Boosting Offline Reinforcement Learning with Action Preference Query.
Qisen YangShenzhi WangMatthieu Gaetan LinShiji SongGao HuangPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- learning algorithm
- action selection
- database
- response time
- query processing
- function approximation
- data sources
- query expansion
- machine learning
- action space
- user interaction
- multi attribute
- partially observable domains
- state action
- model free
- query evaluation
- user preferences
- data structure
- feature selection
- decision problems
- query formulation
- state space
- active learning
- transition model
- reward shaping