Relative Behavioral Attributes: Filling the Gap between Symbolic Goal Specification and Reward Learning from Human Preferences.
Lin GuanKarthik ValmeekamSubbarao KambhampatiPublished in: ICLR (2023)
Keyphrases
- online learning
- language acquisition
- high level
- reinforcement learning
- learning process
- prior knowledge
- learning algorithm
- associative learning
- supervised learning
- collaborative filtering
- learning systems
- human experts
- connectionist systems
- learning tasks
- active learning
- connectionist networks
- preference learning
- data sets