MoMA: Model-based Mirror Ascent for Offline Reinforcement Learning.
Mao HongZhiyue ZhangYue WuYanxun XuPublished in: CoRR (2024)
Keyphrases
- reinforcement learning
- model free
- state space
- reinforcement learning algorithms
- temporal difference
- data driven
- function approximation
- markov decision processes
- relational reinforcement learning
- control problems
- field of view
- data sets
- active learning
- evolutionary algorithm
- website
- decision making
- transition model
- real world