GRD: A Generative Approach for Interpretable Reward Redistribution in Reinforcement Learning.
Yudi ZhangYali DuBiwei HuangZiyan WangJun WangMeng FangMykola PechenizkiyPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- state space
- function approximation
- generative model
- reinforcement learning algorithms
- markov decision processes
- eligibility traces
- reward function
- dynamic programming
- learning algorithm
- temporal difference
- model free
- transfer learning
- total reward
- data driven
- temporal difference learning
- discriminative learning
- partially observable environments
- robotic control
- machine learning
- control policy
- learning agent
- classification rules
- markov decision process
- optimal policy
- optimal control
- learning problems
- decision trees
- supervised learning
- reinforcement learning methods
- action selection
- reward shaping