Datasets:

ArXiv:
License:
JadeHuang commited on
Commit
357f5ea
·
verified ·
1 Parent(s): b952c63

Align README tagline and Core design layout

Browse files
Files changed (1) hide show
  1. README.md +4 -3
README.md CHANGED
@@ -2,7 +2,8 @@
2
  license: apache-2.0
3
  ---
4
 
5
- <p align="center">the first benchmark in embodied simulation for evaluating the continual learning capability of self-evolving agents</p>
 
6
 
7
  <h1 align="center">EmbodiedCLBench: Evaluating Continual Learning for Self-Evolving Embodied Agents</h1>
8
 
@@ -34,12 +35,12 @@ license: apache-2.0
34
 
35
  <p align="center">★ Equal contribution.&emsp;† Corresponding author.</p>
36
 
 
 
37
  <p align="center">
38
  <img src="https://pku-value-lab.github.io/EmbodiedCLBench-homepage/static/images/paper/intro.png" alt="EmbodiedCLBench overview showing compositional generalization from basic to advanced tasks" width="100%">
39
  </p>
40
 
41
- ## Core design
42
-
43
  We operationalize continual learning through compositional generalization: an agent first completes simpler tasks to accumulate experience, then is tested on whether it can recombine the acquired skills to solve more complex, unseen tasks.
44
 
45
  ## Abstract
 
2
  license: apache-2.0
3
  ---
4
 
5
+ <p align="center">the first benchmark in embodied simulation for evaluating<br>
6
+ the continual learning capability of self-evolving agents</p>
7
 
8
  <h1 align="center">EmbodiedCLBench: Evaluating Continual Learning for Self-Evolving Embodied Agents</h1>
9
 
 
35
 
36
  <p align="center">★ Equal contribution.&emsp;† Corresponding author.</p>
37
 
38
+ ## Core design
39
+
40
  <p align="center">
41
  <img src="https://pku-value-lab.github.io/EmbodiedCLBench-homepage/static/images/paper/intro.png" alt="EmbodiedCLBench overview showing compositional generalization from basic to advanced tasks" width="100%">
42
  </p>
43
 
 
 
44
  We operationalize continual learning through compositional generalization: an agent first completes simpler tasks to accumulate experience, then is tested on whether it can recombine the acquired skills to solve more complex, unseen tasks.
45
 
46
  ## Abstract