Policy-Independent Behavioral Metric-Based Representation for Deep Reinforcement Learning.
Weijian LiaoZongzhang ZhangYang YuPublished in: AAAI (2023)
Keyphrases
- reinforcement learning
- optimal policy
- state space
- action selection
- policy search
- action space
- partially observable
- reward function
- function approximation
- reinforcement learning problems
- control policies
- actor critic
- inverse reinforcement learning
- topological map
- neural network
- function approximators
- partially observable environments
- partially observable markov decision processes
- reinforcement learning algorithms
- dynamical systems
- image representation
- distance measure
- supervised learning
- high dimensional
- machine learning