date stringlengths 10 10 | nb_tokens int64 60 629k | text_size int64 234 1.02M | content stringlengths 234 1.02M |
|---|---|---|---|
2019/11/27 | 801 | 2,876 | <issue_start>username_0: Currently, I'm using a Python library, [StellarGraph](https://github.com/stellargraph/stellargraph), to implement GCN. And I now have a situation where I have graphs with weighted edges. Unfortunately, StellarGraph doesn't support those graphs
I'm looking for an open-source implementation for ... |
2019/11/27 | 706 | 3,136 | <issue_start>username_0: What is the difference between genetic algorithms and evolutionary game theory algorithms?<issue_comment>username_1: A genetic algorithm is typically a single population designed to optimise to a specific task, say minimising the distance on the [travelling salesman problem](https://simple.wiki... |
2019/11/28 | 2,507 | 9,630 | <issue_start>username_0: If we were learning or working in the machine learning field, then we frequently come across the term "probability distribution". I know what probability, conditional probability, and probability distribution/density in math mean, but what is its meaning in machine learning?
Take this example ... |
2019/11/29 | 2,201 | 8,839 | <issue_start>username_0: *When using CNNs for non-image (times series) data prediction, what are some constraints or things to look out for as compared to image data?*
To be more precise, I notice there are different types of layers in a CNN model, as described below, which seem to be particularly designed for image d... |
2019/12/01 | 1,307 | 5,660 | <issue_start>username_0: For an AI to represent the world, it would be good if it could translate human sentences into something more precise.
We know, for example, that mathematics can be built up from set theory. So representing statement in language of set theory might be useful.
For example
>
> All grass is gre... |
2019/12/01 | 1,345 | 5,749 | <issue_start>username_0: Suppose you want to predict the price of some stock. Let's say you use the following features.
```
OpenPrice
HighPrice
LowPrice
ClosePrice
```
Is it useful to create new features like the following ones?
```
BodySize = ClosePrice - OpenPrice
```
or the size of the tail
```
TailUp = ... |
2019/12/02 | 449 | 1,990 | <issue_start>username_0: I am planning to enroll for <NAME>'s Machine Learning course <https://www.coursera.org/learn/machine-learning>. I've no background in math. Is it OK if I start the course and learn math as and when required?<issue_comment>username_1: No pre-requisites required for Andrew Ng ML course. There are... |
2019/12/02 | 615 | 2,250 | <issue_start>username_0: I'm doing a paper for a class on the topic of big problems that are still prevalent in AI, specifically in the area of natural language processing and understanding. From what I understand, the areas:
* Text classification
* Entity recognition
* Translation
* POS tagging
are for the most part... |
2019/12/04 | 432 | 1,615 | <issue_start>username_0: I have heard of ensemble methods, such as XGBoost, for binary or categorical machine learning models. However, does this exist for regression? If so, how are the weights for each model in the process of predictions determined?
I am looking to do this manually, as I was planning on training tw... |
2019/12/05 | 582 | 2,519 | <issue_start>username_0: How can I know what each neuron does in NN?
Consider the [Playground from Tensorflow](https://playground.tensorflow.org/), there are some hidden layers with some neurons in each. Each of them shows a line(horizontal or vertical or ...). Where these shapes come from. I think they are understanda... |
2019/12/07 | 1,206 | 4,821 | <issue_start>username_0: It is my understanding that, in Q-learning, you are trying to mimic the optimal $Q$ function $Q\*$, where $Q\*$ is a measure of the predicted reward received from taking action $a$ at state $s$ so that the reward is maximised.
I understand for this to be properly calculated, you must explore a... |
2019/12/07 | 1,250 | 4,887 | <issue_start>username_0: Now I know this might break some StackExchange rules and I am definitely open for taking the thread down if it does!
I am trying to build an AI that can write it's own book and I have no idea where to start or what are the appropriate algorithms and approaches to go with.
How should I start an... |
2019/12/09 | 1,326 | 5,141 | <issue_start>username_0: What is the difference between game theory and machine learning?
I had gone through the papers [Deep Learning for Predicting
Human Strategic Behavior](https://www.cs.ubc.ca/~jasonhar/GameNet-NIPS-2016.pdf), by <NAME> et al., and [When Machine Learning Meets AI and Game
Theory](http://cs229.st... |
2019/12/09 | 1,534 | 6,015 | <issue_start>username_0: Say I have a CNN with this structure:
* input = 1 image (say, 30x30 RGB pixels)
* first convolution layer = 10 5x5 convolution filters
* second convolution layer = 5 3x3 convolution filters
* one dense layer with 1 output
So a graph of the network will look like this:
 and [TensorFlow.js](https://www.tensorflow.org/js).
What is the difference between artificial intelligence and artificial neural networks?<issue_comment>username_1: From wikipedia:
>
> Artificial neural network... |
2019/12/10 | 1,005 | 4,468 | <issue_start>username_0: Suppose a deep neural network is created using Keras or Tensorflow. Usually, when you want to make a prediction, **the user** would invoke `model.predict`. However, how would the actual AI system proactively invoke their own actions (i.e. without the need for me to call `model.predict`)?<issue_... |
2019/12/10 | 996 | 4,037 | <issue_start>username_0: How does a transformer leverage the GPU to be trained faster than RNNs?
I understand the parameter space of the transformer might be significantly larger than that of the RNN. But why does the transformer structure can leverage multiple GPUs, and why does that accelerate its training?<issue_co... |
2019/12/11 | 1,386 | 5,397 | <issue_start>username_0: Why is dropout favored compared to reducing the number of units in hidden layers for the convolutional networks?
If a large set of units leads to overfitting and dropping out "averages" the response units, why not just suppress units?
I have read different questions and answers on the dropout... |
2019/12/11 | 1,184 | 4,757 | <issue_start>username_0: Say a simple neural network's input is a collection of tags (encoded in some way), and the output is an image that corresponds to those tags. Say this network consists of some dense layers and some reverse (transpose) convolution layers.
What is the disadvantage of this network, that directs p... |
2019/12/11 | 2,048 | 6,848 | <issue_start>username_0: I'd like to ask for any kind of assistance regarding the following problem:
I was given the following training data: 100 numbers, each one is a parameter, they together define a number X(also given).This is one instance,I have 20 000 instances for training.Next, I have 5000 lines given, each c... |
2019/12/11 | 924 | 3,902 | <issue_start>username_0: So I’ve been working on my own little dynamic architecture for a deep neural network (any number of hidden layers with any number of nodes in every layer) and got it solving the XOR problem efficiently. I moved on to trying to see if I could train my network on how to classify a number as being... |
2019/12/13 | 384 | 1,657 | <issue_start>username_0: I need to understand whether it is better to use AI algorithms (ML, DL, etc.) instead of the classic parser (based onto grammars with regular expression and automaton) for the following task: structuring in XML an unstructured plain text.
The text is a legal document, so the structure is well... |
2019/12/13 | 1,618 | 6,407 | <issue_start>username_0: I previously asked a question about [How can an AI freely make decisions?](https://ai.stackexchange.com/q/17025/2444). I got a great answer about how current algorithms lack *agency*.
The first thing I thought of was *reinforcement learning*, since the entire concept is oriented around an agen... |
2019/12/14 | 1,694 | 5,560 | <issue_start>username_0: I'm trying to develop a real-time application that, from the sequence of chalkboard images captured by a webcam, recognizes the lines being draw on it.
It must be able of recognize the lines from the chalkboard background, filter the presence in the image of the teacher, and translate these l... |
2019/12/15 | 879 | 2,318 | <issue_start>username_0: While studying backpropagation in CNNs, I can't understand how can we compute the gradient of max pooling with overlapping regions.
That's also a question from [this](http://courses.cs.tau.ac.il/Caffe_workshop/Bootcamp/pdf_lectures/Lecture%203%20CNN%20-%20backpropagation.pdf) quiz and can be a... |
2019/12/15 | 619 | 2,520 | <issue_start>username_0: What i really want to do, is to predict an integer sequence of (5 numbers with values from 1 to 50) for example based on a big dataset of other 5 numbers sequences with same values range created by the same random number generator. I suppose there is a way to train based on the dataset and the ... |
2019/12/18 | 680 | 2,802 | <issue_start>username_0: I have been exploring edge computation for AI, and I came across multiple libraries or frameworks, which can help to convert the model into a lite format, which is suitable for edge devices.
* [TensorFlow Lite](https://www.tensorflow.org/lite) will help us to convert the TensorFlow model into ... |
2019/12/18 | 1,363 | 4,767 | <issue_start>username_0: Consider an optimization problem that involves a set of tasks $T = \{1,2,3,4,5\}$, where the goal is to find a certain order of these tasks.
I would like to solve this problem with a genetic algorithm, where each chromosome $C = [i, j, k, l, m]$ corresponds to a specific order of these five ta... |
2019/12/19 | 424 | 1,653 | <issue_start>username_0: I found the terms **front-end** and **back-end** in the article (or blog post) [How to Develop a CNN for MNIST Handwritten Digit Classification](https://machinelearningmastery.com/how-to-develop-a-convolutional-neural-network-from-scratch-for-mnist-handwritten-digit-classification/). What do th... |
2019/12/19 | 936 | 2,465 | <issue_start>username_0: How can I prove that all the a-cuts of any fuzzy set A defined on $R^n$
are convex if and only if
$$\mu\_A(\lambda r + (1-\lambda)s) \geq min \{\mu\_A(r), \mu\_A(s)\}$$
such that $r, s \in R^n$, $\lambda \in [0, 1]$ ?
That's a fuzzy question on my assignment. Any idea on how to start with?<i... |
2019/12/21 | 829 | 3,612 | <issue_start>username_0: I keep reading about how LSTMs can't remember the "important parts" of a sequence which is why attention-based mechanisms are required. I was trying to use LSTMs to find people's name format.
For example, "<NAME>" can be seen as first\_name middle\_name last\_name format, which I'll denote as... |
2019/12/21 | 1,115 | 4,896 | <issue_start>username_0: I read an interesting [essay](https://aeon.co/essays/how-close-are-we-to-creating-artificial-intelligence) about how far we are from AGI. There were quite a few solid points that made me re-visit the foundation of AI today. A few interesting concepts arose:
>
> imagine that you require a prog... |
2019/12/23 | 444 | 1,705 | <issue_start>username_0: Assuming we have big $m \times n$ input dataset, with $m \times 1$ output vector. It's a **classification** problem with only two possible values: either $1$ or $0$.
Now, the problem is that almost all elements of the output vector are $0$s with a very few $1$s (i.e. it's a sparse vector), suc... |
2019/12/23 | 553 | 1,656 | <issue_start>username_0: I have trained a multi-class CNN model using [fastai](https://www.fast.ai/). The model splits out probabilites for each of the three classes, which, of course, sum up to 1. The class with highest probability becomes the predicted class.
Is there any way I can convert them into 0 to 1 scale, w... |
2019/12/23 | 1,343 | 5,216 | <issue_start>username_0: I understood that we normalize to **input** features in order to bring them on the same scale so that weights won't be learned in arbitrary fashion and training would be faster.
Then I studied about **batch-normalization** and observed that we can do the normalization for outputs of the **hid... |
2019/12/24 | 1,151 | 4,512 | <issue_start>username_0: I am new to self-supervised learning and it all seems a little magical at the moment.
The only way I can get an intuitive understanding is to assume that, for real-world problems, features are still embedded at a per-object level.
For example, to detect cats in unseen images, my self-supervis... |
2019/12/27 | 641 | 3,139 | <issue_start>username_0: Feature extraction is a concept concerning the translation of raw data into the inputs that a particular machine learning algorithm requires. These derived features from the raw data that are actually relevant to tackle the underlying problem. On the other hand, word embeddings are basically di... |
2019/12/30 | 607 | 2,907 | <issue_start>username_0: Is there an AI technology out there or being developed that can predict human behaviour, given that we as humans are irrational decision-makers?
I'm looking at this from an economic standpoint - the issue with current economic models is that they assume that humans are perfectly rational, but ... |
2019/12/31 | 662 | 3,087 | <issue_start>username_0: I want a microphone to pick up sounds around me (let's say beyond a 3 foot radius), but ignore sounds made at my desk, such as the rustling of paper, clicking a mouse and typing, my hands brushing up on the table, putting a pen down, etc.
How hard would it be for AI to be able to distinguish ... |
2019/12/31 | 1,926 | 8,734 | <issue_start>username_0: Neural networks are incredibly good at learning functions. We know by the universal approximation theorem that, theoretically, they can take the form of almost any function - and in practice, they seem particularly apt at learning the right parameters. However, something we often have to combat... |
2020/01/01 | 546 | 2,256 | <issue_start>username_0: Let's assume an extreme case in which the kernel of the convolution layer takes only values 0 or 1. To capture all possible patterns in input of $C$ number of channels, we need $2^{C\*K\_H\*K\_W}$ filters, where $(K\_H, K\_W)$ is the shape of a kernel. So to process a standard RGB image with 3 ... |
2020/01/02 | 697 | 3,150 | <issue_start>username_0: A LSTM model can be trained to generate text sequences by feeding the first word. After feeding the first word, the model will generate a sequence of words (a sentence). Feed the first word to get the second word, feed the first word + the second word to get the third word, and so on.
However,... |
2020/01/03 | 901 | 4,072 | <issue_start>username_0: I've written a program to analyse a given piece of text from a website and make conclusary classifications as to its validity. The code basically vectorizes the description (taken from the HTML of a given webpage in real time) and takes in a few inputs from that as features to make its decision... |
2020/01/04 | 449 | 1,659 | <issue_start>username_0: How to calculate mean speed in FPS for an object detection model like YOLOv3 or YOLOv3-Tiny? Different object detection models are often presented on charts like this:
[](https://i.stack.imgur.com/hUODt.png)
I am using the Dark... |
2020/01/07 | 953 | 3,680 | <issue_start>username_0: For example, I have a paragraph that I want to classify in a binary manner. But because the inputs have to have a fixed length, I need to ensure that every paragraph is represented by a uniform quantity.
One thing I've done is taken every word in the paragraph, vectorized it using GloVe word2v... |
2020/01/09 | 1,400 | 4,680 | <issue_start>username_0: I have a data set that was split using a fixed random seed and I am going to use 80% of the data for training and the rest for validation.
Here are my GPU and batch size configurations
* use `64 batch size` with `one GTX 1080Ti`
* use `128 batch size` with `two GTX 1080Ti`
* use `256 batch si... |
2020/01/09 | 1,293 | 4,424 | <issue_start>username_0: I'm training a multi-label classifier that's supposed to be tested on underwater images. I'm wondering if feeding the model drawings of a certain class plus real images can affect the results badly. Was there a study on this? Or are there any past experiences anyone could share to help?<issue_c... |
2020/01/11 | 1,171 | 4,887 | <issue_start>username_0: I have images that contain lots of elements. Some I know, some I don't. I want to know if it's ok to only label those I do know. Let's take this image for example. I would label the green stuff and the worm but leave the rest unlabeled.
Is that ok?
Another question I would also like to ask is ... |
2020/01/11 | 541 | 2,371 | <issue_start>username_0: In the step of tuning my neural networks I often encounter a problem that every time I train the exact same network, it gives me different final error due to random initialization of the weights. Sometimes the differences are small and negligible, sometimes they are significant, depending on th... |
2020/01/12 | 687 | 2,668 | <issue_start>username_0: I'm building an RL agent using SARSA and Q-Learning for testing its capabilities.
The environment is a 10x10 grid, where it gets a reward of 1 if he reaches the goal while he takes -1 every time he takes a step out of the grid. So, it can freely move out and every time it takes a step outside... |
2020/01/12 | 623 | 2,710 | <issue_start>username_0: I'm working on semantic segmentation tasks in the medical space using the U-Net. Let's say that I train a U-Net model on medical images with the goal of segmenting out, say, ligaments, from a medical image. If I train that model on images that contain just a single labelled ligament, it will be... |
2020/01/13 | 784 | 3,443 | <issue_start>username_0: We know a lot of common sense about the world. Things like "to buy something you need money".
I wonder how much of this common sense comes about through actual someone explicitly telling you the instructions "You need money to buy things". Which we store in our brains as a sort of rule. As opp... |
2020/01/14 | 1,054 | 4,679 | <issue_start>username_0: I'm still on my first steps in the Data Science field. I played with some DL frameworks, like TensorFlow (pure) and Keras (on top) before, and know a little bit of some "classic machine learning" algorithms like decision trees, k-nearest neighbors, etc.
For example, image classification proble... |
2020/01/14 | 1,840 | 7,363 | <issue_start>username_0: I'm struggling to fully understand the stochastic gradient descent algorithm.
I know that gradient descent allows you to find the local minimum of a function. What I don't know is what exactly that function IS.
More specifically, the algorithm should work by initializing the network with rando... |
2020/01/14 | 988 | 4,402 | <issue_start>username_0: While studying genetic algorithms, I've come across different crossover operations used for binary chromosomes, such as the *1-point crossover*, the *uniform crossover*, etc. These methods usually don't use any "intelligence".
I found methods like the *fitness-based crossover* and *Boltzmann c... |
2020/01/15 | 946 | 4,237 | <issue_start>username_0: I am training a convolutional neural network for object detection. Apart from the learning rate, what are the other hyperparameters that I should tune? And in what order of importance? Besides, I read that doing a grid search for hyperparameters is not the best way to go about training and that... |
2020/01/16 | 1,891 | 6,858 | <issue_start>username_0: I'm working on an [advantage actor-critic (A2C)](https://github.com/KokoMind/Actor-Critic-TF/tree/d062227c1fc610139a6f026fa63bb6069230bdc8) reinforcement learning model, but when I test the model after I trained for 3500 episodes, I start to get almost the same action for all testing episodes. ... |
2020/01/16 | 1,218 | 4,859 | <issue_start>username_0: I having trouble finding some starting points for solving an occupancy problem which seems like a good candidate for ai.
Assume the following situation:
In a company I have n cars and m employees. Not every employee can drive any car (f. e. a special driving license is required). A car can onl... |
2020/01/17 | 325 | 1,390 | <issue_start>username_0: The selection of experimental data includes a set of vectors of different dimensions. The input is a 3-dimensional vector, and the output is a 12-dimensional vector. The sample size is 120 pairs of input 3-dimensional and output 12-dimensional vectors.
Is it possible to train such a neural ne... |
2020/01/17 | 489 | 2,065 | <issue_start>username_0: Let us assume we have a general **AI that can improve itself** and is at least as intelligent as humans.
It has wide access to technical systems including the internet, and it can communicate with humans.
The AI could become **malicious**.
**Can we just switch off a rouge AI?**<issue_commen... |
2020/01/18 | 1,859 | 6,500 | <issue_start>username_0: I was wondering if it's possible to get the inverse of a neural network. If we view a NN as a function, can we obtain its inverse?
I tried to build a simple MNIST architecture, with the input of (784,) and output of (10,), train it to reach good accuracy, and then inverse the predicted value t... |
2020/01/19 | 1,386 | 5,115 | <issue_start>username_0: In my code, I usually use the mean squared error (MSE), but the TensorFlow tutorials always use the categorical cross-entropy (CCE). Is the CCE loss function better than MSE? Or is it better only in certain cases?<issue_comment>username_1: As a rule of thumb, [mean squared error (MSE)](https://... |
2020/01/20 | 959 | 3,673 | <issue_start>username_0: I am learning to use tensorflow.js. I am also using the tfvis library to print information about the neural net to the web browser. When I create a create a dense neural net with a layer with 5 neurons and another layer with 2 neurons, each layer has a bias vector of length 5 and 2 respectively... |
2020/01/21 | 769 | 3,157 | <issue_start>username_0: I'd like to better understand temporal-difference learning. In particular, I'm wondering if it is prudent to think about TD($\lambda$) as a type of "truncated" Monte Carlo learning?<issue_comment>username_1: TD($\lambda$) can be thought of as a combination of TD and MC learning, so as to avoid ... |
2020/01/22 | 812 | 3,100 | <issue_start>username_0: I am trying to understand how deep Q learning (DQN) works. To my current understanding, each $Q(s, a)$ functions is estimated to be a function of a feature vector of its state $\phi$(s) and the weight of the network $\theta$.
The loss function to minimise is $||\delta\_{t+1}||^2$ where $\delta... |
2020/01/22 | 1,042 | 3,648 | <issue_start>username_0: In deep learning, is it possible to use discontinuous activation functions (e.g. one with jump discontinuity)?
(My guess: for example, ReLU is non-differentiable at a single point, but it still has a well-defined derivative. If an activation function has a jump discontinuity, then its derivati... |
2020/01/23 | 435 | 1,724 | <issue_start>username_0: During weeks and months of your work, many things may change, for example :
* You may modify the loss function
* Your training or validation datasets may change
* You modify data augmentation
Which tools or processes do you use to track modifications you have made and how did they affected th... |
2020/01/23 | 454 | 1,849 | <issue_start>username_0: Given a pre-trained CNN model, I extract feature vector of images in reference and query dataset with several thousands of elements.
I would like to apply some augmentation techniques to reduce the feature vector dimension to speed up cosine similarity/euclidean distance matrix calculation.
I... |
2020/01/23 | 673 | 2,910 | <issue_start>username_0: It's not clear to me whether or not someone whose work aims to **improve an NLP system** may be called a "*Computational Linguist*" even when she/he **doesn't modify the algorithm directly** by coding.
Let's consider the following activities:
Annotation for Machine Learning:
analysis of Morp... |
2020/01/28 | 2,251 | 9,063 | <issue_start>username_0: I trained a simple CNN on the MNIST database of handwritten digits to 99% accuracy. I'm feeding in a bunch of handwritten digits, and non-digits from a document.
I want the CNN to report errors, so I set a threshold of 90% certainty below which my algorithm assumes that what it's looking at is... |
2020/01/29 | 2,007 | 8,060 | <issue_start>username_0: The famous Nvidia paper [Progressive Growing of GANs for Improved Quality, Stability, and Variation](https://research.nvidia.com/sites/default/files/pubs/2017-10_Progressive-Growing-of/karras2018iclr-paper.pdf), the GAN can generate hyperrealistic human faces. But, in the very same paper, image... |
2020/01/30 | 1,874 | 7,732 | <issue_start>username_0: We have hundreds of thousands of customers records, and we need to take the benefits of our data to train a model that will recognize fake entries or unrealistic ones for our platform, where customers are asked to enter their names, phone number and zip code.
So, our attributes are name, phone... |
2020/01/30 | 660 | 2,471 | <issue_start>username_0: I am looking for a small size dataset on which I can implement object detection, object segmentation and object localization.
Can anyone suggest me a dataset less than 5GB? Or do I need to know anything before implementing these algorithms?<issue_comment>username_1: There are various dataset a... |
2020/01/31 | 827 | 3,609 | <issue_start>username_0: I have trained a XGboost model to predict survival for the [Kaggle Titanic ML competition](https://www.kaggle.com/c/titanic/overview).
As with all Kaggle competitions there is a `train` dataset with the target variable included and a `test` dataset without the target variable which is used by ... |
2020/01/31 | 1,375 | 6,451 | <issue_start>username_0: Why are the terms classification and prediction used as synonyms especially when it comes to deep learning? For example, a CNN predicts the handwritten digit.
To me, a prediction is telling the next step in a sequence, whereas classification is to put labels on (a finite set of) data.<issue_co... |
2020/02/02 | 2,776 | 10,916 | <issue_start>username_0: I am trying to use RealNVP with some data I have (the input size is a 1D vector of size 22). [Here](https://arxiv.org/abs/1605.08803) is the link to the RealNVP paper and [here](https://lilianweng.github.io/lil-log/2018/10/13/flow-based-deep-generative-models.html) is a nice, short explanation ... |
2020/02/02 | 1,585 | 6,927 | <issue_start>username_0: How would you go about training an RL Tic Tac Toe (well, any board game, really) application to learn how to play successfully (win the game), without a human having to play against the RL?
Obviously, it would take a lot longer to train the AI if I have to sit and play "real" games against it... |
2020/02/02 | 2,500 | 9,108 | <issue_start>username_0: I am reading the Understanding Machine Learning book by Shalev-Shwartz and Ben-David and based on the definitions of PAC learnability and No Free Lunch Theorem, and my understanding of them it seems like they contradict themselves. I know this is not the case and I am wrong, but I just don't kn... |
2020/02/03 | 1,015 | 3,699 | <issue_start>username_0: I want to make a model that outputs the centre pixel of objects appearing in an image.
My current method involves using a CNN with L2 loss to output an image of equivalent size to the input where each pixel has a value of 1 if it is the center of an object and 0 otherwise. Each input image has... |
2020/02/03 | 874 | 3,116 | <issue_start>username_0: Consider a feedforward neural network. Suppose you have a layer of inputs, which is feedforward to a hidden layer, and feedforward both the input and hidden layers to an output layer. Is there a name for this architecture? A layer feeds forward around the layer after it?<issue_comment>username_... |
2020/02/04 | 881 | 3,465 | <issue_start>username_0: I am an undergraduate student in applied mathematics with an interest in artificial intelligence. I am currently exploring topics where I could do research. Coming from a mathematical background I am interested in the question: Can we mathematically establish that a certain AI system has the ab... |
2020/02/05 | 837 | 3,228 | <issue_start>username_0: When using convolutional networks on images with multiple channels, do we max pool after we sum the feature map from each channel, or do we max pool each feature map separately and then sum?
What's the intuition behind this, or is there a difference between the two?<issue_comment>username_1: ... |
2020/02/05 | 593 | 2,387 | <issue_start>username_0: I'm training a deep network in `Keras` on some images for a binary classification (I have around 12K images). Once in a while, I collect some false positives and add them to my training sets and re-train for higher accuracy.
I split my training into 20/80 percent for training/validation sets.
... |
2020/02/06 | 2,382 | 9,073 | <issue_start>username_0: Is it possible to estimate the capacity of a neural network model? If so, what are the techniques involved?<issue_comment>username_1: Most methods for measuring the complexity of neural networks are fairly crude. One common measure of complexity is [VC dimension](https://en.wikipedia.org/wiki/V... |
2020/02/06 | 713 | 2,726 | <issue_start>username_0: I have a question regarding features representation for graph convolutional neural network.
For my case, all nodes have a different number of features, and for now, I don't really understand how should I work with these constraints. I can not just reduce the number of features or add meaningle... |
2020/02/07 | 536 | 2,211 | <issue_start>username_0: Why do we split the data into two parts, and then split those segments into training and testing data? Why do we have two sets of data for each training and test data?<issue_comment>username_1: Usually we are splinting the data into 3 chunks for example 70% for training, 10% for validation and ... |
2020/02/07 | 583 | 2,434 | <issue_start>username_0: There is a popular strategy of using a neural network trained on one task to produce features for another related task by "chopping off" the top of the network and sewing the bottom onto some other modeling pipeline.
Word2Vec models employ this strategy, for example.
Is there an industry-popu... |
2020/02/08 | 646 | 2,209 | <issue_start>username_0: This question can seem a little bit too broad, but I am wondering what are the current state-of-the-art works on meta reinforcement learning. Can you provide me with the current state-of-the-art in this field?<issue_comment>username_1: One of the most recent papers on meta-RL is [meta-Q-learnin... |
2020/02/10 | 1,282 | 3,508 | <issue_start>username_0: Given a pre-trained CNN model, I extract feature vector of *3450* reference images `FV_R` as follows:
```
FV_R = [ [-8.2, -52.2, 9.07, -1.1, -0.08, -9.1, ........, -4.11],
[7.8, -3.8, 6.4, -4.27, -2.2, -5.0, ............., 3.6],
[-1.2, -0.8, 49.3, 1.73, -1... |
2020/02/11 | 925 | 3,482 | <issue_start>username_0: There is [this video on pythonprogramming.net](https://pythonprogramming.net/introduction-deep-learning-python-tensorflow-keras/) that trains a network on the MNIST handwriting dataset. At [~9:15](https://youtu.be/wQ8BIBpya2k?t=539), the author explains that the data should be normalized.
The... |
2020/02/11 | 500 | 2,061 | <issue_start>username_0: I am reading through Artificial Intelligence: Modern Approach and it states that the space complexity of the GBFS (tree version) is $\mathcal{O}(b^m)$.
While I am reading, at some points, I found GBFS similar to DFS. It expands the whole branches and goes after one according to the heuristic f... |
2020/02/12 | 537 | 2,102 | <issue_start>username_0: I trained a ResNet20 on Cifar10 and obtained the following learning curves.
[](https://i.stack.imgur.com/O7JVh.png)
From the figures, I see at epoch 52, my validation loss is 0.323 (the lowest), and my validation accuracy is ... |
2020/02/13 | 820 | 3,306 | <issue_start>username_0: In section 3 of the paper [The Limits of Correctness](https://www.student.cs.uwaterloo.ca/~cs492/11public_html/p18-smith.pdf) (1985) [<NAME>](https://en.wikipedia.org/wiki/Brian_Cantwell_Smith) writes
>
> When you design and build a computer system, you first formulate a model of the problem ... |
2020/02/14 | 743 | 2,664 | <issue_start>username_0: I started learning about Q table from this blog post [Introduction to reinforcement learning and OpenAI Gym, by <NAME>](https://www.oreilly.com/radar/introduction-to-reinforcement-learning-and-openai-gym/), which has a line as below -
>
> After so many episodes, the algorithm will converge an... |
2020/02/16 | 408 | 1,728 | <issue_start>username_0: Is there an ideal ratio in reinforcement learning between the positive and negative rewards?
Suppose I have the scenario of moving a robot across the river. There are two options, walk across the bridge or walk across the river. If it walks across the river then the robot breaks so the idea is... |
2020/02/18 | 910 | 3,342 | <issue_start>username_0: It kind of makes sense intuitively but I'm not sure about a formal proof. I'll start with briefly listing definitions from **Intro to Multiagent systems, Wooldridge, 2002** and then give you my reasoning attempts thus far.
$E$ is a finite set of discrete, instantaneous states, $E=(e, e',...)$.... |
2020/02/19 | 573 | 2,285 | <issue_start>username_0: Suppose I want to classify a dataset like the MNIST handwritten dataset, but it has added distractions. For example, here we have a 6 but with extra strokes around it that don't add value.
[](https://i.stack.imgur.com/Z35w2.pn... |
2020/02/20 | 685 | 3,292 | <issue_start>username_0: I am currently studying the textbook *Deep Learning* by Goodfellow, Bengio, and Courville. Chapter **5.1 Learning Algorithms** says the following:
>
> **Classification with missing inputs**: Classification becomes more challenging if the computer program is not guaranteed that every measureme... |
2020/02/21 | 1,021 | 3,918 | <issue_start>username_0: Does a fully convolutional network share the same translation invariance properties we get from networks that use max-pooling?
If not, why do they perform as well as networks which use max-pooling?<issue_comment>username_1: All convolutional networks (with or without max-pooling) are translati... |
2020/02/22 | 1,038 | 3,986 | <issue_start>username_0: I'm working on a project for my college to recognize traffic signs in pictures. I searched a lot but can't find the best method to do it.
Can someone recommend me a paper, article, or even GitHub link that describes the best way to achieve this? It will be helpful.<issue_comment>username_1: Al... |