:0%
678
+ transformation. And the prototypical
679
+ area<00:03:12.640><c> that</c><00:03:12.760><c> we</c><00:03:12.880><c> look</c><00:03:13.120><c> at</c><00:03:13.519><c> is</c><00:03:13.680><c> the</c><00:03:13.800><c> unit</c><00:03:14.120><c> square</c>
680
+
681
+ 00:03:14.630 --> 00:03:14.640 align:start position:0%
682
+ area that we look at is the unit square
683
+
684
+
685
+ 00:03:14.640 --> 00:03:17.390 align:start position:0%
686
+ area that we look at is the unit square
687
+ resting<00:03:15.120><c> on</c><00:03:15.320><c> I</c><00:03:15.440><c> hat</c><00:03:15.760><c> and</c><00:03:15.920><c> J</c><00:03:16.120><c> hat.</c><00:03:16.920><c> After</c><00:03:17.280><c> the</c>
688
+
689
+ 00:03:17.390 --> 00:03:17.400 align:start position:0%
690
+ resting on I hat and J hat. After the
691
+
692
+
693
+ 00:03:17.400 --> 00:03:19.590 align:start position:0%
694
+ resting on I hat and J hat. After the
695
+ transformation,<00:03:18.440><c> that</c><00:03:18.640><c> square</c><00:03:18.920><c> gets</c><00:03:19.200><c> turned</c>
696
+
697
+ 00:03:19.590 --> 00:03:19.600 align:start position:0%
698
+ transformation, that square gets turned
699
+
700
+
701
+ 00:03:19.600 --> 00:03:21.390 align:start position:0%
702
+ transformation, that square gets turned
703
+ into<00:03:20.080><c> the</c><00:03:20.200><c> parallelogram</c><00:03:20.920><c> that</c><00:03:21.040><c> we</c><00:03:21.160><c> care</c>
704
+
705
+ 00:03:21.390 --> 00:03:21.400 align:start position:0%
706
+ into the parallelogram that we care
707
+
708
+
709
+ 00:03:21.400 --> 00:03:23.630 align:start position:0%
710
+ into the parallelogram that we care
711
+ about.<00:03:22.320><c> So,</c><00:03:22.560><c> the</c><00:03:22.640><c> determinant,</c><00:03:23.360><c> which</c>
712
+
713
+ 00:03:23.630 --> 00:03:23.640 align:start position:0%
714
+ about. So, the determinant, which
715
+
716
+
717
+ 00:03:23.640 --> 00:03:25.350 align:start position:0%
718
+ about. So, the determinant, which
719
+ generally<00:03:24.040><c> measures</c><00:03:24.440><c> the</c><00:03:24.560><c> factor</c><00:03:24.959><c> by</c><00:03:25.120><c> which</c>
720
+
721
+ 00:03:25.350 --> 00:03:25.360 align:start position:0%
722
+ generally measures the factor by which
723
+
724
+
725
+ 00:03:25.360 --> 00:03:27.870 align:start position:0%
726
+ generally measures the factor by which
727
+ areas<00:03:25.760><c> are</c><00:03:25.840><c> changed,</c><00:03:26.720><c> gives</c><00:03:26.959><c> the</c><00:03:27.200><c> area</c><00:03:27.680><c> of</c>
728
+
729
+ 00:03:27.870 --> 00:03:27.880 align:start position:0%
730
+ areas are changed, gives the area of
731
+
732
+
733
+ 00:03:27.880 --> 00:03:30.229 align:start position:0%
734
+ areas are changed, gives the area of
735
+ this<00:03:28.120><c> parallelogram</c><00:03:29.239><c> since</c><00:03:29.560><c> it</c><00:03:29.720><c> evolved</c><00:03:30.120><c> from</c>
736
+
737
+ 00:03:30.229 --> 00:03:30.239 align:start position:0%
738
+ this parallelogram since it evolved from
739
+
740
+
741
+ 00:03:30.239 --> 00:03:32.670 align:start position:0%
742
+ this parallelogram since it evolved from
743
+ a<00:03:30.320><c> square</c><00:03:30.640><c> that</c><00:03:30.800><c> started</c><00:03:31.160><c> with</c><00:03:31.320><c> area</c><00:03:31.600><c> one.</c>
744
+
745
+ 00:03:32.670 --> 00:03:32.680 align:start position:0%
746
+ a square that started with area one.
747
+
748
+
749
+ 00:03:32.680 --> 00:03:35.590 align:start position:0%
750
+ a square that started with area one.
751
+ What's<00:03:32.959><c> more,</c><00:03:33.560><c> if</c><00:03:33.800><c> V</c><00:03:34.000><c> is</c><00:03:34.160><c> on</c><00:03:34.280><c> the</c><00:03:34.400><c> left</c><00:03:34.720><c> of</c><00:03:34.800><c> W,</c>
752
+
753
+ 00:03:35.590 --> 00:03:35.600 align:start position:0%
754
+ What's more, if V is on the left of W,
755
+
756
+
757
+ 00:03:35.600 --> 00:03:37.470 align:start position:0%
758
+ What's more, if V is on the left of W,
759
+ it<00:03:35.720><c> means</c><00:03:35.920><c> that</c><00:03:36.160><c> orientation</c><00:03:36.840><c> was</c><00:03:37.040><c> flipped</c>
760
+
761
+ 00:03:37.470 --> 00:03:37.480 align:start position:0%
762
+ it means that orientation was flipped
763
+
764
+
765
+ 00:03:37.480 --> 00:03:39.430 align:start position:0%
766
+ it means that orientation was flipped
767
+ during<00:03:37.760><c> that</c><00:03:37.920><c> transformation,</c><00:03:39.160><c> which</c><00:03:39.320><c> is</c>
768
+
769
+ 00:03:39.430 --> 00:03:39.440 align:start position:0%
770
+ during that transformation, which is
771
+
772
+
773
+ 00:03:39.440 --> 00:03:40.990 align:start position:0%
774
+ during that transformation, which is
775
+ what<00:03:39.600><c> it</c><00:03:39.680><c> means</c><00:03:39.959><c> for</c><00:03:40.080><c> the</c><00:03:40.160><c> determinant</c><00:03:40.840><c> to</c><00:03:40.920><c> be</c>
776
+
777
+ 00:03:40.990 --> 00:03:41.000 align:start position:0%
778
+ what it means for the determinant to be
779
+
780
+
781
+ 00:03:41.000 --> 00:03:44.070 align:start position:0%
782
+ what it means for the determinant to be
783
+ negative.
784
+
785
+ 00:03:44.070 --> 00:03:44.080 align:start position:0%
786
+
787
+
788
+
789
+ 00:03:44.080 --> 00:03:46.070 align:start position:0%
790
+
791
+ As<00:03:44.320><c> an</c><00:03:44.400><c> example,</c><00:03:45.160><c> let's</c><00:03:45.320><c> say</c><00:03:45.480><c> V</c><00:03:45.840><c> has</c>
792
+
793
+ 00:03:46.070 --> 00:03:46.080 align:start position:0%
794
+ As an example, let's say V has
795
+
796
+
797
+ 00:03:46.080 --> 00:03:49.550 align:start position:0%
798
+ As an example, let's say V has
799
+ coordinates<00:03:46.640><c> -3</c><00:03:47.519><c> 1,</c><00:03:48.239><c> and</c><00:03:48.360><c> W</c><00:03:48.800><c> has</c><00:03:49.000><c> coordinates</c>
800
+
801
+ 00:03:49.550 --> 00:03:49.560 align:start position:0%
802
+ coordinates -3 1, and W has coordinates
803
+
804
+
805
+ 00:03:49.560 --> 00:03:52.830 align:start position:0%
806
+ coordinates -3 1, and W has coordinates
807
+ 2<00:03:49.959><c> 1.</c><00:03:50.800><c> The</c><00:03:51.080><c> determinant</c><00:03:51.760><c> of</c><00:03:51.880><c> the</c><00:03:52.000><c> matrix</c><00:03:52.600><c> those</c>
808
+
809
+ 00:03:52.830 --> 00:03:52.840 align:start position:0%
810
+ 2 1. The determinant of the matrix those
811
+
812
+
813
+ 00:03:52.840 --> 00:03:56.830 align:start position:0%
814
+ 2 1. The determinant of the matrix those
815
+ coordinates<00:03:53.280><c> as</c><00:03:53.400><c> columns</c><00:03:54.160><c> is</c><00:03:54.600><c> -3</c><00:03:55.240><c> *</c><00:03:55.560><c> 1</c>
816
+
817
+ 00:03:56.830 --> 00:03:56.840 align:start position:0%
818
+ coordinates as columns is -3 * 1
819
+
820
+
821
+ 00:03:56.840 --> 00:03:59.590 align:start position:0%
822
+ coordinates as columns is -3 * 1
823
+ 2<00:03:57.040><c> *</c><00:03:57.440><c> 1,</c>
824
+
825
+ 00:03:59.590 --> 00:03:59.600 align:start position:0%
826
+ 2 * 1,
827
+
828
+
829
+ 00:03:59.600 --> 00:04:01.550 align:start position:0%
830
+ 2 * 1,
831
+ which<00:03:59.840><c> is</c><00:04:00.080><c> -5.</c>
832
+
833
+ 00:04:01.550 --> 00:04:01.560 align:start position:0%
834
+ which is -5.
835
+
836
+
837
+ 00:04:01.560 --> 00:04:03.150 align:start position:0%
838
+ which is -5.
839
+ So,<00:04:01.680><c> evidently,</c><00:04:02.360><c> the</c><00:04:02.560><c> area</c><00:04:02.920><c> of</c><00:04:03.040><c> the</c>
840
+
841
+ 00:04:03.150 --> 00:04:03.160 align:start position:0%
842
+ So, evidently, the area of the
843
+
844
+
845
+ 00:04:03.160 --> 00:04:05.750 align:start position:0%
846
+ So, evidently, the area of the
847
+ parallelogram<00:04:03.880><c> they</c><00:04:03.960><c> define</c><00:04:04.680><c> is</c><00:04:04.880><c> five.</c><00:04:05.560><c> And</c>
848
+
849
+ 00:04:05.750 --> 00:04:05.760 align:start position:0%
850
+ parallelogram they define is five. And
851
+
852
+
853
+ 00:04:05.760 --> 00:04:08.110 align:start position:0%
854
+ parallelogram they define is five. And
855
+ since<00:04:06.000><c> V</c><00:04:06.240><c> is</c><00:04:06.400><c> on</c><00:04:06.520><c> the</c><00:04:06.640><c> left</c><00:04:07.000><c> of</c><00:04:07.120><c> W,</c><00:04:07.840><c> it</c><00:04:07.959><c> should</c>
856
+
857
+ 00:04:08.110 --> 00:04:08.120 align:start position:0%
858
+ since V is on the left of W, it should
859
+
860
+
861
+ 00:04:08.120 --> 00:04:11.230 align:start position:0%
862
+ since V is on the left of W, it should
863
+ make<00:04:08.280><c> sense</c><00:04:08.560><c> that</c><00:04:08.720><c> this</c><00:04:08.880><c> value</c><00:04:09.200><c> is</c><00:04:09.320><c> negative.</c>
864
+
865
+ 00:04:11.230 --> 00:04:11.240 align:start position:0%
866
+ make sense that this value is negative.
867
+
868
+
869
+ 00:04:11.240 --> 00:04:13.310 align:start position:0%
870
+ make sense that this value is negative.
871
+ As<00:04:11.440><c> with</c><00:04:11.640><c> any</c><00:04:11.840><c> new</c><00:04:12.040><c> operation</c><00:04:12.520><c> you</c><00:04:12.640><c> learn,</c><00:04:13.160><c> I'd</c>
872
+
873
+ 00:04:13.310 --> 00:04:13.320 align:start position:0%
874
+ As with any new operation you learn, I'd
875
+
876
+
877
+ 00:04:13.320 --> 00:04:14.590 align:start position:0%
878
+ As with any new operation you learn, I'd
879
+ recommend<00:04:13.720><c> playing</c><00:04:14.040><c> around</c><00:04:14.360><c> with</c><00:04:14.480><c> this</c>
880
+
881
+ 00:04:14.590 --> 00:04:14.600 align:start position:0%
882
+ recommend playing around with this
883
+
884
+
885
+ 00:04:14.600 --> 00:04:16.069 align:start position:0%
886
+ recommend playing around with this
887
+ notion<00:04:14.920><c> a</c><00:04:14.959><c> bit</c><00:04:15.160><c> in</c><00:04:15.240><c> your</c><00:04:15.360><c> head</c><00:04:15.640><c> just</c><00:04:15.840><c> to</c><00:04:15.920><c> get</c>
888
+
889
+ 00:04:16.069 --> 00:04:16.079 align:start position:0%
890
+ notion a bit in your head just to get
891
+
892
+
893
+ 00:04:16.079 --> 00:04:17.630 align:start position:0%
894
+ notion a bit in your head just to get
895
+ kind<00:04:16.239><c> of</c><00:04:16.359><c> an</c><00:04:16.440><c> intuitive</c><00:04:17.000><c> feel</c><00:04:17.280><c> for</c><00:04:17.440><c> what</c><00:04:17.560><c> the</c>
896
+
897
+ 00:04:17.630 --> 00:04:17.640 align:start position:0%
898
+ kind of an intuitive feel for what the
899
+
900
+
901
+ 00:04:17.640 --> 00:04:19.710 align:start position:0%
902
+ kind of an intuitive feel for what the
903
+ cross<00:04:17.880><c> product</c><00:04:18.239><c> is</c><00:04:18.359><c> all</c><00:04:18.480><c> about.</c>
904
+
905
+ 00:04:19.710 --> 00:04:19.720 align:start position:0%
906
+ cross product is all about.
907
+
908
+
909
+ 00:04:19.720 --> 00:04:21.229 align:start position:0%
910
+ cross product is all about.
911
+ For<00:04:19.799><c> example,</c><00:04:20.359><c> you</c><00:04:20.440><c> might</c><00:04:20.640><c> notice</c><00:04:20.920><c> that</c><00:04:21.079><c> when</c>
912
+
913
+ 00:04:21.229 --> 00:04:21.239 align:start position:0%
914
+ For example, you might notice that when
915
+
916
+
917
+ 00:04:21.239 --> 00:04:23.110 align:start position:0%
918
+ For example, you might notice that when
919
+ two<00:04:21.400><c> vectors</c><00:04:21.799><c> are</c><00:04:21.920><c> perpendicular,</c><00:04:22.760><c> or</c><00:04:23.000><c> at</c>
920
+
921
+ 00:04:23.110 --> 00:04:23.120 align:start position:0%
922
+ two vectors are perpendicular, or at
923
+
924
+
925
+ 00:04:23.120 --> 00:04:24.990 align:start position:0%
926
+ two vectors are perpendicular, or at
927
+ least<00:04:23.360><c> close</c><00:04:23.600><c> to</c><00:04:23.720><c> being</c><00:04:23.919><c> perpendicular,</c>
928
+
929
+ 00:04:24.990 --> 00:04:25.000 align:start position:0%
930
+ least close to being perpendicular,
931
+
932
+
933
+ 00:04:25.000 --> 00:04:26.630 align:start position:0%
934
+ least close to being perpendicular,
935
+ their<00:04:25.160><c> cross</c><00:04:25.400><c> product</c><00:04:25.760><c> is</c><00:04:25.919><c> larger</c><00:04:26.360><c> than</c><00:04:26.520><c> it</c>
936
+
937
+ 00:04:26.630 --> 00:04:26.640 align:start position:0%
938
+ their cross product is larger than it
939
+
940
+
941
+ 00:04:26.640 --> 00:04:28.550 align:start position:0%
942
+ their cross product is larger than it
943
+ would<00:04:26.840><c> be</c><00:04:27.360><c> if</c><00:04:27.480><c> they</c><00:04:27.600><c> were</c><00:04:27.720><c> pointing</c><00:04:28.080><c> in</c><00:04:28.200><c> very</c>
944
+
945
+ 00:04:28.550 --> 00:04:28.560 align:start position:0%
946
+ would be if they were pointing in very
947
+
948
+
949
+ 00:04:28.560 --> 00:04:31.070 align:start position:0%
950
+ would be if they were pointing in very
951
+ similar<00:04:28.960><c> directions,</c><00:04:29.960><c> because</c><00:04:30.280><c> the</c><00:04:30.480><c> area</c><00:04:30.960><c> of</c>
952
+
953
+ 00:04:31.070 --> 00:04:31.080 align:start position:0%
954
+ similar directions, because the area of
955
+
956
+
957
+ 00:04:31.080 --> 00:04:33.550 align:start position:0%
958
+ similar directions, because the area of
959
+ that<00:04:31.280><c> parallelogram</c><00:04:32.440><c> is</c><00:04:32.600><c> larger</c><00:04:33.320><c> when</c><00:04:33.440><c> the</c>
960
+
961
+ 00:04:33.550 --> 00:04:33.560 align:start position:0%
962
+ that parallelogram is larger when the
963
+
964
+
965
+ 00:04:33.560 --> 00:04:37.110 align:start position:0%
966
+ that parallelogram is larger when the
967
+ sides<00:04:33.880><c> are</c><00:04:33.960><c> closer</c><00:04:34.320><c> to</c><00:04:34.400><c> being</c><00:04:34.640><c> perpendicular.</c>
968
+
969
+ 00:04:37.110 --> 00:04:37.120 align:start position:0%
970
+ sides are closer to being perpendicular.
971
+
972
+
973
+ 00:04:37.120 --> 00:04:38.590 align:start position:0%
974
+ sides are closer to being perpendicular.
975
+ Something<00:04:37.360><c> else</c><00:04:37.520><c> you</c><00:04:37.600><c> might</c><00:04:37.760><c> notice</c><00:04:38.280><c> is</c><00:04:38.440><c> that</c>
976
+
977
+ 00:04:38.590 --> 00:04:38.600 align:start position:0%
978
+ Something else you might notice is that
979
+
980
+
981
+ 00:04:38.600 --> 00:04:40.150 align:start position:0%
982
+ Something else you might notice is that
983
+ if<00:04:38.720><c> you</c><00:04:38.800><c> were</c><00:04:38.919><c> to</c><00:04:39.040><c> scale</c><00:04:39.560><c> up</c><00:04:39.720><c> one</c><00:04:39.880><c> of</c><00:04:39.919><c> those</c>
984
+
985
+ 00:04:40.150 --> 00:04:40.160 align:start position:0%
986
+ if you were to scale up one of those
987
+
988
+
989
+ 00:04:40.160 --> 00:04:43.590 align:start position:0%
990
+ if you were to scale up one of those
991
+ vectors,<00:04:40.960><c> perhaps</c><00:04:41.360><c> multiplying</c><00:04:41.960><c> V</c><00:04:42.440><c> by</c><00:04:42.560><c> three,</c>
992
+
993
+ 00:04:43.590 --> 00:04:43.600 align:start position:0%
994
+ vectors, perhaps multiplying V by three,
995
+
996
+
997
+ 00:04:43.600 --> 00:04:46.110 align:start position:0%
998
+ vectors, perhaps multiplying V by three,
999
+ then<00:04:43.680><c> the</c><00:04:43.919><c> area</c><00:04:44.400><c> of</c><00:04:44.520><c> that</c><00:04:44.760><c> parallelogram</c><00:04:45.919><c> is</c>
1000
+
1001
+ 00:04:46.110 --> 00:04:46.120 align:start position:0%
1002
+ then the area of that parallelogram is
1003
+
1004
+
1005
+ 00:04:46.120 --> 00:04:48.950 align:start position:0%
1006
+ then the area of that parallelogram is
1007
+ also<00:04:46.440><c> scaled</c><00:04:46.880><c> up</c><00:04:47.040><c> by</c><00:04:47.200><c> a</c><00:04:47.280><c> factor</c><00:04:47.640><c> of</c><00:04:47.760><c> three.</c>
1008
+
1009
+ 00:04:48.950 --> 00:04:48.960 align:start position:0%
1010
+ also scaled up by a factor of three.
1011
+
1012
+
1013
+ 00:04:48.960 --> 00:04:50.830 align:start position:0%
1014
+ also scaled up by a factor of three.
1015
+ So,<00:04:49.160><c> what</c><00:04:49.360><c> this</c><00:04:49.480><c> means</c><00:04:49.720><c> for</c><00:04:49.800><c> the</c><00:04:49.919><c> operation</c><00:04:50.680><c> is</c>
1016
+
1017
+ 00:04:50.830 --> 00:04:50.840 align:start position:0%
1018
+ So, what this means for the operation is
1019
+
1020
+
1021
+ 00:04:50.840 --> 00:04:54.510 align:start position:0%
1022
+ So, what this means for the operation is
1023
+ that<00:04:51.360><c> 3V</c><00:04:52.200><c> cross</c><00:04:52.520><c> W</c><00:04:53.320><c> will</c><00:04:53.440><c> be</c><00:04:53.600><c> exactly</c><00:04:54.240><c> three</c>
1024
+
1025
+ 00:04:54.510 --> 00:04:54.520 align:start position:0%
1026
+ that 3V cross W will be exactly three
1027
+
1028
+
1029
+ 00:04:54.520 --> 00:04:58.110 align:start position:0%
1030
+ that 3V cross W will be exactly three
1031
+ times<00:04:54.840><c> the</c><00:04:54.919><c> value</c><00:04:55.520><c> of</c><00:04:55.680><c> V</c><00:04:55.880><c> cross</c><00:04:56.160><c> W.</c>
1032
+
1033
+ 00:04:58.110 --> 00:04:58.120 align:start position:0%
1034
+ times the value of V cross W.
1035
+
1036
+
1037
+ 00:04:58.120 --> 00:04:59.510 align:start position:0%
1038
+ times the value of V cross W.
1039
+ Now,<00:04:58.480><c> even</c><00:04:58.720><c> though</c><00:04:58.880><c> all</c><00:04:59.040><c> of</c><00:04:59.120><c> this</c><00:04:59.320><c> is</c><00:04:59.440><c> a</c>
1040
+
1041
+ 00:04:59.510 --> 00:04:59.520 align:start position:0%
1042
+ Now, even though all of this is a
1043
+
1044
+
1045
+ 00:04:59.520 --> 00:05:01.790 align:start position:0%
1046
+ Now, even though all of this is a
1047
+ perfectly<00:05:00.000><c> fine</c><00:05:00.240><c> mathematical</c><00:05:00.800><c> operation,</c>
1048
+
1049
+ 00:05:01.790 --> 00:05:01.800 align:start position:0%
1050
+ perfectly fine mathematical operation,
1051
+
1052
+
1053
+ 00:05:01.800 --> 00:05:04.190 align:start position:0%
1054
+ perfectly fine mathematical operation,
1055
+ what<00:05:02.040><c> I</c><00:05:02.080><c> just</c><00:05:02.320><c> described</c><00:05:02.960><c> is</c><00:05:03.160><c> technically</c><00:05:03.840><c> not</c>
1056
+
1057
+ 00:05:04.190 --> 00:05:04.200 align:start position:0%
1058
+ what I just described is technically not
1059
+
1060
+
1061
+ 00:05:04.200 --> 00:05:06.350 align:start position:0%
1062
+ what I just described is technically not
1063
+ the<00:05:04.320><c> cross</c><00:05:04.600><c> product.</c><00:05:05.600><c> The</c><00:05:05.720><c> true</c><00:05:06.080><c> cross</c>
1064
+
1065
+ 00:05:06.350 --> 00:05:06.360 align:start position:0%
1066
+ the cross product. The true cross
1067
+
1068
+
1069
+ 00:05:06.360 --> 00:05:08.350 align:start position:0%
1070
+ the cross product. The true cross
1071
+ product<00:05:07.000><c> is</c><00:05:07.120><c> something</c><00:05:07.440><c> that</c><00:05:07.560><c> combines</c><00:05:08.120><c> two</c>
1072
+
1073
+ 00:05:08.350 --> 00:05:08.360 align:start position:0%
1074
+ product is something that combines two
1075
+
1076
+
1077
+ 00:05:08.360 --> 00:05:11.230 align:start position:0%
1078
+ product is something that combines two
1079
+ different<00:05:08.720><c> 3D</c><00:05:09.240><c> vectors</c><00:05:10.040><c> to</c><00:05:10.160><c> get</c><00:05:10.320><c> a</c><00:05:10.360><c> new</c><00:05:10.800><c> 3D</c>
1080
+
1081
+ 00:05:11.230 --> 00:05:11.240 align:start position:0%
1082
+ different 3D vectors to get a new 3D
1083
+
1084
+
1085
+ 00:05:11.240 --> 00:05:12.510 align:start position:0%
1086
+ different 3D vectors to get a new 3D
1087
+ vector.
1088
+
1089
+ 00:05:12.510 --> 00:05:12.520 align:start position:0%
1090
+ vector.
1091
+
1092
+
1093
+ 00:05:12.520 --> 00:05:13.990 align:start position:0%
1094
+ vector.
1095
+ Just<00:05:12.720><c> as</c><00:05:12.840><c> before,</c><00:05:13.440><c> we're</c><00:05:13.560><c> still</c><00:05:13.800><c> going</c><00:05:13.919><c> to</c>
1096
+
1097
+ 00:05:13.990 --> 00:05:14.000 align:start position:0%
1098
+ Just as before, we're still going to
1099
+
1100
+
1101
+ 00:05:14.000 --> 00:05:16.190 align:start position:0%
1102
+ Just as before, we're still going to
1103
+ consider<00:05:14.400><c> the</c><00:05:14.520><c> parallelogram</c><00:05:15.560><c> defined</c><00:05:16.080><c> by</c>
1104
+
1105
+ 00:05:16.190 --> 00:05:16.200 align:start position:0%
1106
+ consider the parallelogram defined by
1107
+
1108
+
1109
+ 00:05:16.200 --> 00:05:17.630 align:start position:0%
1110
+ consider the parallelogram defined by
1111
+ the<00:05:16.360><c> two</c><00:05:16.600><c> vectors</c><00:05:17.040><c> that</c><00:05:17.160><c> we're</c><00:05:17.280><c> crossing</c>
1112
+
1113
+ 00:05:17.630 --> 00:05:17.640 align:start position:0%
1114
+ the two vectors that we're crossing
1115
+
1116
+
1117
+ 00:05:17.640 --> 00:05:18.870 align:start position:0%
1118
+ the two vectors that we're crossing
1119
+ together.
1120
+
1121
+ 00:05:18.870 --> 00:05:18.880 align:start position:0%
1122
+ together.
1123
+
1124
+
1125
+ 00:05:18.880 --> 00:05:20.750 align:start position:0%
1126
+ together.
1127
+ And<00:05:18.960><c> the</c><00:05:19.080><c> area</c><00:05:19.440><c> of</c><00:05:19.520><c> this</c><00:05:19.680><c> parallelogram</c><00:05:20.600><c> is</c>
1128
+
1129
+ 00:05:20.750 --> 00:05:20.760 align:start position:0%
1130
+ And the area of this parallelogram is
1131
+
1132
+
1133
+ 00:05:20.760 --> 00:05:22.750 align:start position:0%
1134
+ And the area of this parallelogram is
1135
+ still<00:05:20.960><c> going</c><00:05:21.080><c> to</c><00:05:21.160><c> play</c><00:05:21.360><c> a</c><00:05:21.440><c> big</c><00:05:21.640><c> role.</c><00:05:22.600><c> To</c><00:05:22.640><c> be</c>
1136
+
1137
+ 00:05:22.750 --> 00:05:22.760 align:start position:0%
1138
+ still going to play a big role. To be
1139
+
1140
+
1141
+ 00:05:22.760 --> 00:05:25.270 align:start position:0%
1142
+ still going to play a big role. To be
1143
+ concrete,<00:05:23.200><c> let's</c><00:05:23.360><c> say</c><00:05:23.480><c> that</c><00:05:23.640><c> the</c><00:05:23.760><c> area</c><00:05:24.200><c> is</c><00:05:24.400><c> 2.5</c>
1144
+
1145
+ 00:05:25.270 --> 00:05:25.280 align:start position:0%
1146
+ concrete, let's say that the area is 2.5
1147
+
1148
+
1149
+ 00:05:25.280 --> 00:05:27.150 align:start position:0%
1150
+ concrete, let's say that the area is 2.5
1151
+ for<00:05:25.400><c> the</c><00:05:25.480><c> vectors</c><00:05:25.840><c> shown</c><00:05:26.040><c> here.</c>
1152
+
1153
+ 00:05:27.150 --> 00:05:27.160 align:start position:0%
1154
+ for the vectors shown here.
1155
+
1156
+
1157
+ 00:05:27.160 --> 00:05:28.950 align:start position:0%
1158
+ for the vectors shown here.
1159
+ But<00:05:27.240><c> as</c><00:05:27.360><c> I</c><00:05:27.440><c> said,</c><00:05:27.919><c> the</c><00:05:28.040><c> cross</c><00:05:28.280><c> product</c><00:05:28.600><c> is</c><00:05:28.720><c> not</c>
1160
+
1161
+ 00:05:28.950 --> 00:05:28.960 align:start position:0%
1162
+ But as I said, the cross product is not
1163
+
1164
+
1165
+ 00:05:28.960 --> 00:05:31.030 align:start position:0%
1166
+ But as I said, the cross product is not
1167
+ a<00:05:29.040><c> number,</c><00:05:29.560><c> it's</c><00:05:29.720><c> a</c><00:05:29.760><c> vector.</c>
1168
+
1169
+ 00:05:31.030 --> 00:05:31.040 align:start position:0%
1170
+ a number, it's a vector.
1171
+
1172
+
1173
+ 00:05:31.040 --> 00:05:32.990 align:start position:0%
1174
+ a number, it's a vector.
1175
+ This<00:05:31.200><c> new</c><00:05:31.400><c> vector's</c><00:05:31.800><c> length</c><00:05:32.360><c> will</c><00:05:32.520><c> be</c><00:05:32.840><c> the</c>
1176
+
1177
+ 00:05:32.990 --> 00:05:33.000 align:start position:0%
1178
+ This new vector's length will be the
1179
+
1180
+
1181
+ 00:05:33.000 --> 00:05:34.990 align:start position:0%
1182
+ This new vector's length will be the
1183
+ area<00:05:33.400><c> of</c><00:05:33.520><c> that</c><00:05:33.720><c> parallelogram,</c><00:05:34.800><c> which</c><00:05:34.919><c> in</c>
1184
+
1185
+ 00:05:34.990 --> 00:05:35.000 align:start position:0%
1186
+ area of that parallelogram, which in
1187
+
1188
+
1189
+ 00:05:35.000 --> 00:05:37.030 align:start position:0%
1190
+ area of that parallelogram, which in
1191
+ this<00:05:35.120><c> case</c><00:05:35.360><c> is</c><00:05:35.640><c> 2.5.</c>
1192
+
1193
+ 00:05:37.030 --> 00:05:37.040 align:start position:0%
1194
+ this case is 2.5.
1195
+
1196
+
1197
+ 00:05:37.040 --> 00:05:39.070 align:start position:0%
1198
+ this case is 2.5.
1199
+ And<00:05:37.120><c> the</c><00:05:37.200><c> direction</c><00:05:37.760><c> of</c><00:05:37.840><c> that</c><00:05:38.040><c> new</c><00:05:38.200><c> vector</c><00:05:38.960><c> is</c>
1200
+
1201
+ 00:05:39.070 --> 00:05:39.080 align:start position:0%
1202
+ And the direction of that new vector is
1203
+
1204
+
1205
+ 00:05:39.080 --> 00:05:40.950 align:start position:0%
1206
+ And the direction of that new vector is
1207
+ going<00:05:39.240><c> to</c><00:05:39.280><c> be</c><00:05:39.440><c> perpendicular</c><00:05:40.680><c> to</c><00:05:40.840><c> the</c>
1208
+
1209
+ 00:05:40.950 --> 00:05:40.960 align:start position:0%
1210
+ going to be perpendicular to the
1211
+
1212
+
1213
+ 00:05:40.960 --> 00:05:42.550 align:start position:0%
1214
+ going to be perpendicular to the
1215
+ parallelogram.
1216
+
1217
+ 00:05:42.550 --> 00:05:42.560 align:start position:0%
1218
+ parallelogram.
1219
+
1220
+
1221
+ 00:05:42.560 --> 00:05:44.630 align:start position:0%
1222
+ parallelogram.
1223
+ But<00:05:42.840><c> which</c><00:05:43.000><c> way,</c><00:05:43.440><c> right?</c><00:05:44.000><c> I</c><00:05:44.080><c> mean,</c><00:05:44.360><c> there</c><00:05:44.560><c> are</c>
1224
+
1225
+ 00:05:44.630 --> 00:05:44.640 align:start position:0%
1226
+ But which way, right? I mean, there are
1227
+
1228
+
1229
+ 00:05:44.640 --> 00:05:47.230 align:start position:0%
1230
+ But which way, right? I mean, there are
1231
+ two<00:05:44.919><c> possible</c><00:05:45.360><c> vectors</c><00:05:45.760><c> with</c><00:05:45.919><c> length</c><00:05:46.240><c> 2.5</c>
1232
+
1233
+ 00:05:47.230 --> 00:05:47.240 align:start position:0%
1234
+ two possible vectors with length 2.5
1235
+
1236
+
1237
+ 00:05:47.240 --> 00:05:50.550 align:start position:0%
1238
+ two possible vectors with length 2.5
1239
+ that<00:05:47.440><c> are</c><00:05:47.520><c> perpendicular</c><00:05:48.240><c> to</c><00:05:48.400><c> a</c><00:05:48.480><c> given</c><00:05:48.760><c> plane.</c>
1240
+
1241
+ 00:05:50.550 --> 00:05:50.560 align:start position:0%
1242
+ that are perpendicular to a given plane.
1243
+
1244
+
1245
+ 00:05:50.560 --> 00:05:51.990 align:start position:0%
1246
+ that are perpendicular to a given plane.
1247
+ This<00:05:50.680><c> is</c><00:05:50.800><c> where</c><00:05:50.920><c> the</c><00:05:51.040><c> right-hand</c><00:05:51.560><c> rule</c><00:05:51.720><c> comes</c>
1248
+
1249
+ 00:05:51.990 --> 00:05:52.000 align:start position:0%
1250
+ This is where the right-hand rule comes
1251
+
1252
+
1253
+ 00:05:52.000 --> 00:05:52.910 align:start position:0%
1254
+ This is where the right-hand rule comes
1255
+ in.
1256
+
1257
+ 00:05:52.910 --> 00:05:52.920 align:start position:0%
1258
+ in.
1259
+
1260
+
1261
+ 00:05:52.920 --> 00:05:54.910 align:start position:0%
1262
+ in.
1263
+ Point<00:05:53.160><c> the</c><00:05:53.280><c> forefinger</c><00:05:53.920><c> of</c><00:05:54.080><c> your</c><00:05:54.240><c> right</c><00:05:54.480><c> hand</c>
1264
+
1265
+ 00:05:54.910 --> 00:05:54.920 align:start position:0%
1266
+ Point the forefinger of your right hand
1267
+
1268
+
1269
+ 00:05:54.920 --> 00:05:56.430 align:start position:0%
1270
+ Point the forefinger of your right hand
1271
+ in<00:05:55.040><c> the</c><00:05:55.120><c> direction</c><00:05:55.480><c> of</c><00:05:55.600><c> V.</c>
1272
+
1273
+ 00:05:56.430 --> 00:05:56.440 align:start position:0%
1274
+ in the direction of V.
1275
+
1276
+
1277
+ 00:05:56.440 --> 00:05:57.990 align:start position:0%
1278
+ in the direction of V.
1279
+ Then<00:05:56.560><c> stick</c><00:05:56.840><c> out</c><00:05:56.960><c> your</c><00:05:57.080><c> middle</c><00:05:57.320><c> finger</c><00:05:57.800><c> in</c><00:05:57.920><c> the</c>
1280
+
1281
+ 00:05:57.990 --> 00:05:58.000 align:start position:0%
1282
+ Then stick out your middle finger in the
1283
+
1284
+
1285
+ 00:05:58.000 --> 00:05:59.670 align:start position:0%
1286
+ Then stick out your middle finger in the
1287
+ direction<00:05:58.400><c> of</c><00:05:58.480><c> W.</c>
1288
+
1289
+ 00:05:59.670 --> 00:05:59.680 align:start position:0%
1290
+ direction of W.
1291
+
1292
+
1293
+ 00:05:59.680 --> 00:06:01.670 align:start position:0%
1294
+ direction of W.
1295
+ Then,<00:06:00.200><c> when</c><00:06:00.320><c> you</c><00:06:00.440><c> point</c><00:06:00.680><c> up</c><00:06:00.800><c> your</c><00:06:00.920><c> thumb,</c>
1296
+
1297
+ 00:06:01.670 --> 00:06:01.680 align:start position:0%
1298
+ Then, when you point up your thumb,
1299
+
1300
+
1301
+ 00:06:01.680 --> 00:06:02.910 align:start position:0%
1302
+ Then, when you point up your thumb,
1303
+ that's<00:06:02.040><c> the</c><00:06:02.160><c> direction</c><00:06:02.520><c> of</c><00:06:02.600><c> the</c><00:06:02.680><c> cross</c>
1304
+
1305
+ 00:06:02.910 --> 00:06:02.920 align:start position:0%
1306
+ that's the direction of the cross
1307
+
1308
+
1309
+ 00:06:02.920 --> 00:06:08.230 align:start position:0%
1310
+ that's the direction of the cross
1311
+ product.
1312
+
1313
+ 00:06:08.230 --> 00:06:08.240 align:start position:0%
1314
+
1315
+
1316
+
1317
+ 00:06:08.240 --> 00:06:10.070 align:start position:0%
1318
+
1319
+ For<00:06:08.360><c> example,</c><00:06:09.040><c> let's</c><00:06:09.240><c> say</c><00:06:09.360><c> that</c><00:06:09.560><c> V</c><00:06:09.840><c> was</c><00:06:10.000><c> a</c>
1320
+
1321
+ 00:06:10.070 --> 00:06:10.080 align:start position:0%
1322
+ For example, let's say that V was a
1323
+
1324
+
1325
+ 00:06:10.080 --> 00:06:12.030 align:start position:0%
1326
+ For example, let's say that V was a
1327
+ vector<00:06:10.400><c> with</c><00:06:10.600><c> length</c><00:06:10.920><c> two</c><00:06:11.400><c> pointing</c><00:06:11.720><c> straight</c>
1328
+
1329
+ 00:06:12.030 --> 00:06:12.040 align:start position:0%
1330
+ vector with length two pointing straight
1331
+
1332
+
1333
+ 00:06:12.040 --> 00:06:14.430 align:start position:0%
1334
+ vector with length two pointing straight
1335
+ up<00:06:12.160><c> in</c><00:06:12.240><c> the</c><00:06:12.320><c> Z</c><00:06:12.480><c> direction,</c><00:06:13.360><c> and</c><00:06:13.480><c> W</c><00:06:13.960><c> is</c><00:06:14.120><c> a</c><00:06:14.160><c> vector</c>
1336
+
1337
+ 00:06:14.430 --> 00:06:14.440 align:start position:0%
1338
+ up in the Z direction, and W is a vector
1339
+
1340
+
1341
+ 00:06:14.440 --> 00:06:16.550 align:start position:0%
1342
+ up in the Z direction, and W is a vector
1343
+ with<00:06:14.640><c> length</c><00:06:14.960><c> two</c><00:06:15.480><c> pointing</c><00:06:15.800><c> in</c><00:06:15.880><c> the</c><00:06:16.000><c> pure</c><00:06:16.320><c> Y</c>
1344
+
1345
+ 00:06:16.550 --> 00:06:16.560 align:start position:0%
1346
+ with length two pointing in the pure Y
1347
+
1348
+
1349
+ 00:06:16.560 --> 00:06:17.830 align:start position:0%
1350
+ with length two pointing in the pure Y
1351
+ direction.
1352
+
1353
+ 00:06:17.830 --> 00:06:17.840 align:start position:0%
1354
+ direction.
1355
+
1356
+
1357
+ 00:06:17.840 --> 00:06:19.510 align:start position:0%
1358
+ direction.
1359
+ The<00:06:18.040><c> parallelogram</c><00:06:18.800><c> that</c><00:06:18.919><c> they</c><00:06:19.040><c> define</c><00:06:19.440><c> in</c>
1360
+
1361
+ 00:06:19.510 --> 00:06:19.520 align:start position:0%
1362
+ The parallelogram that they define in
1363
+
1364
+
1365
+ 00:06:19.520 --> 00:06:20.990 align:start position:0%
1366
+ The parallelogram that they define in
1367
+ this<00:06:19.680><c> simple</c><00:06:20.000><c> example</c><00:06:20.480><c> is</c><00:06:20.640><c> actually</c><00:06:20.960><c> a</c>
1368
+
1369
+ 00:06:20.990 --> 00:06:21.000 align:start position:0%
1370
+ this simple example is actually a
1371
+
1372
+
1373
+ 00:06:21.000 --> 00:06:22.870 align:start position:0%
1374
+ this simple example is actually a
1375
+ square,<00:06:21.640><c> since</c><00:06:22.000><c> they're</c><00:06:22.160><c> perpendicular</c><00:06:22.760><c> and</c>
1376
+
1377
+ 00:06:22.870 --> 00:06:22.880 align:start position:0%
1378
+ square, since they're perpendicular and
1379
+
1380
+
1381
+ 00:06:22.880 --> 00:06:24.790 align:start position:0%
1382
+ square, since they're perpendicular and
1383
+ have<00:06:23.000><c> the</c><00:06:23.120><c> same</c><00:06:23.320><c> length.</c><00:06:24.080><c> And</c><00:06:24.200><c> the</c><00:06:24.320><c> area</c><00:06:24.680><c> of</c>
1384
+
1385
+ 00:06:24.790 --> 00:06:24.800 align:start position:0%
1386
+ have the same length. And the area of
1387
+
1388
+
1389
+ 00:06:24.800 --> 00:06:26.990 align:start position:0%
1390
+ have the same length. And the area of
1391
+ that<00:06:25.000><c> square</c><00:06:25.360><c> is</c><00:06:25.560><c> four.</c><00:06:26.440><c> So,</c><00:06:26.560><c> their</c><00:06:26.720><c> cross</c>
1392
+
1393
+ 00:06:26.990 --> 00:06:27.000 align:start position:0%
1394
+ that square is four. So, their cross
1395
+
1396
+
1397
+ 00:06:27.000 --> 00:06:28.430 align:start position:0%
1398
+ that square is four. So, their cross
1399
+ product<00:06:27.320><c> should</c><00:06:27.440><c> be</c><00:06:27.560><c> a</c><00:06:27.600><c> vector</c><00:06:27.960><c> with</c><00:06:28.160><c> length</c>
1400
+
1401
+ 00:06:28.430 --> 00:06:28.440 align:start position:0%
1402
+ product should be a vector with length
1403
+
1404
+
1405
+ 00:06:28.440 --> 00:06:29.710 align:start position:0%
1406
+ product should be a vector with length
1407
+ four.
1408
+
1409
+ 00:06:29.710 --> 00:06:29.720 align:start position:0%
1410
+ four.
1411
+
1412
+
1413
+ 00:06:29.720 --> 00:06:31.550 align:start position:0%
1414
+ four.
1415
+ Using<00:06:30.080><c> the</c><00:06:30.160><c> right-hand</c><00:06:30.640><c> rule,</c><00:06:31.160><c> their</c><00:06:31.320><c> cross</c>
1416
+
1417
+ 00:06:31.550 --> 00:06:31.560 align:start position:0%
1418
+ Using the right-hand rule, their cross
1419
+
1420
+
1421
+ 00:06:31.560 --> 00:06:33.230 align:start position:0%
1422
+ Using the right-hand rule, their cross
1423
+ product<00:06:31.880><c> should</c><00:06:32.040><c> point</c><00:06:32.320><c> in</c><00:06:32.440><c> the</c><00:06:32.560><c> negative</c><00:06:33.040><c> X</c>
1424
+
1425
+ 00:06:33.230 --> 00:06:33.240 align:start position:0%
1426
+ product should point in the negative X
1427
+
1428
+
1429
+ 00:06:33.240 --> 00:06:36.030 align:start position:0%
1430
+ product should point in the negative X
1431
+ direction.
1432
+
1433
+ 00:06:36.030 --> 00:06:36.040 align:start position:0%
1434
+
1435
+
1436
+
1437
+ 00:06:36.040 --> 00:06:37.230 align:start position:0%
1438
+
1439
+ So,<00:06:36.200><c> the</c><00:06:36.280><c> cross</c><00:06:36.520><c> product</c><00:06:36.840><c> of</c><00:06:36.919><c> these</c><00:06:37.120><c> two</c>
1440
+
1441
+ 00:06:37.230 --> 00:06:37.240 align:start position:0%
1442
+ So, the cross product of these two
1443
+
1444
+
1445
+ 00:06:37.240 --> 00:06:45.750 align:start position:0%
1446
+ So, the cross product of these two
1447
+ vectors<00:06:38.000><c> is</c><00:06:38.160><c> -4</c><00:06:38.800><c> *</c><00:06:39.160><c> I-hat.</c>
1448
+
1449
+ 00:06:45.750 --> 00:06:45.760 align:start position:0%
1450
+
1451
+
1452
+
1453
+ 00:06:45.760 --> 00:06:47.830 align:start position:0%
1454
+
1455
+ For<00:06:45.880><c> more</c><00:06:46.040><c> general</c><00:06:46.360><c> computations,</c><00:06:47.280><c> there</c><00:06:47.560><c> is</c>
1456
+
1457
+ 00:06:47.830 --> 00:06:47.840 align:start position:0%
1458
+ For more general computations, there is
1459
+
1460
+
1461
+ 00:06:47.840 --> 00:06:49.510 align:start position:0%
1462
+ For more general computations, there is
1463
+ a<00:06:47.880><c> formula</c><00:06:48.320><c> that</c><00:06:48.520><c> you</c><00:06:48.680><c> could</c><00:06:48.880><c> memorize</c><00:06:49.280><c> if</c><00:06:49.400><c> you</c>
1464
+
1465
+ 00:06:49.510 --> 00:06:49.520 align:start position:0%
1466
+ a formula that you could memorize if you
1467
+
1468
+
1469
+ 00:06:49.520 --> 00:06:51.230 align:start position:0%
1470
+ a formula that you could memorize if you
1471
+ wanted,<00:06:50.000><c> but</c><00:06:50.120><c> it's</c><00:06:50.240><c> common</c><00:06:50.600><c> and</c><00:06:50.760><c> easier</c><00:06:51.120><c> to</c>
1472
+
1473
+ 00:06:51.230 --> 00:06:51.240 align:start position:0%
1474
+ wanted, but it's common and easier to
1475
+
1476
+
1477
+ 00:06:51.240 --> 00:06:52.990 align:start position:0%
1478
+ wanted, but it's common and easier to
1479
+ instead<00:06:51.600><c> remember</c><00:06:51.960><c> a</c><00:06:52.040><c> certain</c><00:06:52.440><c> process</c>
1480
+
1481
+ 00:06:52.990 --> 00:06:53.000 align:start position:0%
1482
+ instead remember a certain process
1483
+
1484
+
1485
+ 00:06:53.000 --> 00:06:55.590 align:start position:0%
1486
+ instead remember a certain process
1487
+ involving<00:06:53.480><c> the</c><00:06:53.600><c> 3D</c><00:06:54.000><c> determinant.</c>
1488
+
1489
+ 00:06:55.590 --> 00:06:55.600 align:start position:0%
1490
+ involving the 3D determinant.
1491
+
1492
+
1493
+ 00:06:55.600 --> 00:06:58.070 align:start position:0%
1494
+ involving the 3D determinant.
1495
+ Now,<00:06:56.160><c> this</c><00:06:56.320><c> process</c><00:06:56.720><c> looks</c><00:06:57.120><c> truly</c><00:06:57.560><c> strange</c><00:06:57.960><c> at</c>
1496
+
1497
+ 00:06:58.070 --> 00:06:58.080 align:start position:0%
1498
+ Now, this process looks truly strange at
1499
+
1500
+
1501
+ 00:06:58.080 --> 00:07:00.670 align:start position:0%
1502
+ Now, this process looks truly strange at
1503
+ first.<00:06:58.880><c> You</c><00:06:59.000><c> write</c><00:06:59.200><c> down</c><00:06:59.400><c> a</c><00:06:59.440><c> 3D</c><00:06:59.800><c> matrix</c><00:07:00.560><c> where</c>
1504
+
1505
+ 00:07:00.670 --> 00:07:00.680 align:start position:0%
1506
+ first. You write down a 3D matrix where
1507
+
1508
+
1509
+ 00:07:00.680 --> 00:07:03.070 align:start position:0%
1510
+ first. You write down a 3D matrix where
1511
+ the<00:07:00.800><c> second</c><00:07:01.400><c> and</c><00:07:01.560><c> third</c><00:07:01.840><c> columns</c><00:07:02.640><c> contain</c><00:07:02.960><c> the</c>
1512
+
1513
+ 00:07:03.070 --> 00:07:03.080 align:start position:0%
1514
+ the second and third columns contain the
1515
+
1516
+
1517
+ 00:07:03.080 --> 00:07:05.909 align:start position:0%
1518
+ the second and third columns contain the
1519
+ coordinates<00:07:03.760><c> of</c><00:07:03.919><c> V</c><00:07:04.320><c> and</c><00:07:04.480><c> W.</c><00:07:05.480><c> But</c><00:07:05.600><c> for</c><00:07:05.720><c> that</c>
1520
+
1521
+ 00:07:05.909 --> 00:07:05.919 align:start position:0%
1522
+ coordinates of V and W. But for that
1523
+
1524
+
1525
+ 00:07:05.919 --> 00:07:07.830 align:start position:0%
1526
+ coordinates of V and W. But for that
1527
+ first<00:07:06.200><c> column,</c><00:07:06.919><c> you</c><00:07:07.040><c> write</c><00:07:07.280><c> the</c><00:07:07.400><c> basis</c>
1528
+
1529
+ 00:07:07.830 --> 00:07:07.840 align:start position:0%
1530
+ first column, you write the basis
1531
+
1532
+
1533
+ 00:07:07.840 --> 00:07:11.430 align:start position:0%
1534
+ first column, you write the basis
1535
+ vectors<00:07:08.320><c> I-hat,</c><00:07:09.080><c> J-hat,</c><00:07:09.760><c> and</c><00:07:09.880><c> K-hat.</c>
1536
+
1537
+ 00:07:11.430 --> 00:07:11.440 align:start position:0%
1538
+ vectors I-hat, J-hat, and K-hat.
1539
+
1540
+
1541
+ 00:07:11.440 --> 00:07:13.550 align:start position:0%
1542
+ vectors I-hat, J-hat, and K-hat.
1543
+ Then,<00:07:12.040><c> you</c><00:07:12.160><c> compute</c><00:07:12.640><c> the</c><00:07:12.760><c> determinant</c><00:07:13.440><c> of</c>
1544
+
1545
+ 00:07:13.550 --> 00:07:13.560 align:start position:0%
1546
+ Then, you compute the determinant of
1547
+
1548
+
1549
+ 00:07:13.560 --> 00:07:15.150 align:start position:0%
1550
+ Then, you compute the determinant of
1551
+ this<00:07:13.800><c> matrix.</c>
1552
+
1553
+ 00:07:15.150 --> 00:07:15.160 align:start position:0%
1554
+ this matrix.
1555
+
1556
+
1557
+ 00:07:15.160 --> 00:07:17.190 align:start position:0%
1558
+ this matrix.
1559
+ The<00:07:15.320><c> silliness</c><00:07:15.800><c> is</c><00:07:15.880><c> probably</c><00:07:16.280><c> clear</c><00:07:16.520><c> here.</c>
1560
+
1561
+ 00:07:17.190 --> 00:07:17.200 align:start position:0%
1562
+ The silliness is probably clear here.
1563
+
1564
+
1565
+ 00:07:17.200 --> 00:07:18.710 align:start position:0%
1566
+ The silliness is probably clear here.
1567
+ What<00:07:17.400><c> on</c><00:07:17.520><c> earth</c><00:07:17.720><c> does</c><00:07:17.880><c> it</c><00:07:17.960><c> mean</c><00:07:18.200><c> to</c><00:07:18.320><c> put</c><00:07:18.520><c> in</c><00:07:18.640><c> a</c>
1568
+
1569
+ 00:07:18.710 --> 00:07:18.720 align:start position:0%
1570
+ What on earth does it mean to put in a
1571
+
1572
+
1573
+ 00:07:18.720 --> 00:07:22.510 align:start position:0%
1574
+ What on earth does it mean to put in a
1575
+ vector<00:07:19.280><c> as</c><00:07:19.480><c> the</c><00:07:19.600><c> entry</c><00:07:20.000><c> of</c><00:07:20.120><c> a</c><00:07:20.200><c> matrix?</c>
1576
+
1577
+ 00:07:22.510 --> 00:07:22.520 align:start position:0%
1578
+ vector as the entry of a matrix?
1579
+
1580
+
1581
+ 00:07:22.520 --> 00:07:23.990 align:start position:0%
1582
+ vector as the entry of a matrix?
1583
+ Students<00:07:22.919><c> are</c><00:07:23.040><c> often</c><00:07:23.280><c> told</c><00:07:23.520><c> that</c><00:07:23.720><c> this</c><00:07:23.880><c> is</c>
1584
+
1585
+ 00:07:23.990 --> 00:07:24.000 align:start position:0%
1586
+ Students are often told that this is
1587
+
1588
+
1589
+ 00:07:24.000 --> 00:07:25.950 align:start position:0%
1590
+ Students are often told that this is
1591
+ just<00:07:24.200><c> a</c><00:07:24.240><c> notational</c><00:07:24.760><c> trick.</c><00:07:25.400><c> When</c><00:07:25.560><c> you</c><00:07:25.640><c> carry</c>
1592
+
1593
+ 00:07:25.950 --> 00:07:25.960 align:start position:0%
1594
+ just a notational trick. When you carry
1595
+
1596
+
1597
+ 00:07:25.960 --> 00:07:28.550 align:start position:0%
1598
+ just a notational trick. When you carry
1599
+ out<00:07:26.080><c> the</c><00:07:26.200><c> computations</c><00:07:27.000><c> as</c><00:07:27.280><c> if</c><00:07:27.680><c> I-hat,</c><00:07:28.120><c> J-hat,</c>
1600
+
1601
+ 00:07:28.550 --> 00:07:28.560 align:start position:0%
1602
+ out the computations as if I-hat, J-hat,
1603
+
1604
+
1605
+ 00:07:28.560 --> 00:07:30.470 align:start position:0%
1606
+ out the computations as if I-hat, J-hat,
1607
+ and<00:07:28.680><c> K-hat</c><00:07:29.080><c> were</c><00:07:29.200><c> numbers,</c><00:07:30.000><c> then</c><00:07:30.160><c> you</c><00:07:30.240><c> get</c>
1608
+
1609
+ 00:07:30.470 --> 00:07:30.480 align:start position:0%
1610
+ and K-hat were numbers, then you get
1611
+
1612
+
1613
+ 00:07:30.480 --> 00:07:32.390 align:start position:0%
1614
+ and K-hat were numbers, then you get
1615
+ some<00:07:30.840><c> linear</c><00:07:31.160><c> combination</c><00:07:31.720><c> of</c><00:07:31.800><c> those</c><00:07:32.040><c> basis</c>
1616
+
1617
+ 00:07:32.390 --> 00:07:32.400 align:start position:0%
1618
+ some linear combination of those basis
1619
+
1620
+
1621
+ 00:07:32.400 --> 00:07:36.070 align:start position:0%
1622
+ some linear combination of those basis
1623
+ vectors.
1624
+
1625
+ 00:07:36.070 --> 00:07:36.080 align:start position:0%
1626
+
1627
+
1628
+
1629
+ 00:07:36.080 --> 00:07:38.070 align:start position:0%
1630
+
1631
+ And<00:07:36.280><c> the</c><00:07:36.360><c> vector</c><00:07:36.720><c> defined</c><00:07:37.280><c> by</c><00:07:37.440><c> that</c><00:07:37.760><c> linear</c>
1632
+
1633
+ 00:07:38.070 --> 00:07:38.080 align:start position:0%
1634
+ And the vector defined by that linear
1635
+
1636
+
1637
+ 00:07:38.080 --> 00:07:40.030 align:start position:0%
1638
+ And the vector defined by that linear
1639
+ combination,<00:07:38.880><c> students</c><00:07:39.280><c> are</c><00:07:39.360><c> told</c><00:07:39.640><c> to</c><00:07:39.760><c> just</c>
1640
+
1641
+ 00:07:40.030 --> 00:07:40.040 align:start position:0%
1642
+ combination, students are told to just
1643
+
1644
+
1645
+ 00:07:40.040 --> 00:07:42.030 align:start position:0%
1646
+ combination, students are told to just
1647
+ believe,<00:07:40.960><c> is</c><00:07:41.160><c> the</c><00:07:41.280><c> unique</c><00:07:41.640><c> vector</c>
1648
+
1649
+ 00:07:42.030 --> 00:07:42.040 align:start position:0%
1650
+ believe, is the unique vector
1651
+
1652
+
1653
+ 00:07:42.040 --> 00:07:44.670 align:start position:0%
1654
+ believe, is the unique vector
1655
+ perpendicular<00:07:42.720><c> to</c><00:07:42.840><c> V</c><00:07:43.000><c> and</c><00:07:43.120><c> W</c><00:07:43.840><c> whose</c><00:07:44.040><c> magnitude</c>
1656
+
1657
+ 00:07:44.670 --> 00:07:44.680 align:start position:0%
1658
+ perpendicular to V and W whose magnitude
1659
+
1660
+
1661
+ 00:07:44.680 --> 00:07:45.950 align:start position:0%
1662
+ perpendicular to V and W whose magnitude
1663
+ is<00:07:44.800><c> the</c><00:07:44.960><c> area</c><00:07:45.280><c> of</c><00:07:45.400><c> the</c><00:07:45.520><c> appropriate</c>
1664
+
1665
+ 00:07:45.950 --> 00:07:45.960 align:start position:0%
1666
+ is the area of the appropriate
1667
+
1668
+
1669
+ 00:07:45.960 --> 00:07:48.230 align:start position:0%
1670
+ is the area of the appropriate
1671
+ parallelogram<00:07:47.000><c> and</c><00:07:47.120><c> whose</c><00:07:47.320><c> direction</c><00:07:47.840><c> obeys</c>
1672
+
1673
+ 00:07:48.230 --> 00:07:48.240 align:start position:0%
1674
+ parallelogram and whose direction obeys
1675
+
1676
+
1677
+ 00:07:48.240 --> 00:07:51.230 align:start position:0%
1678
+ parallelogram and whose direction obeys
1679
+ the<00:07:48.360><c> right-hand</c><00:07:48.800><c> rule.</c>
1680
+
1681
+ 00:07:51.230 --> 00:07:51.240 align:start position:0%
1682
+
1683
+
1684
+
1685
+ 00:07:51.240 --> 00:07:53.510 align:start position:0%
1686
+
1687
+ And<00:07:51.800><c> sure,</c><00:07:52.400><c> in</c><00:07:52.520><c> some</c><00:07:52.720><c> sense,</c><00:07:52.960><c> this</c><00:07:53.160><c> is</c><00:07:53.280><c> just</c><00:07:53.480><c> a</c>
1688
+
1689
+ 00:07:53.510 --> 00:07:53.520 align:start position:0%
1690
+ And sure, in some sense, this is just a
1691
+
1692
+
1693
+ 00:07:53.520 --> 00:07:55.790 align:start position:0%
1694
+ And sure, in some sense, this is just a
1695
+ notational<00:07:54.080><c> trick,</c><00:07:54.760><c> but</c><00:07:54.880><c> there</c><00:07:55.040><c> is</c><00:07:55.200><c> a</c><00:07:55.280><c> reason</c>
1696
+
1697
+ 00:07:55.790 --> 00:07:55.800 align:start position:0%
1698
+ notational trick, but there is a reason
1699
+
1700
+
1701
+ 00:07:55.800 --> 00:07:57.830 align:start position:0%
1702
+ notational trick, but there is a reason
1703
+ for<00:07:55.920><c> doing</c><00:07:56.240><c> it.</c>
1704
+
1705
+ 00:07:57.830 --> 00:07:57.840 align:start position:0%
1706
+ for doing it.
1707
+
1708
+
1709
+ 00:07:57.840 --> 00:07:59.350 align:start position:0%
1710
+ for doing it.
1711
+ It's<00:07:58.120><c> not</c><00:07:58.280><c> just</c><00:07:58.480><c> a</c><00:07:58.520><c> coincidence</c><00:07:59.160><c> that</c><00:07:59.280><c> the</c>
1712
+
1713
+ 00:07:59.350 --> 00:07:59.360 align:start position:0%
1714
+ It's not just a coincidence that the
1715
+
1716
+
1717
+ 00:07:59.360 --> 00:08:01.870 align:start position:0%
1718
+ It's not just a coincidence that the
1719
+ determinant<00:08:00.000><c> is</c><00:08:00.160><c> once</c><00:08:00.400><c> again</c><00:08:00.600><c> important.</c><00:08:01.680><c> And</c>
1720
+
1721
+ 00:08:01.870 --> 00:08:01.880 align:start position:0%
1722
+ determinant is once again important. And
1723
+
1724
+
1725
+ 00:08:01.880 --> 00:08:03.910 align:start position:0%
1726
+ determinant is once again important. And
1727
+ putting<00:08:02.120><c> the</c><00:08:02.200><c> basis</c><00:08:02.600><c> vectors</c><00:08:03.040><c> in</c><00:08:03.240><c> those</c><00:08:03.480><c> slots</c>
1728
+
1729
+ 00:08:03.910 --> 00:08:03.920 align:start position:0%
1730
+ putting the basis vectors in those slots
1731
+
1732
+
1733
+ 00:08:03.920 --> 00:08:06.150 align:start position:0%
1734
+ putting the basis vectors in those slots
1735
+ is<00:08:04.080><c> not</c><00:08:04.240><c> just</c><00:08:04.440><c> a</c><00:08:04.480><c> random</c><00:08:04.840><c> thing</c><00:08:05.000><c> to</c><00:08:05.120><c> do.</c>
1736
+
1737
+ 00:08:06.150 --> 00:08:06.160 align:start position:0%
1738
+ is not just a random thing to do.
1739
+
1740
+
1741
+ 00:08:06.160 --> 00:08:07.630 align:start position:0%
1742
+ is not just a random thing to do.
1743
+ To<00:08:06.360><c> understand</c><00:08:06.880><c> where</c><00:08:07.000><c> all</c><00:08:07.160><c> of</c><00:08:07.240><c> this</c><00:08:07.360><c> comes</c>
1744
+
1745
+ 00:08:07.630 --> 00:08:07.640 align:start position:0%
1746
+ To understand where all of this comes
1747
+
1748
+
1749
+ 00:08:07.640 --> 00:08:09.630 align:start position:0%
1750
+ To understand where all of this comes
1751
+ from,<00:08:08.160><c> it</c><00:08:08.320><c> helps</c><00:08:08.600><c> to</c><00:08:08.720><c> use</c><00:08:08.960><c> the</c><00:08:09.120><c> idea</c><00:08:09.480><c> of</c>
1752
+
1753
+ 00:08:09.630 --> 00:08:09.640 align:start position:0%
1754
+ from, it helps to use the idea of
1755
+
1756
+
1757
+ 00:08:09.640 --> 00:08:11.510 align:start position:0%
1758
+ from, it helps to use the idea of
1759
+ duality<00:08:10.440><c> that</c><00:08:10.600><c> I</c><00:08:10.680><c> introduced</c><00:08:11.120><c> in</c><00:08:11.200><c> the</c><00:08:11.280><c> last</c>
1760
+
1761
+ 00:08:11.510 --> 00:08:11.520 align:start position:0%
1762
+ duality that I introduced in the last
1763
+
1764
+
1765
+ 00:08:11.520 --> 00:08:12.670 align:start position:0%
1766
+ duality that I introduced in the last
1767
+ video.
1768
+
1769
+ 00:08:12.670 --> 00:08:12.680 align:start position:0%
1770
+ video.
1771
+
1772
+
1773
+ 00:08:12.680 --> 00:08:14.030 align:start position:0%
1774
+ video.
1775
+ This<00:08:12.880><c> concept</c><00:08:13.280><c> is</c><00:08:13.400><c> a</c><00:08:13.440><c> little</c><00:08:13.640><c> bit</c><00:08:13.760><c> heavy,</c>
1776
+
1777
+ 00:08:14.030 --> 00:08:14.040 align:start position:0%
1778
+ This concept is a little bit heavy,
1779
+
1780
+
1781
+ 00:08:14.040 --> 00:08:16.070 align:start position:0%
1782
+ This concept is a little bit heavy,
1783
+ though.<00:08:14.440><c> So,</c><00:08:14.919><c> I'm</c><00:08:15.040><c> putting</c><00:08:15.280><c> it</c><00:08:15.400><c> in</c><00:08:15.520><c> a</c><00:08:15.600><c> separate</c>
1784
+
1785
+ 00:08:16.070 --> 00:08:16.080 align:start position:0%
1786
+ though. So, I'm putting it in a separate
1787
+
1788
+
1789
+ 00:08:16.080 --> 00:08:17.950 align:start position:0%
1790
+ though. So, I'm putting it in a separate
1791
+ follow-on<00:08:16.600><c> video</c><00:08:17.200><c> for</c><00:08:17.360><c> any</c><00:08:17.520><c> of</c><00:08:17.600><c> you</c><00:08:17.720><c> who</c><00:08:17.840><c> are</c>
1792
+
1793
+ 00:08:17.950 --> 00:08:17.960 align:start position:0%
1794
+ follow-on video for any of you who are
1795
+
1796
+
1797
+ 00:08:17.960 --> 00:08:19.550 align:start position:0%
1798
+ follow-on video for any of you who are
1799
+ curious<00:08:18.360><c> to</c><00:08:18.480><c> learn</c><00:08:18.680><c> more.</c>
1800
+
1801
+ 00:08:19.550 --> 00:08:19.560 align:start position:0%
1802
+ curious to learn more.
1803
+
1804
+
1805
+ 00:08:19.560 --> 00:08:21.990 align:start position:0%
1806
+ curious to learn more.
1807
+ Arguably,<00:08:20.440><c> it</c><00:08:20.600><c> falls</c><00:08:20.960><c> outside</c><00:08:21.360><c> the</c><00:08:21.520><c> essence</c>
1808
+
1809
+ 00:08:21.990 --> 00:08:22.000 align:start position:0%
1810
+ Arguably, it falls outside the essence
1811
+
1812
+
1813
+ 00:08:22.000 --> 00:08:23.830 align:start position:0%
1814
+ Arguably, it falls outside the essence
1815
+ of<00:08:22.120><c> linear</c><00:08:22.440><c> algebra.</c><00:08:23.240><c> The</c><00:08:23.320><c> important</c><00:08:23.680><c> part</c>
1816
+
1817
+ 00:08:23.830 --> 00:08:23.840 align:start position:0%
1818
+ of linear algebra. The important part
1819
+
1820
+
1821
+ 00:08:23.840 --> 00:08:25.630 align:start position:0%
1822
+ of linear algebra. The important part
1823
+ here<00:08:24.120><c> is</c><00:08:24.280><c> to</c><00:08:24.400><c> know</c><00:08:24.640><c> what</c><00:08:24.840><c> that</c><00:08:25.000><c> cross</c><00:08:25.280><c> product</c>
1824
+
1825
+ 00:08:25.630 --> 00:08:25.640 align:start position:0%
1826
+ here is to know what that cross product
1827
+
1828
+
1829
+ 00:08:25.640 --> 00:08:27.909 align:start position:0%
1830
+ here is to know what that cross product
1831
+ vector<00:08:25.919><c> geometrically</c><00:08:26.640><c> represents.</c>
1832
+
1833
+ 00:08:27.909 --> 00:08:27.919 align:start position:0%
1834
+ vector geometrically represents.
1835
+
1836
+
1837
+ 00:08:27.919 --> 00:08:29.350 align:start position:0%
1838
+ vector geometrically represents.
1839
+ So,<00:08:28.080><c> if</c><00:08:28.160><c> you</c><00:08:28.280><c> want</c><00:08:28.400><c> to</c><00:08:28.480><c> skip</c><00:08:28.680><c> that</c><00:08:28.840><c> next</c><00:08:29.040><c> video,</c>
1840
+
1841
+ 00:08:29.350 --> 00:08:29.360 align:start position:0%
1842
+ So, if you want to skip that next video,
1843
+
1844
+
1845
+ 00:08:29.360 --> 00:08:31.350 align:start position:0%
1846
+ So, if you want to skip that next video,
1847
+ feel<00:08:29.720><c> free.</c><00:08:30.440><c> But</c><00:08:30.600><c> for</c><00:08:30.720><c> those</c><00:08:30.919><c> of</c><00:08:31.000><c> you</c><00:08:31.120><c> who</c><00:08:31.240><c> are</c>
1848
+
1849
+ 00:08:31.350 --> 00:08:31.360 align:start position:0%
1850
+ feel free. But for those of you who are
1851
+
1852
+
1853
+ 00:08:31.360 --> 00:08:32.990 align:start position:0%
1854
+ feel free. But for those of you who are
1855
+ willing<00:08:31.640><c> to</c><00:08:31.720><c> go</c><00:08:31.840><c> a</c><00:08:31.919><c> bit</c><00:08:32.080><c> deeper</c><00:08:32.719><c> and</c><00:08:32.840><c> who</c><00:08:32.919><c> are</c>
1856
+
1857
+ 00:08:32.990 --> 00:08:33.000 align:start position:0%
1858
+ willing to go a bit deeper and who are
1859
+
1860
+
1861
+ 00:08:33.000 --> 00:08:34.430 align:start position:0%
1862
+ willing to go a bit deeper and who are
1863
+ curious<00:08:33.440><c> about</c><00:08:33.680><c> the</c><00:08:33.760><c> connection</c><00:08:34.159><c> between</c>
1864
+
1865
+ 00:08:34.430 --> 00:08:34.440 align:start position:0%
1866
+ curious about the connection between
1867
+
1868
+
1869
+ 00:08:34.440 --> 00:08:36.110 align:start position:0%
1870
+ curious about the connection between
1871
+ this<00:08:34.640><c> computation</c><00:08:35.320><c> and</c><00:08:35.520><c> the</c><00:08:35.640><c> underlying</c>
1872
+
1873
+ 00:08:36.110 --> 00:08:36.120 align:start position:0%
1874
+ this computation and the underlying
1875
+
1876
+
1877
+ 00:08:36.120 --> 00:08:38.149 align:start position:0%
1878
+ this computation and the underlying
1879
+ geometry,<00:08:37.039><c> the</c><00:08:37.159><c> ideas</c><00:08:37.520><c> that</c><00:08:37.640><c> I'll</c><00:08:37.760><c> talk</c><00:08:37.960><c> about</c>
1880
+
1881
+ 00:08:38.149 --> 00:08:38.159 align:start position:0%
1882
+ geometry, the ideas that I'll talk about
1883
+
1884
+
1885
+ 00:08:38.159 --> 00:08:40.070 align:start position:0%
1886
+ geometry, the ideas that I'll talk about
1887
+ in<00:08:38.240><c> the</c><00:08:38.320><c> next</c><00:08:38.560><c> video</c><00:08:38.880><c> are</c><00:08:39.000><c> just</c><00:08:39.360><c> a</c><00:08:39.479><c> really</c>
1888
+
1889
+ 00:08:40.070 --> 00:08:40.080 align:start position:0%
1890
+ in the next video are just a really
1891
+
1892
+
1893
+ 00:08:40.080 --> 00:08:43.240 align:start position:0%
1894
+ in the next video are just a really
1895
+ elegant<00:08:40.520><c> piece</c><00:08:40.719><c> of</c><00:08:40.840><c> math.</c>
1896
+
3B1B_LinearAlgebra/10 - Cross products | Chapter 10, Essence of linear algebra.en.vtt ADDED
@@ -0,0 +1,1896 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:09.200 --> 00:00:10.830 align:start position:0%
6
+
7
+ Last<00:00:09.440><c> video,</c><00:00:09.880><c> I</c><00:00:09.960><c> talked</c><00:00:10.200><c> about</c><00:00:10.440><c> the</c><00:00:10.560><c> dot</c>
8
+
9
+ 00:00:10.830 --> 00:00:10.840 align:start position:0%
10
+ Last video, I talked about the dot
11
+
12
+
13
+ 00:00:10.840 --> 00:00:12.630 align:start position:0%
14
+ Last video, I talked about the dot
15
+ product,<00:00:11.560><c> showing</c><00:00:11.880><c> both</c><00:00:12.120><c> the</c><00:00:12.200><c> standard</c>
16
+
17
+ 00:00:12.630 --> 00:00:12.640 align:start position:0%
18
+ product, showing both the standard
19
+
20
+
21
+ 00:00:12.640 --> 00:00:14.630 align:start position:0%
22
+ product, showing both the standard
23
+ introduction<00:00:13.160><c> to</c><00:00:13.240><c> the</c><00:00:13.320><c> topic,</c><00:00:14.080><c> as</c><00:00:14.280><c> well</c><00:00:14.440><c> as</c><00:00:14.560><c> a</c>
24
+
25
+ 00:00:14.630 --> 00:00:14.640 align:start position:0%
26
+ introduction to the topic, as well as a
27
+
28
+
29
+ 00:00:14.640 --> 00:00:16.510 align:start position:0%
30
+ introduction to the topic, as well as a
31
+ deeper<00:00:15.040><c> view</c><00:00:15.240><c> of</c><00:00:15.400><c> how</c><00:00:15.520><c> it</c><00:00:15.640><c> relates</c><00:00:16.000><c> to</c><00:00:16.160><c> linear</c>
32
+
33
+ 00:00:16.510 --> 00:00:16.520 align:start position:0%
34
+ deeper view of how it relates to linear
35
+
36
+
37
+ 00:00:16.520 --> 00:00:18.750 align:start position:0%
38
+ deeper view of how it relates to linear
39
+ transformations.
40
+
41
+ 00:00:18.750 --> 00:00:18.760 align:start position:0%
42
+ transformations.
43
+
44
+
45
+ 00:00:18.760 --> 00:00:20.070 align:start position:0%
46
+ transformations.
47
+ I'd<00:00:18.880><c> like</c><00:00:19.040><c> to</c><00:00:19.120><c> do</c><00:00:19.240><c> the</c><00:00:19.320><c> same</c><00:00:19.560><c> thing</c><00:00:19.720><c> for</c><00:00:19.840><c> cross</c>
48
+
49
+ 00:00:20.070 --> 00:00:20.080 align:start position:0%
50
+ I'd like to do the same thing for cross
51
+
52
+
53
+ 00:00:20.080 --> 00:00:21.750 align:start position:0%
54
+ I'd like to do the same thing for cross
55
+ products,<00:00:20.560><c> which</c><00:00:20.880><c> also</c><00:00:21.160><c> have</c><00:00:21.320><c> a</c><00:00:21.360><c> standard</c>
56
+
57
+ 00:00:21.750 --> 00:00:21.760 align:start position:0%
58
+ products, which also have a standard
59
+
60
+
61
+ 00:00:21.760 --> 00:00:23.670 align:start position:0%
62
+ products, which also have a standard
63
+ introduction<00:00:22.680><c> along</c><00:00:22.960><c> with</c><00:00:23.160><c> a</c><00:00:23.280><c> deeper</c>
64
+
65
+ 00:00:23.670 --> 00:00:23.680 align:start position:0%
66
+ introduction along with a deeper
67
+
68
+
69
+ 00:00:23.680 --> 00:00:25.230 align:start position:0%
70
+ introduction along with a deeper
71
+ understanding<00:00:24.360><c> in</c><00:00:24.520><c> the</c><00:00:24.600><c> light</c><00:00:24.840><c> of</c><00:00:24.960><c> linear</c>
72
+
73
+ 00:00:25.230 --> 00:00:25.240 align:start position:0%
74
+ understanding in the light of linear
75
+
76
+
77
+ 00:00:25.240 --> 00:00:27.150 align:start position:0%
78
+ understanding in the light of linear
79
+ transformations,<00:00:26.400><c> but</c><00:00:26.560><c> this</c><00:00:26.680><c> time</c><00:00:27.040><c> I'm</c>
80
+
81
+ 00:00:27.150 --> 00:00:27.160 align:start position:0%
82
+ transformations, but this time I'm
83
+
84
+
85
+ 00:00:27.160 --> 00:00:29.670 align:start position:0%
86
+ transformations, but this time I'm
87
+ dividing<00:00:27.560><c> it</c><00:00:27.680><c> into</c><00:00:27.920><c> two</c><00:00:28.080><c> separate</c><00:00:28.400><c> videos.</c>
88
+
89
+ 00:00:29.670 --> 00:00:29.680 align:start position:0%
90
+ dividing it into two separate videos.
91
+
92
+
93
+ 00:00:29.680 --> 00:00:31.350 align:start position:0%
94
+ dividing it into two separate videos.
95
+ Here,<00:00:30.120><c> I'll</c><00:00:30.240><c> try</c><00:00:30.400><c> to</c><00:00:30.520><c> hit</c><00:00:30.640><c> the</c><00:00:30.760><c> main</c><00:00:31.040><c> points</c>
96
+
97
+ 00:00:31.350 --> 00:00:31.360 align:start position:0%
98
+ Here, I'll try to hit the main points
99
+
100
+
101
+ 00:00:31.360 --> 00:00:32.950 align:start position:0%
102
+ Here, I'll try to hit the main points
103
+ that<00:00:31.520><c> students</c><00:00:31.880><c> are</c><00:00:31.960><c> usually</c><00:00:32.279><c> shown</c><00:00:32.640><c> about</c>
104
+
105
+ 00:00:32.950 --> 00:00:32.960 align:start position:0%
106
+ that students are usually shown about
107
+
108
+
109
+ 00:00:32.960 --> 00:00:34.590 align:start position:0%
110
+ that students are usually shown about
111
+ the<00:00:33.040><c> cross</c><00:00:33.320><c> product,</c><00:00:34.040><c> and</c><00:00:34.160><c> in</c><00:00:34.240><c> the</c><00:00:34.320><c> next</c>
112
+
113
+ 00:00:34.590 --> 00:00:34.600 align:start position:0%
114
+ the cross product, and in the next
115
+
116
+
117
+ 00:00:34.600 --> 00:00:36.750 align:start position:0%
118
+ the cross product, and in the next
119
+ video,<00:00:35.360><c> I'll</c><00:00:35.440><c> be</c><00:00:35.560><c> showing</c><00:00:35.840><c> a</c><00:00:35.880><c> view</c><00:00:36.120><c> which</c><00:00:36.360><c> is</c>
120
+
121
+ 00:00:36.750 --> 00:00:36.760 align:start position:0%
122
+ video, I'll be showing a view which is
123
+
124
+
125
+ 00:00:36.760 --> 00:00:38.830 align:start position:0%
126
+ video, I'll be showing a view which is
127
+ less<00:00:37.080><c> commonly</c><00:00:37.640><c> taught,</c><00:00:38.160><c> but</c><00:00:38.520><c> really</c>
128
+
129
+ 00:00:38.830 --> 00:00:38.840 align:start position:0%
130
+ less commonly taught, but really
131
+
132
+
133
+ 00:00:38.840 --> 00:00:40.830 align:start position:0%
134
+ less commonly taught, but really
135
+ satisfying<00:00:39.480><c> when</c><00:00:39.600><c> you</c><00:00:39.720><c> learn</c><00:00:39.960><c> it.</c><00:00:40.720><c> We'll</c>
136
+
137
+ 00:00:40.830 --> 00:00:40.840 align:start position:0%
138
+ satisfying when you learn it. We'll
139
+
140
+
141
+ 00:00:40.840 --> 00:00:42.870 align:start position:0%
142
+ satisfying when you learn it. We'll
143
+ start<00:00:41.080><c> in</c><00:00:41.200><c> two</c><00:00:41.320><c> dimensions.</c><00:00:42.280><c> If</c><00:00:42.400><c> you</c><00:00:42.480><c> have</c><00:00:42.720><c> two</c>
144
+
145
+ 00:00:42.870 --> 00:00:42.880 align:start position:0%
146
+ start in two dimensions. If you have two
147
+
148
+
149
+ 00:00:42.880 --> 00:00:45.430 align:start position:0%
150
+ start in two dimensions. If you have two
151
+ vectors,<00:00:43.560><c> V</c><00:00:43.920><c> and</c><00:00:44.080><c> W,</c><00:00:44.920><c> think</c><00:00:45.080><c> about</c><00:00:45.320><c> the</c>
152
+
153
+ 00:00:45.430 --> 00:00:45.440 align:start position:0%
154
+ vectors, V and W, think about the
155
+
156
+
157
+ 00:00:45.440 --> 00:00:47.990 align:start position:0%
158
+ vectors, V and W, think about the
159
+ parallelogram<00:00:46.360><c> that</c><00:00:46.520><c> they</c><00:00:46.640><c> span</c><00:00:47.040><c> out.</c><00:00:47.880><c> What</c><00:00:47.960><c> I</c>
160
+
161
+ 00:00:47.990 --> 00:00:48.000 align:start position:0%
162
+ parallelogram that they span out. What I
163
+
164
+
165
+ 00:00:48.000 --> 00:00:50.070 align:start position:0%
166
+ parallelogram that they span out. What I
167
+ mean<00:00:48.200><c> by</c><00:00:48.320><c> that</c><00:00:48.880><c> is</c><00:00:49.000><c> that</c><00:00:49.120><c> if</c><00:00:49.240><c> you</c><00:00:49.360><c> take</c><00:00:49.560><c> a</c><00:00:49.600><c> copy</c>
168
+
169
+ 00:00:50.070 --> 00:00:50.080 align:start position:0%
170
+ mean by that is that if you take a copy
171
+
172
+
173
+ 00:00:50.080 --> 00:00:53.270 align:start position:0%
174
+ mean by that is that if you take a copy
175
+ of<00:00:50.200><c> V</c><00:00:50.840><c> and</c><00:00:51.040><c> move</c><00:00:51.320><c> its</c><00:00:51.520><c> tail</c><00:00:51.880><c> to</c><00:00:52.000><c> the</c><00:00:52.120><c> tip</c><00:00:52.320><c> of</c><00:00:52.440><c> W,</c>
176
+
177
+ 00:00:53.270 --> 00:00:53.280 align:start position:0%
178
+ of V and move its tail to the tip of W,
179
+
180
+
181
+ 00:00:53.280 --> 00:00:55.030 align:start position:0%
182
+ of V and move its tail to the tip of W,
183
+ and<00:00:53.400><c> you</c><00:00:53.480><c> take</c><00:00:53.640><c> a</c><00:00:53.720><c> copy</c><00:00:54.000><c> of</c><00:00:54.120><c> W</c><00:00:54.560><c> and</c><00:00:54.680><c> move</c><00:00:54.840><c> its</c>
184
+
185
+ 00:00:55.030 --> 00:00:55.040 align:start position:0%
186
+ and you take a copy of W and move its
187
+
188
+
189
+ 00:00:55.040 --> 00:00:57.510 align:start position:0%
190
+ and you take a copy of W and move its
191
+ tail<00:00:55.360><c> to</c><00:00:55.480><c> the</c><00:00:55.560><c> tip</c><00:00:55.760><c> of</c><00:00:55.880><c> V,</c><00:00:56.800><c> the</c><00:00:56.880><c> four</c><00:00:57.120><c> vectors</c>
192
+
193
+ 00:00:57.510 --> 00:00:57.520 align:start position:0%
194
+ tail to the tip of V, the four vectors
195
+
196
+
197
+ 00:00:57.520 --> 00:00:59.630 align:start position:0%
198
+ tail to the tip of V, the four vectors
199
+ now<00:00:57.720><c> on</c><00:00:57.800><c> the</c><00:00:57.880><c> screen</c><00:00:58.640><c> enclose</c><00:00:59.240><c> a</c><00:00:59.320><c> certain</c>
200
+
201
+ 00:00:59.630 --> 00:00:59.640 align:start position:0%
202
+ now on the screen enclose a certain
203
+
204
+
205
+ 00:00:59.640 --> 00:01:01.910 align:start position:0%
206
+ now on the screen enclose a certain
207
+ parallelogram.
208
+
209
+ 00:01:01.910 --> 00:01:01.920 align:start position:0%
210
+ parallelogram.
211
+
212
+
213
+ 00:01:01.920 --> 00:01:04.350 align:start position:0%
214
+ parallelogram.
215
+ The<00:01:02.080><c> cross</c><00:01:02.480><c> product</c><00:01:02.880><c> of</c><00:01:03.000><c> V</c><00:01:03.160><c> and</c><00:01:03.280><c> W,</c><00:01:04.080><c> written</c>
216
+
217
+ 00:01:04.350 --> 00:01:04.360 align:start position:0%
218
+ The cross product of V and W, written
219
+
220
+
221
+ 00:01:04.360 --> 00:01:06.790 align:start position:0%
222
+ The cross product of V and W, written
223
+ with<00:01:04.480><c> the</c><00:01:04.720><c> X-shaped</c><00:01:05.320><c> multiplication</c><00:01:06.000><c> symbol,</c>
224
+
225
+ 00:01:06.790 --> 00:01:06.800 align:start position:0%
226
+ with the X-shaped multiplication symbol,
227
+
228
+
229
+ 00:01:06.800 --> 00:01:10.830 align:start position:0%
230
+ with the X-shaped multiplication symbol,
231
+ is<00:01:06.960><c> the</c><00:01:07.160><c> area</c><00:01:07.680><c> of</c><00:01:07.800><c> this</c><00:01:08.000><c> parallelogram.</c>
232
+
233
+ 00:01:10.830 --> 00:01:10.840 align:start position:0%
234
+ is the area of this parallelogram.
235
+
236
+
237
+ 00:01:10.840 --> 00:01:12.670 align:start position:0%
238
+ is the area of this parallelogram.
239
+ Well,<00:01:11.200><c> almost.</c><00:01:11.760><c> We</c><00:01:11.840><c> also</c><00:01:12.040><c> need</c><00:01:12.200><c> to</c><00:01:12.280><c> consider</c>
240
+
241
+ 00:01:12.670 --> 00:01:12.680 align:start position:0%
242
+ Well, almost. We also need to consider
243
+
244
+
245
+ 00:01:12.680 --> 00:01:15.390 align:start position:0%
246
+ Well, almost. We also need to consider
247
+ orientation.<00:01:13.880><c> Basically,</c><00:01:14.400><c> if</c><00:01:14.760><c> V</c><00:01:15.000><c> is</c><00:01:15.120><c> on</c><00:01:15.240><c> the</c>
248
+
249
+ 00:01:15.390 --> 00:01:15.400 align:start position:0%
250
+ orientation. Basically, if V is on the
251
+
252
+
253
+ 00:01:15.400 --> 00:01:18.630 align:start position:0%
254
+ orientation. Basically, if V is on the
255
+ right<00:01:15.760><c> of</c><00:01:15.880><c> W,</c><00:01:16.640><c> then</c><00:01:16.880><c> V</c><00:01:17.120><c> cross</c><00:01:17.400><c> W</c><00:01:17.840><c> is</c><00:01:18.040><c> positive</c>
256
+
257
+ 00:01:18.630 --> 00:01:18.640 align:start position:0%
258
+ right of W, then V cross W is positive
259
+
260
+
261
+ 00:01:18.640 --> 00:01:19.870 align:start position:0%
262
+ right of W, then V cross W is positive
263
+ and<00:01:18.840><c> equal</c><00:01:19.080><c> to</c><00:01:19.200><c> the</c><00:01:19.320><c> area</c><00:01:19.680><c> of</c><00:01:19.800><c> the</c>
264
+
265
+ 00:01:19.870 --> 00:01:19.880 align:start position:0%
266
+ and equal to the area of the
267
+
268
+
269
+ 00:01:19.880 --> 00:01:21.510 align:start position:0%
270
+ and equal to the area of the
271
+ parallelogram.
272
+
273
+ 00:01:21.510 --> 00:01:21.520 align:start position:0%
274
+ parallelogram.
275
+
276
+
277
+ 00:01:21.520 --> 00:01:23.870 align:start position:0%
278
+ parallelogram.
279
+ But<00:01:21.640><c> if</c><00:01:21.800><c> V</c><00:01:22.040><c> is</c><00:01:22.160><c> on</c><00:01:22.280><c> the</c><00:01:22.400><c> left</c><00:01:22.800><c> of</c><00:01:22.880><c> W,</c><00:01:23.680><c> then</c><00:01:23.800><c> the</c>
280
+
281
+ 00:01:23.870 --> 00:01:23.880 align:start position:0%
282
+ But if V is on the left of W, then the
283
+
284
+
285
+ 00:01:23.880 --> 00:01:25.910 align:start position:0%
286
+ But if V is on the left of W, then the
287
+ cross<00:01:24.120><c> product</c><00:01:24.440><c> is</c><00:01:24.600><c> negative,</c><00:01:25.280><c> namely</c><00:01:25.640><c> the</c>
288
+
289
+ 00:01:25.910 --> 00:01:25.920 align:start position:0%
290
+ cross product is negative, namely the
291
+
292
+
293
+ 00:01:25.920 --> 00:01:28.550 align:start position:0%
294
+ cross product is negative, namely the
295
+ negative<00:01:26.400><c> area</c><00:01:26.800><c> of</c><00:01:26.920><c> that</c><00:01:27.120><c> parallelogram.</c>
296
+
297
+ 00:01:28.550 --> 00:01:28.560 align:start position:0%
298
+ negative area of that parallelogram.
299
+
300
+
301
+ 00:01:28.560 --> 00:01:30.990 align:start position:0%
302
+ negative area of that parallelogram.
303
+ Notice,<00:01:29.040><c> this</c><00:01:29.280><c> means</c><00:01:29.520><c> that</c><00:01:29.800><c> order</c><00:01:30.120><c> matters.</c>
304
+
305
+ 00:01:30.990 --> 00:01:31.000 align:start position:0%
306
+ Notice, this means that order matters.
307
+
308
+
309
+ 00:01:31.000 --> 00:01:33.790 align:start position:0%
310
+ Notice, this means that order matters.
311
+ If<00:01:31.160><c> you</c><00:01:31.320><c> swapped</c><00:01:31.760><c> V</c><00:01:31.920><c> and</c><00:01:32.040><c> W,</c><00:01:32.720><c> instead</c><00:01:33.120><c> taking</c><00:01:33.400><c> W</c>
312
+
313
+ 00:01:33.790 --> 00:01:33.800 align:start position:0%
314
+ If you swapped V and W, instead taking W
315
+
316
+
317
+ 00:01:33.800 --> 00:01:35.830 align:start position:0%
318
+ If you swapped V and W, instead taking W
319
+ cross<00:01:34.080><c> V,</c><00:01:34.760><c> the</c><00:01:34.840><c> cross</c><00:01:35.080><c> product</c><00:01:35.440><c> would</c><00:01:35.560><c> become</c>
320
+
321
+ 00:01:35.830 --> 00:01:35.840 align:start position:0%
322
+ cross V, the cross product would become
323
+
324
+
325
+ 00:01:35.840 --> 00:01:38.950 align:start position:0%
326
+ cross V, the cross product would become
327
+ the<00:01:35.960><c> negative</c><00:01:36.520><c> of</c><00:01:36.680><c> whatever</c><00:01:37.040><c> it</c><00:01:37.160><c> was</c><00:01:37.360><c> before.</c>
328
+
329
+ 00:01:38.950 --> 00:01:38.960 align:start position:0%
330
+ the negative of whatever it was before.
331
+
332
+
333
+ 00:01:38.960 --> 00:01:40.470 align:start position:0%
334
+ the negative of whatever it was before.
335
+ The<00:01:39.040><c> way</c><00:01:39.240><c> I</c><00:01:39.360><c> always</c><00:01:39.640><c> remember</c><00:01:39.960><c> the</c><00:01:40.080><c> ordering</c>
336
+
337
+ 00:01:40.470 --> 00:01:40.480 align:start position:0%
338
+ The way I always remember the ordering
339
+
340
+
341
+ 00:01:40.480 --> 00:01:42.110 align:start position:0%
342
+ The way I always remember the ordering
343
+ here<00:01:41.040><c> is</c><00:01:41.160><c> that</c><00:01:41.360><c> when</c><00:01:41.480><c> you</c><00:01:41.560><c> take</c><00:01:41.760><c> the</c><00:01:41.880><c> cross</c>
344
+
345
+ 00:01:42.110 --> 00:01:42.120 align:start position:0%
346
+ here is that when you take the cross
347
+
348
+
349
+ 00:01:42.120 --> 00:01:44.150 align:start position:0%
350
+ here is that when you take the cross
351
+ product<00:01:42.480><c> of</c><00:01:42.600><c> the</c><00:01:42.760><c> two</c><00:01:43.120><c> basis</c><00:01:43.560><c> vectors</c><00:01:44.000><c> in</c>
352
+
353
+ 00:01:44.150 --> 00:01:44.160 align:start position:0%
354
+ product of the two basis vectors in
355
+
356
+
357
+ 00:01:44.160 --> 00:01:46.830 align:start position:0%
358
+ product of the two basis vectors in
359
+ order,<00:01:44.760><c> I</c><00:01:44.920><c> hat</c><00:01:45.240><c> cross</c><00:01:45.520><c> J</c><00:01:45.760><c> hat,</c><00:01:46.360><c> the</c><00:01:46.480><c> result</c>
360
+
361
+ 00:01:46.830 --> 00:01:46.840 align:start position:0%
362
+ order, I hat cross J hat, the result
363
+
364
+
365
+ 00:01:46.840 --> 00:01:49.150 align:start position:0%
366
+ order, I hat cross J hat, the result
367
+ should<00:01:46.960><c> be</c><00:01:47.080><c> positive.</c><00:01:48.160><c> In</c><00:01:48.240><c> fact,</c><00:01:48.760><c> the</c><00:01:48.880><c> order</c>
368
+
369
+ 00:01:49.150 --> 00:01:49.160 align:start position:0%
370
+ should be positive. In fact, the order
371
+
372
+
373
+ 00:01:49.160 --> 00:01:51.230 align:start position:0%
374
+ should be positive. In fact, the order
375
+ of<00:01:49.240><c> your</c><00:01:49.360><c> basis</c><00:01:49.720><c> vectors</c><00:01:50.160><c> is</c><00:01:50.360><c> what</c><00:01:50.520><c> defines</c>
376
+
377
+ 00:01:51.230 --> 00:01:51.240 align:start position:0%
378
+ of your basis vectors is what defines
379
+
380
+
381
+ 00:01:51.240 --> 00:01:52.710 align:start position:0%
382
+ of your basis vectors is what defines
383
+ orientation.
384
+
385
+ 00:01:52.710 --> 00:01:52.720 align:start position:0%
386
+ orientation.
387
+
388
+
389
+ 00:01:52.720 --> 00:01:54.630 align:start position:0%
390
+ orientation.
391
+ So,<00:01:52.960><c> since</c><00:01:53.320><c> I</c><00:01:53.440><c> hat</c><00:01:53.720><c> is</c><00:01:53.840><c> on</c><00:01:53.960><c> the</c><00:01:54.040><c> right</c><00:01:54.320><c> of</c><00:01:54.440><c> J</c>
392
+
393
+ 00:01:54.630 --> 00:01:54.640 align:start position:0%
394
+ So, since I hat is on the right of J
395
+
396
+
397
+ 00:01:54.640 --> 00:01:57.470 align:start position:0%
398
+ So, since I hat is on the right of J
399
+ hat,<00:01:55.240><c> I</c><00:01:55.280><c> remember</c><00:01:55.600><c> that</c><00:01:55.840><c> V</c><00:01:56.040><c> cross</c><00:01:56.320><c> W</c><00:01:57.040><c> has</c><00:01:57.240><c> to</c><00:01:57.320><c> be</c>
400
+
401
+ 00:01:57.470 --> 00:01:57.480 align:start position:0%
402
+ hat, I remember that V cross W has to be
403
+
404
+
405
+ 00:01:57.480 --> 00:01:59.430 align:start position:0%
406
+ hat, I remember that V cross W has to be
407
+ positive<00:01:57.960><c> whenever</c><00:01:58.320><c> V</c><00:01:58.680><c> is</c><00:01:58.800><c> on</c><00:01:58.920><c> the</c><00:01:59.040><c> right</c><00:01:59.320><c> of</c>
408
+
409
+ 00:01:59.430 --> 00:01:59.440 align:start position:0%
410
+ positive whenever V is on the right of
411
+
412
+
413
+ 00:01:59.440 --> 00:02:01.550 align:start position:0%
414
+ positive whenever V is on the right of
415
+ W.
416
+
417
+ 00:02:01.550 --> 00:02:01.560 align:start position:0%
418
+ W.
419
+
420
+
421
+ 00:02:01.560 --> 00:02:03.430 align:start position:0%
422
+ W.
423
+ So,<00:02:01.920><c> for</c><00:02:02.080><c> example,</c><00:02:02.720><c> with</c><00:02:02.840><c> the</c><00:02:02.920><c> vectors</c><00:02:03.200><c> shown</c>
424
+
425
+ 00:02:03.430 --> 00:02:03.440 align:start position:0%
426
+ So, for example, with the vectors shown
427
+
428
+
429
+ 00:02:03.440 --> 00:02:05.430 align:start position:0%
430
+ So, for example, with the vectors shown
431
+ here,<00:02:04.000><c> I'll</c><00:02:04.120><c> just</c><00:02:04.400><c> tell</c><00:02:04.600><c> you</c><00:02:04.760><c> that</c><00:02:04.920><c> the</c><00:02:05.040><c> area</c>
432
+
433
+ 00:02:05.430 --> 00:02:05.440 align:start position:0%
434
+ here, I'll just tell you that the area
435
+
436
+
437
+ 00:02:05.440 --> 00:02:07.830 align:start position:0%
438
+ here, I'll just tell you that the area
439
+ of<00:02:05.560><c> that</c><00:02:05.760><c> parallelogram</c><00:02:06.479><c> is</c><00:02:06.680><c> seven.</c><00:02:07.600><c> And</c>
440
+
441
+ 00:02:07.830 --> 00:02:07.840 align:start position:0%
442
+ of that parallelogram is seven. And
443
+
444
+
445
+ 00:02:07.840 --> 00:02:10.350 align:start position:0%
446
+ of that parallelogram is seven. And
447
+ since<00:02:08.080><c> V</c><00:02:08.399><c> is</c><00:02:08.560><c> on</c><00:02:08.679><c> the</c><00:02:08.800><c> left</c><00:02:09.160><c> of</c><00:02:09.280><c> W,</c><00:02:10.000><c> the</c><00:02:10.119><c> cross</c>
448
+
449
+ 00:02:10.350 --> 00:02:10.360 align:start position:0%
450
+ since V is on the left of W, the cross
451
+
452
+
453
+ 00:02:10.360 --> 00:02:12.390 align:start position:0%
454
+ since V is on the left of W, the cross
455
+ product<00:02:10.720><c> should</c><00:02:10.840><c> be</c><00:02:10.960><c> negative.</c><00:02:11.600><c> So,</c><00:02:12.000><c> V</c><00:02:12.160><c> cross</c>
456
+
457
+ 00:02:12.390 --> 00:02:12.400 align:start position:0%
458
+ product should be negative. So, V cross
459
+
460
+
461
+ 00:02:12.400 --> 00:02:15.830 align:start position:0%
462
+ product should be negative. So, V cross
463
+ W<00:02:12.880><c> is</c><00:02:13.120><c> -7.</c>
464
+
465
+ 00:02:15.830 --> 00:02:15.840 align:start position:0%
466
+ W is -7.
467
+
468
+
469
+ 00:02:15.840 --> 00:02:17.350 align:start position:0%
470
+ W is -7.
471
+ But,<00:02:16.040><c> of</c><00:02:16.120><c> course,</c><00:02:16.640><c> you</c><00:02:16.760><c> want</c><00:02:16.920><c> to</c><00:02:16.959><c> be</c><00:02:17.080><c> able</c><00:02:17.280><c> to</c>
472
+
473
+ 00:02:17.350 --> 00:02:17.360 align:start position:0%
474
+ But, of course, you want to be able to
475
+
476
+
477
+ 00:02:17.360 --> 00:02:19.070 align:start position:0%
478
+ But, of course, you want to be able to
479
+ compute<00:02:17.720><c> this</c><00:02:17.959><c> without</c><00:02:18.320><c> someone</c><00:02:18.640><c> telling</c><00:02:18.959><c> you</c>
480
+
481
+ 00:02:19.070 --> 00:02:19.080 align:start position:0%
482
+ compute this without someone telling you
483
+
484
+
485
+ 00:02:19.080 --> 00:02:20.430 align:start position:0%
486
+ compute this without someone telling you
487
+ the<00:02:19.200><c> area.</c>
488
+
489
+ 00:02:20.430 --> 00:02:20.440 align:start position:0%
490
+ the area.
491
+
492
+
493
+ 00:02:20.440 --> 00:02:22.830 align:start position:0%
494
+ the area.
495
+ This<00:02:20.800><c> is</c><00:02:20.959><c> where</c><00:02:21.120><c> the</c><00:02:21.240><c> determinant</c><00:02:21.840><c> comes</c><00:02:22.120><c> in.</c>
496
+
497
+ 00:02:22.830 --> 00:02:22.840 align:start position:0%
498
+ This is where the determinant comes in.
499
+
500
+
501
+ 00:02:22.840 --> 00:02:24.630 align:start position:0%
502
+ This is where the determinant comes in.
503
+ So,<00:02:23.200><c> if</c><00:02:23.360><c> you</c><00:02:23.480><c> didn't</c><00:02:23.800><c> see</c><00:02:23.959><c> chapter</c><00:02:24.280><c> five</c><00:02:24.520><c> of</c>
504
+
505
+ 00:02:24.630 --> 00:02:24.640 align:start position:0%
506
+ So, if you didn't see chapter five of
507
+
508
+
509
+ 00:02:24.640 --> 00:02:25.949 align:start position:0%
510
+ So, if you didn't see chapter five of
511
+ this<00:02:24.800><c> series</c><00:02:25.280><c> where</c><00:02:25.400><c> I</c><00:02:25.480><c> talk</c><00:02:25.680><c> about</c><00:02:25.880><c> the</c>
512
+
513
+ 00:02:25.949 --> 00:02:25.959 align:start position:0%
514
+ this series where I talk about the
515
+
516
+
517
+ 00:02:25.959 --> 00:02:27.990 align:start position:0%
518
+ this series where I talk about the
519
+ determinant,<00:02:26.840><c> now</c><00:02:27.120><c> would</c><00:02:27.280><c> be</c><00:02:27.440><c> a</c><00:02:27.520><c> really</c><00:02:27.800><c> good</c>
520
+
521
+ 00:02:27.990 --> 00:02:28.000 align:start position:0%
522
+ determinant, now would be a really good
523
+
524
+
525
+ 00:02:28.000 --> 00:02:30.190 align:start position:0%
526
+ determinant, now would be a really good
527
+ time<00:02:28.280><c> to</c><00:02:28.400><c> go</c><00:02:28.560><c> take</c><00:02:28.760><c> a</c><00:02:28.800><c> look.</c><00:02:29.520><c> Even</c><00:02:29.760><c> if</c><00:02:29.880><c> you</c><00:02:30.000><c> did</c>
528
+
529
+ 00:02:30.190 --> 00:02:30.200 align:start position:0%
530
+ time to go take a look. Even if you did
531
+
532
+
533
+ 00:02:30.200 --> 00:02:31.750 align:start position:0%
534
+ time to go take a look. Even if you did
535
+ see<00:02:30.320><c> it,</c><00:02:30.480><c> but</c><00:02:30.600><c> it</c><00:02:30.680><c> was</c><00:02:30.800><c> a</c><00:02:30.840><c> while</c><00:02:31.080><c> ago,</c><00:02:31.640><c> I'd</c>
536
+
537
+ 00:02:31.750 --> 00:02:31.760 align:start position:0%
538
+ see it, but it was a while ago, I'd
539
+
540
+
541
+ 00:02:31.760 --> 00:02:33.310 align:start position:0%
542
+ see it, but it was a while ago, I'd
543
+ recommend<00:02:32.200><c> taking</c><00:02:32.520><c> another</c><00:02:32.800><c> look</c><00:02:33.040><c> just</c><00:02:33.240><c> to</c>
544
+
545
+ 00:02:33.310 --> 00:02:33.320 align:start position:0%
546
+ recommend taking another look just to
547
+
548
+
549
+ 00:02:33.320 --> 00:02:34.670 align:start position:0%
550
+ recommend taking another look just to
551
+ make<00:02:33.480><c> sure</c><00:02:33.600><c> those</c><00:02:33.800><c> ideas</c><00:02:34.120><c> are</c><00:02:34.200><c> fresh</c><00:02:34.440><c> in</c><00:02:34.560><c> your</c>
552
+
553
+ 00:02:34.670 --> 00:02:34.680 align:start position:0%
554
+ make sure those ideas are fresh in your
555
+
556
+
557
+ 00:02:34.680 --> 00:02:37.230 align:start position:0%
558
+ make sure those ideas are fresh in your
559
+ mind.
560
+
561
+ 00:02:37.230 --> 00:02:37.240 align:start position:0%
562
+
563
+
564
+
565
+ 00:02:37.240 --> 00:02:40.070 align:start position:0%
566
+
567
+ For<00:02:37.440><c> the</c><00:02:37.600><c> 2D</c><00:02:37.920><c> cross</c><00:02:38.160><c> product,</c><00:02:38.720><c> V</c><00:02:38.880><c> cross</c><00:02:39.160><c> W,</c>
568
+
569
+ 00:02:40.070 --> 00:02:40.080 align:start position:0%
570
+ For the 2D cross product, V cross W,
571
+
572
+
573
+ 00:02:40.080 --> 00:02:41.870 align:start position:0%
574
+ For the 2D cross product, V cross W,
575
+ what<00:02:40.320><c> you</c><00:02:40.400><c> do</c><00:02:40.760><c> is</c><00:02:40.920><c> you</c><00:02:41.080><c> write</c><00:02:41.320><c> the</c><00:02:41.440><c> coordinates</c>
576
+
577
+ 00:02:41.870 --> 00:02:41.880 align:start position:0%
578
+ what you do is you write the coordinates
579
+
580
+
581
+ 00:02:41.880 --> 00:02:44.350 align:start position:0%
582
+ what you do is you write the coordinates
583
+ of<00:02:42.000><c> V</c><00:02:42.400><c> as</c><00:02:42.560><c> the</c><00:02:42.680><c> first</c><00:02:42.959><c> column</c><00:02:43.240><c> of</c><00:02:43.360><c> a</c><00:02:43.400><c> matrix,</c>
584
+
585
+ 00:02:44.350 --> 00:02:44.360 align:start position:0%
586
+ of V as the first column of a matrix,
587
+
588
+
589
+ 00:02:44.360 --> 00:02:46.030 align:start position:0%
590
+ of V as the first column of a matrix,
591
+ and<00:02:44.480><c> you</c><00:02:44.560><c> take</c><00:02:44.720><c> the</c><00:02:44.800><c> coordinates</c><00:02:45.280><c> of</c><00:02:45.400><c> W</c><00:02:45.920><c> and</c>
592
+
593
+ 00:02:46.030 --> 00:02:46.040 align:start position:0%
594
+ and you take the coordinates of W and
595
+
596
+
597
+ 00:02:46.040 --> 00:02:47.790 align:start position:0%
598
+ and you take the coordinates of W and
599
+ make<00:02:46.240><c> them</c><00:02:46.360><c> the</c><00:02:46.440><c> second</c><00:02:46.800><c> column.</c><00:02:47.560><c> Then,</c><00:02:47.720><c> you</c>
600
+
601
+ 00:02:47.790 --> 00:02:47.800 align:start position:0%
602
+ make them the second column. Then, you
603
+
604
+
605
+ 00:02:47.800 --> 00:02:51.270 align:start position:0%
606
+ make them the second column. Then, you
607
+ just<00:02:48.120><c> compute</c><00:02:48.440><c> the</c><00:02:48.560><c> determinant.</c>
608
+
609
+ 00:02:51.270 --> 00:02:51.280 align:start position:0%
610
+ just compute the determinant.
611
+
612
+
613
+ 00:02:51.280 --> 00:02:53.350 align:start position:0%
614
+ just compute the determinant.
615
+ This<00:02:51.560><c> is</c><00:02:51.680><c> because</c><00:02:52.120><c> a</c><00:02:52.160><c> matrix</c><00:02:52.640><c> whose</c><00:02:52.840><c> columns</c>
616
+
617
+ 00:02:53.350 --> 00:02:53.360 align:start position:0%
618
+ This is because a matrix whose columns
619
+
620
+
621
+ 00:02:53.360 --> 00:02:55.630 align:start position:0%
622
+ This is because a matrix whose columns
623
+ represent<00:02:53.800><c> V</c><00:02:54.000><c> and</c><00:02:54.120><c> W</c><00:02:54.959><c> corresponds</c><00:02:55.440><c> with</c><00:02:55.560><c> a</c>
624
+
625
+ 00:02:55.630 --> 00:02:55.640 align:start position:0%
626
+ represent V and W corresponds with a
627
+
628
+
629
+ 00:02:55.640 --> 00:02:57.910 align:start position:0%
630
+ represent V and W corresponds with a
631
+ linear<00:02:56.080><c> transformation</c><00:02:57.280><c> that</c><00:02:57.440><c> moves</c><00:02:57.840><c> the</c>
632
+
633
+ 00:02:57.910 --> 00:02:57.920 align:start position:0%
634
+ linear transformation that moves the
635
+
636
+
637
+ 00:02:57.920 --> 00:03:00.870 align:start position:0%
638
+ linear transformation that moves the
639
+ basis<00:02:58.320><c> vectors</c><00:02:58.880><c> I</c><00:02:59.000><c> hat</c><00:02:59.280><c> and</c><00:02:59.400><c> J</c><00:02:59.600><c> hat</c><00:03:00.160><c> to</c><00:03:00.280><c> V</c><00:03:00.720><c> and</c>
640
+
641
+ 00:03:00.870 --> 00:03:00.880 align:start position:0%
642
+ basis vectors I hat and J hat to V and
643
+
644
+
645
+ 00:03:00.880 --> 00:03:06.550 align:start position:0%
646
+ basis vectors I hat and J hat to V and
647
+ W.
648
+
649
+ 00:03:06.550 --> 00:03:06.560 align:start position:0%
650
+
651
+
652
+
653
+ 00:03:06.560 --> 00:03:08.470 align:start position:0%
654
+
655
+ The<00:03:06.760><c> determinant</c><00:03:07.360><c> is</c><00:03:07.560><c> all</c><00:03:07.760><c> about</c><00:03:08.040><c> measuring</c>
656
+
657
+ 00:03:08.470 --> 00:03:08.480 align:start position:0%
658
+ The determinant is all about measuring
659
+
660
+
661
+ 00:03:08.480 --> 00:03:10.030 align:start position:0%
662
+ The determinant is all about measuring
663
+ how<00:03:08.680><c> areas</c><00:03:09.160><c> change</c><00:03:09.640><c> due</c><00:03:09.840><c> to</c><00:03:10.000><c> a</c>
664
+
665
+ 00:03:10.030 --> 00:03:10.040 align:start position:0%
666
+ how areas change due to a
667
+
668
+
669
+ 00:03:10.040 --> 00:03:12.310 align:start position:0%
670
+ how areas change due to a
671
+ transformation.<00:03:11.280><c> And</c><00:03:11.440><c> the</c><00:03:11.560><c> prototypical</c>
672
+
673
+ 00:03:12.310 --> 00:03:12.320 align:start position:0%
674
+ transformation. And the prototypical
675
+
676
+
677
+ 00:03:12.320 --> 00:03:14.630 align:start position:0%
678
+ transformation. And the prototypical
679
+ area<00:03:12.640><c> that</c><00:03:12.760><c> we</c><00:03:12.880><c> look</c><00:03:13.120><c> at</c><00:03:13.519><c> is</c><00:03:13.680><c> the</c><00:03:13.800><c> unit</c><00:03:14.120><c> square</c>
680
+
681
+ 00:03:14.630 --> 00:03:14.640 align:start position:0%
682
+ area that we look at is the unit square
683
+
684
+
685
+ 00:03:14.640 --> 00:03:17.390 align:start position:0%
686
+ area that we look at is the unit square
687
+ resting<00:03:15.120><c> on</c><00:03:15.320><c> I</c><00:03:15.440><c> hat</c><00:03:15.760><c> and</c><00:03:15.920><c> J</c><00:03:16.120><c> hat.</c><00:03:16.920><c> After</c><00:03:17.280><c> the</c>
688
+
689
+ 00:03:17.390 --> 00:03:17.400 align:start position:0%
690
+ resting on I hat and J hat. After the
691
+
692
+
693
+ 00:03:17.400 --> 00:03:19.590 align:start position:0%
694
+ resting on I hat and J hat. After the
695
+ transformation,<00:03:18.440><c> that</c><00:03:18.640><c> square</c><00:03:18.920><c> gets</c><00:03:19.200><c> turned</c>
696
+
697
+ 00:03:19.590 --> 00:03:19.600 align:start position:0%
698
+ transformation, that square gets turned
699
+
700
+
701
+ 00:03:19.600 --> 00:03:21.390 align:start position:0%
702
+ transformation, that square gets turned
703
+ into<00:03:20.080><c> the</c><00:03:20.200><c> parallelogram</c><00:03:20.920><c> that</c><00:03:21.040><c> we</c><00:03:21.160><c> care</c>
704
+
705
+ 00:03:21.390 --> 00:03:21.400 align:start position:0%
706
+ into the parallelogram that we care
707
+
708
+
709
+ 00:03:21.400 --> 00:03:23.630 align:start position:0%
710
+ into the parallelogram that we care
711
+ about.<00:03:22.320><c> So,</c><00:03:22.560><c> the</c><00:03:22.640><c> determinant,</c><00:03:23.360><c> which</c>
712
+
713
+ 00:03:23.630 --> 00:03:23.640 align:start position:0%
714
+ about. So, the determinant, which
715
+
716
+
717
+ 00:03:23.640 --> 00:03:25.350 align:start position:0%
718
+ about. So, the determinant, which
719
+ generally<00:03:24.040><c> measures</c><00:03:24.440><c> the</c><00:03:24.560><c> factor</c><00:03:24.959><c> by</c><00:03:25.120><c> which</c>
720
+
721
+ 00:03:25.350 --> 00:03:25.360 align:start position:0%
722
+ generally measures the factor by which
723
+
724
+
725
+ 00:03:25.360 --> 00:03:27.870 align:start position:0%
726
+ generally measures the factor by which
727
+ areas<00:03:25.760><c> are</c><00:03:25.840><c> changed,</c><00:03:26.720><c> gives</c><00:03:26.959><c> the</c><00:03:27.200><c> area</c><00:03:27.680><c> of</c>
728
+
729
+ 00:03:27.870 --> 00:03:27.880 align:start position:0%
730
+ areas are changed, gives the area of
731
+
732
+
733
+ 00:03:27.880 --> 00:03:30.229 align:start position:0%
734
+ areas are changed, gives the area of
735
+ this<00:03:28.120><c> parallelogram</c><00:03:29.239><c> since</c><00:03:29.560><c> it</c><00:03:29.720><c> evolved</c><00:03:30.120><c> from</c>
736
+
737
+ 00:03:30.229 --> 00:03:30.239 align:start position:0%
738
+ this parallelogram since it evolved from
739
+
740
+
741
+ 00:03:30.239 --> 00:03:32.670 align:start position:0%
742
+ this parallelogram since it evolved from
743
+ a<00:03:30.320><c> square</c><00:03:30.640><c> that</c><00:03:30.800><c> started</c><00:03:31.160><c> with</c><00:03:31.320><c> area</c><00:03:31.600><c> one.</c>
744
+
745
+ 00:03:32.670 --> 00:03:32.680 align:start position:0%
746
+ a square that started with area one.
747
+
748
+
749
+ 00:03:32.680 --> 00:03:35.590 align:start position:0%
750
+ a square that started with area one.
751
+ What's<00:03:32.959><c> more,</c><00:03:33.560><c> if</c><00:03:33.800><c> V</c><00:03:34.000><c> is</c><00:03:34.160><c> on</c><00:03:34.280><c> the</c><00:03:34.400><c> left</c><00:03:34.720><c> of</c><00:03:34.800><c> W,</c>
752
+
753
+ 00:03:35.590 --> 00:03:35.600 align:start position:0%
754
+ What's more, if V is on the left of W,
755
+
756
+
757
+ 00:03:35.600 --> 00:03:37.470 align:start position:0%
758
+ What's more, if V is on the left of W,
759
+ it<00:03:35.720><c> means</c><00:03:35.920><c> that</c><00:03:36.160><c> orientation</c><00:03:36.840><c> was</c><00:03:37.040><c> flipped</c>
760
+
761
+ 00:03:37.470 --> 00:03:37.480 align:start position:0%
762
+ it means that orientation was flipped
763
+
764
+
765
+ 00:03:37.480 --> 00:03:39.430 align:start position:0%
766
+ it means that orientation was flipped
767
+ during<00:03:37.760><c> that</c><00:03:37.920><c> transformation,</c><00:03:39.160><c> which</c><00:03:39.320><c> is</c>
768
+
769
+ 00:03:39.430 --> 00:03:39.440 align:start position:0%
770
+ during that transformation, which is
771
+
772
+
773
+ 00:03:39.440 --> 00:03:40.990 align:start position:0%
774
+ during that transformation, which is
775
+ what<00:03:39.600><c> it</c><00:03:39.680><c> means</c><00:03:39.959><c> for</c><00:03:40.080><c> the</c><00:03:40.160><c> determinant</c><00:03:40.840><c> to</c><00:03:40.920><c> be</c>
776
+
777
+ 00:03:40.990 --> 00:03:41.000 align:start position:0%
778
+ what it means for the determinant to be
779
+
780
+
781
+ 00:03:41.000 --> 00:03:44.070 align:start position:0%
782
+ what it means for the determinant to be
783
+ negative.
784
+
785
+ 00:03:44.070 --> 00:03:44.080 align:start position:0%
786
+
787
+
788
+
789
+ 00:03:44.080 --> 00:03:46.070 align:start position:0%
790
+
791
+ As<00:03:44.320><c> an</c><00:03:44.400><c> example,</c><00:03:45.160><c> let's</c><00:03:45.320><c> say</c><00:03:45.480><c> V</c><00:03:45.840><c> has</c>
792
+
793
+ 00:03:46.070 --> 00:03:46.080 align:start position:0%
794
+ As an example, let's say V has
795
+
796
+
797
+ 00:03:46.080 --> 00:03:49.550 align:start position:0%
798
+ As an example, let's say V has
799
+ coordinates<00:03:46.640><c> -3</c><00:03:47.519><c> 1,</c><00:03:48.239><c> and</c><00:03:48.360><c> W</c><00:03:48.800><c> has</c><00:03:49.000><c> coordinates</c>
800
+
801
+ 00:03:49.550 --> 00:03:49.560 align:start position:0%
802
+ coordinates -3 1, and W has coordinates
803
+
804
+
805
+ 00:03:49.560 --> 00:03:52.830 align:start position:0%
806
+ coordinates -3 1, and W has coordinates
807
+ 2<00:03:49.959><c> 1.</c><00:03:50.800><c> The</c><00:03:51.080><c> determinant</c><00:03:51.760><c> of</c><00:03:51.880><c> the</c><00:03:52.000><c> matrix</c><00:03:52.600><c> those</c>
808
+
809
+ 00:03:52.830 --> 00:03:52.840 align:start position:0%
810
+ 2 1. The determinant of the matrix those
811
+
812
+
813
+ 00:03:52.840 --> 00:03:56.830 align:start position:0%
814
+ 2 1. The determinant of the matrix those
815
+ coordinates<00:03:53.280><c> as</c><00:03:53.400><c> columns</c><00:03:54.160><c> is</c><00:03:54.600><c> -3</c><00:03:55.240><c> *</c><00:03:55.560><c> 1</c>
816
+
817
+ 00:03:56.830 --> 00:03:56.840 align:start position:0%
818
+ coordinates as columns is -3 * 1
819
+
820
+
821
+ 00:03:56.840 --> 00:03:59.590 align:start position:0%
822
+ coordinates as columns is -3 * 1
823
+ 2<00:03:57.040><c> *</c><00:03:57.440><c> 1,</c>
824
+
825
+ 00:03:59.590 --> 00:03:59.600 align:start position:0%
826
+ 2 * 1,
827
+
828
+
829
+ 00:03:59.600 --> 00:04:01.550 align:start position:0%
830
+ 2 * 1,
831
+ which<00:03:59.840><c> is</c><00:04:00.080><c> -5.</c>
832
+
833
+ 00:04:01.550 --> 00:04:01.560 align:start position:0%
834
+ which is -5.
835
+
836
+
837
+ 00:04:01.560 --> 00:04:03.150 align:start position:0%
838
+ which is -5.
839
+ So,<00:04:01.680><c> evidently,</c><00:04:02.360><c> the</c><00:04:02.560><c> area</c><00:04:02.920><c> of</c><00:04:03.040><c> the</c>
840
+
841
+ 00:04:03.150 --> 00:04:03.160 align:start position:0%
842
+ So, evidently, the area of the
843
+
844
+
845
+ 00:04:03.160 --> 00:04:05.750 align:start position:0%
846
+ So, evidently, the area of the
847
+ parallelogram<00:04:03.880><c> they</c><00:04:03.960><c> define</c><00:04:04.680><c> is</c><00:04:04.880><c> five.</c><00:04:05.560><c> And</c>
848
+
849
+ 00:04:05.750 --> 00:04:05.760 align:start position:0%
850
+ parallelogram they define is five. And
851
+
852
+
853
+ 00:04:05.760 --> 00:04:08.110 align:start position:0%
854
+ parallelogram they define is five. And
855
+ since<00:04:06.000><c> V</c><00:04:06.240><c> is</c><00:04:06.400><c> on</c><00:04:06.520><c> the</c><00:04:06.640><c> left</c><00:04:07.000><c> of</c><00:04:07.120><c> W,</c><00:04:07.840><c> it</c><00:04:07.959><c> should</c>
856
+
857
+ 00:04:08.110 --> 00:04:08.120 align:start position:0%
858
+ since V is on the left of W, it should
859
+
860
+
861
+ 00:04:08.120 --> 00:04:11.230 align:start position:0%
862
+ since V is on the left of W, it should
863
+ make<00:04:08.280><c> sense</c><00:04:08.560><c> that</c><00:04:08.720><c> this</c><00:04:08.880><c> value</c><00:04:09.200><c> is</c><00:04:09.320><c> negative.</c>
864
+
865
+ 00:04:11.230 --> 00:04:11.240 align:start position:0%
866
+ make sense that this value is negative.
867
+
868
+
869
+ 00:04:11.240 --> 00:04:13.310 align:start position:0%
870
+ make sense that this value is negative.
871
+ As<00:04:11.440><c> with</c><00:04:11.640><c> any</c><00:04:11.840><c> new</c><00:04:12.040><c> operation</c><00:04:12.520><c> you</c><00:04:12.640><c> learn,</c><00:04:13.160><c> I'd</c>
872
+
873
+ 00:04:13.310 --> 00:04:13.320 align:start position:0%
874
+ As with any new operation you learn, I'd
875
+
876
+
877
+ 00:04:13.320 --> 00:04:14.590 align:start position:0%
878
+ As with any new operation you learn, I'd
879
+ recommend<00:04:13.720><c> playing</c><00:04:14.040><c> around</c><00:04:14.360><c> with</c><00:04:14.480><c> this</c>
880
+
881
+ 00:04:14.590 --> 00:04:14.600 align:start position:0%
882
+ recommend playing around with this
883
+
884
+
885
+ 00:04:14.600 --> 00:04:16.069 align:start position:0%
886
+ recommend playing around with this
887
+ notion<00:04:14.920><c> a</c><00:04:14.959><c> bit</c><00:04:15.160><c> in</c><00:04:15.240><c> your</c><00:04:15.360><c> head</c><00:04:15.640><c> just</c><00:04:15.840><c> to</c><00:04:15.920><c> get</c>
888
+
889
+ 00:04:16.069 --> 00:04:16.079 align:start position:0%
890
+ notion a bit in your head just to get
891
+
892
+
893
+ 00:04:16.079 --> 00:04:17.630 align:start position:0%
894
+ notion a bit in your head just to get
895
+ kind<00:04:16.239><c> of</c><00:04:16.359><c> an</c><00:04:16.440><c> intuitive</c><00:04:17.000><c> feel</c><00:04:17.280><c> for</c><00:04:17.440><c> what</c><00:04:17.560><c> the</c>
896
+
897
+ 00:04:17.630 --> 00:04:17.640 align:start position:0%
898
+ kind of an intuitive feel for what the
899
+
900
+
901
+ 00:04:17.640 --> 00:04:19.710 align:start position:0%
902
+ kind of an intuitive feel for what the
903
+ cross<00:04:17.880><c> product</c><00:04:18.239><c> is</c><00:04:18.359><c> all</c><00:04:18.480><c> about.</c>
904
+
905
+ 00:04:19.710 --> 00:04:19.720 align:start position:0%
906
+ cross product is all about.
907
+
908
+
909
+ 00:04:19.720 --> 00:04:21.229 align:start position:0%
910
+ cross product is all about.
911
+ For<00:04:19.799><c> example,</c><00:04:20.359><c> you</c><00:04:20.440><c> might</c><00:04:20.640><c> notice</c><00:04:20.920><c> that</c><00:04:21.079><c> when</c>
912
+
913
+ 00:04:21.229 --> 00:04:21.239 align:start position:0%
914
+ For example, you might notice that when
915
+
916
+
917
+ 00:04:21.239 --> 00:04:23.110 align:start position:0%
918
+ For example, you might notice that when
919
+ two<00:04:21.400><c> vectors</c><00:04:21.799><c> are</c><00:04:21.920><c> perpendicular,</c><00:04:22.760><c> or</c><00:04:23.000><c> at</c>
920
+
921
+ 00:04:23.110 --> 00:04:23.120 align:start position:0%
922
+ two vectors are perpendicular, or at
923
+
924
+
925
+ 00:04:23.120 --> 00:04:24.990 align:start position:0%
926
+ two vectors are perpendicular, or at
927
+ least<00:04:23.360><c> close</c><00:04:23.600><c> to</c><00:04:23.720><c> being</c><00:04:23.919><c> perpendicular,</c>
928
+
929
+ 00:04:24.990 --> 00:04:25.000 align:start position:0%
930
+ least close to being perpendicular,
931
+
932
+
933
+ 00:04:25.000 --> 00:04:26.630 align:start position:0%
934
+ least close to being perpendicular,
935
+ their<00:04:25.160><c> cross</c><00:04:25.400><c> product</c><00:04:25.760><c> is</c><00:04:25.919><c> larger</c><00:04:26.360><c> than</c><00:04:26.520><c> it</c>
936
+
937
+ 00:04:26.630 --> 00:04:26.640 align:start position:0%
938
+ their cross product is larger than it
939
+
940
+
941
+ 00:04:26.640 --> 00:04:28.550 align:start position:0%
942
+ their cross product is larger than it
943
+ would<00:04:26.840><c> be</c><00:04:27.360><c> if</c><00:04:27.480><c> they</c><00:04:27.600><c> were</c><00:04:27.720><c> pointing</c><00:04:28.080><c> in</c><00:04:28.200><c> very</c>
944
+
945
+ 00:04:28.550 --> 00:04:28.560 align:start position:0%
946
+ would be if they were pointing in very
947
+
948
+
949
+ 00:04:28.560 --> 00:04:31.070 align:start position:0%
950
+ would be if they were pointing in very
951
+ similar<00:04:28.960><c> directions,</c><00:04:29.960><c> because</c><00:04:30.280><c> the</c><00:04:30.480><c> area</c><00:04:30.960><c> of</c>
952
+
953
+ 00:04:31.070 --> 00:04:31.080 align:start position:0%
954
+ similar directions, because the area of
955
+
956
+
957
+ 00:04:31.080 --> 00:04:33.550 align:start position:0%
958
+ similar directions, because the area of
959
+ that<00:04:31.280><c> parallelogram</c><00:04:32.440><c> is</c><00:04:32.600><c> larger</c><00:04:33.320><c> when</c><00:04:33.440><c> the</c>
960
+
961
+ 00:04:33.550 --> 00:04:33.560 align:start position:0%
962
+ that parallelogram is larger when the
963
+
964
+
965
+ 00:04:33.560 --> 00:04:37.110 align:start position:0%
966
+ that parallelogram is larger when the
967
+ sides<00:04:33.880><c> are</c><00:04:33.960><c> closer</c><00:04:34.320><c> to</c><00:04:34.400><c> being</c><00:04:34.640><c> perpendicular.</c>
968
+
969
+ 00:04:37.110 --> 00:04:37.120 align:start position:0%
970
+ sides are closer to being perpendicular.
971
+
972
+
973
+ 00:04:37.120 --> 00:04:38.590 align:start position:0%
974
+ sides are closer to being perpendicular.
975
+ Something<00:04:37.360><c> else</c><00:04:37.520><c> you</c><00:04:37.600><c> might</c><00:04:37.760><c> notice</c><00:04:38.280><c> is</c><00:04:38.440><c> that</c>
976
+
977
+ 00:04:38.590 --> 00:04:38.600 align:start position:0%
978
+ Something else you might notice is that
979
+
980
+
981
+ 00:04:38.600 --> 00:04:40.150 align:start position:0%
982
+ Something else you might notice is that
983
+ if<00:04:38.720><c> you</c><00:04:38.800><c> were</c><00:04:38.919><c> to</c><00:04:39.040><c> scale</c><00:04:39.560><c> up</c><00:04:39.720><c> one</c><00:04:39.880><c> of</c><00:04:39.919><c> those</c>
984
+
985
+ 00:04:40.150 --> 00:04:40.160 align:start position:0%
986
+ if you were to scale up one of those
987
+
988
+
989
+ 00:04:40.160 --> 00:04:43.590 align:start position:0%
990
+ if you were to scale up one of those
991
+ vectors,<00:04:40.960><c> perhaps</c><00:04:41.360><c> multiplying</c><00:04:41.960><c> V</c><00:04:42.440><c> by</c><00:04:42.560><c> three,</c>
992
+
993
+ 00:04:43.590 --> 00:04:43.600 align:start position:0%
994
+ vectors, perhaps multiplying V by three,
995
+
996
+
997
+ 00:04:43.600 --> 00:04:46.110 align:start position:0%
998
+ vectors, perhaps multiplying V by three,
999
+ then<00:04:43.680><c> the</c><00:04:43.919><c> area</c><00:04:44.400><c> of</c><00:04:44.520><c> that</c><00:04:44.760><c> parallelogram</c><00:04:45.919><c> is</c>
1000
+
1001
+ 00:04:46.110 --> 00:04:46.120 align:start position:0%
1002
+ then the area of that parallelogram is
1003
+
1004
+
1005
+ 00:04:46.120 --> 00:04:48.950 align:start position:0%
1006
+ then the area of that parallelogram is
1007
+ also<00:04:46.440><c> scaled</c><00:04:46.880><c> up</c><00:04:47.040><c> by</c><00:04:47.200><c> a</c><00:04:47.280><c> factor</c><00:04:47.640><c> of</c><00:04:47.760><c> three.</c>
1008
+
1009
+ 00:04:48.950 --> 00:04:48.960 align:start position:0%
1010
+ also scaled up by a factor of three.
1011
+
1012
+
1013
+ 00:04:48.960 --> 00:04:50.830 align:start position:0%
1014
+ also scaled up by a factor of three.
1015
+ So,<00:04:49.160><c> what</c><00:04:49.360><c> this</c><00:04:49.480><c> means</c><00:04:49.720><c> for</c><00:04:49.800><c> the</c><00:04:49.919><c> operation</c><00:04:50.680><c> is</c>
1016
+
1017
+ 00:04:50.830 --> 00:04:50.840 align:start position:0%
1018
+ So, what this means for the operation is
1019
+
1020
+
1021
+ 00:04:50.840 --> 00:04:54.510 align:start position:0%
1022
+ So, what this means for the operation is
1023
+ that<00:04:51.360><c> 3V</c><00:04:52.200><c> cross</c><00:04:52.520><c> W</c><00:04:53.320><c> will</c><00:04:53.440><c> be</c><00:04:53.600><c> exactly</c><00:04:54.240><c> three</c>
1024
+
1025
+ 00:04:54.510 --> 00:04:54.520 align:start position:0%
1026
+ that 3V cross W will be exactly three
1027
+
1028
+
1029
+ 00:04:54.520 --> 00:04:58.110 align:start position:0%
1030
+ that 3V cross W will be exactly three
1031
+ times<00:04:54.840><c> the</c><00:04:54.919><c> value</c><00:04:55.520><c> of</c><00:04:55.680><c> V</c><00:04:55.880><c> cross</c><00:04:56.160><c> W.</c>
1032
+
1033
+ 00:04:58.110 --> 00:04:58.120 align:start position:0%
1034
+ times the value of V cross W.
1035
+
1036
+
1037
+ 00:04:58.120 --> 00:04:59.510 align:start position:0%
1038
+ times the value of V cross W.
1039
+ Now,<00:04:58.480><c> even</c><00:04:58.720><c> though</c><00:04:58.880><c> all</c><00:04:59.040><c> of</c><00:04:59.120><c> this</c><00:04:59.320><c> is</c><00:04:59.440><c> a</c>
1040
+
1041
+ 00:04:59.510 --> 00:04:59.520 align:start position:0%
1042
+ Now, even though all of this is a
1043
+
1044
+
1045
+ 00:04:59.520 --> 00:05:01.790 align:start position:0%
1046
+ Now, even though all of this is a
1047
+ perfectly<00:05:00.000><c> fine</c><00:05:00.240><c> mathematical</c><00:05:00.800><c> operation,</c>
1048
+
1049
+ 00:05:01.790 --> 00:05:01.800 align:start position:0%
1050
+ perfectly fine mathematical operation,
1051
+
1052
+
1053
+ 00:05:01.800 --> 00:05:04.190 align:start position:0%
1054
+ perfectly fine mathematical operation,
1055
+ what<00:05:02.040><c> I</c><00:05:02.080><c> just</c><00:05:02.320><c> described</c><00:05:02.960><c> is</c><00:05:03.160><c> technically</c><00:05:03.840><c> not</c>
1056
+
1057
+ 00:05:04.190 --> 00:05:04.200 align:start position:0%
1058
+ what I just described is technically not
1059
+
1060
+
1061
+ 00:05:04.200 --> 00:05:06.350 align:start position:0%
1062
+ what I just described is technically not
1063
+ the<00:05:04.320><c> cross</c><00:05:04.600><c> product.</c><00:05:05.600><c> The</c><00:05:05.720><c> true</c><00:05:06.080><c> cross</c>
1064
+
1065
+ 00:05:06.350 --> 00:05:06.360 align:start position:0%
1066
+ the cross product. The true cross
1067
+
1068
+
1069
+ 00:05:06.360 --> 00:05:08.350 align:start position:0%
1070
+ the cross product. The true cross
1071
+ product<00:05:07.000><c> is</c><00:05:07.120><c> something</c><00:05:07.440><c> that</c><00:05:07.560><c> combines</c><00:05:08.120><c> two</c>
1072
+
1073
+ 00:05:08.350 --> 00:05:08.360 align:start position:0%
1074
+ product is something that combines two
1075
+
1076
+
1077
+ 00:05:08.360 --> 00:05:11.230 align:start position:0%
1078
+ product is something that combines two
1079
+ different<00:05:08.720><c> 3D</c><00:05:09.240><c> vectors</c><00:05:10.040><c> to</c><00:05:10.160><c> get</c><00:05:10.320><c> a</c><00:05:10.360><c> new</c><00:05:10.800><c> 3D</c>
1080
+
1081
+ 00:05:11.230 --> 00:05:11.240 align:start position:0%
1082
+ different 3D vectors to get a new 3D
1083
+
1084
+
1085
+ 00:05:11.240 --> 00:05:12.510 align:start position:0%
1086
+ different 3D vectors to get a new 3D
1087
+ vector.
1088
+
1089
+ 00:05:12.510 --> 00:05:12.520 align:start position:0%
1090
+ vector.
1091
+
1092
+
1093
+ 00:05:12.520 --> 00:05:13.990 align:start position:0%
1094
+ vector.
1095
+ Just<00:05:12.720><c> as</c><00:05:12.840><c> before,</c><00:05:13.440><c> we're</c><00:05:13.560><c> still</c><00:05:13.800><c> going</c><00:05:13.919><c> to</c>
1096
+
1097
+ 00:05:13.990 --> 00:05:14.000 align:start position:0%
1098
+ Just as before, we're still going to
1099
+
1100
+
1101
+ 00:05:14.000 --> 00:05:16.190 align:start position:0%
1102
+ Just as before, we're still going to
1103
+ consider<00:05:14.400><c> the</c><00:05:14.520><c> parallelogram</c><00:05:15.560><c> defined</c><00:05:16.080><c> by</c>
1104
+
1105
+ 00:05:16.190 --> 00:05:16.200 align:start position:0%
1106
+ consider the parallelogram defined by
1107
+
1108
+
1109
+ 00:05:16.200 --> 00:05:17.630 align:start position:0%
1110
+ consider the parallelogram defined by
1111
+ the<00:05:16.360><c> two</c><00:05:16.600><c> vectors</c><00:05:17.040><c> that</c><00:05:17.160><c> we're</c><00:05:17.280><c> crossing</c>
1112
+
1113
+ 00:05:17.630 --> 00:05:17.640 align:start position:0%
1114
+ the two vectors that we're crossing
1115
+
1116
+
1117
+ 00:05:17.640 --> 00:05:18.870 align:start position:0%
1118
+ the two vectors that we're crossing
1119
+ together.
1120
+
1121
+ 00:05:18.870 --> 00:05:18.880 align:start position:0%
1122
+ together.
1123
+
1124
+
1125
+ 00:05:18.880 --> 00:05:20.750 align:start position:0%
1126
+ together.
1127
+ And<00:05:18.960><c> the</c><00:05:19.080><c> area</c><00:05:19.440><c> of</c><00:05:19.520><c> this</c><00:05:19.680><c> parallelogram</c><00:05:20.600><c> is</c>
1128
+
1129
+ 00:05:20.750 --> 00:05:20.760 align:start position:0%
1130
+ And the area of this parallelogram is
1131
+
1132
+
1133
+ 00:05:20.760 --> 00:05:22.750 align:start position:0%
1134
+ And the area of this parallelogram is
1135
+ still<00:05:20.960><c> going</c><00:05:21.080><c> to</c><00:05:21.160><c> play</c><00:05:21.360><c> a</c><00:05:21.440><c> big</c><00:05:21.640><c> role.</c><00:05:22.600><c> To</c><00:05:22.640><c> be</c>
1136
+
1137
+ 00:05:22.750 --> 00:05:22.760 align:start position:0%
1138
+ still going to play a big role. To be
1139
+
1140
+
1141
+ 00:05:22.760 --> 00:05:25.270 align:start position:0%
1142
+ still going to play a big role. To be
1143
+ concrete,<00:05:23.200><c> let's</c><00:05:23.360><c> say</c><00:05:23.480><c> that</c><00:05:23.640><c> the</c><00:05:23.760><c> area</c><00:05:24.200><c> is</c><00:05:24.400><c> 2.5</c>
1144
+
1145
+ 00:05:25.270 --> 00:05:25.280 align:start position:0%
1146
+ concrete, let's say that the area is 2.5
1147
+
1148
+
1149
+ 00:05:25.280 --> 00:05:27.150 align:start position:0%
1150
+ concrete, let's say that the area is 2.5
1151
+ for<00:05:25.400><c> the</c><00:05:25.480><c> vectors</c><00:05:25.840><c> shown</c><00:05:26.040><c> here.</c>
1152
+
1153
+ 00:05:27.150 --> 00:05:27.160 align:start position:0%
1154
+ for the vectors shown here.
1155
+
1156
+
1157
+ 00:05:27.160 --> 00:05:28.950 align:start position:0%
1158
+ for the vectors shown here.
1159
+ But<00:05:27.240><c> as</c><00:05:27.360><c> I</c><00:05:27.440><c> said,</c><00:05:27.919><c> the</c><00:05:28.040><c> cross</c><00:05:28.280><c> product</c><00:05:28.600><c> is</c><00:05:28.720><c> not</c>
1160
+
1161
+ 00:05:28.950 --> 00:05:28.960 align:start position:0%
1162
+ But as I said, the cross product is not
1163
+
1164
+
1165
+ 00:05:28.960 --> 00:05:31.030 align:start position:0%
1166
+ But as I said, the cross product is not
1167
+ a<00:05:29.040><c> number,</c><00:05:29.560><c> it's</c><00:05:29.720><c> a</c><00:05:29.760><c> vector.</c>
1168
+
1169
+ 00:05:31.030 --> 00:05:31.040 align:start position:0%
1170
+ a number, it's a vector.
1171
+
1172
+
1173
+ 00:05:31.040 --> 00:05:32.990 align:start position:0%
1174
+ a number, it's a vector.
1175
+ This<00:05:31.200><c> new</c><00:05:31.400><c> vector's</c><00:05:31.800><c> length</c><00:05:32.360><c> will</c><00:05:32.520><c> be</c><00:05:32.840><c> the</c>
1176
+
1177
+ 00:05:32.990 --> 00:05:33.000 align:start position:0%
1178
+ This new vector's length will be the
1179
+
1180
+
1181
+ 00:05:33.000 --> 00:05:34.990 align:start position:0%
1182
+ This new vector's length will be the
1183
+ area<00:05:33.400><c> of</c><00:05:33.520><c> that</c><00:05:33.720><c> parallelogram,</c><00:05:34.800><c> which</c><00:05:34.919><c> in</c>
1184
+
1185
+ 00:05:34.990 --> 00:05:35.000 align:start position:0%
1186
+ area of that parallelogram, which in
1187
+
1188
+
1189
+ 00:05:35.000 --> 00:05:37.030 align:start position:0%
1190
+ area of that parallelogram, which in
1191
+ this<00:05:35.120><c> case</c><00:05:35.360><c> is</c><00:05:35.640><c> 2.5.</c>
1192
+
1193
+ 00:05:37.030 --> 00:05:37.040 align:start position:0%
1194
+ this case is 2.5.
1195
+
1196
+
1197
+ 00:05:37.040 --> 00:05:39.070 align:start position:0%
1198
+ this case is 2.5.
1199
+ And<00:05:37.120><c> the</c><00:05:37.200><c> direction</c><00:05:37.760><c> of</c><00:05:37.840><c> that</c><00:05:38.040><c> new</c><00:05:38.200><c> vector</c><00:05:38.960><c> is</c>
1200
+
1201
+ 00:05:39.070 --> 00:05:39.080 align:start position:0%
1202
+ And the direction of that new vector is
1203
+
1204
+
1205
+ 00:05:39.080 --> 00:05:40.950 align:start position:0%
1206
+ And the direction of that new vector is
1207
+ going<00:05:39.240><c> to</c><00:05:39.280><c> be</c><00:05:39.440><c> perpendicular</c><00:05:40.680><c> to</c><00:05:40.840><c> the</c>
1208
+
1209
+ 00:05:40.950 --> 00:05:40.960 align:start position:0%
1210
+ going to be perpendicular to the
1211
+
1212
+
1213
+ 00:05:40.960 --> 00:05:42.550 align:start position:0%
1214
+ going to be perpendicular to the
1215
+ parallelogram.
1216
+
1217
+ 00:05:42.550 --> 00:05:42.560 align:start position:0%
1218
+ parallelogram.
1219
+
1220
+
1221
+ 00:05:42.560 --> 00:05:44.630 align:start position:0%
1222
+ parallelogram.
1223
+ But<00:05:42.840><c> which</c><00:05:43.000><c> way,</c><00:05:43.440><c> right?</c><00:05:44.000><c> I</c><00:05:44.080><c> mean,</c><00:05:44.360><c> there</c><00:05:44.560><c> are</c>
1224
+
1225
+ 00:05:44.630 --> 00:05:44.640 align:start position:0%
1226
+ But which way, right? I mean, there are
1227
+
1228
+
1229
+ 00:05:44.640 --> 00:05:47.230 align:start position:0%
1230
+ But which way, right? I mean, there are
1231
+ two<00:05:44.919><c> possible</c><00:05:45.360><c> vectors</c><00:05:45.760><c> with</c><00:05:45.919><c> length</c><00:05:46.240><c> 2.5</c>
1232
+
1233
+ 00:05:47.230 --> 00:05:47.240 align:start position:0%
1234
+ two possible vectors with length 2.5
1235
+
1236
+
1237
+ 00:05:47.240 --> 00:05:50.550 align:start position:0%
1238
+ two possible vectors with length 2.5
1239
+ that<00:05:47.440><c> are</c><00:05:47.520><c> perpendicular</c><00:05:48.240><c> to</c><00:05:48.400><c> a</c><00:05:48.480><c> given</c><00:05:48.760><c> plane.</c>
1240
+
1241
+ 00:05:50.550 --> 00:05:50.560 align:start position:0%
1242
+ that are perpendicular to a given plane.
1243
+
1244
+
1245
+ 00:05:50.560 --> 00:05:51.990 align:start position:0%
1246
+ that are perpendicular to a given plane.
1247
+ This<00:05:50.680><c> is</c><00:05:50.800><c> where</c><00:05:50.920><c> the</c><00:05:51.040><c> right-hand</c><00:05:51.560><c> rule</c><00:05:51.720><c> comes</c>
1248
+
1249
+ 00:05:51.990 --> 00:05:52.000 align:start position:0%
1250
+ This is where the right-hand rule comes
1251
+
1252
+
1253
+ 00:05:52.000 --> 00:05:52.910 align:start position:0%
1254
+ This is where the right-hand rule comes
1255
+ in.
1256
+
1257
+ 00:05:52.910 --> 00:05:52.920 align:start position:0%
1258
+ in.
1259
+
1260
+
1261
+ 00:05:52.920 --> 00:05:54.910 align:start position:0%
1262
+ in.
1263
+ Point<00:05:53.160><c> the</c><00:05:53.280><c> forefinger</c><00:05:53.920><c> of</c><00:05:54.080><c> your</c><00:05:54.240><c> right</c><00:05:54.480><c> hand</c>
1264
+
1265
+ 00:05:54.910 --> 00:05:54.920 align:start position:0%
1266
+ Point the forefinger of your right hand
1267
+
1268
+
1269
+ 00:05:54.920 --> 00:05:56.430 align:start position:0%
1270
+ Point the forefinger of your right hand
1271
+ in<00:05:55.040><c> the</c><00:05:55.120><c> direction</c><00:05:55.480><c> of</c><00:05:55.600><c> V.</c>
1272
+
1273
+ 00:05:56.430 --> 00:05:56.440 align:start position:0%
1274
+ in the direction of V.
1275
+
1276
+
1277
+ 00:05:56.440 --> 00:05:57.990 align:start position:0%
1278
+ in the direction of V.
1279
+ Then<00:05:56.560><c> stick</c><00:05:56.840><c> out</c><00:05:56.960><c> your</c><00:05:57.080><c> middle</c><00:05:57.320><c> finger</c><00:05:57.800><c> in</c><00:05:57.920><c> the</c>
1280
+
1281
+ 00:05:57.990 --> 00:05:58.000 align:start position:0%
1282
+ Then stick out your middle finger in the
1283
+
1284
+
1285
+ 00:05:58.000 --> 00:05:59.670 align:start position:0%
1286
+ Then stick out your middle finger in the
1287
+ direction<00:05:58.400><c> of</c><00:05:58.480><c> W.</c>
1288
+
1289
+ 00:05:59.670 --> 00:05:59.680 align:start position:0%
1290
+ direction of W.
1291
+
1292
+
1293
+ 00:05:59.680 --> 00:06:01.670 align:start position:0%
1294
+ direction of W.
1295
+ Then,<00:06:00.200><c> when</c><00:06:00.320><c> you</c><00:06:00.440><c> point</c><00:06:00.680><c> up</c><00:06:00.800><c> your</c><00:06:00.920><c> thumb,</c>
1296
+
1297
+ 00:06:01.670 --> 00:06:01.680 align:start position:0%
1298
+ Then, when you point up your thumb,
1299
+
1300
+
1301
+ 00:06:01.680 --> 00:06:02.910 align:start position:0%
1302
+ Then, when you point up your thumb,
1303
+ that's<00:06:02.040><c> the</c><00:06:02.160><c> direction</c><00:06:02.520><c> of</c><00:06:02.600><c> the</c><00:06:02.680><c> cross</c>
1304
+
1305
+ 00:06:02.910 --> 00:06:02.920 align:start position:0%
1306
+ that's the direction of the cross
1307
+
1308
+
1309
+ 00:06:02.920 --> 00:06:08.230 align:start position:0%
1310
+ that's the direction of the cross
1311
+ product.
1312
+
1313
+ 00:06:08.230 --> 00:06:08.240 align:start position:0%
1314
+
1315
+
1316
+
1317
+ 00:06:08.240 --> 00:06:10.070 align:start position:0%
1318
+
1319
+ For<00:06:08.360><c> example,</c><00:06:09.040><c> let's</c><00:06:09.240><c> say</c><00:06:09.360><c> that</c><00:06:09.560><c> V</c><00:06:09.840><c> was</c><00:06:10.000><c> a</c>
1320
+
1321
+ 00:06:10.070 --> 00:06:10.080 align:start position:0%
1322
+ For example, let's say that V was a
1323
+
1324
+
1325
+ 00:06:10.080 --> 00:06:12.030 align:start position:0%
1326
+ For example, let's say that V was a
1327
+ vector<00:06:10.400><c> with</c><00:06:10.600><c> length</c><00:06:10.920><c> two</c><00:06:11.400><c> pointing</c><00:06:11.720><c> straight</c>
1328
+
1329
+ 00:06:12.030 --> 00:06:12.040 align:start position:0%
1330
+ vector with length two pointing straight
1331
+
1332
+
1333
+ 00:06:12.040 --> 00:06:14.430 align:start position:0%
1334
+ vector with length two pointing straight
1335
+ up<00:06:12.160><c> in</c><00:06:12.240><c> the</c><00:06:12.320><c> Z</c><00:06:12.480><c> direction,</c><00:06:13.360><c> and</c><00:06:13.480><c> W</c><00:06:13.960><c> is</c><00:06:14.120><c> a</c><00:06:14.160><c> vector</c>
1336
+
1337
+ 00:06:14.430 --> 00:06:14.440 align:start position:0%
1338
+ up in the Z direction, and W is a vector
1339
+
1340
+
1341
+ 00:06:14.440 --> 00:06:16.550 align:start position:0%
1342
+ up in the Z direction, and W is a vector
1343
+ with<00:06:14.640><c> length</c><00:06:14.960><c> two</c><00:06:15.480><c> pointing</c><00:06:15.800><c> in</c><00:06:15.880><c> the</c><00:06:16.000><c> pure</c><00:06:16.320><c> Y</c>
1344
+
1345
+ 00:06:16.550 --> 00:06:16.560 align:start position:0%
1346
+ with length two pointing in the pure Y
1347
+
1348
+
1349
+ 00:06:16.560 --> 00:06:17.830 align:start position:0%
1350
+ with length two pointing in the pure Y
1351
+ direction.
1352
+
1353
+ 00:06:17.830 --> 00:06:17.840 align:start position:0%
1354
+ direction.
1355
+
1356
+
1357
+ 00:06:17.840 --> 00:06:19.510 align:start position:0%
1358
+ direction.
1359
+ The<00:06:18.040><c> parallelogram</c><00:06:18.800><c> that</c><00:06:18.919><c> they</c><00:06:19.040><c> define</c><00:06:19.440><c> in</c>
1360
+
1361
+ 00:06:19.510 --> 00:06:19.520 align:start position:0%
1362
+ The parallelogram that they define in
1363
+
1364
+
1365
+ 00:06:19.520 --> 00:06:20.990 align:start position:0%
1366
+ The parallelogram that they define in
1367
+ this<00:06:19.680><c> simple</c><00:06:20.000><c> example</c><00:06:20.480><c> is</c><00:06:20.640><c> actually</c><00:06:20.960><c> a</c>
1368
+
1369
+ 00:06:20.990 --> 00:06:21.000 align:start position:0%
1370
+ this simple example is actually a
1371
+
1372
+
1373
+ 00:06:21.000 --> 00:06:22.870 align:start position:0%
1374
+ this simple example is actually a
1375
+ square,<00:06:21.640><c> since</c><00:06:22.000><c> they're</c><00:06:22.160><c> perpendicular</c><00:06:22.760><c> and</c>
1376
+
1377
+ 00:06:22.870 --> 00:06:22.880 align:start position:0%
1378
+ square, since they're perpendicular and
1379
+
1380
+
1381
+ 00:06:22.880 --> 00:06:24.790 align:start position:0%
1382
+ square, since they're perpendicular and
1383
+ have<00:06:23.000><c> the</c><00:06:23.120><c> same</c><00:06:23.320><c> length.</c><00:06:24.080><c> And</c><00:06:24.200><c> the</c><00:06:24.320><c> area</c><00:06:24.680><c> of</c>
1384
+
1385
+ 00:06:24.790 --> 00:06:24.800 align:start position:0%
1386
+ have the same length. And the area of
1387
+
1388
+
1389
+ 00:06:24.800 --> 00:06:26.990 align:start position:0%
1390
+ have the same length. And the area of
1391
+ that<00:06:25.000><c> square</c><00:06:25.360><c> is</c><00:06:25.560><c> four.</c><00:06:26.440><c> So,</c><00:06:26.560><c> their</c><00:06:26.720><c> cross</c>
1392
+
1393
+ 00:06:26.990 --> 00:06:27.000 align:start position:0%
1394
+ that square is four. So, their cross
1395
+
1396
+
1397
+ 00:06:27.000 --> 00:06:28.430 align:start position:0%
1398
+ that square is four. So, their cross
1399
+ product<00:06:27.320><c> should</c><00:06:27.440><c> be</c><00:06:27.560><c> a</c><00:06:27.600><c> vector</c><00:06:27.960><c> with</c><00:06:28.160><c> length</c>
1400
+
1401
+ 00:06:28.430 --> 00:06:28.440 align:start position:0%
1402
+ product should be a vector with length
1403
+
1404
+
1405
+ 00:06:28.440 --> 00:06:29.710 align:start position:0%
1406
+ product should be a vector with length
1407
+ four.
1408
+
1409
+ 00:06:29.710 --> 00:06:29.720 align:start position:0%
1410
+ four.
1411
+
1412
+
1413
+ 00:06:29.720 --> 00:06:31.550 align:start position:0%
1414
+ four.
1415
+ Using<00:06:30.080><c> the</c><00:06:30.160><c> right-hand</c><00:06:30.640><c> rule,</c><00:06:31.160><c> their</c><00:06:31.320><c> cross</c>
1416
+
1417
+ 00:06:31.550 --> 00:06:31.560 align:start position:0%
1418
+ Using the right-hand rule, their cross
1419
+
1420
+
1421
+ 00:06:31.560 --> 00:06:33.230 align:start position:0%
1422
+ Using the right-hand rule, their cross
1423
+ product<00:06:31.880><c> should</c><00:06:32.040><c> point</c><00:06:32.320><c> in</c><00:06:32.440><c> the</c><00:06:32.560><c> negative</c><00:06:33.040><c> X</c>
1424
+
1425
+ 00:06:33.230 --> 00:06:33.240 align:start position:0%
1426
+ product should point in the negative X
1427
+
1428
+
1429
+ 00:06:33.240 --> 00:06:36.030 align:start position:0%
1430
+ product should point in the negative X
1431
+ direction.
1432
+
1433
+ 00:06:36.030 --> 00:06:36.040 align:start position:0%
1434
+
1435
+
1436
+
1437
+ 00:06:36.040 --> 00:06:37.230 align:start position:0%
1438
+
1439
+ So,<00:06:36.200><c> the</c><00:06:36.280><c> cross</c><00:06:36.520><c> product</c><00:06:36.840><c> of</c><00:06:36.919><c> these</c><00:06:37.120><c> two</c>
1440
+
1441
+ 00:06:37.230 --> 00:06:37.240 align:start position:0%
1442
+ So, the cross product of these two
1443
+
1444
+
1445
+ 00:06:37.240 --> 00:06:45.750 align:start position:0%
1446
+ So, the cross product of these two
1447
+ vectors<00:06:38.000><c> is</c><00:06:38.160><c> -4</c><00:06:38.800><c> *</c><00:06:39.160><c> I-hat.</c>
1448
+
1449
+ 00:06:45.750 --> 00:06:45.760 align:start position:0%
1450
+
1451
+
1452
+
1453
+ 00:06:45.760 --> 00:06:47.830 align:start position:0%
1454
+
1455
+ For<00:06:45.880><c> more</c><00:06:46.040><c> general</c><00:06:46.360><c> computations,</c><00:06:47.280><c> there</c><00:06:47.560><c> is</c>
1456
+
1457
+ 00:06:47.830 --> 00:06:47.840 align:start position:0%
1458
+ For more general computations, there is
1459
+
1460
+
1461
+ 00:06:47.840 --> 00:06:49.510 align:start position:0%
1462
+ For more general computations, there is
1463
+ a<00:06:47.880><c> formula</c><00:06:48.320><c> that</c><00:06:48.520><c> you</c><00:06:48.680><c> could</c><00:06:48.880><c> memorize</c><00:06:49.280><c> if</c><00:06:49.400><c> you</c>
1464
+
1465
+ 00:06:49.510 --> 00:06:49.520 align:start position:0%
1466
+ a formula that you could memorize if you
1467
+
1468
+
1469
+ 00:06:49.520 --> 00:06:51.230 align:start position:0%
1470
+ a formula that you could memorize if you
1471
+ wanted,<00:06:50.000><c> but</c><00:06:50.120><c> it's</c><00:06:50.240><c> common</c><00:06:50.600><c> and</c><00:06:50.760><c> easier</c><00:06:51.120><c> to</c>
1472
+
1473
+ 00:06:51.230 --> 00:06:51.240 align:start position:0%
1474
+ wanted, but it's common and easier to
1475
+
1476
+
1477
+ 00:06:51.240 --> 00:06:52.990 align:start position:0%
1478
+ wanted, but it's common and easier to
1479
+ instead<00:06:51.600><c> remember</c><00:06:51.960><c> a</c><00:06:52.040><c> certain</c><00:06:52.440><c> process</c>
1480
+
1481
+ 00:06:52.990 --> 00:06:53.000 align:start position:0%
1482
+ instead remember a certain process
1483
+
1484
+
1485
+ 00:06:53.000 --> 00:06:55.590 align:start position:0%
1486
+ instead remember a certain process
1487
+ involving<00:06:53.480><c> the</c><00:06:53.600><c> 3D</c><00:06:54.000><c> determinant.</c>
1488
+
1489
+ 00:06:55.590 --> 00:06:55.600 align:start position:0%
1490
+ involving the 3D determinant.
1491
+
1492
+
1493
+ 00:06:55.600 --> 00:06:58.070 align:start position:0%
1494
+ involving the 3D determinant.
1495
+ Now,<00:06:56.160><c> this</c><00:06:56.320><c> process</c><00:06:56.720><c> looks</c><00:06:57.120><c> truly</c><00:06:57.560><c> strange</c><00:06:57.960><c> at</c>
1496
+
1497
+ 00:06:58.070 --> 00:06:58.080 align:start position:0%
1498
+ Now, this process looks truly strange at
1499
+
1500
+
1501
+ 00:06:58.080 --> 00:07:00.670 align:start position:0%
1502
+ Now, this process looks truly strange at
1503
+ first.<00:06:58.880><c> You</c><00:06:59.000><c> write</c><00:06:59.200><c> down</c><00:06:59.400><c> a</c><00:06:59.440><c> 3D</c><00:06:59.800><c> matrix</c><00:07:00.560><c> where</c>
1504
+
1505
+ 00:07:00.670 --> 00:07:00.680 align:start position:0%
1506
+ first. You write down a 3D matrix where
1507
+
1508
+
1509
+ 00:07:00.680 --> 00:07:03.070 align:start position:0%
1510
+ first. You write down a 3D matrix where
1511
+ the<00:07:00.800><c> second</c><00:07:01.400><c> and</c><00:07:01.560><c> third</c><00:07:01.840><c> columns</c><00:07:02.640><c> contain</c><00:07:02.960><c> the</c>
1512
+
1513
+ 00:07:03.070 --> 00:07:03.080 align:start position:0%
1514
+ the second and third columns contain the
1515
+
1516
+
1517
+ 00:07:03.080 --> 00:07:05.909 align:start position:0%
1518
+ the second and third columns contain the
1519
+ coordinates<00:07:03.760><c> of</c><00:07:03.919><c> V</c><00:07:04.320><c> and</c><00:07:04.480><c> W.</c><00:07:05.480><c> But</c><00:07:05.600><c> for</c><00:07:05.720><c> that</c>
1520
+
1521
+ 00:07:05.909 --> 00:07:05.919 align:start position:0%
1522
+ coordinates of V and W. But for that
1523
+
1524
+
1525
+ 00:07:05.919 --> 00:07:07.830 align:start position:0%
1526
+ coordinates of V and W. But for that
1527
+ first<00:07:06.200><c> column,</c><00:07:06.919><c> you</c><00:07:07.040><c> write</c><00:07:07.280><c> the</c><00:07:07.400><c> basis</c>
1528
+
1529
+ 00:07:07.830 --> 00:07:07.840 align:start position:0%
1530
+ first column, you write the basis
1531
+
1532
+
1533
+ 00:07:07.840 --> 00:07:11.430 align:start position:0%
1534
+ first column, you write the basis
1535
+ vectors<00:07:08.320><c> I-hat,</c><00:07:09.080><c> J-hat,</c><00:07:09.760><c> and</c><00:07:09.880><c> K-hat.</c>
1536
+
1537
+ 00:07:11.430 --> 00:07:11.440 align:start position:0%
1538
+ vectors I-hat, J-hat, and K-hat.
1539
+
1540
+
1541
+ 00:07:11.440 --> 00:07:13.550 align:start position:0%
1542
+ vectors I-hat, J-hat, and K-hat.
1543
+ Then,<00:07:12.040><c> you</c><00:07:12.160><c> compute</c><00:07:12.640><c> the</c><00:07:12.760><c> determinant</c><00:07:13.440><c> of</c>
1544
+
1545
+ 00:07:13.550 --> 00:07:13.560 align:start position:0%
1546
+ Then, you compute the determinant of
1547
+
1548
+
1549
+ 00:07:13.560 --> 00:07:15.150 align:start position:0%
1550
+ Then, you compute the determinant of
1551
+ this<00:07:13.800><c> matrix.</c>
1552
+
1553
+ 00:07:15.150 --> 00:07:15.160 align:start position:0%
1554
+ this matrix.
1555
+
1556
+
1557
+ 00:07:15.160 --> 00:07:17.190 align:start position:0%
1558
+ this matrix.
1559
+ The<00:07:15.320><c> silliness</c><00:07:15.800><c> is</c><00:07:15.880><c> probably</c><00:07:16.280><c> clear</c><00:07:16.520><c> here.</c>
1560
+
1561
+ 00:07:17.190 --> 00:07:17.200 align:start position:0%
1562
+ The silliness is probably clear here.
1563
+
1564
+
1565
+ 00:07:17.200 --> 00:07:18.710 align:start position:0%
1566
+ The silliness is probably clear here.
1567
+ What<00:07:17.400><c> on</c><00:07:17.520><c> earth</c><00:07:17.720><c> does</c><00:07:17.880><c> it</c><00:07:17.960><c> mean</c><00:07:18.200><c> to</c><00:07:18.320><c> put</c><00:07:18.520><c> in</c><00:07:18.640><c> a</c>
1568
+
1569
+ 00:07:18.710 --> 00:07:18.720 align:start position:0%
1570
+ What on earth does it mean to put in a
1571
+
1572
+
1573
+ 00:07:18.720 --> 00:07:22.510 align:start position:0%
1574
+ What on earth does it mean to put in a
1575
+ vector<00:07:19.280><c> as</c><00:07:19.480><c> the</c><00:07:19.600><c> entry</c><00:07:20.000><c> of</c><00:07:20.120><c> a</c><00:07:20.200><c> matrix?</c>
1576
+
1577
+ 00:07:22.510 --> 00:07:22.520 align:start position:0%
1578
+ vector as the entry of a matrix?
1579
+
1580
+
1581
+ 00:07:22.520 --> 00:07:23.990 align:start position:0%
1582
+ vector as the entry of a matrix?
1583
+ Students<00:07:22.919><c> are</c><00:07:23.040><c> often</c><00:07:23.280><c> told</c><00:07:23.520><c> that</c><00:07:23.720><c> this</c><00:07:23.880><c> is</c>
1584
+
1585
+ 00:07:23.990 --> 00:07:24.000 align:start position:0%
1586
+ Students are often told that this is
1587
+
1588
+
1589
+ 00:07:24.000 --> 00:07:25.950 align:start position:0%
1590
+ Students are often told that this is
1591
+ just<00:07:24.200><c> a</c><00:07:24.240><c> notational</c><00:07:24.760><c> trick.</c><00:07:25.400><c> When</c><00:07:25.560><c> you</c><00:07:25.640><c> carry</c>
1592
+
1593
+ 00:07:25.950 --> 00:07:25.960 align:start position:0%
1594
+ just a notational trick. When you carry
1595
+
1596
+
1597
+ 00:07:25.960 --> 00:07:28.550 align:start position:0%
1598
+ just a notational trick. When you carry
1599
+ out<00:07:26.080><c> the</c><00:07:26.200><c> computations</c><00:07:27.000><c> as</c><00:07:27.280><c> if</c><00:07:27.680><c> I-hat,</c><00:07:28.120><c> J-hat,</c>
1600
+
1601
+ 00:07:28.550 --> 00:07:28.560 align:start position:0%
1602
+ out the computations as if I-hat, J-hat,
1603
+
1604
+
1605
+ 00:07:28.560 --> 00:07:30.470 align:start position:0%
1606
+ out the computations as if I-hat, J-hat,
1607
+ and<00:07:28.680><c> K-hat</c><00:07:29.080><c> were</c><00:07:29.200><c> numbers,</c><00:07:30.000><c> then</c><00:07:30.160><c> you</c><00:07:30.240><c> get</c>
1608
+
1609
+ 00:07:30.470 --> 00:07:30.480 align:start position:0%
1610
+ and K-hat were numbers, then you get
1611
+
1612
+
1613
+ 00:07:30.480 --> 00:07:32.390 align:start position:0%
1614
+ and K-hat were numbers, then you get
1615
+ some<00:07:30.840><c> linear</c><00:07:31.160><c> combination</c><00:07:31.720><c> of</c><00:07:31.800><c> those</c><00:07:32.040><c> basis</c>
1616
+
1617
+ 00:07:32.390 --> 00:07:32.400 align:start position:0%
1618
+ some linear combination of those basis
1619
+
1620
+
1621
+ 00:07:32.400 --> 00:07:36.070 align:start position:0%
1622
+ some linear combination of those basis
1623
+ vectors.
1624
+
1625
+ 00:07:36.070 --> 00:07:36.080 align:start position:0%
1626
+
1627
+
1628
+
1629
+ 00:07:36.080 --> 00:07:38.070 align:start position:0%
1630
+
1631
+ And<00:07:36.280><c> the</c><00:07:36.360><c> vector</c><00:07:36.720><c> defined</c><00:07:37.280><c> by</c><00:07:37.440><c> that</c><00:07:37.760><c> linear</c>
1632
+
1633
+ 00:07:38.070 --> 00:07:38.080 align:start position:0%
1634
+ And the vector defined by that linear
1635
+
1636
+
1637
+ 00:07:38.080 --> 00:07:40.030 align:start position:0%
1638
+ And the vector defined by that linear
1639
+ combination,<00:07:38.880><c> students</c><00:07:39.280><c> are</c><00:07:39.360><c> told</c><00:07:39.640><c> to</c><00:07:39.760><c> just</c>
1640
+
1641
+ 00:07:40.030 --> 00:07:40.040 align:start position:0%
1642
+ combination, students are told to just
1643
+
1644
+
1645
+ 00:07:40.040 --> 00:07:42.030 align:start position:0%
1646
+ combination, students are told to just
1647
+ believe,<00:07:40.960><c> is</c><00:07:41.160><c> the</c><00:07:41.280><c> unique</c><00:07:41.640><c> vector</c>
1648
+
1649
+ 00:07:42.030 --> 00:07:42.040 align:start position:0%
1650
+ believe, is the unique vector
1651
+
1652
+
1653
+ 00:07:42.040 --> 00:07:44.670 align:start position:0%
1654
+ believe, is the unique vector
1655
+ perpendicular<00:07:42.720><c> to</c><00:07:42.840><c> V</c><00:07:43.000><c> and</c><00:07:43.120><c> W</c><00:07:43.840><c> whose</c><00:07:44.040><c> magnitude</c>
1656
+
1657
+ 00:07:44.670 --> 00:07:44.680 align:start position:0%
1658
+ perpendicular to V and W whose magnitude
1659
+
1660
+
1661
+ 00:07:44.680 --> 00:07:45.950 align:start position:0%
1662
+ perpendicular to V and W whose magnitude
1663
+ is<00:07:44.800><c> the</c><00:07:44.960><c> area</c><00:07:45.280><c> of</c><00:07:45.400><c> the</c><00:07:45.520><c> appropriate</c>
1664
+
1665
+ 00:07:45.950 --> 00:07:45.960 align:start position:0%
1666
+ is the area of the appropriate
1667
+
1668
+
1669
+ 00:07:45.960 --> 00:07:48.230 align:start position:0%
1670
+ is the area of the appropriate
1671
+ parallelogram<00:07:47.000><c> and</c><00:07:47.120><c> whose</c><00:07:47.320><c> direction</c><00:07:47.840><c> obeys</c>
1672
+
1673
+ 00:07:48.230 --> 00:07:48.240 align:start position:0%
1674
+ parallelogram and whose direction obeys
1675
+
1676
+
1677
+ 00:07:48.240 --> 00:07:51.230 align:start position:0%
1678
+ parallelogram and whose direction obeys
1679
+ the<00:07:48.360><c> right-hand</c><00:07:48.800><c> rule.</c>
1680
+
1681
+ 00:07:51.230 --> 00:07:51.240 align:start position:0%
1682
+
1683
+
1684
+
1685
+ 00:07:51.240 --> 00:07:53.510 align:start position:0%
1686
+
1687
+ And<00:07:51.800><c> sure,</c><00:07:52.400><c> in</c><00:07:52.520><c> some</c><00:07:52.720><c> sense,</c><00:07:52.960><c> this</c><00:07:53.160><c> is</c><00:07:53.280><c> just</c><00:07:53.480><c> a</c>
1688
+
1689
+ 00:07:53.510 --> 00:07:53.520 align:start position:0%
1690
+ And sure, in some sense, this is just a
1691
+
1692
+
1693
+ 00:07:53.520 --> 00:07:55.790 align:start position:0%
1694
+ And sure, in some sense, this is just a
1695
+ notational<00:07:54.080><c> trick,</c><00:07:54.760><c> but</c><00:07:54.880><c> there</c><00:07:55.040><c> is</c><00:07:55.200><c> a</c><00:07:55.280><c> reason</c>
1696
+
1697
+ 00:07:55.790 --> 00:07:55.800 align:start position:0%
1698
+ notational trick, but there is a reason
1699
+
1700
+
1701
+ 00:07:55.800 --> 00:07:57.830 align:start position:0%
1702
+ notational trick, but there is a reason
1703
+ for<00:07:55.920><c> doing</c><00:07:56.240><c> it.</c>
1704
+
1705
+ 00:07:57.830 --> 00:07:57.840 align:start position:0%
1706
+ for doing it.
1707
+
1708
+
1709
+ 00:07:57.840 --> 00:07:59.350 align:start position:0%
1710
+ for doing it.
1711
+ It's<00:07:58.120><c> not</c><00:07:58.280><c> just</c><00:07:58.480><c> a</c><00:07:58.520><c> coincidence</c><00:07:59.160><c> that</c><00:07:59.280><c> the</c>
1712
+
1713
+ 00:07:59.350 --> 00:07:59.360 align:start position:0%
1714
+ It's not just a coincidence that the
1715
+
1716
+
1717
+ 00:07:59.360 --> 00:08:01.870 align:start position:0%
1718
+ It's not just a coincidence that the
1719
+ determinant<00:08:00.000><c> is</c><00:08:00.160><c> once</c><00:08:00.400><c> again</c><00:08:00.600><c> important.</c><00:08:01.680><c> And</c>
1720
+
1721
+ 00:08:01.870 --> 00:08:01.880 align:start position:0%
1722
+ determinant is once again important. And
1723
+
1724
+
1725
+ 00:08:01.880 --> 00:08:03.910 align:start position:0%
1726
+ determinant is once again important. And
1727
+ putting<00:08:02.120><c> the</c><00:08:02.200><c> basis</c><00:08:02.600><c> vectors</c><00:08:03.040><c> in</c><00:08:03.240><c> those</c><00:08:03.480><c> slots</c>
1728
+
1729
+ 00:08:03.910 --> 00:08:03.920 align:start position:0%
1730
+ putting the basis vectors in those slots
1731
+
1732
+
1733
+ 00:08:03.920 --> 00:08:06.150 align:start position:0%
1734
+ putting the basis vectors in those slots
1735
+ is<00:08:04.080><c> not</c><00:08:04.240><c> just</c><00:08:04.440><c> a</c><00:08:04.480><c> random</c><00:08:04.840><c> thing</c><00:08:05.000><c> to</c><00:08:05.120><c> do.</c>
1736
+
1737
+ 00:08:06.150 --> 00:08:06.160 align:start position:0%
1738
+ is not just a random thing to do.
1739
+
1740
+
1741
+ 00:08:06.160 --> 00:08:07.630 align:start position:0%
1742
+ is not just a random thing to do.
1743
+ To<00:08:06.360><c> understand</c><00:08:06.880><c> where</c><00:08:07.000><c> all</c><00:08:07.160><c> of</c><00:08:07.240><c> this</c><00:08:07.360><c> comes</c>
1744
+
1745
+ 00:08:07.630 --> 00:08:07.640 align:start position:0%
1746
+ To understand where all of this comes
1747
+
1748
+
1749
+ 00:08:07.640 --> 00:08:09.630 align:start position:0%
1750
+ To understand where all of this comes
1751
+ from,<00:08:08.160><c> it</c><00:08:08.320><c> helps</c><00:08:08.600><c> to</c><00:08:08.720><c> use</c><00:08:08.960><c> the</c><00:08:09.120><c> idea</c><00:08:09.480><c> of</c>
1752
+
1753
+ 00:08:09.630 --> 00:08:09.640 align:start position:0%
1754
+ from, it helps to use the idea of
1755
+
1756
+
1757
+ 00:08:09.640 --> 00:08:11.510 align:start position:0%
1758
+ from, it helps to use the idea of
1759
+ duality<00:08:10.440><c> that</c><00:08:10.600><c> I</c><00:08:10.680><c> introduced</c><00:08:11.120><c> in</c><00:08:11.200><c> the</c><00:08:11.280><c> last</c>
1760
+
1761
+ 00:08:11.510 --> 00:08:11.520 align:start position:0%
1762
+ duality that I introduced in the last
1763
+
1764
+
1765
+ 00:08:11.520 --> 00:08:12.670 align:start position:0%
1766
+ duality that I introduced in the last
1767
+ video.
1768
+
1769
+ 00:08:12.670 --> 00:08:12.680 align:start position:0%
1770
+ video.
1771
+
1772
+
1773
+ 00:08:12.680 --> 00:08:14.030 align:start position:0%
1774
+ video.
1775
+ This<00:08:12.880><c> concept</c><00:08:13.280><c> is</c><00:08:13.400><c> a</c><00:08:13.440><c> little</c><00:08:13.640><c> bit</c><00:08:13.760><c> heavy,</c>
1776
+
1777
+ 00:08:14.030 --> 00:08:14.040 align:start position:0%
1778
+ This concept is a little bit heavy,
1779
+
1780
+
1781
+ 00:08:14.040 --> 00:08:16.070 align:start position:0%
1782
+ This concept is a little bit heavy,
1783
+ though.<00:08:14.440><c> So,</c><00:08:14.919><c> I'm</c><00:08:15.040><c> putting</c><00:08:15.280><c> it</c><00:08:15.400><c> in</c><00:08:15.520><c> a</c><00:08:15.600><c> separate</c>
1784
+
1785
+ 00:08:16.070 --> 00:08:16.080 align:start position:0%
1786
+ though. So, I'm putting it in a separate
1787
+
1788
+
1789
+ 00:08:16.080 --> 00:08:17.950 align:start position:0%
1790
+ though. So, I'm putting it in a separate
1791
+ follow-on<00:08:16.600><c> video</c><00:08:17.200><c> for</c><00:08:17.360><c> any</c><00:08:17.520><c> of</c><00:08:17.600><c> you</c><00:08:17.720><c> who</c><00:08:17.840><c> are</c>
1792
+
1793
+ 00:08:17.950 --> 00:08:17.960 align:start position:0%
1794
+ follow-on video for any of you who are
1795
+
1796
+
1797
+ 00:08:17.960 --> 00:08:19.550 align:start position:0%
1798
+ follow-on video for any of you who are
1799
+ curious<00:08:18.360><c> to</c><00:08:18.480><c> learn</c><00:08:18.680><c> more.</c>
1800
+
1801
+ 00:08:19.550 --> 00:08:19.560 align:start position:0%
1802
+ curious to learn more.
1803
+
1804
+
1805
+ 00:08:19.560 --> 00:08:21.990 align:start position:0%
1806
+ curious to learn more.
1807
+ Arguably,<00:08:20.440><c> it</c><00:08:20.600><c> falls</c><00:08:20.960><c> outside</c><00:08:21.360><c> the</c><00:08:21.520><c> essence</c>
1808
+
1809
+ 00:08:21.990 --> 00:08:22.000 align:start position:0%
1810
+ Arguably, it falls outside the essence
1811
+
1812
+
1813
+ 00:08:22.000 --> 00:08:23.830 align:start position:0%
1814
+ Arguably, it falls outside the essence
1815
+ of<00:08:22.120><c> linear</c><00:08:22.440><c> algebra.</c><00:08:23.240><c> The</c><00:08:23.320><c> important</c><00:08:23.680><c> part</c>
1816
+
1817
+ 00:08:23.830 --> 00:08:23.840 align:start position:0%
1818
+ of linear algebra. The important part
1819
+
1820
+
1821
+ 00:08:23.840 --> 00:08:25.630 align:start position:0%
1822
+ of linear algebra. The important part
1823
+ here<00:08:24.120><c> is</c><00:08:24.280><c> to</c><00:08:24.400><c> know</c><00:08:24.640><c> what</c><00:08:24.840><c> that</c><00:08:25.000><c> cross</c><00:08:25.280><c> product</c>
1824
+
1825
+ 00:08:25.630 --> 00:08:25.640 align:start position:0%
1826
+ here is to know what that cross product
1827
+
1828
+
1829
+ 00:08:25.640 --> 00:08:27.909 align:start position:0%
1830
+ here is to know what that cross product
1831
+ vector<00:08:25.919><c> geometrically</c><00:08:26.640><c> represents.</c>
1832
+
1833
+ 00:08:27.909 --> 00:08:27.919 align:start position:0%
1834
+ vector geometrically represents.
1835
+
1836
+
1837
+ 00:08:27.919 --> 00:08:29.350 align:start position:0%
1838
+ vector geometrically represents.
1839
+ So,<00:08:28.080><c> if</c><00:08:28.160><c> you</c><00:08:28.280><c> want</c><00:08:28.400><c> to</c><00:08:28.480><c> skip</c><00:08:28.680><c> that</c><00:08:28.840><c> next</c><00:08:29.040><c> video,</c>
1840
+
1841
+ 00:08:29.350 --> 00:08:29.360 align:start position:0%
1842
+ So, if you want to skip that next video,
1843
+
1844
+
1845
+ 00:08:29.360 --> 00:08:31.350 align:start position:0%
1846
+ So, if you want to skip that next video,
1847
+ feel<00:08:29.720><c> free.</c><00:08:30.440><c> But</c><00:08:30.600><c> for</c><00:08:30.720><c> those</c><00:08:30.919><c> of</c><00:08:31.000><c> you</c><00:08:31.120><c> who</c><00:08:31.240><c> are</c>
1848
+
1849
+ 00:08:31.350 --> 00:08:31.360 align:start position:0%
1850
+ feel free. But for those of you who are
1851
+
1852
+
1853
+ 00:08:31.360 --> 00:08:32.990 align:start position:0%
1854
+ feel free. But for those of you who are
1855
+ willing<00:08:31.640><c> to</c><00:08:31.720><c> go</c><00:08:31.840><c> a</c><00:08:31.919><c> bit</c><00:08:32.080><c> deeper</c><00:08:32.719><c> and</c><00:08:32.840><c> who</c><00:08:32.919><c> are</c>
1856
+
1857
+ 00:08:32.990 --> 00:08:33.000 align:start position:0%
1858
+ willing to go a bit deeper and who are
1859
+
1860
+
1861
+ 00:08:33.000 --> 00:08:34.430 align:start position:0%
1862
+ willing to go a bit deeper and who are
1863
+ curious<00:08:33.440><c> about</c><00:08:33.680><c> the</c><00:08:33.760><c> connection</c><00:08:34.159><c> between</c>
1864
+
1865
+ 00:08:34.430 --> 00:08:34.440 align:start position:0%
1866
+ curious about the connection between
1867
+
1868
+
1869
+ 00:08:34.440 --> 00:08:36.110 align:start position:0%
1870
+ curious about the connection between
1871
+ this<00:08:34.640><c> computation</c><00:08:35.320><c> and</c><00:08:35.520><c> the</c><00:08:35.640><c> underlying</c>
1872
+
1873
+ 00:08:36.110 --> 00:08:36.120 align:start position:0%
1874
+ this computation and the underlying
1875
+
1876
+
1877
+ 00:08:36.120 --> 00:08:38.149 align:start position:0%
1878
+ this computation and the underlying
1879
+ geometry,<00:08:37.039><c> the</c><00:08:37.159><c> ideas</c><00:08:37.520><c> that</c><00:08:37.640><c> I'll</c><00:08:37.760><c> talk</c><00:08:37.960><c> about</c>
1880
+
1881
+ 00:08:38.149 --> 00:08:38.159 align:start position:0%
1882
+ geometry, the ideas that I'll talk about
1883
+
1884
+
1885
+ 00:08:38.159 --> 00:08:40.070 align:start position:0%
1886
+ geometry, the ideas that I'll talk about
1887
+ in<00:08:38.240><c> the</c><00:08:38.320><c> next</c><00:08:38.560><c> video</c><00:08:38.880><c> are</c><00:08:39.000><c> just</c><00:08:39.360><c> a</c><00:08:39.479><c> really</c>
1888
+
1889
+ 00:08:40.070 --> 00:08:40.080 align:start position:0%
1890
+ in the next video are just a really
1891
+
1892
+
1893
+ 00:08:40.080 --> 00:08:43.240 align:start position:0%
1894
+ in the next video are just a really
1895
+ elegant<00:08:40.520><c> piece</c><00:08:40.719><c> of</c><00:08:40.840><c> math.</c>
1896
+
3B1B_LinearAlgebra/10 - Cross products | Chapter 10, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a990f66bf057b69f095a55f240bb9d6550659fa1982597153a8e757fe7976da5
3
+ size 23411098
3B1B_LinearAlgebra/11 - Cross products in the light of linear transformations | Chapter 11, Essence of linear algebra.en.vtt ADDED
@@ -0,0 +1,532 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:00.000 --> 00:00:16.540
6
+ [Submit subtitle corrections at criblate.com]
7
+
8
+ 00:00:16.540 --> 00:00:20.183
9
+ Hey folks, where we left off I was talking about how to compute
10
+
11
+ 00:00:20.183 --> 00:00:24.000
12
+ a three-dimensional cross product between two vectors, v cross w.
13
+
14
+ 00:00:25.280 --> 00:00:29.468
15
+ It's this funny thing where you write a matrix whose second column has the
16
+
17
+ 00:00:29.468 --> 00:00:33.034
18
+ coordinates of v, whose third column has the coordinates of w,
19
+
20
+ 00:00:33.034 --> 00:00:37.393
21
+ but the entries of that first column, weirdly, are the symbols i-hat, j-hat,
22
+
23
+ 00:00:37.393 --> 00:00:41.638
24
+ and k-hat, where you just pretend like those guys are numbers for the sake
25
+
26
+ 00:00:41.638 --> 00:00:42.600
27
+ of computations.
28
+
29
+ 00:00:43.780 --> 00:00:47.460
30
+ Then with that funky matrix in hand, you compute its determinant.
31
+
32
+ 00:00:48.080 --> 00:00:52.073
33
+ If you just chug along with those computations, ignoring the weirdness,
34
+
35
+ 00:00:52.073 --> 00:00:55.840
36
+ you get some constant times i-hat, plus some constant times j-hat,
37
+
38
+ 00:00:55.840 --> 00:00:57.640
39
+ plus some constant times k-hat.
40
+
41
+ 00:00:59.800 --> 00:01:04.300
42
+ How specifically you think about computing that determinant is kind of beside the point.
43
+
44
+ 00:01:04.820 --> 00:01:07.983
45
+ All that really matters here is that you'll end up with three different
46
+
47
+ 00:01:07.983 --> 00:01:11.280
48
+ numbers that are interpreted as the coordinates of some resulting vector.
49
+
50
+ 00:01:13.760 --> 00:01:16.848
51
+ From here, students are typically told to just believe that
52
+
53
+ 00:01:16.848 --> 00:01:20.040
54
+ the resulting vector has the following geometric properties.
55
+
56
+ 00:01:20.040 --> 00:01:24.760
57
+ Its length equals the area of the parallelogram defined by v and w.
58
+
59
+ 00:01:25.640 --> 00:01:28.729
60
+ It points in a direction perpendicular to both v and w,
61
+
62
+ 00:01:28.729 --> 00:01:31.312
63
+ and this direction obeys the right-hand rule,
64
+
65
+ 00:01:31.312 --> 00:01:36.199
66
+ in the sense that if you point your forefinger along v and your middle finger along w,
67
+
68
+ 00:01:36.199 --> 00:01:40.860
69
+ then when you stick up your thumb, it'll point in the direction of the new vector.
70
+
71
+ 00:01:43.660 --> 00:01:47.422
72
+ There are some brute force computations that you could do to confirm these facts,
73
+
74
+ 00:01:47.422 --> 00:01:50.440
75
+ but I want to share with you a really elegant line of reasoning.
76
+
77
+ 00:01:51.120 --> 00:01:54.196
78
+ It leverages a bit of background, though, so for this video
79
+
80
+ 00:01:54.196 --> 00:01:58.575
81
+ I'm assuming that everybody has watched chapter 5 on the determinant and chapter 7,
82
+
83
+ 00:01:58.575 --> 00:02:00.660
84
+ where I introduced the idea of duality.
85
+
86
+ 00:02:04.580 --> 00:02:08.518
87
+ As a quick reminder, the idea of duality is that any time you have a
88
+
89
+ 00:02:08.518 --> 00:02:11.877
90
+ linear transformation from some space to the number line,
91
+
92
+ 00:02:11.877 --> 00:02:14.888
93
+ it's associated with a unique vector in that space,
94
+
95
+ 00:02:14.888 --> 00:02:18.942
96
+ in the sense that performing the linear transformation is the same as
97
+
98
+ 00:02:18.942 --> 00:02:21.200
99
+ taking a dot product with that vector.
100
+
101
+ 00:02:22.080 --> 00:02:27.059
102
+ Numerically, this is because one of those transformations is described by a matrix with
103
+
104
+ 00:02:27.059 --> 00:02:31.980
105
+ just one row, where each column tells you the number that each basis vector lands on.
106
+
107
+ 00:02:35.240 --> 00:02:40.062
108
+ And multiplying this matrix by some vector v is computationally identical to taking
109
+
110
+ 00:02:40.062 --> 00:02:45.000
111
+ the dot product between v and the vector you get by turning that matrix on its side.
112
+
113
+ 00:02:46.580 --> 00:02:50.779
114
+ The takeaway is that whenever you're out in the mathematical wild and you find a
115
+
116
+ 00:02:50.779 --> 00:02:55.345
117
+ linear transformation to the number line, you will be able to match it to some vector,
118
+
119
+ 00:02:55.345 --> 00:02:58.284
120
+ which is called the dual vector of that transformation,
121
+
122
+ 00:02:58.284 --> 00:03:02.535
123
+ so that performing the linear transformation is the same as taking a dot product
124
+
125
+ 00:03:02.535 --> 00:03:03.480
126
+ with that vector.
127
+
128
+ 00:03:06.360 --> 00:03:10.040
129
+ The cross product gives us a really slick example of this process in action.
130
+
131
+ 00:03:10.360 --> 00:03:13.040
132
+ It takes some effort, but it's definitely worth it.
133
+
134
+ 00:03:13.640 --> 00:03:18.099
135
+ What I'm going to do is define a certain linear transformation from three dimensions
136
+
137
+ 00:03:18.099 --> 00:03:22.240
138
+ to the number line, and it'll be defined in terms of the two vectors v and w.
139
+
140
+ 00:03:23.140 --> 00:03:28.328
141
+ Then when we associate that transformation with its dual vector in 3D space,
142
+
143
+ 00:03:28.328 --> 00:03:32.560
144
+ that dual vector is going to be the cross product of v and w.
145
+
146
+ 00:03:33.220 --> 00:03:37.774
147
+ The reason for doing this will be that understanding that transformation is going to
148
+
149
+ 00:03:37.774 --> 00:03:42.600
150
+ make clear the connection between the computation and the geometry of the cross product.
151
+
152
+ 00:03:44.720 --> 00:03:47.485
153
+ So to back up a bit, remember in two dimensions what it
154
+
155
+ 00:03:47.485 --> 00:03:50.200
156
+ meant to compute the 2D version of the cross product?
157
+
158
+ 00:03:50.820 --> 00:03:55.074
159
+ When you have two vectors v and w, you put the coordinates of v as the first
160
+
161
+ 00:03:55.074 --> 00:03:59.440
162
+ column of a matrix and the coordinates of w as the second column of a matrix.
163
+
164
+ 00:03:59.720 --> 00:04:01.300
165
+ Then you just compute the determinant.
166
+
167
+ 00:04:01.680 --> 00:04:05.593
168
+ There's no nonsense with basis vectors stuck in a matrix or anything like that,
169
+
170
+ 00:04:05.593 --> 00:04:08.020
171
+ just an ordinary determinant returning a number.
172
+
173
+ 00:04:09.380 --> 00:04:13.765
174
+ Geometrically, this gives us the area of a parallelogram spanned out by those two
175
+
176
+ 00:04:13.765 --> 00:04:18.313
177
+ vectors, with the possibility of being negative depending on the orientation of the
178
+
179
+ 00:04:18.313 --> 00:04:18.800
180
+ vectors.
181
+
182
+ 00:04:19.780 --> 00:04:25.356
183
+ Now, if you didn't already know the 3D cross product and you're trying to extrapolate,
184
+
185
+ 00:04:25.356 --> 00:04:29.829
186
+ you might imagine that it involves taking three separate 3D vectors,
187
+
188
+ 00:04:29.829 --> 00:04:34.433
189
+ u, v, and w, and making their coordinates the columns of a 3x3 matrix,
190
+
191
+ 00:04:34.433 --> 00:04:37.480
192
+ then computing the determinant of that matrix.
193
+
194
+ 00:04:38.840 --> 00:04:43.306
195
+ And as you know from chapter 5, geometrically this would give you the volume
196
+
197
+ 00:04:43.306 --> 00:04:46.597
198
+ of a parallelepiped spanned out by those three vectors,
199
+
200
+ 00:04:46.597 --> 00:04:50.946
201
+ with a plus or minus sign depending on the right hand rule orientation of
202
+
203
+ 00:04:50.946 --> 00:04:52.180
204
+ those three vectors.
205
+
206
+ 00:04:53.060 --> 00:04:55.920
207
+ Of course, you all know that this is not the 3D cross product.
208
+
209
+ 00:04:56.920 --> 00:05:01.100
210
+ The actual 3D cross product takes in two vectors and spits out a vector.
211
+
212
+ 00:05:02.640 --> 00:05:05.060
213
+ It doesn't take in three vectors and spit out a number.
214
+
215
+ 00:05:05.660 --> 00:05:09.640
216
+ But this idea actually gets us really close to what the real cross product is.
217
+
218
+ 00:05:10.900 --> 00:05:14.092
219
+ Consider that first vector u to be a variable,
220
+
221
+ 00:05:14.092 --> 00:05:18.740
222
+ say with variable entries x, y, and z, while v and w remain fixed.
223
+
224
+ 00:05:22.760 --> 00:05:26.600
225
+ What we have then is a function from three dimensions to the number line.
226
+
227
+ 00:05:27.180 --> 00:05:31.569
228
+ You input some vector (x, y, z) and you get out a number by taking the
229
+
230
+ 00:05:31.569 --> 00:05:34.893
231
+ determinant of a matrix whose first column is (x, y,
232
+
233
+ 00:05:34.893 --> 00:05:40.160
234
+ z) and whose other two columns are the coordinates of the constant vectors v and w.
235
+
236
+ 00:05:40.920 --> 00:05:45.985
237
+ Geometrically, the meaning of this function is that for any input vector x,
238
+
239
+ 00:05:45.985 --> 00:05:50.780
240
+ y, z, you consider the parallelepiped defined by this vector, v and w.
241
+
242
+ 00:05:51.420 --> 00:05:55.380
243
+ Then you return its volume with a plus or minus sign depending on orientation.
244
+
245
+ 00:05:57.500 --> 00:05:59.740
246
+ Now, this might feel like kind of a random thing to do.
247
+
248
+ 00:06:00.160 --> 00:06:01.700
249
+ I mean, where does this function come from?
250
+
251
+ 00:06:01.760 --> 00:06:03.040
252
+ Why are we defining it this way?
253
+
254
+ 00:06:03.860 --> 00:06:06.680
255
+ And I'll admit, at this stage it might kind of feel like it's coming out of the blue.
256
+
257
+ 00:06:06.980 --> 00:06:10.368
258
+ But if you're willing to go along with it and play around with the properties
259
+
260
+ 00:06:10.368 --> 00:06:13.360
261
+ that this guy has, it's the key to understanding the cross product.
262
+
263
+ 00:06:15.340 --> 00:06:19.160
264
+ One really important fact about this function is that it's linear.
265
+
266
+ 00:06:20.020 --> 00:06:22.563
267
+ I'll actually leave it to you to work through the details
268
+
269
+ 00:06:22.563 --> 00:06:25.240
270
+ of why this is true based on properties of the determinant.
271
+
272
+ 00:06:26.380 --> 00:06:30.780
273
+ But once you know that it's linear, we can start bringing in the idea of duality.
274
+
275
+ 00:06:35.060 --> 00:06:37.950
276
+ Once you know that it's linear, you know that there's some
277
+
278
+ 00:06:37.950 --> 00:06:40.740
279
+ way to describe this function as matrix multiplication.
280
+
281
+ 00:06:41.320 --> 00:06:46.036
282
+ Specifically, since it's a function that goes from three dimensions to one dimension,
283
+
284
+ 00:06:46.036 --> 00:06:49.920
285
+ there will be a one by three matrix that encodes this transformation.
286
+
287
+ 00:06:53.360 --> 00:06:57.612
288
+ And the whole idea of duality is that the special thing about transformations from
289
+
290
+ 00:06:57.612 --> 00:07:02.020
291
+ several dimensions to one dimension is that you can turn that matrix on its side and
292
+
293
+ 00:07:02.020 --> 00:07:06.480
294
+ instead interpret the entire transformation as the dot product with a certain vector.
295
+
296
+ 00:07:07.900 --> 00:07:13.070
297
+ What we're looking for is the special 3D vector that I'll call p such that taking
298
+
299
+ 00:07:13.070 --> 00:07:16.516
300
+ the dot product between p and any other vector (x, y,
301
+
302
+ 00:07:16.516 --> 00:07:19.452
303
+ z) gives the same result as plugging in x, y,
304
+
305
+ 00:07:19.452 --> 00:07:24.431
306
+ z as the first column of a three by three matrix whose other two columns have
307
+
308
+ 00:07:24.431 --> 00:07:28.260
309
+ the coordinates of v and w, then computing the determinant.
310
+
311
+ 00:07:29.160 --> 00:07:31.382
312
+ I'll get to the geometry of this in just a moment,
313
+
314
+ 00:07:31.382 --> 00:07:34.760
315
+ but right now let's dig in and think about what this means computationally.
316
+
317
+ 00:07:35.780 --> 00:07:41.150
318
+ Taking the dot product between p and (x, y, z) will give us something times x plus
319
+
320
+ 00:07:41.150 --> 00:07:46.847
321
+ something times y plus something times z, where those "somethings" are the coordinates
322
+
323
+ 00:07:46.847 --> 00:07:47.240
324
+ of p.
325
+
326
+ 00:07:47.980 --> 00:07:51.650
327
+ But on the right side here, when you compute the determinant,
328
+
329
+ 00:07:51.650 --> 00:07:56.583
330
+ you can organize it to look like some constant times x plus some constant times y
331
+
332
+ 00:07:56.583 --> 00:08:01.516
333
+ plus some constant times z, where those constants involve certain combinations of
334
+
335
+ 00:08:01.516 --> 00:08:03.140
336
+ the components of v and w.
337
+
338
+ 00:08:03.880 --> 00:08:08.357
339
+ So those constants, those particular combinations of the coordinates of v
340
+
341
+ 00:08:08.357 --> 00:08:13.140
342
+ and w are going to be the coordinates of the vector p that we're looking for.
343
+
344
+ 00:08:18.260 --> 00:08:21.420
345
+ But what's going on on the right here should feel very familiar to
346
+
347
+ 00:08:21.420 --> 00:08:24.580
348
+ anyone who's actually worked through a cross product computation.
349
+
350
+ 00:08:25.900 --> 00:08:29.287
351
+ Collecting the constant terms that are multiplied by x, y,
352
+
353
+ 00:08:29.287 --> 00:08:33.841
354
+ and by z like this is no different from plugging in the symbols i-hat, j-hat,
355
+
356
+ 00:08:33.841 --> 00:08:38.512
357
+ and k-hat to that first column, and seeing which coefficients aggregate on each
358
+
359
+ 00:08:38.512 --> 00:08:39.680
360
+ one of those terms.
361
+
362
+ 00:08:40.960 --> 00:08:44.998
363
+ It's just that plugging in i-hat, j-hat, and k-hat is a way of signaling
364
+
365
+ 00:08:44.998 --> 00:08:49.260
366
+ that we should interpret those coefficients as the coordinates of a vector.
367
+
368
+ 00:08:51.280 --> 00:08:54.245
369
+ So what all of this is saying is that this funky computation
370
+
371
+ 00:08:54.245 --> 00:08:57.260
372
+ can be thought of as a way to answer the following question.
373
+
374
+ 00:08:57.740 --> 00:09:02.183
375
+ What vector p has the special property that when you take a dot product
376
+
377
+ 00:09:02.183 --> 00:09:07.252
378
+ between p and some vector (x, y, z), it gives the same result as plugging in (x,
379
+
380
+ 00:09:07.252 --> 00:09:11.445
381
+ y, z) to the first column of a matrix whose other two columns have
382
+
383
+ 00:09:11.445 --> 00:09:15.200
384
+ the coordinates of v and w, then computing the determinant.
385
+
386
+ 00:09:15.960 --> 00:09:19.780
387
+ That's a bit of a mouthful, but it's an important question to digest for this video.
388
+
389
+ 00:09:21.220 --> 00:09:24.507
390
+ Now for the cool part, which ties all this together with the geometric
391
+
392
+ 00:09:24.507 --> 00:09:27.560
393
+ understanding of the cross product that I introduced last video.
394
+
395
+ 00:09:28.920 --> 00:09:31.804
396
+ I'm going to ask the same question again, but this time we're
397
+
398
+ 00:09:31.804 --> 00:09:35.020
399
+ going to try to answer it geometrically instead of computationally.
400
+
401
+ 00:09:36.420 --> 00:09:42.279
402
+ What 3D vector p has the special property that when you take a dot product between
403
+
404
+ 00:09:42.279 --> 00:09:47.924
405
+ p and some other vector (x, y, z), it gives the same result as if you took the
406
+
407
+ 00:09:47.924 --> 00:09:54.140
408
+ signed volume of a parallelepiped defined by this vector (x, y, z) along with v and w.
409
+
410
+ 00:09:57.140 --> 00:10:01.418
411
+ Remember, the geometric interpretation of a dot product between a
412
+
413
+ 00:10:01.418 --> 00:10:06.091
414
+ vector p and some other vector is to project that other vector onto p,
415
+
416
+ 00:10:06.091 --> 00:10:10.500
417
+ then to multiply the length of that projection by the length of p.
418
+
419
+ 00:10:13.460 --> 00:10:16.290
420
+ With that in mind, let me show a certain way to think
421
+
422
+ 00:10:16.290 --> 00:10:19.440
423
+ about the volume of the parallelepiped that we care about.
424
+
425
+ 00:10:20.240 --> 00:10:24.479
426
+ Start by taking the area of the parallelogram defined by v and w,
427
+
428
+ 00:10:24.479 --> 00:10:29.500
429
+ then multiply it not by the length of (x, y, z), but by the component of (x,
430
+
431
+ 00:10:29.500 --> 00:10:32.760
432
+ y, z) that's perpendicular to that parallelogram.
433
+
434
+ 00:10:34.280 --> 00:10:39.560
435
+ In other words, the way our linear function works on a given vector is to project
436
+
437
+ 00:10:39.560 --> 00:10:43.601
438
+ that vector onto a line that's perpendicular to both v and w,
439
+
440
+ 00:10:43.601 --> 00:10:48.816
441
+ then to multiply the length of that projection by the area of the parallelogram
442
+
443
+ 00:10:48.816 --> 00:10:50.120
444
+ spanned by v and w.
445
+
446
+ 00:10:51.560 --> 00:10:55.568
447
+ But this is the same thing as taking a dot product between (x, y,
448
+
449
+ 00:10:55.568 --> 00:11:00.687
450
+ z) and a vector that's perpendicular to v and w with a length equal to the area of
451
+
452
+ 00:11:00.687 --> 00:11:01.920
453
+ that parallelogram.
454
+
455
+ 00:11:03.200 --> 00:11:07.001
456
+ What's more, if you choose the appropriate direction for that vector,
457
+
458
+ 00:11:07.001 --> 00:11:10.968
459
+ the cases where the dot product is negative will line up with the cases
460
+
461
+ 00:11:10.968 --> 00:11:15.320
462
+ where the right hand rule for the orientation of x, y, z, v and w is negative.
463
+
464
+ 00:11:19.600 --> 00:11:24.697
465
+ This means that we just found a vector p so that taking a dot product between
466
+
467
+ 00:11:24.697 --> 00:11:29.728
468
+ p and some vector (x, y, z) is the same thing as computing that determinant
469
+
470
+ 00:11:29.728 --> 00:11:34.560
471
+ of a 3x3 matrix whose columns are (x, y, z), the coordinates of v and w.
472
+
473
+ 00:11:35.480 --> 00:11:39.032
474
+ So the answer that we found earlier computationally using that
475
+
476
+ 00:11:39.032 --> 00:11:43.100
477
+ special notational trick must correspond geometrically to this vector.
478
+
479
+ 00:11:43.900 --> 00:11:47.073
480
+ This is the fundamental reason why the computation and the
481
+
482
+ 00:11:47.073 --> 00:11:50.300
483
+ geometric interpretation of the cross product are related.
484
+
485
+ 00:11:52.640 --> 00:11:57.648
486
+ Just to sum up what happened here, I started by defining a linear transformation from
487
+
488
+ 00:11:57.648 --> 00:12:02.420
489
+ 3D space to the number line, and it was defined in terms of the vectors v and w.
490
+
491
+ 00:12:03.280 --> 00:12:06.913
492
+ Then I went through two separate ways to think about the dual vector
493
+
494
+ 00:12:06.913 --> 00:12:10.814
495
+ of this transformation, the vector such that applying the transformation
496
+
497
+ 00:12:10.814 --> 00:12:14.020
498
+ is the same thing as taking a dot product with that vector.
499
+
500
+ 00:12:14.860 --> 00:12:17.945
501
+ On the one hand, a computational approach will lead you to
502
+
503
+ 00:12:17.945 --> 00:12:20.657
504
+ the trick of plugging in the symbols i-hat, j-hat,
505
+
506
+ 00:12:20.657 --> 00:12:24.540
507
+ and k-hat to the first column of a matrix and computing the determinant.
508
+
509
+ 00:12:26.020 --> 00:12:29.814
510
+ But thinking geometrically, we can deduce that this dual vector
511
+
512
+ 00:12:29.814 --> 00:12:33.427
513
+ must be perpendicular to v and w with a length equal to the
514
+
515
+ 00:12:33.427 --> 00:12:37.040
516
+ area of the parallelogram spanned out by those two vectors.
517
+
518
+ 00:12:39.100 --> 00:12:43.405
519
+ Since both of these approaches give us a dual vector to the same transformation,
520
+
521
+ 00:12:43.405 --> 00:12:45.020
522
+ they must be the same vector.
523
+
524
+ 00:12:47.400 --> 00:12:49.885
525
+ So that wraps up dot products and cross products,
526
+
527
+ 00:12:49.885 --> 00:12:53.638
528
+ and the next video will be a really important concept for linear algebra,
529
+
530
+ 00:12:53.638 --> 00:12:54.500
531
+ change of basis.
532
+
3B1B_LinearAlgebra/11 - Cross products in the light of linear transformations | Chapter 11, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6e1b14c621aa7759f5a97405538c8bb86d794e8280ad4653651ed899c1b355d7
3
+ size 31547777
3B1B_LinearAlgebra/12 - Cramer's rule, explained geometrically | Chapter 12, Essence of linear algebra.en.vtt ADDED
@@ -0,0 +1,562 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:00.000 --> 00:00:11.200
6
+ [Submit subtitle corrections at criblate.com]
7
+
8
+ 00:00:11.200 --> 00:00:14.496
9
+ In a previous video I've talked about linear systems of equations,
10
+
11
+ 00:00:14.496 --> 00:00:18.391
12
+ and I sort of brushed aside the discussion of actually computing solutions to
13
+
14
+ 00:00:18.391 --> 00:00:19.140
15
+ these systems.
16
+
17
+ 00:00:19.700 --> 00:00:22.580
18
+ And while it's true that number crunching is typically something we
19
+
20
+ 00:00:22.580 --> 00:00:25.374
21
+ leave to the computers, digging into some of these computational
22
+
23
+ 00:00:25.374 --> 00:00:28.511
24
+ methods is a good litmus test for whether or not you actually understand
25
+
26
+ 00:00:28.511 --> 00:00:31.520
27
+ what's going on, since that's really where the rubber meets the road.
28
+
29
+ 00:00:32.120 --> 00:00:35.554
30
+ Here I want to describe the geometry behind a certain method for
31
+
32
+ 00:00:35.554 --> 00:00:38.880
33
+ computing solutions to these systems, known as Cramer's rule.
34
+
35
+ 00:00:39.680 --> 00:00:42.573
36
+ The relevant background here is understanding determinants,
37
+
38
+ 00:00:42.573 --> 00:00:46.153
39
+ a little bit of dot products, and of course linear systems of equations,
40
+
41
+ 00:00:46.153 --> 00:00:50.420
42
+ so be sure to watch the relevant videos on those topics if you're unfamiliar or rusty.
43
+
44
+ 00:00:51.020 --> 00:00:54.259
45
+ But first I should say up front that this Cramer's rule is not
46
+
47
+ 00:00:54.259 --> 00:00:58.334
48
+ actually the best way for computing solutions to linear systems of equations,
49
+
50
+ 00:00:58.334 --> 00:01:01.260
51
+ Gaussian elimination for example will always be faster.
52
+
53
+ 00:01:01.980 --> 00:01:03.520
54
+ So why learn it?
55
+
56
+ 00:01:03.780 --> 00:01:05.840
57
+ Well think of it as a sort of cultural excursion.
58
+
59
+ 00:01:06.420 --> 00:01:10.460
60
+ It's a helpful exercise in deepening your knowledge of the theory behind these systems.
61
+
62
+ 00:01:11.040 --> 00:01:15.184
63
+ Wrapping your mind around this concept is going to help consolidate ideas from linear
64
+
65
+ 00:01:15.184 --> 00:01:17.621
66
+ algebra, like the determinant and linear systems,
67
+
68
+ 00:01:17.621 --> 00:01:19.620
69
+ by seeing how they relate to each other.
70
+
71
+ 00:01:20.100 --> 00:01:23.538
72
+ Also from a purely artistic standpoint the ultimate result here is just
73
+
74
+ 00:01:23.538 --> 00:01:26.880
75
+ really pretty to think about, way more so than Gaussian elimination.
76
+
77
+ 00:01:28.680 --> 00:01:32.663
78
+ Alright so the setup here will be some linear system of equations,
79
+
80
+ 00:01:32.663 --> 00:01:35.620
81
+ say with two unknowns x and y and two equations.
82
+
83
+ 00:01:36.300 --> 00:01:39.423
84
+ In principle everything we're talking about will also work for systems
85
+
86
+ 00:01:39.423 --> 00:01:42.323
87
+ with larger number of unknowns and the same number of equations,
88
+
89
+ 00:01:42.323 --> 00:01:45.580
90
+ but for simplicity a smaller example is just nicer to hold in our heads.
91
+
92
+ 00:01:46.320 --> 00:01:50.639
93
+ So as I talked about in a previous video you can think of this setup
94
+
95
+ 00:01:50.639 --> 00:01:55.403
96
+ geometrically as a certain known matrix transforming an unknown vector (x,
97
+
98
+ 00:01:55.403 --> 00:02:00.040
99
+ y) where you know what the output is going to be, in this case (-4, -2).
100
+
101
+ 00:02:00.800 --> 00:02:05.969
102
+ Remember the columns of this matrix are telling you how that matrix acts as a transform,
103
+
104
+ 00:02:05.969 --> 00:02:10.080
105
+ each one telling you where the basis vectors of the input space land.
106
+
107
+ 00:02:10.860 --> 00:02:15.251
108
+ So what we have is a sort of puzzle, which input vector (x,
109
+
110
+ 00:02:15.251 --> 00:02:18.600
111
+ y) is going to land on this output (-4, -2).
112
+
113
+ 00:02:19.700 --> 00:02:23.953
114
+ One way to think about our little puzzle here is that we know the given output
115
+
116
+ 00:02:23.953 --> 00:02:28.042
117
+ vector is some linear combination of the columns of the matrix x times the
118
+
119
+ 00:02:28.042 --> 00:02:31.749
120
+ vector where i-hat lands plus y times the vector where j-hat lands,
121
+
122
+ 00:02:31.749 --> 00:02:36.220
123
+ but what we want is to figure out what exactly that linear combination should be.
124
+
125
+ 00:02:37.240 --> 00:02:41.773
126
+ Remember the type of answer you get here can depend on whether or not the transformation
127
+
128
+ 00:02:41.773 --> 00:02:46.100
129
+ squishes all of space into a lower dimension, that is if it has a zero determinant.
130
+
131
+ 00:02:46.100 --> 00:02:50.159
132
+ In that case either none of the inputs land on our given output,
133
+
134
+ 00:02:50.159 --> 00:02:53.900
135
+ or there's a whole bunch of inputs landing on that output.
136
+
137
+ 00:02:57.060 --> 00:03:01.404
138
+ But for this video we'll limit our view to the case of a non-zero determinant,
139
+
140
+ 00:03:01.404 --> 00:03:04.634
141
+ meaning the outputs of this transformation still span the
142
+
143
+ 00:03:04.634 --> 00:03:07.140
144
+ full n-dimensional space that it started in.
145
+
146
+ 00:03:07.500 --> 00:03:12.700
147
+ Every input lands on one and only one output, and every output has one and only one input.
148
+
149
+ 00:03:14.180 --> 00:03:18.160
150
+ As a first pass let me show you an idea that's wrong but in the right direction.
151
+
152
+ 00:03:18.800 --> 00:03:22.093
153
+ The x coordinate of this mystery input vector is what you get
154
+
155
+ 00:03:22.093 --> 00:03:25.440
156
+ by taking its dot product with the first basis vector (1, 0).
157
+
158
+ 00:03:26.160 --> 00:03:30.949
159
+ Likewise the y coordinate is what you get by dotting it with the second basis vector,
160
+
161
+ 00:03:30.949 --> 00:03:31.400
162
+ (0, 1).
163
+
164
+ 00:03:31.900 --> 00:03:35.717
165
+ So maybe you hope that after the transformation the dot products with
166
+
167
+ 00:03:35.717 --> 00:03:39.423
168
+ the transformed version of the mystery vector with the transformed
169
+
170
+ 00:03:39.423 --> 00:03:43.240
171
+ version of the basis vectors will also be these coordinates x and y.
172
+
173
+ 00:03:43.940 --> 00:03:46.800
174
+ That'd be fantastic because we know what the transformed
175
+
176
+ 00:03:46.800 --> 00:03:48.740
177
+ version of each of those vectors are.
178
+
179
+ 00:03:51.180 --> 00:03:54.200
180
+ There's just one problem with it, it's not at all true.
181
+
182
+ 00:03:54.640 --> 00:03:57.414
183
+ For most linear transformations the dot product before
184
+
185
+ 00:03:57.414 --> 00:04:00.240
186
+ and after the transformation will look very different.
187
+
188
+ 00:04:00.800 --> 00:04:04.434
189
+ For example, you could have two vectors generally pointing in the same direction
190
+
191
+ 00:04:04.434 --> 00:04:08.022
192
+ with a positive dot product, which get pulled apart from each other during the
193
+
194
+ 00:04:08.022 --> 00:04:11.520
195
+ transformation in such a way that they end up having a negative dot product.
196
+
197
+ 00:04:12.220 --> 00:04:15.541
198
+ Likewise things that start off perpendicular with dot product 0,
199
+
200
+ 00:04:15.541 --> 00:04:19.329
201
+ like the two basis vectors, quite often don't stay perpendicular to each
202
+
203
+ 00:04:19.329 --> 00:04:23.480
204
+ other after the transformation, that is they don't preserve that 0 dot product.
205
+
206
+ 00:04:24.100 --> 00:04:27.426
207
+ And looking at the example I just showed dot products certainly aren't preserved,
208
+
209
+ 00:04:27.426 --> 00:04:30.300
210
+ they tend to get bigger since most vectors are getting stretched out.
211
+
212
+ 00:04:31.040 --> 00:04:34.861
213
+ In fact, worthwhile side note here, transformations which do preserve dot
214
+
215
+ 00:04:34.861 --> 00:04:39.100
216
+ products are special enough to have their own name, orthonormal transformations.
217
+
218
+ 00:04:39.720 --> 00:04:42.123
219
+ These are the ones that leave all of the basis vectors
220
+
221
+ 00:04:42.123 --> 00:04:44.660
222
+ perpendicular to each other and still with unit lengths.
223
+
224
+ 00:04:45.740 --> 00:04:48.283
225
+ You often think of these as the rotation matrices,
226
+
227
+ 00:04:48.283 --> 00:04:52.200
228
+ they correspond to rigid motion with no stretching or squishing or morphing.
229
+
230
+ 00:04:53.000 --> 00:04:56.780
231
+ Solving a linear system with an orthonormal matrix is actually super easy.
232
+
233
+ 00:04:57.260 --> 00:05:01.322
234
+ Because dot products are preserved, taking the dot product between the output
235
+
236
+ 00:05:01.322 --> 00:05:05.173
237
+ vector and all the columns of your matrix will be the same as taking the
238
+
239
+ 00:05:05.173 --> 00:05:09.129
240
+ dot product between the mystery input vector and all of the basis vectors,
241
+
242
+ 00:05:09.129 --> 00:05:12.980
243
+ which is the same as just finding the coordinates of that mystery input.
244
+
245
+ 00:05:13.680 --> 00:05:18.548
246
+ So in that very special case, x would be the dot product of the first column with the
247
+
248
+ 00:05:18.548 --> 00:05:23.302
249
+ output vector, and y would be the dot product of the second column with the output
250
+
251
+ 00:05:23.302 --> 00:05:23.760
252
+ vector.
253
+
254
+ 00:05:26.820 --> 00:05:30.860
255
+ Why am I bringing this up when this idea breaks down for almost all linear systems?
256
+
257
+ 00:05:31.420 --> 00:05:34.080
258
+ Well, it points us in a direction of something to look for.
259
+
260
+ 00:05:34.320 --> 00:05:38.045
261
+ Is there an alternate geometric understanding for the coordinates of
262
+
263
+ 00:05:38.045 --> 00:05:41.660
264
+ our input vector that remains unchanged after the transformation?
265
+
266
+ 00:05:42.360 --> 00:05:44.627
267
+ If your mind has been mulling over determinants,
268
+
269
+ 00:05:44.627 --> 00:05:46.800
270
+ you might think of the following clever idea.
271
+
272
+ 00:05:47.360 --> 00:05:50.661
273
+ Take the parallelogram defined by the first basis
274
+
275
+ 00:05:50.661 --> 00:05:53.760
276
+ vector i-hat and the mystery input vector xy.
277
+
278
+ 00:05:54.440 --> 00:05:57.260
279
+ The area of this parallelogram is the base, 1,
280
+
281
+ 00:05:57.260 --> 00:06:01.612
282
+ times the height perpendicular to that base, which is the y-coordinate
283
+
284
+ 00:06:01.612 --> 00:06:02.960
285
+ of that input vector.
286
+
287
+ 00:06:03.680 --> 00:06:06.791
288
+ So the area of that parallelogram is a sort of screwy
289
+
290
+ 00:06:06.791 --> 00:06:09.960
291
+ roundabout way to describe the vector's y-coordinate.
292
+
293
+ 00:06:10.420 --> 00:06:13.140
294
+ It's a wacky way to talk about coordinates, but run with me.
295
+
296
+ 00:06:13.700 --> 00:06:17.795
297
+ And actually, to be a little more accurate, you should think of this as the signed
298
+
299
+ 00:06:17.795 --> 00:06:21.640
300
+ area of that parallelogram, in the sense described in the determinant video.
301
+
302
+ 00:06:22.200 --> 00:06:26.483
303
+ That way, a vector with a negative y-coordinate would correspond to a negative
304
+
305
+ 00:06:26.483 --> 00:06:30.382
306
+ area for this parallelogram, at least if you think of i-hat as in some
307
+
308
+ 00:06:30.382 --> 00:06:34.500
309
+ sense being the first out of these two vectors defining the parallelogram.
310
+
311
+ 00:06:35.160 --> 00:06:38.433
312
+ And symmetrically, if you look at the parallelogram spanned
313
+
314
+ 00:06:38.433 --> 00:06:41.594
315
+ by our mystery input vector and the second basis, j-hat,
316
+
317
+ 00:06:41.594 --> 00:06:45.200
318
+ its area is going to be the x-coordinate of that mystery vector.
319
+
320
+ 00:06:45.780 --> 00:06:48.369
321
+ Again, it's a strange way to represent the x-coordinate,
322
+
323
+ 00:06:48.369 --> 00:06:50.080
324
+ but see what it buys us in a moment.
325
+
326
+ 00:06:50.680 --> 00:06:52.677
327
+ And just to make sure it's clear how this might generalize,
328
+
329
+ 00:06:52.677 --> 00:06:53.760
330
+ let's look in three dimensions.
331
+
332
+ 00:06:54.300 --> 00:06:57.904
333
+ Ordinarily, the way you might think about one of a vector's coordinates,
334
+
335
+ 00:06:57.904 --> 00:07:00.907
336
+ say its z-coordinate, would be to take its dot product with
337
+
338
+ 00:07:00.907 --> 00:07:03.560
339
+ the third standard basis vector, often called k-hat.
340
+
341
+ 00:07:04.280 --> 00:07:08.987
342
+ But an alternate geometric interpretation would be to consider the parallelepiped
343
+
344
+ 00:07:08.987 --> 00:07:12.880
345
+ that it creates with the other two basis vectors, i-hat and j-hat.
346
+
347
+ 00:07:13.420 --> 00:07:16.643
348
+ If you think of the square with area 1 spanned by i-hat and
349
+
350
+ 00:07:16.643 --> 00:07:21.122
351
+ j-hat as the base of this whole shape, then its volume is the same as its height,
352
+
353
+ 00:07:21.122 --> 00:07:23.580
354
+ which is the third coordinate of our vector.
355
+
356
+ 00:07:24.340 --> 00:07:28.056
357
+ And likewise, the wacky way to think about the other coordinates of the vector
358
+
359
+ 00:07:28.056 --> 00:07:31.724
360
+ would be to form a parallelepiped using the vector and then all of the basis
361
+
362
+ 00:07:31.724 --> 00:07:35.440
363
+ vectors other than the one corresponding to the direction you're looking for.
364
+
365
+ 00:07:35.900 --> 00:07:37.900
366
+ Then the volume of this gives you the coordinate.
367
+
368
+ 00:07:38.440 --> 00:07:41.676
369
+ Or rather, we should be talking about the signed volume of parallelepiped
370
+
371
+ 00:07:41.676 --> 00:07:45.000
372
+ in the sense described in the determinant video using the right-hand rule.
373
+
374
+ 00:07:45.560 --> 00:07:48.140
375
+ So the order in which you list these three vectors matters.
376
+
377
+ 00:07:48.800 --> 00:07:51.100
378
+ That way, negative coordinates still make sense.
379
+
380
+ 00:07:52.040 --> 00:07:55.240
381
+ Okay, so why think of coordinates as areas and volumes like this?
382
+
383
+ 00:07:55.720 --> 00:08:00.348
384
+ Well, as you apply some sort of matrix transformation, the areas of these parallelograms,
385
+
386
+ 00:08:00.348 --> 00:08:03.780
387
+ well, they don't stay the same, they might get scaled up or down.
388
+
389
+ 00:08:04.160 --> 00:08:06.814
390
+ But, and this is the key idea of determinants,
391
+
392
+ 00:08:06.814 --> 00:08:09.583
393
+ all of the areas get scaled by the same amount,
394
+
395
+ 00:08:09.583 --> 00:08:12.640
396
+ namely the determinant of our transformation matrix.
397
+
398
+ 00:08:13.520 --> 00:08:17.188
399
+ For example, if you look at the parallelogram spanned by the vector
400
+
401
+ 00:08:17.188 --> 00:08:21.459
402
+ where your first basis vector lands, which is the first column of the matrix,
403
+
404
+ 00:08:21.459 --> 00:08:24.580
405
+ and the transformed version of (x, y), what is its area?
406
+
407
+ 00:08:25.580 --> 00:08:29.908
408
+ Well, this is the transformed version of the parallelogram we were looking at earlier,
409
+
410
+ 00:08:29.908 --> 00:08:33.380
411
+ the one whose area was the y-coordinate of the mystery input vector.
412
+
413
+ 00:08:33.700 --> 00:08:36.435
414
+ So its area is just going to be the determinant of
415
+
416
+ 00:08:36.435 --> 00:08:39.280
417
+ the transformation multiplied by that y-coordinate.
418
+
419
+ 00:08:40.180 --> 00:08:45.094
420
+ So that means we can solve for y by taking the area of this new parallelogram
421
+
422
+ 00:08:45.094 --> 00:08:49.880
423
+ in the output space divided by the determinant of the full transformation.
424
+
425
+ 00:08:50.900 --> 00:08:52.420
426
+ And how do you get that area?
427
+
428
+ 00:08:53.240 --> 00:08:56.550
429
+ Well, we know the coordinates for where the mystery input vector lands,
430
+
431
+ 00:08:56.550 --> 00:08:59.160
432
+ that's the whole point of a linear system of equations.
433
+
434
+ 00:08:59.720 --> 00:09:05.020
435
+ So what you might do is create a new matrix whose first column is the same as that of our
436
+
437
+ 00:09:05.020 --> 00:09:10.320
438
+ matrix, but whose second column is the output vector, and then you take its determinant.
439
+
440
+ 00:09:11.260 --> 00:09:15.037
441
+ So look at that, just using data from the output of the transformation,
442
+
443
+ 00:09:15.037 --> 00:09:19.027
444
+ namely the columns of the matrix and the coordinates of our output vector,
445
+
446
+ 00:09:19.027 --> 00:09:22.272
447
+ we can recover the y-coordinate of the mystery input vector,
448
+
449
+ 00:09:22.272 --> 00:09:24.400
450
+ which is halfway to solving the system.
451
+
452
+ 00:09:25.120 --> 00:09:27.540
453
+ Likewise, the same idea can give us the x-coordinate.
454
+
455
+ 00:09:28.000 --> 00:09:31.533
456
+ Look at the parallelogram we defined earlier, which encodes the
457
+
458
+ 00:09:31.533 --> 00:09:35.740
459
+ x-coordinate of the mystery input vector spanned by that vector and j-hat.
460
+
461
+ 00:09:36.400 --> 00:09:41.660
462
+ The transformed version of this guy is spanned by the output vector and the second column
463
+
464
+ 00:09:41.660 --> 00:09:46.920
465
+ of the matrix, and its area will have been multiplied by the determinant of that matrix.
466
+
467
+ 00:09:47.700 --> 00:09:50.477
468
+ So to solve for x, you can take this new area divided
469
+
470
+ 00:09:50.477 --> 00:09:52.940
471
+ by the determinant of the full transformation.
472
+
473
+ 00:09:53.860 --> 00:09:57.960
474
+ And similar to what we did before, you can compute the area of that output
475
+
476
+ 00:09:57.960 --> 00:10:01.948
477
+ parallelogram by creating a new matrix whose first column is the output
478
+
479
+ 00:10:01.948 --> 00:10:05.660
480
+ vector and whose second column is the same as the original matrix.
481
+
482
+ 00:10:06.240 --> 00:10:08.666
483
+ So again, just using data from the output space,
484
+
485
+ 00:10:08.666 --> 00:10:12.860
486
+ the numbers we see in our original linear system, we can solve for what x must be.
487
+
488
+ 00:10:13.420 --> 00:10:15.997
489
+ This formula for finding the solutions to a linear
490
+
491
+ 00:10:15.997 --> 00:10:18.420
492
+ system of equations is known as Cramer's rule.
493
+
494
+ 00:10:19.120 --> 00:10:21.900
495
+ Here, just to sanity check ourselves, plug in some numbers here.
496
+
497
+ 00:10:22.260 --> 00:10:26.572
498
+ The determinant of that top altered matrix is 4 plus 2, which is 6,
499
+
500
+ 00:10:26.572 --> 00:10:30.820
501
+ and the bottom determinant is 2, so the x-coordinate should be 3.
502
+
503
+ 00:10:31.440 --> 00:10:35.460
504
+ And indeed, looking back at the input vector we started with, the x-coordinate is 3.
505
+
506
+ 00:10:36.320 --> 00:10:41.226
507
+ Likewise, Cramer's rule suggests that the y-coordinate should be 4 divided by 2,
508
+
509
+ 00:10:41.226 --> 00:10:46.500
510
+ or 2, and that is in fact the y-coordinate of the input vector we were starting with.
511
+
512
+ 00:10:47.380 --> 00:10:49.708
513
+ The case with three dimensions or more is similar,
514
+
515
+ 00:10:49.708 --> 00:10:53.480
516
+ and I highly recommend you take a moment to pause and think through it yourself.
517
+
518
+ 00:10:54.180 --> 00:10:55.900
519
+ Here, I'll give you a little bit of momentum.
520
+
521
+ 00:10:56.340 --> 00:11:00.328
522
+ What we have is a known transformation given by some 3x3 matrix
523
+
524
+ 00:11:00.328 --> 00:11:04.885
525
+ and a known output vector given by the right side of our linear system,
526
+
527
+ 00:11:04.885 --> 00:11:08.240
528
+ and we want to know what input lands on that output.
529
+
530
+ 00:11:09.100 --> 00:11:13.893
531
+ And if you think of, say, the z-coordinate of that input vector as the volume of
532
+
533
+ 00:11:13.893 --> 00:11:18.747
534
+ that special parallelepiped we were looking at earlier, spanned by i-hat, j-hat,
535
+
536
+ 00:11:18.747 --> 00:11:23.780
537
+ and the mystery input vector, what happens to that volume after the transformation?
538
+
539
+ 00:11:24.800 --> 00:11:27.480
540
+ And what are the various ways you can compute that volume?
541
+
542
+ 00:11:28.340 --> 00:11:31.383
543
+ Really, pause and take a moment to think through the details of
544
+
545
+ 00:11:31.383 --> 00:11:34.377
546
+ generalizing this to higher dimensions, finding an expression
547
+
548
+ 00:11:34.377 --> 00:11:37.420
549
+ for each coordinate of the solution to a larger linear system.
550
+
551
+ 00:11:38.060 --> 00:11:41.452
552
+ Thinking through more general cases like this and convincing yourself that it
553
+
554
+ 00:11:41.452 --> 00:11:44.315
555
+ works and why it works is where all the learning really happens,
556
+
557
+ 00:11:44.315 --> 00:11:47.751
558
+ much more so than listening to some dude on YouTube walk you through the same
559
+
560
+ 00:11:47.751 --> 00:11:48.500
561
+ reasoning again.
562
+
3B1B_LinearAlgebra/12 - Cramer's rule, explained geometrically | Chapter 12, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6eb02b47e12d20019d459c64580e33529665a31e8a3a0bcbb0ce7c4b4f1a4e99
3
+ size 82340068
3B1B_LinearAlgebra/13 - Change of basis | Chapter 13, Essence of linear algebra.en.vtt ADDED
@@ -0,0 +1,520 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:00.000 --> 00:00:12.760
6
+ [Submit subtitle corrections at criblate.com]
7
+
8
+ 00:00:12.760 --> 00:00:15.370
9
+ If I have a vector sitting here in 2D space, we have
10
+
11
+ 00:00:15.370 --> 00:00:17.780
12
+ a standard way to describe it with coordinates.
13
+
14
+ 00:00:18.560 --> 00:00:21.203
15
+ In this case, the vector has coordinates (3, 2),
16
+
17
+ 00:00:21.203 --> 00:00:25.278
18
+ which means going from its tail to its tip involves moving three units to
19
+
20
+ 00:00:25.278 --> 00:00:26.820
21
+ the right and two units up.
22
+
23
+ 00:00:27.510 --> 00:00:31.862
24
+ Now, the more linear algebra-oriented way to describe coordinates is to think of
25
+
26
+ 00:00:31.862 --> 00:00:36.160
27
+ each of these numbers as a scalar, a thing that stretches or squishes vectors.
28
+
29
+ 00:00:37.020 --> 00:00:40.322
30
+ You think of that first coordinate as scaling i-hat,
31
+
32
+ 00:00:40.322 --> 00:00:43.369
33
+ the vector with length 1 pointing to the right,
34
+
35
+ 00:00:43.369 --> 00:00:49.020
36
+ while the second coordinate scales j-hat, the vector with length 1 pointing straight up.
37
+
38
+ 00:00:49.660 --> 00:00:52.200
39
+ The tip-to-tail sum of those two scaled vectors
40
+
41
+ 00:00:52.200 --> 00:00:54.740
42
+ is what the coordinates are meant to describe.
43
+
44
+ 00:00:55.700 --> 00:00:58.837
45
+ You can think of these two special vectors as encapsulating
46
+
47
+ 00:00:58.837 --> 00:01:01.920
48
+ all of the implicit assumptions of our coordinate system.
49
+
50
+ 00:01:02.860 --> 00:01:05.848
51
+ The fact that the first number indicates rightward motion,
52
+
53
+ 00:01:05.848 --> 00:01:10.175
54
+ that the second one indicates upward motion, exactly how far a unit of distance is,
55
+
56
+ 00:01:10.175 --> 00:01:13.420
57
+ all of that is tied up in the choice of i-hat and j-hat as the
58
+
59
+ 00:01:13.420 --> 00:01:16.820
60
+ vectors which our scalar coordinates are meant to actually scale.
61
+
62
+ 00:01:17.920 --> 00:01:22.970
63
+ Any way to translate between vectors and sets of numbers is called a coordinate system,
64
+
65
+ 00:01:22.970 --> 00:01:26.744
66
+ and the two special vectors i-hat and j-hat are called the basis
67
+
68
+ 00:01:26.744 --> 00:01:29.240
69
+ vectors of our standard coordinate system.
70
+
71
+ 00:01:30.240 --> 00:01:34.640
72
+ What I'd like to talk about here is the idea of using a different set of basis vectors.
73
+
74
+ 00:01:35.920 --> 00:01:39.106
75
+ For example, let's say you have a friend, Jennifer,
76
+
77
+ 00:01:39.106 --> 00:01:43.480
78
+ who uses a different set of basis vectors, which I'll call b1 and b2.
79
+
80
+ 00:01:44.080 --> 00:01:48.266
81
+ Her first basis vector, b1, points up and to the right a little bit,
82
+
83
+ 00:01:48.266 --> 00:01:51.160
84
+ and her second vector, b2, points left and up.
85
+
86
+ 00:01:51.760 --> 00:01:54.773
87
+ Now take another look at that vector that I showed earlier,
88
+
89
+ 00:01:54.773 --> 00:01:58.246
90
+ the one that you and I would describe using the coordinates (3, 2),
91
+
92
+ 00:01:58.246 --> 00:02:00.340
93
+ using our basis vectors i-hat and j-hat.
94
+
95
+ 00:02:01.480 --> 00:02:05.900
96
+ Jennifer would actually describe this vector with the coordinates (5/3, 1/3).
97
+
98
+ 00:02:06.820 --> 00:02:13.176
99
+ What this means is that the particular way to get to that vector using her two
100
+
101
+ 00:02:13.176 --> 00:02:19.940
102
+ basis vectors is to scale b1 by 5/3, scale b2 by 1/3, then add them both together.
103
+
104
+ 00:02:20.560 --> 00:02:24.276
105
+ In a little bit, I'll show you how you could have figured out those two numbers,
106
+
107
+ 00:02:24.276 --> 00:02:24.880
108
+ 5/3 and 1/3.
109
+
110
+ 00:02:25.980 --> 00:02:29.669
111
+ In general, whenever Jennifer uses coordinates to describe a vector,
112
+
113
+ 00:02:29.669 --> 00:02:32.382
114
+ she thinks of her first coordinate as scaling b1,
115
+
116
+ 00:02:32.382 --> 00:02:35.800
117
+ the second coordinate as scaling b2, and she adds the results.
118
+
119
+ 00:02:37.820 --> 00:02:40.270
120
+ What she gets will typically be completely different from the
121
+
122
+ 00:02:40.270 --> 00:02:42.920
123
+ vector that you and I would think of as having those coordinates.
124
+
125
+ 00:02:44.960 --> 00:02:49.366
126
+ To be a little more precise about the setup here, her first basis vector,
127
+
128
+ 00:02:49.366 --> 00:02:53.531
129
+ b1, is something that we would describe with the coordinates (2, 1),
130
+
131
+ 00:02:53.531 --> 00:02:58.420
132
+ and her second basis vector, b2, is something that we would describe as (-1, 1).
133
+
134
+ 00:02:59.080 --> 00:03:02.856
135
+ But it's important to realize from her perspective in her system,
136
+
137
+ 00:03:02.856 --> 00:03:05.760
138
+ those vectors have coordinates (1, 0) and (0, 1).
139
+
140
+ 00:03:06.060 --> 00:03:11.600
141
+ They are what define the meaning of the coordinates (1, 0) and (0, 1) in her world.
142
+
143
+ 00:03:12.760 --> 00:03:15.340
144
+ So in effect, we're speaking different languages.
145
+
146
+ 00:03:15.340 --> 00:03:18.033
147
+ We're all looking at the same vectors in space,
148
+
149
+ 00:03:18.033 --> 00:03:21.700
150
+ but Jennifer uses different words and numbers to describe them.
151
+
152
+ 00:03:23.600 --> 00:03:26.060
153
+ Let me say a quick word about how I'm representing things here.
154
+
155
+ 00:03:26.540 --> 00:03:29.780
156
+ When I animate 2D space, I typically use this square grid.
157
+
158
+ 00:03:30.300 --> 00:03:34.229
159
+ But that grid is just a construct, a way to visualize our coordinate system,
160
+
161
+ 00:03:34.229 --> 00:03:36.400
162
+ and so it depends on our choice of basis.
163
+
164
+ 00:03:38.060 --> 00:03:40.500
165
+ Space itself has no intrinsic grid.
166
+
167
+ 00:03:41.560 --> 00:03:45.534
168
+ Jennifer might draw her own grid, which would be an equally made up construct meant
169
+
170
+ 00:03:45.534 --> 00:03:49.460
171
+ as nothing more than a visual tool to help follow the meaning of her coordinates.
172
+
173
+ 00:03:53.300 --> 00:03:55.859
174
+ Her origin though would actually line up with ours,
175
+
176
+ 00:03:55.859 --> 00:03:59.220
177
+ since everybody agrees on what the coordinates (0, 0) should mean.
178
+
179
+ 00:03:59.800 --> 00:04:02.740
180
+ It's the thing that you get when you scale any vector by 0.
181
+
182
+ 00:04:03.620 --> 00:04:07.680
183
+ But the direction of her axes and the spacing of her grid lines will be different,
184
+
185
+ 00:04:07.680 --> 00:04:09.760
186
+ depending on her choice of basis vectors.
187
+
188
+ 00:04:11.440 --> 00:04:14.264
189
+ So after all this is set up, a pretty natural question
190
+
191
+ 00:04:14.264 --> 00:04:17.140
192
+ to ask is how we translate between coordinate systems.
193
+
194
+ 00:04:18.400 --> 00:04:22.665
195
+ If for example, Jennifer describes a vector with coordinates (-1,
196
+
197
+ 00:04:22.665 --> 00:04:25.880
198
+ 2), what would that be in our coordinate system?
199
+
200
+ 00:04:26.320 --> 00:04:28.520
201
+ How do you translate from her language to ours?
202
+
203
+ 00:04:29.380 --> 00:04:36.260
204
+ Well, what her coordinates are saying is that this vector is -1 times b1 plus 2 times b2.
205
+
206
+ 00:04:39.480 --> 00:04:46.700
207
+ And from our perspective, b1 has coordinates (2, 1), and b2 has coordinates (-1, 1).
208
+
209
+ 00:04:49.300 --> 00:04:53.172
210
+ So we can actually compute -1 times b1 plus 2 times
211
+
212
+ 00:04:53.172 --> 00:04:57.120
213
+ b2 as they're represented in our coordinate system.
214
+
215
+ 00:04:57.760 --> 00:05:01.340
216
+ And working this out, you get a vector with coordinates (-4, 1).
217
+
218
+ 00:05:03.440 --> 00:05:07.540
219
+ So that's how we would describe the vector that she thinks of as (-1, 2).
220
+
221
+ 00:05:08.360 --> 00:05:12.673
222
+ This process here of scaling each of her basis vectors by the corresponding
223
+
224
+ 00:05:12.673 --> 00:05:17.560
225
+ coordinates of some vector, then adding them together, might feel somewhat familiar.
226
+
227
+ 00:05:18.440 --> 00:05:21.604
228
+ It's matrix-vector multiplication, with a matrix whose
229
+
230
+ 00:05:21.604 --> 00:05:25.120
231
+ columns represent Jennifer's basis vectors in our language.
232
+
233
+ 00:05:25.720 --> 00:05:30.094
234
+ In fact, once you understand matrix-vector multiplication as applying a certain linear
235
+
236
+ 00:05:30.094 --> 00:05:34.315
237
+ transformation, say by watching what I view to be the most important video in this
238
+
239
+ 00:05:34.315 --> 00:05:38.740
240
+ series, chapter 3, there's a pretty intuitive way to think about what's going on here.
241
+
242
+ 00:05:39.680 --> 00:05:44.107
243
+ A matrix whose columns represent Jennifer's basis vectors can be thought
244
+
245
+ 00:05:44.107 --> 00:05:48.411
246
+ of as a transformation that moves our basis vectors, i-hat and j-hat,
247
+
248
+ 00:05:48.411 --> 00:05:53.514
249
+ the things we think of when we say (1, 0) and (0, 1), to Jennifer's basis vectors,
250
+
251
+ 00:05:53.514 --> 00:05:57.080
252
+ the things she thinks of when she says (1, 0) and (0, 1).
253
+
254
+ 00:05:58.600 --> 00:06:02.814
255
+ To show how this works, let's walk through what it would mean to take the vector
256
+
257
+ 00:06:02.814 --> 00:06:07.080
258
+ that we think of as having coordinates (-1, 2) and applying that transformation.
259
+
260
+ 00:06:08.200 --> 00:06:12.450
261
+ Before the linear transformation, we're thinking of this vector as a certain
262
+
263
+ 00:06:12.450 --> 00:06:16.700
264
+ linear combination of our basis vectors, -1 times i-hat plus 2 times j-hat.
265
+
266
+ 00:06:17.300 --> 00:06:22.133
267
+ And the key feature of a linear transformation is that the resulting vector
268
+
269
+ 00:06:22.133 --> 00:06:26.450
270
+ will be that same linear combination but of the new basis vectors,
271
+
272
+ 00:06:26.450 --> 00:06:31.540
273
+ -1 times the place where i-hat lands plus 2 times the place where j-hat lands.
274
+
275
+ 00:06:33.460 --> 00:06:37.620
276
+ So what this matrix does is transform our misconception of what
277
+
278
+ 00:06:37.620 --> 00:06:41.780
279
+ Jennifer means into the actual vector that she's referring to.
280
+
281
+ 00:06:44.280 --> 00:06:47.800
282
+ I remember that when I was first learning this, it always felt kind of backwards to me.
283
+
284
+ 00:06:47.800 --> 00:06:53.121
285
+ Geometrically, this matrix transforms our grid into Jennifer's grid, but numerically,
286
+
287
+ 00:06:53.121 --> 00:06:57.440
288
+ it's translating a vector described in her language to our language.
289
+
290
+ 00:07:00.880 --> 00:07:05.087
291
+ What made it finally click for me was thinking about how it takes our misconception
292
+
293
+ 00:07:05.087 --> 00:07:09.548
294
+ of what Jennifer means, the vector we get using the same coordinates but in our system,
295
+
296
+ 00:07:09.548 --> 00:07:12.640
297
+ then it transforms it into the vector that she really meant.
298
+
299
+ 00:07:14.940 --> 00:07:16.820
300
+ What about going the other way around?
301
+
302
+ 00:07:18.120 --> 00:07:22.786
303
+ In the example I used earlier this video, when I had the vector with coordinates (3,
304
+
305
+ 00:07:22.786 --> 00:07:26.840
306
+ 2) in our system, how did I compute that it would have coordinates (5/3,
307
+
308
+ 00:07:26.840 --> 00:07:28.340
309
+ 1/3) in Jennifer's system?
310
+
311
+ 00:07:32.220 --> 00:07:36.854
312
+ You start with that change of basis matrix that translates Jennifer's language into ours,
313
+
314
+ 00:07:36.854 --> 00:07:38.260
315
+ then you take its inverse.
316
+
317
+ 00:07:43.440 --> 00:07:46.914
318
+ Remember, the inverse of a transformation is a new transformation
319
+
320
+ 00:07:46.914 --> 00:07:49.800
321
+ that corresponds to playing that first one backwards.
322
+
323
+ 00:07:50.300 --> 00:07:53.479
324
+ In practice, especially when you're working in more than two dimensions,
325
+
326
+ 00:07:53.479 --> 00:07:57.100
327
+ you'd use a computer to compute the matrix that actually represents this inverse.
328
+
329
+ 00:07:58.180 --> 00:08:03.805
330
+ In this case, the inverse of the change of basis matrix that has Jennifer's basis
331
+
332
+ 00:08:03.805 --> 00:08:09.360
333
+ as its columns ends up working out to have columns (1/3, -1/3), and (1/3, 2/3).
334
+
335
+ 00:08:11.280 --> 00:08:16.685
336
+ So, for example, to see what the vector (3, 2) looks like in Jennifer's system,
337
+
338
+ 00:08:16.685 --> 00:08:21.200
339
+ we multiply this inverse change of basis matrix by the vector (3,
340
+
341
+ 00:08:21.200 --> 00:08:23.800
342
+ 2), which works out to be (5/3, 1/3).
343
+
344
+ 00:08:27.280 --> 00:08:30.910
345
+ So that, in a nutshell, is how to translate the description of
346
+
347
+ 00:08:30.910 --> 00:08:34.540
348
+ individual vectors back and forth between coordinate systems.
349
+
350
+ 00:08:35.400 --> 00:08:39.647
351
+ The matrix whose columns represent Jennifer's basis vectors but written in
352
+
353
+ 00:08:39.647 --> 00:08:43.780
354
+ our coordinates translates vectors from her language into our language.
355
+
356
+ 00:08:46.460 --> 00:08:48.700
357
+ And the inverse matrix does the opposite.
358
+
359
+ 00:08:50.840 --> 00:08:54.360
360
+ But vectors aren't the only thing that we describe using coordinates.
361
+
362
+ 00:08:55.660 --> 00:08:58.723
363
+ For this next part, it's important that you're all comfortable
364
+
365
+ 00:08:58.723 --> 00:09:01.935
366
+ representing transformations with matrices and that you know how
367
+
368
+ 00:09:01.935 --> 00:09:05.640
369
+ matrix multiplication corresponds to composing successive transformations.
370
+
371
+ 00:09:06.440 --> 00:09:10.780
372
+ Definitely pause and take a look at chapters 3 and 4 if any of that feels uneasy.
373
+
374
+ 00:09:12.840 --> 00:09:17.280
375
+ Consider some linear transformation, like a 90 degree counterclockwise rotation.
376
+
377
+ 00:09:17.940 --> 00:09:21.442
378
+ When you and I represent this with a matrix, we follow
379
+
380
+ 00:09:21.442 --> 00:09:24.620
381
+ where the basis vectors i-hat and j-hat each go.
382
+
383
+ 00:09:25.580 --> 00:09:28.561
384
+ i-hat ends up at the spot with coordinates (0, 1),
385
+
386
+ 00:09:28.561 --> 00:09:31.900
387
+ and j-hat ends up at the spot with coordinates (-1, 0).
388
+
389
+ 00:09:32.240 --> 00:09:35.400
390
+ So those coordinates become the columns of our matrix.
391
+
392
+ 00:09:36.360 --> 00:09:40.252
393
+ But this representation is heavily tied up in our choice of basis vectors,
394
+
395
+ 00:09:40.252 --> 00:09:44.092
396
+ from the fact that we're following i-hat and j-hat in the first place to
397
+
398
+ 00:09:44.092 --> 00:09:48.300
399
+ the fact that we're recording their landing spots in our own coordinate system.
400
+
401
+ 00:09:50.240 --> 00:09:54.440
402
+ How would Jennifer describe this same 90 degree rotation of space?
403
+
404
+ 00:09:59.580 --> 00:10:02.641
405
+ You might be tempted to just translate the columns
406
+
407
+ 00:10:02.641 --> 00:10:05.640
408
+ of our rotation matrix into Jennifer's language.
409
+
410
+ 00:10:06.100 --> 00:10:07.140
411
+ But that's not quite right.
412
+
413
+ 00:10:07.580 --> 00:10:11.460
414
+ Those columns represent where our basis vectors i-hat and j-hat go.
415
+
416
+ 00:10:11.460 --> 00:10:16.014
417
+ But the matrix that Jennifer wants should represent where her basis vectors land,
418
+
419
+ 00:10:16.014 --> 00:10:19.500
420
+ and it needs to describe those landing spots in her language.
421
+
422
+ 00:10:20.780 --> 00:10:22.760
423
+ Here's a common way to think of how this is done.
424
+
425
+ 00:10:23.600 --> 00:10:26.120
426
+ Start with any vector written in Jennifer's language.
427
+
428
+ 00:10:27.060 --> 00:10:30.968
429
+ Rather than trying to follow what happens to it in terms of her language,
430
+
431
+ 00:10:30.968 --> 00:10:35.573
432
+ first we're going to translate it into our language using the change of basis matrix,
433
+
434
+ 00:10:35.573 --> 00:10:39.160
435
+ the one whose columns represent her basis vectors in our language.
436
+
437
+ 00:10:39.680 --> 00:10:42.680
438
+ This gives us the same vector but now written in our language.
439
+
440
+ 00:10:43.840 --> 00:10:48.380
441
+ Then apply the transformation matrix to what you get by multiplying it on the left.
442
+
443
+ 00:10:49.220 --> 00:10:52.560
444
+ This tells us where that vector lands, but still in our language.
445
+
446
+ 00:10:53.580 --> 00:10:57.025
447
+ So as a last step, apply the inverse change of basis matrix,
448
+
449
+ 00:10:57.025 --> 00:11:01.274
450
+ multiplied on the left as usual, to get the transformed vector but now in
451
+
452
+ 00:11:01.274 --> 00:11:02.480
453
+ Jennifer's language.
454
+
455
+ 00:11:04.300 --> 00:11:07.980
456
+ Since we could do this with any vector written in her language,
457
+
458
+ 00:11:07.980 --> 00:11:11.543
459
+ first applying the change of basis, then the transformation,
460
+
461
+ 00:11:11.543 --> 00:11:16.333
462
+ then the inverse change of basis, that composition of three matrices gives us the
463
+
464
+ 00:11:16.333 --> 00:11:19.020
465
+ transformation matrix in Jennifer's language.
466
+
467
+ 00:11:19.580 --> 00:11:22.545
468
+ It takes in a vector of her language and spits out the
469
+
470
+ 00:11:22.545 --> 00:11:25.400
471
+ transformed version of that vector in her language.
472
+
473
+ 00:11:26.340 --> 00:11:31.877
474
+ For this specific example, when Jennifer's basis vectors look like (2, 1) and (-1,
475
+
476
+ 00:11:31.877 --> 00:11:36.806
477
+ 1) in our language, and when the transformation is a 90 degree rotation,
478
+
479
+ 00:11:36.806 --> 00:11:42.545
480
+ the product of these three matrices, if you work through it, has columns (1/3, 5/3),
481
+
482
+ 00:11:42.545 --> 00:11:43.760
483
+ and (-2/3, -1/3).
484
+
485
+ 00:11:44.400 --> 00:11:49.565
486
+ So if Jennifer multiplies that matrix by the coordinates of a vector in her system,
487
+
488
+ 00:11:49.565 --> 00:11:55.042
489
+ it will return the 90 degree rotated version of that vector expressed in her coordinate
490
+
491
+ 00:11:55.042 --> 00:11:55.540
492
+ system.
493
+
494
+ 00:11:58.760 --> 00:12:03.689
495
+ In general, whenever you see an expression like A inverse times M times A,
496
+
497
+ 00:12:03.689 --> 00:12:06.620
498
+ it suggests a mathematical sort of empathy.
499
+
500
+ 00:12:07.140 --> 00:12:11.333
501
+ That middle matrix represents a transformation of some kind as you see it,
502
+
503
+ 00:12:11.333 --> 00:12:15.640
504
+ and the outer two matrices represent the empathy, the shift in perspective.
505
+
506
+ 00:12:16.560 --> 00:12:19.357
507
+ And the full matrix product represents that same
508
+
509
+ 00:12:19.357 --> 00:12:21.920
510
+ transformation but as someone else sees it.
511
+
512
+ 00:12:22.700 --> 00:12:26.715
513
+ For those of you wondering why we care about alternate coordinate systems,
514
+
515
+ 00:12:26.715 --> 00:12:31.272
516
+ the next video on eigenvectors and eigenvalues will give a really important example
517
+
518
+ 00:12:31.272 --> 00:12:31.760
519
+ of this.
520
+
3B1B_LinearAlgebra/13 - Change of basis | Chapter 13, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c8e7c5306e594b4b70e013bcab5d550f27861cd089e9bb3d4f632de046d1177e
3
+ size 33998299
3B1B_LinearAlgebra/14 - Eigenvectors and eigenvalues | Chapter 14, Essence of linear algebra.en.vtt ADDED
@@ -0,0 +1,694 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:00.000 --> 00:00:19.920
6
+ [Submit subtitle corrections at criblate.com]
7
+
8
+ 00:00:19.920 --> 00:00:22.757
9
+ Eigenvectors and eigenvalues is one of those topics
10
+
11
+ 00:00:22.757 --> 00:00:25.760
12
+ that a lot of students find particularly unintuitive.
13
+
14
+ 00:00:25.760 --> 00:00:29.382
15
+ Questions like, why are we doing this and what does this actually mean,
16
+
17
+ 00:00:29.382 --> 00:00:33.260
18
+ are too often left just floating away in an unanswered sea of computations.
19
+
20
+ 00:00:33.920 --> 00:00:35.982
21
+ And as I've put out the videos of this series,
22
+
23
+ 00:00:35.982 --> 00:00:39.119
24
+ a lot of you have commented about looking forward to visualizing this
25
+
26
+ 00:00:39.119 --> 00:00:40.060
27
+ topic in particular.
28
+
29
+ 00:00:40.680 --> 00:00:43.324
30
+ I suspect that the reason for this is not so much that
31
+
32
+ 00:00:43.324 --> 00:00:46.360
33
+ eigenthings are particularly complicated or poorly explained.
34
+
35
+ 00:00:46.860 --> 00:00:49.020
36
+ In fact, it's comparatively straightforward, and
37
+
38
+ 00:00:49.020 --> 00:00:51.180
39
+ I think most books do a fine job explaining it.
40
+
41
+ 00:00:51.520 --> 00:00:55.084
42
+ The issue is that it only really makes sense if you have a solid
43
+
44
+ 00:00:55.084 --> 00:00:58.480
45
+ visual understanding for many of the topics that precede it.
46
+
47
+ 00:00:59.060 --> 00:01:02.565
48
+ Most important here is that you know how to think about matrices as
49
+
50
+ 00:01:02.565 --> 00:01:06.331
51
+ linear transformations, but you also need to be comfortable with things
52
+
53
+ 00:01:06.331 --> 00:01:09.940
54
+ like determinants, linear systems of equations, and change of basis.
55
+
56
+ 00:01:10.720 --> 00:01:14.925
57
+ Confusion about eigenstuffs usually has more to do with a shaky foundation in
58
+
59
+ 00:01:14.925 --> 00:01:19.240
60
+ one of these topics than it does with eigenvectors and eigenvalues themselves.
61
+
62
+ 00:01:19.980 --> 00:01:24.840
63
+ To start, consider some linear transformation in two dimensions, like the one shown here.
64
+
65
+ 00:01:25.460 --> 00:01:31.040
66
+ It moves the basis vector i-hat to the coordinates (3, 0), and j-hat to (1, 2).
67
+
68
+ 00:01:31.780 --> 00:01:35.640
69
+ So it's represented with a matrix whose columns are (3, 0), and (1, 2).
70
+
71
+ 00:01:36.600 --> 00:01:39.300
72
+ Focus in on what it does to one particular vector,
73
+
74
+ 00:01:39.300 --> 00:01:44.160
75
+ and think about the span of that vector, the line passing through its origin and its tip.
76
+
77
+ 00:01:44.920 --> 00:01:48.380
78
+ Most vectors are going to get knocked off their span during the transformation.
79
+
80
+ 00:01:48.780 --> 00:01:51.996
81
+ I mean, it would seem pretty coincidental if the place where
82
+
83
+ 00:01:51.996 --> 00:01:55.320
84
+ the vector landed also happened to be somewhere on that line.
85
+
86
+ 00:01:57.400 --> 00:02:00.593
87
+ But some special vectors do remain on their own span,
88
+
89
+ 00:02:00.593 --> 00:02:05.474
90
+ meaning the effect that the matrix has on such a vector is just to stretch it or
91
+
92
+ 00:02:05.474 --> 00:02:07.040
93
+ squish it, like a scalar.
94
+
95
+ 00:02:09.460 --> 00:02:14.100
96
+ For this specific example, the basis vector i-hat is one such special vector.
97
+
98
+ 00:02:14.640 --> 00:02:19.348
99
+ The span of i-hat is the x-axis, and from the first column of the matrix,
100
+
101
+ 00:02:19.348 --> 00:02:24.120
102
+ we can see that i-hat moves over to 3 times itself, still on that x-axis.
103
+
104
+ 00:02:26.320 --> 00:02:29.970
105
+ What's more, because of the way linear transformations work,
106
+
107
+ 00:02:29.970 --> 00:02:34.351
108
+ any other vector on the x-axis is also just stretched by a factor of 3,
109
+
110
+ 00:02:34.351 --> 00:02:36.480
111
+ and hence remains on its own span.
112
+
113
+ 00:02:38.500 --> 00:02:41.211
114
+ A slightly sneakier vector that remains on its
115
+
116
+ 00:02:41.211 --> 00:02:44.040
117
+ own span during this transformation is (-1, 1).
118
+
119
+ 00:02:44.660 --> 00:02:47.140
120
+ It ends up getting stretched by a factor of 2.
121
+
122
+ 00:02:49.000 --> 00:02:53.550
123
+ And again, linearity is going to imply that any other vector on the diagonal
124
+
125
+ 00:02:53.550 --> 00:02:58.220
126
+ line spanned by this guy is just going to get stretched out by a factor of 2.
127
+
128
+ 00:02:59.820 --> 00:03:02.525
129
+ And for this transformation, those are all the vectors
130
+
131
+ 00:03:02.525 --> 00:03:05.180
132
+ with this special property of staying on their span.
133
+
134
+ 00:03:05.620 --> 00:03:08.575
135
+ Those on the x-axis getting stretched out by a factor of 3,
136
+
137
+ 00:03:08.575 --> 00:03:11.980
138
+ and those on this diagonal line getting stretched by a factor of 2.
139
+
140
+ 00:03:12.760 --> 00:03:16.370
141
+ Any other vector is going to get rotated somewhat during the transformation,
142
+
143
+ 00:03:16.370 --> 00:03:18.080
144
+ knocked off the line that it spans.
145
+
146
+ 00:03:22.520 --> 00:03:27.306
147
+ As you might have guessed by now, these special vectors are called the eigenvectors of
148
+
149
+ 00:03:27.306 --> 00:03:31.814
150
+ the transformation, and each eigenvector has associated with it what's called an
151
+
152
+ 00:03:31.814 --> 00:03:36.490
153
+ eigenvalue, which is just the factor by which it's stretched or squished during the
154
+
155
+ 00:03:36.490 --> 00:03:37.380
156
+ transformation.
157
+
158
+ 00:03:40.280 --> 00:03:43.355
159
+ Of course, there's nothing special about stretching versus squishing,
160
+
161
+ 00:03:43.355 --> 00:03:45.940
162
+ or the fact that these eigenvalues happen to be positive.
163
+
164
+ 00:03:46.380 --> 00:03:50.781
165
+ In another example, you could have an eigenvector with eigenvalue -1/2,
166
+
167
+ 00:03:50.781 --> 00:03:55.120
168
+ meaning that the vector gets flipped and squished by a factor of 1/2.
169
+
170
+ 00:03:56.980 --> 00:03:59.950
171
+ But the important part here is that it stays on the line
172
+
173
+ 00:03:59.950 --> 00:04:02.760
174
+ that it spans out without getting rotated off of it.
175
+
176
+ 00:04:04.460 --> 00:04:07.704
177
+ For a glimpse of why this might be a useful thing to think about,
178
+
179
+ 00:04:07.704 --> 00:04:09.800
180
+ consider some three-dimensional rotation.
181
+
182
+ 00:04:11.660 --> 00:04:14.917
183
+ If you can find an eigenvector for that rotation,
184
+
185
+ 00:04:14.917 --> 00:04:20.500
186
+ a vector that remains on its own span, what you have found is the axis of rotation.
187
+
188
+ 00:04:22.600 --> 00:04:26.528
189
+ And it's much easier to think about a 3D rotation in terms of some
190
+
191
+ 00:04:26.528 --> 00:04:29.741
192
+ axis of rotation and an angle by which it's rotating,
193
+
194
+ 00:04:29.741 --> 00:04:34.740
195
+ rather than thinking about the full 3x3 matrix associated with that transformation.
196
+
197
+ 00:04:37.000 --> 00:04:40.747
198
+ In this case, by the way, the corresponding eigenvalue would have to be 1,
199
+
200
+ 00:04:40.747 --> 00:04:43.278
201
+ since rotations never stretch or squish anything,
202
+
203
+ 00:04:43.278 --> 00:04:45.860
204
+ so the length of the vector would remain the same.
205
+
206
+ 00:04:48.080 --> 00:04:50.020
207
+ This pattern shows up a lot in linear algebra.
208
+
209
+ 00:04:50.440 --> 00:04:53.201
210
+ With any linear transformation described by a matrix,
211
+
212
+ 00:04:53.201 --> 00:04:57.681
213
+ you could understand what it's doing by reading off the columns of this matrix as the
214
+
215
+ 00:04:57.681 --> 00:04:59.400
216
+ landing spots for basis vectors.
217
+
218
+ 00:05:00.020 --> 00:05:03.544
219
+ But often, a better way to get at the heart of what the linear
220
+
221
+ 00:05:03.544 --> 00:05:08.262
222
+ transformation actually does, less dependent on your particular coordinate system,
223
+
224
+ 00:05:08.262 --> 00:05:10.820
225
+ is to find the eigenvectors and eigenvalues.
226
+
227
+ 00:05:15.460 --> 00:05:18.946
228
+ I won't cover the full details on methods for computing eigenvectors
229
+
230
+ 00:05:18.946 --> 00:05:22.124
231
+ and eigenvalues here, but I'll try to give an overview of the
232
+
233
+ 00:05:22.124 --> 00:05:26.020
234
+ computational ideas that are most important for a conceptual understanding.
235
+
236
+ 00:05:27.180 --> 00:05:30.480
237
+ Symbolically, here's what the idea of an eigenvector looks like.
238
+
239
+ 00:05:31.040 --> 00:05:35.866
240
+ A is the matrix representing some transformation, with v as the eigenvector,
241
+
242
+ 00:05:35.866 --> 00:05:39.740
243
+ and lambda is a number, namely the corresponding eigenvalue.
244
+
245
+ 00:05:40.680 --> 00:05:45.230
246
+ What this expression is saying is that the matrix-vector product, A times v,
247
+
248
+ 00:05:45.230 --> 00:05:49.900
249
+ gives the same result as just scaling the eigenvector v by some value lambda.
250
+
251
+ 00:05:51.000 --> 00:05:55.676
252
+ So finding the eigenvectors and their eigenvalues of a matrix A comes down
253
+
254
+ 00:05:55.676 --> 00:06:00.100
255
+ to finding the values of v and lambda that make this expression true.
256
+
257
+ 00:06:01.920 --> 00:06:06.008
258
+ It's a little awkward to work with at first, because that left-hand side represents
259
+
260
+ 00:06:06.008 --> 00:06:09.752
261
+ matrix-vector multiplication, but the right-hand side here is scalar-vector
262
+
263
+ 00:06:09.752 --> 00:06:10.540
264
+ multiplication.
265
+
266
+ 00:06:11.120 --> 00:06:15.354
267
+ So let's start by rewriting that right-hand side as some kind of matrix-vector
268
+
269
+ 00:06:15.354 --> 00:06:20.023
270
+ multiplication, using a matrix which has the effect of scaling any vector by a factor
271
+
272
+ 00:06:20.023 --> 00:06:20.620
273
+ of lambda.
274
+
275
+ 00:06:21.680 --> 00:06:26.121
276
+ The columns of such a matrix will represent what happens to each basis vector,
277
+
278
+ 00:06:26.121 --> 00:06:29.196
279
+ and each basis vector is simply multiplied by lambda,
280
+
281
+ 00:06:29.196 --> 00:06:34.320
282
+ so this matrix will have the number lambda down the diagonal, with zeros everywhere else.
283
+
284
+ 00:06:36.180 --> 00:06:40.434
285
+ The common way to write this guy is to factor that lambda out and write it
286
+
287
+ 00:06:40.434 --> 00:06:44.860
288
+ as lambda times I, where I is the identity matrix with 1s down the diagonal.
289
+
290
+ 00:06:45.860 --> 00:06:48.736
291
+ With both sides looking like matrix-vector multiplication,
292
+
293
+ 00:06:48.736 --> 00:06:51.860
294
+ we can subtract off that right-hand side and factor out the v.
295
+
296
+ 00:06:54.160 --> 00:06:58.905
297
+ So what we now have is a new matrix, A minus lambda times the identity,
298
+
299
+ 00:06:58.905 --> 00:07:04.920
300
+ and we're looking for a vector v such that this new matrix times v gives the zero vector.
301
+
302
+ 00:07:06.380 --> 00:07:11.100
303
+ Now, this will always be true if v itself is the zero vector, but that's boring.
304
+
305
+ 00:07:11.340 --> 00:07:13.640
306
+ What we want is a non-zero eigenvector.
307
+
308
+ 00:07:14.420 --> 00:07:18.876
309
+ And if you watch chapter 5 and 6, you'll know that the only way it's possible
310
+
311
+ 00:07:18.876 --> 00:07:23.274
312
+ for the product of a matrix with a non-zero vector to become zero is if the
313
+
314
+ 00:07:23.274 --> 00:07:28.020
315
+ transformation associated with that matrix squishes space into a lower dimension.
316
+
317
+ 00:07:29.300 --> 00:07:34.220
318
+ And that squishification corresponds to a zero determinant for the matrix.
319
+
320
+ 00:07:35.480 --> 00:07:40.014
321
+ To be concrete, let's say your matrix A has columns (2, 1) and (2, 3),
322
+
323
+ 00:07:40.014 --> 00:07:45.520
324
+ and think about subtracting off a variable amount, lambda, from each diagonal entry.
325
+
326
+ 00:07:46.480 --> 00:07:50.280
327
+ Now imagine tweaking lambda, turning a knob to change its value.
328
+
329
+ 00:07:50.940 --> 00:07:54.480
330
+ As that value of lambda changes, the matrix itself changes,
331
+
332
+ 00:07:54.480 --> 00:07:57.240
333
+ and so the determinant of the matrix changes.
334
+
335
+ 00:07:58.220 --> 00:08:02.906
336
+ The goal here is to find a value of lambda that will make this determinant zero,
337
+
338
+ 00:08:02.906 --> 00:08:07.240
339
+ meaning the tweaked transformation squishes space into a lower dimension.
340
+
341
+ 00:08:08.160 --> 00:08:11.160
342
+ In this case, the sweet spot comes when lambda equals 1.
343
+
344
+ 00:08:12.180 --> 00:08:16.120
345
+ Of course, if we had chosen some other matrix, the eigenvalue might not necessarily be 1.
346
+
347
+ 00:08:16.240 --> 00:08:18.600
348
+ The sweet spot might be hit at some other value of lambda.
349
+
350
+ 00:08:20.080 --> 00:08:22.960
351
+ So this is kind of a lot, but let's unravel what this is saying.
352
+
353
+ 00:08:22.960 --> 00:08:26.260
354
+ When lambda equals 1, the matrix A minus lambda
355
+
356
+ 00:08:26.260 --> 00:08:29.560
357
+ times the identity squishes space onto a line.
358
+
359
+ 00:08:30.440 --> 00:08:34.429
360
+ That means there's a non-zero vector v such that A minus
361
+
362
+ 00:08:34.429 --> 00:08:38.560
363
+ lambda times the identity times v equals the zero vector.
364
+
365
+ 00:08:40.480 --> 00:08:45.955
366
+ And remember, the reason we care about that is because it means A times v
367
+
368
+ 00:08:45.955 --> 00:08:51.505
369
+ equals lambda times v, which you can read off as saying that the vector v
370
+
371
+ 00:08:51.505 --> 00:08:57.280
372
+ is an eigenvector of A, staying on its own span during the transformation A.
373
+
374
+ 00:08:58.320 --> 00:09:01.317
375
+ In this example, the corresponding eigenvalue is 1,
376
+
377
+ 00:09:01.317 --> 00:09:04.020
378
+ so v would actually just stay fixed in place.
379
+
380
+ 00:09:06.220 --> 00:09:09.500
381
+ Pause and ponder if you need to make sure that that line of reasoning feels good.
382
+
383
+ 00:09:13.380 --> 00:09:15.640
384
+ This is the kind of thing I mentioned in the introduction.
385
+
386
+ 00:09:16.220 --> 00:09:19.473
387
+ If you didn't have a solid grasp of determinants and why they
388
+
389
+ 00:09:19.473 --> 00:09:22.940
390
+ relate to linear systems of equations having non-zero solutions,
391
+
392
+ 00:09:22.940 --> 00:09:26.300
393
+ an expression like this would feel completely out of the blue.
394
+
395
+ 00:09:28.320 --> 00:09:31.782
396
+ To see this in action, let's revisit the example from the start,
397
+
398
+ 00:09:31.782 --> 00:09:34.540
399
+ with a matrix whose columns are (3, 0) and (1, 2).
400
+
401
+ 00:09:35.350 --> 00:09:39.443
402
+ To find if a value lambda is an eigenvalue, subtract it from
403
+
404
+ 00:09:39.443 --> 00:09:43.400
405
+ the diagonals of this matrix and compute the determinant.
406
+
407
+ 00:09:50.580 --> 00:09:54.378
408
+ Doing this, we get a certain quadratic polynomial in lambda,
409
+
410
+ 00:09:54.378 --> 00:09:56.720
411
+ 3 minus lambda times 2 minus lambda.
412
+
413
+ 00:09:57.800 --> 00:10:02.836
414
+ Since lambda can only be an eigenvalue if this determinant happens to be zero,
415
+
416
+ 00:10:02.836 --> 00:10:08.194
417
+ you can conclude that the only possible eigenvalues are lambda equals 2 and lambda
418
+
419
+ 00:10:08.194 --> 00:10:08.840
420
+ equals 3.
421
+
422
+ 00:10:09.640 --> 00:10:14.917
423
+ To figure out what the eigenvectors are that actually have one of these eigenvalues,
424
+
425
+ 00:10:14.917 --> 00:10:19.503
426
+ say lambda equals 2, plug in that value of lambda to the matrix and then
427
+
428
+ 00:10:19.503 --> 00:10:23.900
429
+ solve for which vectors this diagonally altered matrix sends to zero.
430
+
431
+ 00:10:24.940 --> 00:10:28.794
432
+ If you computed this the way you would any other linear system,
433
+
434
+ 00:10:28.794 --> 00:10:34.300
435
+ you'd see that the solutions are all the vectors on the diagonal line spanned by (-1, 1).
436
+
437
+ 00:10:35.220 --> 00:10:39.279
438
+ This corresponds to the fact that the unaltered matrix, (3, 0), (1,
439
+
440
+ 00:10:39.279 --> 00:10:43.460
441
+ 2), has the effect of stretching all those vectors by a factor of 2.
442
+
443
+ 00:10:46.320 --> 00:10:50.200
444
+ Now, a 2D transformation doesn't have to have eigenvectors.
445
+
446
+ 00:10:50.720 --> 00:10:53.400
447
+ For example, consider a rotation by 90 degrees.
448
+
449
+ 00:10:53.660 --> 00:10:58.200
450
+ This doesn't have any eigenvectors since it rotates every vector off of its own span.
451
+
452
+ 00:11:00.800 --> 00:11:04.462
453
+ If you actually try computing the eigenvalues of a rotation like this,
454
+
455
+ 00:11:04.462 --> 00:11:05.560
456
+ notice what happens.
457
+
458
+ 00:11:06.300 --> 00:11:10.140
459
+ Its matrix has columns (0, 1) and (-1, 0).
460
+
461
+ 00:11:11.100 --> 00:11:15.800
462
+ Subtract off lambda from the diagonal elements and look for when the determinant is zero.
463
+
464
+ 00:11:18.140 --> 00:11:21.940
465
+ In this case, you get the polynomial lambda squared plus 1.
466
+
467
+ 00:11:22.680 --> 00:11:27.920
468
+ The only roots of that polynomial are the imaginary numbers, i and -i.
469
+
470
+ 00:11:28.840 --> 00:11:33.600
471
+ The fact that there are no real number solutions indicates that there are no eigenvectors.
472
+
473
+ 00:11:35.540 --> 00:11:39.820
474
+ Another pretty interesting example worth holding in the back of your mind is a shear.
475
+
476
+ 00:11:40.560 --> 00:11:44.355
477
+ This fixes i-hat in place and moves j-hat 1 over,
478
+
479
+ 00:11:44.355 --> 00:11:47.840
480
+ so its matrix has columns (1, 0) and (1, 1).
481
+
482
+ 00:11:48.740 --> 00:11:51.554
483
+ All of the vectors on the x-axis are eigenvectors
484
+
485
+ 00:11:51.554 --> 00:11:54.540
486
+ with eigenvalue 1 since they remain fixed in place.
487
+
488
+ 00:11:55.680 --> 00:11:57.820
489
+ In fact, these are the only eigenvectors.
490
+
491
+ 00:11:58.760 --> 00:12:03.857
492
+ When you subtract off lambda from the diagonals and compute the determinant,
493
+
494
+ 00:12:03.857 --> 00:12:06.540
495
+ what you get is 1 minus lambda squared.
496
+
497
+ 00:12:09.320 --> 00:12:12.860
498
+ And the only root of this expression is lambda equals 1.
499
+
500
+ 00:12:14.560 --> 00:12:17.057
501
+ This lines up with what we see geometrically,
502
+
503
+ 00:12:17.057 --> 00:12:19.720
504
+ that all of the eigenvectors have eigenvalue 1.
505
+
506
+ 00:12:21.080 --> 00:12:24.980
507
+ Keep in mind though, it's also possible to have just one eigenvalue,
508
+
509
+ 00:12:24.980 --> 00:12:28.020
510
+ but with more than just a line full of eigenvectors.
511
+
512
+ 00:12:29.900 --> 00:12:33.180
513
+ A simple example is a matrix that scales everything by 2.
514
+
515
+ 00:12:33.900 --> 00:12:37.135
516
+ The only eigenvalue is 2, but every vector in the
517
+
518
+ 00:12:37.135 --> 00:12:40.700
519
+ plane gets to be an eigenvector with that eigenvalue.
520
+
521
+ 00:12:42.000 --> 00:12:44.613
522
+ Now is another good time to pause and ponder some
523
+
524
+ 00:12:44.613 --> 00:12:46.960
525
+ of this before I move on to the last topic.
526
+
527
+ 00:13:03.540 --> 00:13:06.886
528
+ I want to finish off here with the idea of an eigenbasis,
529
+
530
+ 00:13:06.886 --> 00:13:09.880
531
+ which relies heavily on ideas from the last video.
532
+
533
+ 00:13:11.480 --> 00:13:16.380
534
+ Take a look at what happens if our basis vectors just so happen to be eigenvectors.
535
+
536
+ 00:13:17.120 --> 00:13:22.380
537
+ For example, maybe i-hat is scaled by -1 and j-hat is scaled by 2.
538
+
539
+ 00:13:23.420 --> 00:13:27.085
540
+ Writing their new coordinates as the columns of a matrix,
541
+
542
+ 00:13:27.085 --> 00:13:30.043
543
+ notice that those scalar multiples, -1 and 2,
544
+
545
+ 00:13:30.043 --> 00:13:35.251
546
+ which are the eigenvalues of i-hat and j-hat, sit on the diagonal of our matrix,
547
+
548
+ 00:13:35.251 --> 00:13:37.180
549
+ and every other entry is a 0.
550
+
551
+ 00:13:38.880 --> 00:13:42.494
552
+ Any time a matrix has zeros everywhere other than the diagonal,
553
+
554
+ 00:13:42.494 --> 00:13:45.420
555
+ it's called, reasonably enough, a diagonal matrix.
556
+
557
+ 00:13:45.840 --> 00:13:50.449
558
+ And the way to interpret this is that all the basis vectors are eigenvectors,
559
+
560
+ 00:13:50.449 --> 00:13:54.400
561
+ with the diagonal entries of this matrix being their eigenvalues.
562
+
563
+ 00:13:57.100 --> 00:14:01.060
564
+ There are a lot of things that make diagonal matrices much nicer to work with.
565
+
566
+ 00:14:01.780 --> 00:14:04.979
567
+ One big one is that it's easier to compute what will happen
568
+
569
+ 00:14:04.979 --> 00:14:08.340
570
+ if you multiply this matrix by itself a whole bunch of times.
571
+
572
+ 00:14:09.420 --> 00:14:14.670
573
+ Since all one of these matrices does is scale each basis vector by some eigenvalue,
574
+
575
+ 00:14:14.670 --> 00:14:17.706
576
+ applying that matrix many times, say 100 times,
577
+
578
+ 00:14:17.706 --> 00:14:22.702
579
+ is just going to correspond to scaling each basis vector by the 100th power of
580
+
581
+ 00:14:22.702 --> 00:14:24.600
582
+ the corresponding eigenvalue.
583
+
584
+ 00:14:25.700 --> 00:14:29.680
585
+ In contrast, try computing the 100th power of a non-diagonal matrix.
586
+
587
+ 00:14:29.680 --> 00:14:31.320
588
+ Really, try it for a moment.
589
+
590
+ 00:14:31.740 --> 00:14:32.440
591
+ It's a nightmare.
592
+
593
+ 00:14:36.080 --> 00:14:41.260
594
+ Of course, you'll rarely be so lucky as to have your basis vectors also be eigenvectors.
595
+
596
+ 00:14:42.040 --> 00:14:45.054
597
+ But if your transformation has a lot of eigenvectors,
598
+
599
+ 00:14:45.054 --> 00:14:49.830
600
+ like the one from the start of this video, enough so that you can choose a set that
601
+
602
+ 00:14:49.830 --> 00:14:54.436
603
+ spans the full space, then you could change your coordinate system so that these
604
+
605
+ 00:14:54.436 --> 00:14:56.540
606
+ eigenvectors are your basis vectors.
607
+
608
+ 00:14:57.140 --> 00:15:00.423
609
+ I talked about change of basis last video, but I'll go through a
610
+
611
+ 00:15:00.423 --> 00:15:03.552
612
+ super quick reminder here of how to express a transformation
613
+
614
+ 00:15:03.552 --> 00:15:07.040
615
+ currently written in our coordinate system into a different system.
616
+
617
+ 00:15:08.440 --> 00:15:12.229
618
+ Take the coordinates of the vectors that you want to use as a new basis,
619
+
620
+ 00:15:12.229 --> 00:15:14.703
621
+ which in this case means our two eigenvectors,
622
+
623
+ 00:15:14.703 --> 00:15:19.440
624
+ then make those coordinates the columns of a matrix, known as the change of basis matrix.
625
+
626
+ 00:15:20.180 --> 00:15:22.778
627
+ When you sandwich the original transformation,
628
+
629
+ 00:15:22.778 --> 00:15:26.787
630
+ putting the change of basis matrix on its right and the inverse of the
631
+
632
+ 00:15:26.787 --> 00:15:31.135
633
+ change of basis matrix on its left, the result will be a matrix representing
634
+
635
+ 00:15:31.135 --> 00:15:34.975
636
+ that same transformation, but from the perspective of the new basis
637
+
638
+ 00:15:34.975 --> 00:15:36.500
639
+ vectors coordinate system.
640
+
641
+ 00:15:37.440 --> 00:15:41.851
642
+ The whole point of doing this with eigenvectors is that this new matrix is
643
+
644
+ 00:15:41.851 --> 00:15:46.680
645
+ guaranteed to be diagonal with its corresponding eigenvalues down that diagonal.
646
+
647
+ 00:15:46.860 --> 00:15:50.368
648
+ This is because it represents working in a coordinate system where what
649
+
650
+ 00:15:50.368 --> 00:15:54.320
651
+ happens to the basis vectors is that they get scaled during the transformation.
652
+
653
+ 00:15:55.800 --> 00:15:59.245
654
+ A set of basis vectors which are also eigenvectors is called,
655
+
656
+ 00:15:59.245 --> 00:16:01.560
657
+ again, reasonably enough, an eigenbasis.
658
+
659
+ 00:16:02.340 --> 00:16:07.044
660
+ So if, for example, you needed to compute the 100th power of this matrix,
661
+
662
+ 00:16:07.044 --> 00:16:10.396
663
+ it would be much easier to change to an eigenbasis,
664
+
665
+ 00:16:10.396 --> 00:16:15.680
666
+ compute the 100th power in that system, then convert back to our standard system.
667
+
668
+ 00:16:16.620 --> 00:16:18.320
669
+ You can't do this with all transformations.
670
+
671
+ 00:16:18.320 --> 00:16:22.980
672
+ A shear, for example, doesn't have enough eigenvectors to span the full space.
673
+
674
+ 00:16:23.460 --> 00:16:28.160
675
+ But if you can find an eigenbasis, it makes matrix operations really lovely.
676
+
677
+ 00:16:29.120 --> 00:16:31.894
678
+ For those of you willing to work through a pretty neat puzzle to see
679
+
680
+ 00:16:31.894 --> 00:16:34.546
681
+ what this looks like in action and how it can be used to produce
682
+
683
+ 00:16:34.546 --> 00:16:37.320
684
+ some surprising results, I'll leave up a prompt here on the screen.
685
+
686
+ 00:16:37.600 --> 00:16:40.280
687
+ It takes a bit of work, but I think you'll enjoy it.
688
+
689
+ 00:16:40.840 --> 00:16:45.342
690
+ The next and final video of this series is going to be on abstract vector spaces.
691
+
692
+ 00:16:45.342 --> 00:16:46.120
693
+ See you then!
694
+
3B1B_LinearAlgebra/14 - Eigenvectors and eigenvalues | Chapter 14, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6b3188620c705f5d137b95d380b53fd387652397d1faa8099ec698553df01fb4
3
+ size 54345962
3B1B_LinearAlgebra/15 - A quick trick for computing eigenvalues | Chapter 15, Essence of linear algebra.en.vtt ADDED
@@ -0,0 +1,619 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:00.000 --> 00:00:03.922
6
+ This is a video for anyone who already knows what eigenvalues and eigenvectors are,
7
+
8
+ 00:00:03.922 --> 00:00:07.560
9
+ and who might enjoy a quick way to compute them in the case of 2x2 matrices.
10
+
11
+ 00:00:08.580 --> 00:00:11.953
12
+ If you're unfamiliar with eigenvalues, go ahead and take a look at this video here,
13
+
14
+ 00:00:11.953 --> 00:00:13.700
15
+ which is actually meant to introduce them.
16
+
17
+ 00:00:14.680 --> 00:00:17.344
18
+ You can skip ahead if all you want to do is see the trick,
19
+
20
+ 00:00:17.344 --> 00:00:20.100
21
+ but if possible I'd like you to rediscover it for yourself.
22
+
23
+ 00:00:20.580 --> 00:00:22.380
24
+ So for that, let's lay out a little background.
25
+
26
+ 00:00:23.260 --> 00:00:26.902
27
+ As a quick reminder, if the effect of a linear transformation on a
28
+
29
+ 00:00:26.902 --> 00:00:29.937
30
+ given vector is to scale that vector by some constant,
31
+
32
+ 00:00:29.937 --> 00:00:32.641
33
+ we call it an eigenvector of the transformation,
34
+
35
+ 00:00:32.641 --> 00:00:36.503
36
+ and we call the relevant scaling factor the corresponding eigenvalue,
37
+
38
+ 00:00:36.503 --> 00:00:38.600
39
+ often denoted with the letter lambda.
40
+
41
+ 00:00:39.840 --> 00:00:44.444
42
+ When you write this as an equation, and you rearrange a little bit,
43
+
44
+ 00:00:44.444 --> 00:00:49.530
45
+ what you see is that if the number lambda is an eigenvalue of a matrix A,
46
+
47
+ 00:00:49.530 --> 00:00:55.165
48
+ then the matrix A minus lambda times the identity must send some non-zero vector,
49
+
50
+ 00:00:55.165 --> 00:00:59.151
51
+ namely the corresponding eigenvector, to the zero vector,
52
+
53
+ 00:00:59.151 --> 00:01:04.580
54
+ which in turn means that the determinant of this modified matrix must be zero.
55
+
56
+ 00:01:06.120 --> 00:01:08.766
57
+ Okay, that's all a little bit of a mouthful to say, but again,
58
+
59
+ 00:01:08.766 --> 00:01:11.540
60
+ I'm assuming that all of this is review for any of you watching.
61
+
62
+ 00:01:12.820 --> 00:01:17.241
63
+ So, the usual way to compute eigenvalues, how I used to do it and how I believe
64
+
65
+ 00:01:17.241 --> 00:01:21.439
66
+ most students are taught to carry it out, is to subtract the unknown value
67
+
68
+ 00:01:21.439 --> 00:01:25.860
69
+ lambda off the diagonals, and then solve for the determinant is equal to zero.
70
+
71
+ 00:01:27.760 --> 00:01:32.187
72
+ Doing this always involves a few extra steps to expand out and simplify to get a clean
73
+
74
+ 00:01:32.187 --> 00:01:36.460
75
+ quadratic polynomial, what's known as the characteristic polynomial of the matrix.
76
+
77
+ 00:01:37.360 --> 00:01:39.873
78
+ The eigenvalues are the roots of this polynomial,
79
+
80
+ 00:01:39.873 --> 00:01:42.796
81
+ so to find them you have to apply the quadratic formula,
82
+
83
+ 00:01:42.796 --> 00:01:46.540
84
+ which itself typically requires one or two more steps of simplification.
85
+
86
+ 00:01:47.760 --> 00:01:51.633
87
+ Honestly, the process isn't terrible, but at least for two by two matrices,
88
+
89
+ 00:01:51.633 --> 00:01:54.680
90
+ there is a much more direct way you can get at the answer.
91
+
92
+ 00:01:55.400 --> 00:01:57.819
93
+ And if you want to rediscover this trick, there's only three
94
+
95
+ 00:01:57.819 --> 00:02:00.400
96
+ relevant facts you need to know, each of which is worth knowing
97
+
98
+ 00:02:00.400 --> 00:02:02.900
99
+ in its own right and can help you with other problem solving.
100
+
101
+ 00:02:03.820 --> 00:02:08.592
102
+ Number one, the trace of a matrix, which is the sum of these two diagonal entries,
103
+
104
+ 00:02:08.592 --> 00:02:10.920
105
+ is equal to the sum of the eigenvalues.
106
+
107
+ 00:02:11.700 --> 00:02:14.673
108
+ Or, another way to phrase it, more useful for our purposes,
109
+
110
+ 00:02:14.673 --> 00:02:18.553
111
+ is that the mean of the two eigenvalues is the same as the mean of these two
112
+
113
+ 00:02:18.553 --> 00:02:19.460
114
+ diagonal entries.
115
+
116
+ 00:02:21.000 --> 00:02:25.579
117
+ Number two, the determinant of a matrix, our usual ad-bc formula,
118
+
119
+ 00:02:25.579 --> 00:02:28.960
120
+ is equal to the product of the two eigenvalues.
121
+
122
+ 00:02:30.060 --> 00:02:33.927
123
+ And this should kind of make sense if you understand that eigenvalues describe
124
+
125
+ 00:02:33.927 --> 00:02:37.100
126
+ how much an operator stretches space in a particular direction,
127
+
128
+ 00:02:37.100 --> 00:02:41.165
129
+ and that the determinant describes how much an operator scales areas, or volumes,
130
+
131
+ 00:02:41.165 --> 00:02:41.760
132
+ as a whole.
133
+
134
+ 00:02:42.800 --> 00:02:45.937
135
+ Now before getting to the third fact, notice how you can essentially read
136
+
137
+ 00:02:45.937 --> 00:02:49.160
138
+ these first two values out of the matrix without really writing much down.
139
+
140
+ 00:02:49.760 --> 00:02:51.320
141
+ Take this matrix here as an example.
142
+
143
+ 00:02:51.820 --> 00:02:54.487
144
+ Straight away, you can know that the mean of the
145
+
146
+ 00:02:54.487 --> 00:02:57.820
147
+ eigenvalues is the same as the mean of 8 and 6, which is 7.
148
+
149
+ 00:02:59.580 --> 00:03:03.195
150
+ Likewise, most linear algebra students are pretty well practiced at
151
+
152
+ 00:03:03.195 --> 00:03:07.080
153
+ finding the determinant, which in this case works out to be 48 minus 8.
154
+
155
+ 00:03:08.240 --> 00:03:11.700
156
+ So right away, you know that the product of the two eigenvalues is 40.
157
+
158
+ 00:03:12.780 --> 00:03:16.639
159
+ Now take a moment to see if you can derive what will be our third relevant fact,
160
+
161
+ 00:03:16.639 --> 00:03:19.437
162
+ which is how you can quickly recover two numbers when you
163
+
164
+ 00:03:19.437 --> 00:03:21.560
165
+ know their mean and you know their product.
166
+
167
+ 00:03:22.460 --> 00:03:23.720
168
+ Here, let's focus on this example.
169
+
170
+ 00:03:24.200 --> 00:03:27.933
171
+ You know that the two values are evenly spaced around the number 7,
172
+
173
+ 00:03:27.933 --> 00:03:32.780
174
+ so they look like 7 plus or minus something, let's call that something d for distance.
175
+
176
+ 00:03:33.560 --> 00:03:36.380
177
+ You also know that the product of these two numbers is 40.
178
+
179
+ 00:03:38.600 --> 00:03:41.670
180
+ Now to find d, notice that this product expands really nicely,
181
+
182
+ 00:03:41.670 --> 00:03:43.700
183
+ it works out as a difference of squares.
184
+
185
+ 00:03:44.560 --> 00:03:46.860
186
+ So from there, you can find d.
187
+
188
+ 00:03:48.200 --> 00:03:53.400
189
+ d squared is 7 squared minus 40, or 9, which means that d itself is 3.
190
+
191
+ 00:03:56.380 --> 00:04:01.100
192
+ In other words, the two values for this very specific example work out to be 4 and 10.
193
+
194
+ 00:04:01.680 --> 00:04:05.562
195
+ But our goal is a quick trick, and you wouldn't want to think through this each time,
196
+
197
+ 00:04:05.562 --> 00:04:08.120
198
+ so let's wrap up what we just did in a general formula.
199
+
200
+ 00:04:08.640 --> 00:04:12.508
201
+ For any mean m and product p, the distance squared
202
+
203
+ 00:04:12.508 --> 00:04:15.680
204
+ is always going to be m squared minus p.
205
+
206
+ 00:04:17.560 --> 00:04:21.233
207
+ This gives the third key fact, which is that when two numbers
208
+
209
+ 00:04:21.233 --> 00:04:25.027
210
+ have a mean m and a product p, you can write those two numbers
211
+
212
+ 00:04:25.027 --> 00:04:28.460
213
+ as m plus or minus the square root of m squared minus p.
214
+
215
+ 00:04:30.100 --> 00:04:33.373
216
+ This is decently fast to re-derive on the fly if you ever forget it,
217
+
218
+ 00:04:33.373 --> 00:04:37.080
219
+ and it's essentially just a rephrasing of the difference of squares formula.
220
+
221
+ 00:04:37.860 --> 00:04:41.220
222
+ But even still, it's a fact that's worth memorizing so it's at the tip of your fingers.
223
+
224
+ 00:04:41.220 --> 00:04:44.173
225
+ In fact, my friend Tim from the channel A Capella Science wrote us a nice quick jingle to
226
+
227
+ 00:04:44.173 --> 00:04:47.160
228
+ make it a little bit more memorable: "m plus or minus squared root of m squared minus p".
229
+
230
+ 00:04:51.900 --> 00:04:57.620
231
+ Let me show you how this works, say for the matrix ((3, 1), (4, 1)).
232
+
233
+ 00:04:58.100 --> 00:04:59.974
234
+ You start by bringing to mind the formula, maybe stating it all in
235
+
236
+ 00:04:59.974 --> 00:05:01.820
237
+ your head ("m plus or minus squared root of m squared minus p").
238
+
239
+ 00:05:06.200 --> 00:05:11.620
240
+ But when you write it down, you fill in the appropriate values for m and p as you go.
241
+
242
+ 00:05:12.340 --> 00:05:17.091
243
+ So in this example, the mean of the eigenvalues is the same as the mean of 3 and 1,
244
+
245
+ 00:05:17.091 --> 00:05:20.639
246
+ which is 2, so the thing you start writing is 2 plus or minus
247
+
248
+ 00:05:20.639 --> 00:05:22.700
249
+ the square root of 2 squared minus.
250
+
251
+ 00:05:23.540 --> 00:05:27.164
252
+ Then the product of the eigenvalues is the determinant,
253
+
254
+ 00:05:27.164 --> 00:05:31.578
255
+ which in this example is 3 times 1 minus 1 times 4, or negative 1,
256
+
257
+ 00:05:31.578 --> 00:05:36.717
258
+ so that's the final thing you fill in, which means the eigenvalues are 2 plus
259
+
260
+ 00:05:36.717 --> 00:05:38.760
261
+ or minus the square root of 5.
262
+
263
+ 00:05:40.300 --> 00:05:43.804
264
+ You might recognize that this is the same matrix I was using at the beginning,
265
+
266
+ 00:05:43.804 --> 00:05:46.500
267
+ but notice how much more directly we can get at the answer.
268
+
269
+ 00:05:48.140 --> 00:05:49.180
270
+ Here, try another one.
271
+
272
+ 00:05:49.440 --> 00:05:54.480
273
+ This time, the mean of the eigenvalues is the same as the mean of 2 and 8, which is 5.
274
+
275
+ 00:05:55.100 --> 00:05:57.116
276
+ So again, you start writing out the formula, but this time writing 5
277
+
278
+ 00:05:57.116 --> 00:05:59.220
279
+ in place of m ("m plus or minus squared root of m squared minus ...").
280
+
281
+ 00:06:02.980 --> 00:06:08.300
282
+ And then the determinant is 2 times 8 minus 7 times 1, or 9 ("... p").
283
+
284
+ 00:06:09.520 --> 00:06:15.333
285
+ So in this example, the eigenvalues look like 5 plus or minus the square root of 16,
286
+
287
+ 00:06:15.333 --> 00:06:18.240
288
+ which simplifies even further as 9 and 1.
289
+
290
+ 00:06:19.420 --> 00:06:21.872
291
+ You see what I mean about how you can basically just start
292
+
293
+ 00:06:21.872 --> 00:06:24.620
294
+ writing down the eigenvalues while you're staring at the matrix?
295
+
296
+ 00:06:25.280 --> 00:06:28.160
297
+ It's typically just the tiniest bit of simplification at the end.
298
+
299
+ 00:06:29.060 --> 00:06:32.368
300
+ Honestly, I've found myself using this trick a lot when I'm sketching quick
301
+
302
+ 00:06:32.368 --> 00:06:35.720
303
+ notes related to linear algebra and want to use small matrices as examples.
304
+
305
+ 00:06:36.180 --> 00:06:38.567
306
+ I've been working on a video about matrix exponents,
307
+
308
+ 00:06:38.567 --> 00:06:41.643
309
+ where eigenvalues pop up a lot, and I realize it's just very handy
310
+
311
+ 00:06:41.643 --> 00:06:44.810
312
+ if students can read out the eigenvalues from small examples without
313
+
314
+ 00:06:44.810 --> 00:06:48.620
315
+ losing the main line of thought by getting bogged down in a different calculation.
316
+
317
+ 00:06:49.740 --> 00:06:53.393
318
+ As another fun example, take a look at this set of three different matrices,
319
+
320
+ 00:06:53.393 --> 00:06:55.460
321
+ which comes up a lot in quantum mechanics.
322
+
323
+ 00:06:55.760 --> 00:06:57.520
324
+ They're known as the Pauli spin matrices.
325
+
326
+ 00:06:58.600 --> 00:07:01.555
327
+ If you know quantum mechanics, you'll know that the eigenvalues of
328
+
329
+ 00:07:01.555 --> 00:07:04.420
330
+ matrices are highly relevant to the physics that they describe.
331
+
332
+ 00:07:05.220 --> 00:07:08.200
333
+ And if you don't know quantum mechanics, let this just be a little glimpse
334
+
335
+ 00:07:08.200 --> 00:07:11.220
336
+ of how these computations are actually very relevant to real applications.
337
+
338
+ 00:07:12.540 --> 00:07:15.880
339
+ The mean of the diagonal entries in all three cases is zero.
340
+
341
+ 00:07:17.560 --> 00:07:20.638
342
+ So the mean of the eigenvalues in all of these cases is zero,
343
+
344
+ 00:07:20.638 --> 00:07:23.060
345
+ which makes our formula look especially simple.
346
+
347
+ 00:07:25.380 --> 00:07:28.800
348
+ What about the products of the eigenvalues, the determinants of these matrices?
349
+
350
+ 00:07:29.700 --> 00:07:32.560
351
+ For the first one, it's 0 minus 1, or negative 1.
352
+
353
+ 00:07:33.200 --> 00:07:35.723
354
+ The second one also looks like 0 minus 1, but it takes
355
+
356
+ 00:07:35.723 --> 00:07:38.200
357
+ a moment more to see because of the complex numbers.
358
+
359
+ 00:07:38.840 --> 00:07:41.360
360
+ And the final one looks like -1 minus 0.
361
+
362
+ 00:07:42.060 --> 00:07:45.920
363
+ So in all cases, the eigenvalues simplify to be plus and minus 1.
364
+
365
+ 00:07:46.720 --> 00:07:49.957
366
+ Although in this case, you really don't need a formula to find two values if
367
+
368
+ 00:07:49.957 --> 00:07:53.280
369
+ you know that they're evenly spaced around 0 and their product is negative 1.
370
+
371
+ 00:07:54.640 --> 00:07:57.555
372
+ If you're curious, in the context of quantum mechanics,
373
+
374
+ 00:07:57.555 --> 00:08:02.113
375
+ these matrices describe observations you might make about a particle's spin in the x,
376
+
377
+ 00:08:02.113 --> 00:08:03.120
378
+ y, or z direction.
379
+
380
+ 00:08:03.560 --> 00:08:08.116
381
+ And the fact that their eigenvalues are plus and minus 1 corresponds with the idea that
382
+
383
+ 00:08:08.116 --> 00:08:12.778
384
+ the values for the spin that you would observe would be either entirely in one direction
385
+
386
+ 00:08:12.778 --> 00:08:17.020
387
+ or entirely in another, as opposed to something continuously ranging in between.
388
+
389
+ 00:08:18.320 --> 00:08:21.766
390
+ Maybe you'd wonder how exactly this works, or why you would use 2x2
391
+
392
+ 00:08:21.766 --> 00:08:25.520
393
+ matrices that have complex numbers to describe spin in three dimensions.
394
+
395
+ 00:08:26.100 --> 00:08:29.760
396
+ Those would be fair questions, just outside the scope of what I want to talk about here.
397
+
398
+ 00:08:30.480 --> 00:08:34.236
399
+ You know, it's funny, I wrote this section because I wanted some case where you
400
+
401
+ 00:08:34.236 --> 00:08:37.564
402
+ have 2x2 matrices that aren't just toy examples or homework problems,
403
+
404
+ 00:08:37.564 --> 00:08:41.700
405
+ ones where they actually come up in practice, and quantum mechanics is great for that.
406
+
407
+ 00:08:41.700 --> 00:08:44.830
408
+ The thing is, after I made it, I realized that the whole
409
+
410
+ 00:08:44.830 --> 00:08:48.240
411
+ example kind of undercuts the point that I'm trying to make.
412
+
413
+ 00:08:48.740 --> 00:08:52.232
414
+ For these specific matrices, when you use the traditional method,
415
+
416
+ 00:08:52.232 --> 00:08:56.100
417
+ the one with characteristic polynomials, it's essentially just as fast.
418
+
419
+ 00:08:56.220 --> 00:08:57.640
420
+ It might actually be faster.
421
+
422
+ 00:08:58.240 --> 00:08:59.400
423
+ I mean, take a look at the first one.
424
+
425
+ 00:08:59.680 --> 00:09:03.998
426
+ The relevant determinant directly gives you a characteristic polynomial of
427
+
428
+ 00:09:03.998 --> 00:09:08.200
429
+ lambda squared minus 1, and clearly that has roots of plus and minus 1.
430
+
431
+ 00:09:08.840 --> 00:09:11.760
432
+ Same answer when you do the second matrix, lambda squared minus 1.
433
+
434
+ 00:09:13.880 --> 00:09:17.235
435
+ And as for the last matrix, forget about doing any computations,
436
+
437
+ 00:09:17.235 --> 00:09:20.276
438
+ traditional or otherwise, it's already a diagonal matrix,
439
+
440
+ 00:09:20.276 --> 00:09:22.740
441
+ so those diagonal entries are the eigenvalues.
442
+
443
+ 00:09:24.300 --> 00:09:26.920
444
+ However, the example is not totally lost to our cause.
445
+
446
+ 00:09:27.380 --> 00:09:30.903
447
+ Where you will actually feel the speedup is in the more general case,
448
+
449
+ 00:09:30.903 --> 00:09:35.192
450
+ where you take a linear combination of these three matrices and then try to compute
451
+
452
+ 00:09:35.192 --> 00:09:36.060
453
+ the eigenvalues.
454
+
455
+ 00:09:36.820 --> 00:09:39.532
456
+ You might write this as a times the first one,
457
+
458
+ 00:09:39.532 --> 00:09:42.420
459
+ plus b times the second, plus c times the third.
460
+
461
+ 00:09:43.020 --> 00:09:46.201
462
+ In quantum mechanics, this would describe spin observations in
463
+
464
+ 00:09:46.201 --> 00:09:49.280
465
+ a general direction of a vector with coordinates (a, b, c).
466
+
467
+ 00:09:50.900 --> 00:09:54.430
468
+ More specifically, you should assume that this vector is normalized,
469
+
470
+ 00:09:54.430 --> 00:09:57.700
471
+ meaning a squared plus b squared plus c squared is equal to 1.
472
+
473
+ 00:09:58.600 --> 00:10:01.405
474
+ When you look at this new matrix, it's immediate to
475
+
476
+ 00:10:01.405 --> 00:10:04.100
477
+ see that the mean of the eigenvalues is still 0.
478
+
479
+ 00:10:04.600 --> 00:10:07.611
480
+ And you might also enjoy pausing for a brief moment to
481
+
482
+ 00:10:07.611 --> 00:10:10.900
483
+ confirm that the product of those eigenvalues is still -1.
484
+
485
+ 00:10:13.260 --> 00:10:15.920
486
+ And then from there, concluding what the eigenvalues must be.
487
+
488
+ 00:10:17.220 --> 00:10:20.238
489
+ And this time, the characteristic polynomial approach would be by
490
+
491
+ 00:10:20.238 --> 00:10:23.580
492
+ comparison a lot more cumbersome, definitely harder to do in your head.
493
+
494
+ 00:10:25.080 --> 00:10:28.043
495
+ To be clear, using the mean product formula is not fundamentally
496
+
497
+ 00:10:28.043 --> 00:10:30.960
498
+ different from finding roots of the characteristic polynomial.
499
+
500
+ 00:10:31.340 --> 00:10:33.440
501
+ I mean, it can't be, they're solving the same problem.
502
+
503
+ 00:10:34.160 --> 00:10:36.400
504
+ One way to think about this actually is that the mean
505
+
506
+ 00:10:36.400 --> 00:10:39.020
507
+ product formula is a nice way to solve quadratics in general.
508
+
509
+ 00:10:39.600 --> 00:10:41.660
510
+ And some viewers of the channel may recognize this.
511
+
512
+ 00:10:42.540 --> 00:10:45.788
513
+ Think about it, when you're trying to find the roots of a quadratic,
514
+
515
+ 00:10:45.788 --> 00:10:49.944
516
+ given the coefficients, that's another situation where you know the sum of two values,
517
+
518
+ 00:10:49.944 --> 00:10:54.100
519
+ and you also know their product, but you're trying to recover the original two values.
520
+
521
+ 00:10:55.560 --> 00:11:00.420
522
+ Specifically, if the polynomial is normalized, so that this leading coefficient is 1,
523
+
524
+ 00:11:00.420 --> 00:11:04.479
525
+ then the mean of the roots will be -1/2 times this linear coefficient,
526
+
527
+ 00:11:04.479 --> 00:11:06.880
528
+ which is -1 times the sum of those roots.
529
+
530
+ 00:11:08.020 --> 00:11:10.180
531
+ With the example on the screen, that makes the mean 5.
532
+
533
+ 00:11:11.980 --> 00:11:15.432
534
+ And the product of the roots is even easier, it's just the constant term,
535
+
536
+ 00:11:15.432 --> 00:11:16.520
537
+ no adjustments needed.
538
+
539
+ 00:11:17.340 --> 00:11:18.764
540
+ So from there, you would apply the mean product formula,
541
+
542
+ 00:11:18.764 --> 00:11:20.900
543
+ and that gives you the roots ("m plus or minus squared root of m squared minus p").
544
+
545
+ 00:11:25.140 --> 00:11:27.772
546
+ And on the one hand, you could think of this as a lighter
547
+
548
+ 00:11:27.772 --> 00:11:30.220
549
+ weight version of the traditional quadratic formula.
550
+
551
+ 00:11:30.960 --> 00:11:34.000
552
+ But the real advantage is not just that it's fewer symbols to memorize,
553
+
554
+ 00:11:34.000 --> 00:11:36.440
555
+ it's that each one of them carries more meaning with it.
556
+
557
+ 00:11:36.940 --> 00:11:40.564
558
+ I mean, the whole point of this eigenvalue trick is that because you can read
559
+
560
+ 00:11:40.564 --> 00:11:43.482
561
+ out the mean and product directly from looking at the matrix,
562
+
563
+ 00:11:43.482 --> 00:11:47.435
564
+ you don't need to go through the intermediate step of setting up the characteristic
565
+
566
+ 00:11:47.435 --> 00:11:48.000
567
+ polynomial.
568
+
569
+ 00:11:48.420 --> 00:11:51.074
570
+ You can jump straight to writing down the roots without ever
571
+
572
+ 00:11:51.074 --> 00:11:53.640
573
+ explicitly thinking about what the polynomial looks like.
574
+
575
+ 00:11:53.840 --> 00:11:56.281
576
+ But to do that, we need a version of the quadratic
577
+
578
+ 00:11:56.281 --> 00:11:58.820
579
+ formula where the terms carry some kind of meaning.
580
+
581
+ 00:12:00.380 --> 00:12:03.413
582
+ I realize this is a very specific trick for a very specific audience,
583
+
584
+ 00:12:03.413 --> 00:12:06.491
585
+ but it's something I wish I knew in college, so if you happen to know
586
+
587
+ 00:12:06.491 --> 00:12:09.700
588
+ any students who might benefit from this, consider sharing it with them.
589
+
590
+ 00:12:10.280 --> 00:12:13.201
591
+ The hope is that it's not just one more thing that you memorize,
592
+
593
+ 00:12:13.201 --> 00:12:16.762
594
+ but that the framing reinforces some other nice facts that are worth knowing,
595
+
596
+ 00:12:16.762 --> 00:12:19.820
597
+ like how the trace and the determinant are related to eigenvalues.
598
+
599
+ 00:12:20.560 --> 00:12:23.455
600
+ If you want to prove those facts, by the way, take a moment to
601
+
602
+ 00:12:23.455 --> 00:12:26.398
603
+ expand out the characteristic polynomial for a general matrix,
604
+
605
+ 00:12:26.398 --> 00:12:29.620
606
+ and then think hard about the meaning of each of these coefficients.
607
+
608
+ 00:12:32.400 --> 00:12:35.139
609
+ Many thanks to Tim for ensuring that this mean product formula will stay stuck in all of
610
+
611
+ 00:12:35.139 --> 00:12:37.940
612
+ our heads for at least a few months: "m plus or minus squared root of m squared minus p".
613
+
614
+ 00:12:41.700 --> 00:12:46.000
615
+ If you don't know about A Cappella Science, please do check it out.
616
+
617
+ 00:12:46.280 --> 00:12:49.580
618
+ The molecular shape of you in particular is one of the greatest things on the internet.
619
+
3B1B_LinearAlgebra/15 - A quick trick for computing eigenvalues | Chapter 15, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bdee28ccb5ad031cd14c4d1d2cffb47361b19514b082ff2ff60beb7892103632
3
+ size 36227733
3B1B_LinearAlgebra/16 - Abstract vector spaces | Chapter 16, Essence of linear algebra.en.vtt ADDED
@@ -0,0 +1,670 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WEBVTT
2
+ Kind: captions
3
+ Language: en
4
+
5
+ 00:00:00.000 --> 00:00:16.880
6
+ [Submit subtitle corrections at criblate.com]
7
+
8
+ 00:00:16.880 --> 00:00:19.340
9
+ I'd like to revisit a deceptively simple question
10
+
11
+ 00:00:19.340 --> 00:00:22.000
12
+ that I asked in the very first video of this series.
13
+
14
+ 00:00:22.700 --> 00:00:23.560
15
+ What are vectors?
16
+
17
+ 00:00:24.480 --> 00:00:27.517
18
+ Is a two-dimensional vector, for example, fundamentally an arrow on
19
+
20
+ 00:00:27.517 --> 00:00:30.600
21
+ a flat plane that we can describe with coordinates for convenience?
22
+
23
+ 00:00:30.860 --> 00:00:34.259
24
+ Or is it fundamentally that pair of real numbers which
25
+
26
+ 00:00:34.259 --> 00:00:37.720
27
+ is just nicely visualized as an arrow on a flat plane?
28
+
29
+ 00:00:38.480 --> 00:00:41.360
30
+ Or are both of these just manifestations of something deeper?
31
+
32
+ 00:00:42.300 --> 00:00:45.420
33
+ On the one hand, defining vectors as primarily being
34
+
35
+ 00:00:45.420 --> 00:00:48.480
36
+ a list of numbers feels clear-cut and unambiguous.
37
+
38
+ 00:00:49.060 --> 00:00:53.387
39
+ It makes things like four-dimensional vectors or 100-dimensional vectors sound like real,
40
+
41
+ 00:00:53.387 --> 00:00:55.720
42
+ concrete ideas that you can actually work with.
43
+
44
+ 00:00:55.720 --> 00:00:59.719
45
+ When otherwise, an idea like four dimensions is just a vague geometric
46
+
47
+ 00:00:59.719 --> 00:01:03.660
48
+ notion that's difficult to describe without waving your hands a bit.
49
+
50
+ 00:01:05.540 --> 00:01:09.198
51
+ But on the other hand, a common sensation for those who actually work with
52
+
53
+ 00:01:09.198 --> 00:01:12.954
54
+ linear algebra, especially as you get more fluent with changing your basis,
55
+
56
+ 00:01:12.954 --> 00:01:16.464
57
+ is that you're dealing with a space that exists independently from the
58
+
59
+ 00:01:16.464 --> 00:01:20.615
60
+ coordinates that you give it, and that coordinates are actually somewhat arbitrary,
61
+
62
+ 00:01:20.615 --> 00:01:23.680
63
+ depending on what you happen to choose as your basis vectors.
64
+
65
+ 00:01:24.480 --> 00:01:27.840
66
+ Core topics in linear algebra, like determinants and eigenvectors,
67
+
68
+ 00:01:27.840 --> 00:01:30.640
69
+ seem indifferent to your choice of coordinate systems.
70
+
71
+ 00:01:31.440 --> 00:01:34.901
72
+ The determinant tells you how much a transformation scales areas,
73
+
74
+ 00:01:34.901 --> 00:01:39.320
75
+ and eigenvectors are the ones that stay on their own span during a transformation.
76
+
77
+ 00:01:40.000 --> 00:01:42.798
78
+ But both of these properties are inherently spatial,
79
+
80
+ 00:01:42.798 --> 00:01:47.156
81
+ and you can freely change your coordinate system without changing the underlying
82
+
83
+ 00:01:47.156 --> 00:01:48.340
84
+ values of either one.
85
+
86
+ 00:01:50.760 --> 00:01:53.921
87
+ But if vectors are not fundamentally lists of real numbers,
88
+
89
+ 00:01:53.921 --> 00:01:57.082
90
+ and if their underlying essence is something more spatial,
91
+
92
+ 00:01:57.082 --> 00:02:00.940
93
+ that just begs the question of what mathematicians mean when they use a
94
+
95
+ 00:02:00.940 --> 00:02:02.440
96
+ word like space or spatial.
97
+
98
+ 00:02:03.400 --> 00:02:06.523
99
+ To build up to where this is going, I'd actually like to spend the
100
+
101
+ 00:02:06.523 --> 00:02:09.788
102
+ bulk of this video talking about something which is neither an arrow
103
+
104
+ 00:02:09.788 --> 00:02:13.100
105
+ nor a list of numbers, but also has vector-ish qualities – functions.
106
+
107
+ 00:02:13.740 --> 00:02:17.880
108
+ You see, there's a sense in which functions are actually just another type of vector.
109
+
110
+ 00:02:19.760 --> 00:02:22.541
111
+ In the same way that you can add two vectors together,
112
+
113
+ 00:02:22.541 --> 00:02:25.940
114
+ there's also a sensible notion for adding two functions, f and g,
115
+
116
+ 00:02:25.940 --> 00:02:27.640
117
+ to get a new function, f plus g.
118
+
119
+ 00:02:28.200 --> 00:02:31.465
120
+ It's one of those things where you kind of already know what it's going to be,
121
+
122
+ 00:02:31.465 --> 00:02:33.140
123
+ but actually phrasing it is a mouthful.
124
+
125
+ 00:02:33.960 --> 00:02:38.345
126
+ The output of this new function at any given input, like negative four,
127
+
128
+ 00:02:38.345 --> 00:02:43.594
129
+ is the sum of the outputs of f and g when you evaluate them each at that same input,
130
+
131
+ 00:02:43.594 --> 00:02:44.520
132
+ negative four.
133
+
134
+ 00:02:45.420 --> 00:02:49.580
135
+ Or more generally, the value of the sum function at any
136
+
137
+ 00:02:49.580 --> 00:02:53.740
138
+ given input x is the sum of the values f(x) plus g(x).
139
+
140
+ 00:03:00.700 --> 00:03:04.226
141
+ This is pretty similar to adding vectors coordinate by coordinate,
142
+
143
+ 00:03:04.226 --> 00:03:08.500
144
+ it's just that there are, in a sense, infinitely many coordinates to deal with.
145
+
146
+ 00:03:11.100 --> 00:03:15.532
147
+ Similarly, there's a sensible notion for scaling a function by a real number,
148
+
149
+ 00:03:15.532 --> 00:03:18.180
150
+ just scale all of the outputs by that number.
151
+
152
+ 00:03:20.240 --> 00:03:23.644
153
+ And again, this is analogous to scaling a vector coordinate by coordinate,
154
+
155
+ 00:03:23.644 --> 00:03:26.220
156
+ it just feels like there's infinitely many coordinates.
157
+
158
+ 00:03:28.900 --> 00:03:33.492
159
+ Now, given that the only thing vectors can really do is get added together or scaled,
160
+
161
+ 00:03:33.492 --> 00:03:37.760
162
+ it feels like we should be able to take the same useful constructs and problem
163
+
164
+ 00:03:37.760 --> 00:03:41.812
165
+ solving techniques of linear algebra that were originally thought about in
166
+
167
+ 00:03:41.812 --> 00:03:45.540
168
+ the context of arrows and space and apply them to functions as well.
169
+
170
+ 00:03:46.540 --> 00:03:51.012
171
+ For example, there's a perfectly reasonable notion of a linear transformation
172
+
173
+ 00:03:51.012 --> 00:03:55.600
174
+ for functions, something that takes in one function and turns it into another.
175
+
176
+ 00:03:59.820 --> 00:04:02.780
177
+ One familiar example comes from calculus, the derivative.
178
+
179
+ 00:04:03.420 --> 00:04:07.140
180
+ It's something which transforms one function into another function.
181
+
182
+ 00:04:08.720 --> 00:04:12.676
183
+ Sometimes in this context you'll hear these called operators instead of transformations,
184
+
185
+ 00:04:12.676 --> 00:04:13.980
186
+ but the meaning is the same.
187
+
188
+ 00:04:16.240 --> 00:04:18.814
189
+ A natural question you might want to ask is what it
190
+
191
+ 00:04:18.814 --> 00:04:21.540
192
+ means for a transformation of functions to be linear.
193
+
194
+ 00:04:22.440 --> 00:04:26.515
195
+ The formal definition of linearity is relatively abstract and symbolically driven
196
+
197
+ 00:04:26.515 --> 00:04:30.440
198
+ compared to the way that I first talked about it in chapter 3 of this series.
199
+
200
+ 00:04:30.440 --> 00:04:33.696
201
+ But the reward of abstractness is that we'll get something
202
+
203
+ 00:04:33.696 --> 00:04:36.840
204
+ general enough to apply to functions as well as arrows.
205
+
206
+ 00:04:39.180 --> 00:04:42.624
207
+ A transformation is linear if it satisfies two properties,
208
+
209
+ 00:04:42.624 --> 00:04:45.000
210
+ commonly called additivity and scaling.
211
+
212
+ 00:04:46.040 --> 00:04:50.472
213
+ Additivity means that if you add two vectors, v and w,
214
+
215
+ 00:04:50.472 --> 00:04:57.531
216
+ then apply a transformation to their sum, you get the same result as if you added the
217
+
218
+ 00:04:57.531 --> 00:05:00.240
219
+ transformed versions of v and w.
220
+
221
+ 00:05:04.520 --> 00:05:09.535
222
+ The scaling property is that when you scale a vector v by some number,
223
+
224
+ 00:05:09.535 --> 00:05:14.335
225
+ then apply the transformation, you get the same ultimate vector as
226
+
227
+ 00:05:14.335 --> 00:05:18.920
228
+ if you scaled the transformed version of v by that same amount.
229
+
230
+ 00:05:21.700 --> 00:05:25.426
231
+ The way you'll often hear this described is that linear transformations
232
+
233
+ 00:05:25.426 --> 00:05:29.100
234
+ preserve the operations of vector addition and scalar multiplication.
235
+
236
+ 00:05:32.200 --> 00:05:36.114
237
+ The idea of gridlines remaining parallel and evenly spaced that I've
238
+
239
+ 00:05:36.114 --> 00:05:39.971
240
+ talked about in past videos is really just an illustration of what
241
+
242
+ 00:05:39.971 --> 00:05:44.000
243
+ these two properties mean in the specific case of points in 2D space.
244
+
245
+ 00:05:44.880 --> 00:05:48.112
246
+ One of the most important consequences of these properties,
247
+
248
+ 00:05:48.112 --> 00:05:50.906
249
+ which makes matrix-vector multiplication possible,
250
+
251
+ 00:05:50.906 --> 00:05:54.631
252
+ is that a linear transformation is completely described by where it
253
+
254
+ 00:05:54.631 --> 00:05:56.000
255
+ takes the basis vectors.
256
+
257
+ 00:05:57.720 --> 00:06:02.539
258
+ Since any vector can be expressed by scaling and adding the basis vectors in some way,
259
+
260
+ 00:06:02.539 --> 00:06:06.854
261
+ finding the transformed version of a vector comes down to scaling and adding
262
+
263
+ 00:06:06.854 --> 00:06:10.440
264
+ the transformed versions of the basis vectors in that same way.
265
+
266
+ 00:06:12.280 --> 00:06:16.780
267
+ As you'll see in just a moment, this is as true for functions as it is for arrows.
268
+
269
+ 00:06:18.360 --> 00:06:22.386
270
+ For example, calculus students are always using the fact that the derivative is
271
+
272
+ 00:06:22.386 --> 00:06:26.820
273
+ additive and has the scaling property, even if they haven't heard it phrased that way.
274
+
275
+ 00:06:28.140 --> 00:06:31.129
276
+ If you add two functions, then take the derivative,
277
+
278
+ 00:06:31.129 --> 00:06:35.173
279
+ it's the same as first taking the derivative of each one separately,
280
+
281
+ 00:06:35.173 --> 00:06:36.580
282
+ then adding the result.
283
+
284
+ 00:06:40.140 --> 00:06:43.255
285
+ Similarly, if you scale a function, then take the derivative,
286
+
287
+ 00:06:43.255 --> 00:06:46.880
288
+ it's the same as first taking the derivative, then scaling the result.
289
+
290
+ 00:06:50.280 --> 00:06:53.308
291
+ To really drill in the parallel, let's see what it might
292
+
293
+ 00:06:53.308 --> 00:06:56.120
294
+ look like to describe the derivative with a matrix.
295
+
296
+ 00:06:56.980 --> 00:07:00.287
297
+ This will be a little tricky, since function spaces have a tendency to be
298
+
299
+ 00:07:00.287 --> 00:07:03.820
300
+ infinite dimensional, but I think this exercise is actually quite satisfying.
301
+
302
+ 00:07:04.840 --> 00:07:09.458
303
+ Let's limit ourselves to polynomials, things like x squared plus 3x plus 5,
304
+
305
+ 00:07:09.458 --> 00:07:11.860
306
+ or 4x to the seventh minus 5x squared.
307
+
308
+ 00:07:12.330 --> 00:07:16.384
309
+ Each of the polynomials in our space will only have finitely many terms,
310
+
311
+ 00:07:16.384 --> 00:07:21.000
312
+ but the full space is going to include polynomials with arbitrarily large degree.
313
+
314
+ 00:07:22.220 --> 00:07:25.545
315
+ The first thing we need to do is give coordinates to this space,
316
+
317
+ 00:07:25.545 --> 00:07:27.260
318
+ which requires choosing a basis.
319
+
320
+ 00:07:28.180 --> 00:07:33.335
321
+ Since polynomials are already written down as the sum of scaled powers of the variable x,
322
+
323
+ 00:07:33.335 --> 00:07:37.680
324
+ it's pretty natural to just choose pure powers of x as the basis function.
325
+
326
+ 00:07:38.280 --> 00:07:43.700
327
+ In other words, our first basis function will be the constant function, b0(x) equals 1.
328
+
329
+ 00:07:44.180 --> 00:07:48.041
330
+ The second basis function will be b1(x) equals x,
331
+
332
+ 00:07:48.041 --> 00:07:53.320
333
+ then b2(x) equals x squared, then b3(x) equals x cubed, and so on.
334
+
335
+ 00:07:53.860 --> 00:07:58.175
336
+ The role that these basis functions serve will be similar to the roles of i-hat,
337
+
338
+ 00:07:58.175 --> 00:08:00.980
339
+ j-hat, and k-hat in the world of vectors as arrows.
340
+
341
+ 00:08:02.120 --> 00:08:05.214
342
+ Since our polynomials can have arbitrarily large degree,
343
+
344
+ 00:08:05.214 --> 00:08:07.480
345
+ this set of basis functions is infinite.
346
+
347
+ 00:08:08.240 --> 00:08:11.777
348
+ But that's okay, it just means that when we treat our polynomials as vectors,
349
+
350
+ 00:08:11.777 --> 00:08:14.120
351
+ they're going to have infinitely many coordinates.
352
+
353
+ 00:08:15.600 --> 00:08:19.768
354
+ A polynomial like x squared plus 3x plus 5, for example,
355
+
356
+ 00:08:19.768 --> 00:08:25.500
357
+ would be described with the coordinates 5, 3, 1, then infinitely many zeros.
358
+
359
+ 00:08:26.100 --> 00:08:30.064
360
+ You'd read this as saying that it's 5 times the first basis function,
361
+
362
+ 00:08:30.064 --> 00:08:34.661
363
+ plus 3 times that second basis function, plus 1 times the third basis function,
364
+
365
+ 00:08:34.661 --> 00:08:39.200
366
+ and then none of the other basis functions should be added from that point on.
367
+
368
+ 00:08:40.620 --> 00:08:47.095
369
+ The polynomial 4x to the seventh minus 5x squared would have the coordinates 0,
370
+
371
+ 00:08:47.095 --> 00:08:52.340
372
+ 0, negative 5, 0, 0, 0, 0, 4, then an infinite string of zeros.
373
+
374
+ 00:08:53.260 --> 00:08:57.797
375
+ In general, since every individual polynomial has only finitely many terms,
376
+
377
+ 00:08:57.797 --> 00:09:03.000
378
+ its coordinates will be some finite string of numbers with an infinite tail of zeros.
379
+
380
+ 00:09:06.900 --> 00:09:10.369
381
+ In this coordinate system, the derivative is described with
382
+
383
+ 00:09:10.369 --> 00:09:13.191
384
+ an infinite matrix that's mostly full of zeros,
385
+
386
+ 00:09:13.191 --> 00:09:17.600
387
+ but which has the positive integers counting down on this offset diagonal.
388
+
389
+ 00:09:18.400 --> 00:09:21.268
390
+ I'll talk about how you could find this matrix in just a moment,
391
+
392
+ 00:09:21.268 --> 00:09:24.360
393
+ but the best way to get a feel for it is to just watch it in action.
394
+
395
+ 00:09:24.970 --> 00:09:31.105
396
+ Take the coordinates representing the polynomial x cubed plus 5x squared plus 4x plus 5,
397
+
398
+ 00:09:31.105 --> 00:09:34.940
399
+ then put those coordinates on the right of the matrix.
400
+
401
+ 00:09:40.410 --> 00:09:45.180
402
+ The only term that contributes to the first coordinate of the result is 1 times 4,
403
+
404
+ 00:09:45.180 --> 00:09:48.380
405
+ which means the constant term in the result will be 4.
406
+
407
+ 00:09:50.100 --> 00:09:54.380
408
+ This corresponds to the fact that the derivative of 4x is the constant 4.
409
+
410
+ 00:09:55.640 --> 00:10:00.659
411
+ The only term contributing to the second coordinate of the matrix-vector product
412
+
413
+ 00:10:00.659 --> 00:10:05.740
414
+ is 2 times 5, which means the coefficient in front of x in the derivative is 10.
415
+
416
+ 00:10:06.500 --> 00:10:09.280
417
+ That one corresponds to the derivative of 5x squared.
418
+
419
+ 00:10:10.780 --> 00:10:13.776
420
+ Similarly, the third coordinate in the matrix-vector
421
+
422
+ 00:10:13.776 --> 00:10:16.080
423
+ product comes down to taking 3 times 1.
424
+
425
+ 00:10:17.660 --> 00:10:21.740
426
+ This one corresponds to the derivative of x cubed being 3x squared.
427
+
428
+ 00:10:23.080 --> 00:10:25.020
429
+ And after that, it'll be nothing but zeros.
430
+
431
+ 00:10:26.880 --> 00:10:29.800
432
+ What makes this possible is that the derivative is linear.
433
+
434
+ 00:10:31.640 --> 00:10:34.248
435
+ And for those of you who like to pause and ponder,
436
+
437
+ 00:10:34.248 --> 00:10:37.639
438
+ you could construct this matrix by taking the derivative of each
439
+
440
+ 00:10:37.639 --> 00:10:41.500
441
+ basis function and putting the coordinates of the results in each column.
442
+
443
+ 00:10:59.780 --> 00:11:03.861
444
+ So, surprisingly, matrix-vector multiplication and taking a derivative,
445
+
446
+ 00:11:03.861 --> 00:11:07.023
447
+ which at first seem like completely different animals,
448
+
449
+ 00:11:07.023 --> 00:11:09.840
450
+ are both just really members of the same family.
451
+
452
+ 00:11:11.220 --> 00:11:14.796
453
+ In fact, most of the concepts I've talked about in this series with
454
+
455
+ 00:11:14.796 --> 00:11:19.280
456
+ respect to vectors as arrows in space, things like the dot product or eigenvectors,
457
+
458
+ 00:11:19.280 --> 00:11:21.789
459
+ have direct analogs in the world of functions,
460
+
461
+ 00:11:21.789 --> 00:11:26.540
462
+ though sometimes they go by different names, things like inner product or eigenfunction.
463
+
464
+ 00:11:28.400 --> 00:11:30.880
465
+ So back to the question of what is a vector.
466
+
467
+ 00:11:31.560 --> 00:11:35.840
468
+ The point I want to make here is that there are lots of vector-ish things in math.
469
+
470
+ 00:11:35.840 --> 00:11:40.481
471
+ As long as you're dealing with a set of objects where there's a reasonable notion of
472
+
473
+ 00:11:40.481 --> 00:11:44.846
474
+ scaling and adding, whether that's a set of arrows in space, lists of numbers,
475
+
476
+ 00:11:44.846 --> 00:11:48.327
477
+ functions, or whatever other crazy thing you choose to define,
478
+
479
+ 00:11:48.327 --> 00:11:51.863
480
+ all of the tools developed in linear algebra regarding vectors,
481
+
482
+ 00:11:51.863 --> 00:11:55.620
483
+ linear transformations and all that stuff, should be able to apply.
484
+
485
+ 00:11:57.480 --> 00:11:59.791
486
+ Take a moment to imagine yourself right now as a
487
+
488
+ 00:11:59.791 --> 00:12:02.440
489
+ mathematician developing the theory of linear algebra.
490
+
491
+ 00:12:02.440 --> 00:12:06.870
492
+ You want all of the definitions and discoveries of your work to apply to all
493
+
494
+ 00:12:06.870 --> 00:12:11.300
495
+ of the vector-ish things in full generality, not just to one specific case.
496
+
497
+ 00:12:13.400 --> 00:12:18.140
498
+ These sets of vector-ish things, like arrows or lists of numbers or functions,
499
+
500
+ 00:12:18.140 --> 00:12:19.720
501
+ are called vector spaces.
502
+
503
+ 00:12:20.580 --> 00:12:23.258
504
+ And what you as the mathematician might want to do is say,
505
+
506
+ 00:12:23.258 --> 00:12:25.936
507
+ hey everyone, I don't want to have to think about all the
508
+
509
+ 00:12:25.936 --> 00:12:29.260
510
+ different types of crazy vector spaces that you all might come up with.
511
+
512
+ 00:12:29.260 --> 00:12:32.324
513
+ So what you do is establish a list of rules that
514
+
515
+ 00:12:32.324 --> 00:12:35.260
516
+ vector addition and scaling have to abide by.
517
+
518
+ 00:12:36.400 --> 00:12:40.149
519
+ These rules are called axioms, and in the modern theory of linear algebra,
520
+
521
+ 00:12:40.149 --> 00:12:43.595
522
+ there are eight axioms that any vector space must satisfy if all of
523
+
524
+ 00:12:43.595 --> 00:12:47.040
525
+ the theory and constructs that we've discovered are going to apply.
526
+
527
+ 00:12:47.700 --> 00:12:51.135
528
+ I'll leave them on the screen here for anyone who wants to pause and ponder,
529
+
530
+ 00:12:51.135 --> 00:12:54.570
531
+ but basically it's just a checklist to make sure that the notions of vector
532
+
533
+ 00:12:54.570 --> 00:12:58.140
534
+ addition and scalar multiplication do the things that you'd expect them to do.
535
+
536
+ 00:12:58.720 --> 00:13:02.515
537
+ These axioms are not so much fundamental rules of nature as they are an
538
+
539
+ 00:13:02.515 --> 00:13:05.883
540
+ interface between you, the mathematician, discovering results,
541
+
542
+ 00:13:05.883 --> 00:13:10.480
543
+ and other people who might want to apply those results to new sorts of vector spaces.
544
+
545
+ 00:13:11.420 --> 00:13:14.926
546
+ If, for example, someone defines some crazy type of vector space,
547
+
548
+ 00:13:14.926 --> 00:13:19.025
549
+ like the set of all pi creatures with some definition of adding and scaling
550
+
551
+ 00:13:19.025 --> 00:13:23.124
552
+ pi creatures, these axioms are like a checklist of things that they need to
553
+
554
+ 00:13:23.124 --> 00:13:27.277
555
+ verify about their definitions before they can start applying the results of
556
+
557
+ 00:13:27.277 --> 00:13:28.140
558
+ linear algebra.
559
+
560
+ 00:13:28.820 --> 00:13:31.532
561
+ And you, as the mathematician, never have to think about
562
+
563
+ 00:13:31.532 --> 00:13:34.340
564
+ all the possible crazy vector spaces people might define.
565
+
566
+ 00:13:34.860 --> 00:13:38.302
567
+ You just have to prove your results in terms of these axioms so
568
+
569
+ 00:13:38.302 --> 00:13:42.563
570
+ anyone whose definitions satisfy those axioms can happily apply your results,
571
+
572
+ 00:13:42.563 --> 00:13:45.240
573
+ even if you never thought about their situation.
574
+
575
+ 00:13:46.520 --> 00:13:50.791
576
+ As a consequence, you'd tend to phrase all of your results pretty abstractly,
577
+
578
+ 00:13:50.791 --> 00:13:53.454
579
+ which is to say, only in terms of these axioms,
580
+
581
+ 00:13:53.454 --> 00:13:58.280
582
+ rather than centering on a specific type of vector, like arrows in space or functions.
583
+
584
+ 00:14:01.860 --> 00:14:05.550
585
+ For example, this is why just about every textbook you'll find will
586
+
587
+ 00:14:05.550 --> 00:14:09.185
588
+ define linear transformations in terms of additivity and scaling,
589
+
590
+ 00:14:09.185 --> 00:14:13.260
591
+ rather than talking about gridlines remaining parallel and evenly spaced.
592
+
593
+ 00:14:13.260 --> 00:14:16.901
594
+ Even though the latter is more intuitive, and at least in my view,
595
+
596
+ 00:14:16.901 --> 00:14:21.260
597
+ more helpful for first-time learners, even if it is specific to one situation.
598
+
599
+ 00:14:22.620 --> 00:14:26.920
600
+ So the mathematician's answer to what are vectors is to just ignore the question.
601
+
602
+ 00:14:27.500 --> 00:14:31.260
603
+ In the modern theory, the form that vectors take doesn't really matter.
604
+
605
+ 00:14:31.860 --> 00:14:36.285
606
+ Arrows, lists of numbers, functions, pi creatures, really, it can be anything,
607
+
608
+ 00:14:36.285 --> 00:14:41.220
609
+ so long as there's some notion of adding and scaling vectors that follows these rules.
610
+
611
+ 00:14:41.860 --> 00:14:44.880
612
+ It's like asking what the number 3 really is.
613
+
614
+ 00:14:45.380 --> 00:14:49.761
615
+ Whenever it comes up concretely, it's in the context of some triplet of things,
616
+
617
+ 00:14:49.761 --> 00:14:54.253
618
+ but in math, it's treated as an abstraction for all possible triplets of things,
619
+
620
+ 00:14:54.253 --> 00:14:58.080
621
+ and lets you reason about all possible triplets using a single idea.
622
+
623
+ 00:14:59.120 --> 00:15:02.224
624
+ Same goes with vectors, which have many embodiments,
625
+
626
+ 00:15:02.224 --> 00:15:07.000
627
+ but math abstracts them all into a single, intangible notion of a vector space.
628
+
629
+ 00:15:08.860 --> 00:15:12.299
630
+ But, as anyone watching this series knows, I think it's better
631
+
632
+ 00:15:12.299 --> 00:15:16.182
633
+ to begin reasoning about vectors in a concrete, visualizable setting,
634
+
635
+ 00:15:16.182 --> 00:15:18.900
636
+ like 2D space, with arrows rooted at the origin.
637
+
638
+ 00:15:19.660 --> 00:15:24.381
639
+ But as you learn more linear algebra, know that these tools apply much more generally,
640
+
641
+ 00:15:24.381 --> 00:15:29.102
642
+ and that this is the underlying reason why textbooks and lectures tend to be phrased,
643
+
644
+ 00:15:29.102 --> 00:15:30.090
645
+ well, abstractly.
646
+
647
+ 00:15:31.940 --> 00:15:36.140
648
+ So with that, folks, I think I'll call it an in to this essence of linear algebra series.
649
+
650
+ 00:15:36.140 --> 00:15:39.996
651
+ If you've watched and understood the videos, I really do believe that you
652
+
653
+ 00:15:39.996 --> 00:15:43.800
654
+ have a solid foundation in the underlying intuitions of linear algebra.
655
+
656
+ 00:15:44.640 --> 00:15:47.453
657
+ This is not the same thing as learning the full topic, of course,
658
+
659
+ 00:15:47.453 --> 00:15:50.655
660
+ that's something that can only really come from working through problems,
661
+
662
+ 00:15:50.655 --> 00:15:54.506
663
+ but the learning you do moving forward could be substantially more efficient if you have
664
+
665
+ 00:15:54.506 --> 00:15:56.020
666
+ all the right intuitions in place.
667
+
668
+ 00:15:56.660 --> 00:16:00.000
669
+ So, have fun applying those intuitions, and best of luck with your future learning.
670
+
3B1B_LinearAlgebra/16 - Abstract vector spaces | Chapter 16, Essence of linear algebra.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:21a44db838bfd08c24232646d98fe053e786cba552ddc7f0440e662b6f067e28
3
+ size 40687086
3B1B_LinearAlgebra/ids.tsv ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 01 fNk_zzaMoSs
2
+ 02 k7RM-ot2NWY
3
+ 03 kYB8IZa5AuE
4
+ 04 XkY2DOUCWMU
5
+ 05 rHLEWRxRGiM
6
+ 06 Ip3X9LOh2dk
7
+ 07 uQhTuRlWMxw
8
+ 08 v8VSDg_WQlA
9
+ 09 LyGKycYT2v0
10
+ 10 eu6i7WJeinw
11
+ 11 BaM7OCEm3G0
12
+ 12 jBsC34PxzoM
13
+ 13 P2LTAUO1TdA
14
+ 14 PFDu9oVAE-g
15
+ 15 e50Bj7jn9IQ
16
+ 16 TgKwz5Ikpc8