Provable Offline Preference-Based Reinforcement Learning.
Wenhao ZhanMasatoshi UeharaNathan KallusJason D. LeeWen SunPublished in: ICLR (2024)
Keyphrases
- reinforcement learning
- function approximation
- state space
- multi agent
- reinforcement learning algorithms
- direct policy search
- control problems
- temporal difference
- model free
- supervised learning
- transfer learning
- optimal control
- markov decision processes
- real time
- optimal policy
- user preferences
- dynamic programming
- machine learning
- special case
- learning process
- preference relations
- search algorithm
- learning algorithm
- function approximators
- transition model
- robotic control