Phasic Self-Imitative Reduction for Sparse-Reward Goal-Conditioned Reinforcement Learning.
Yunfei LiTian GaoJiaqi YangHuazhe XuYi WuPublished in: ICML (2022)
Keyphrases
- reinforcement learning
- state space
- function approximation
- learning algorithm
- agent learns
- reinforcement learning algorithms
- reward function
- model free
- sparse representation
- markov decision processes
- temporal difference
- learning process
- compressive sensing
- eligibility traces
- transfer learning
- dynamic programming
- sparse data
- partially observable
- partially observable environments
- learning problems
- optimal policy
- action selection
- reinforcement learning methods
- data sets