Reachability-Aware Laplacian Representation in Reinforcement Learning.
Kaixin WangKuangqi ZhouJiashi FengBryan HooiXinchao WangPublished in: CoRR (2022)
Keyphrases
- reinforcement learning
- markov decision processes
- image processing
- state space
- image representation
- learning problems
- function approximation
- spectral decomposition
- feature selection
- dynamic programming
- neural network
- model free
- edge detection
- reinforcement learning algorithms
- transitive closure
- temporal difference
- random walk
- mobile robot
- learning process
- multiscale
- machine learning