Relative Behavioral Attributes: Filling the Gap between Symbolic Goal Specification and Reward Learning from Human Preferences.
Lin GuanKarthik ValmeekamSubbarao KambhampatiPublished in: CoRR (2022)
Keyphrases
- learning algorithm
- learning process
- reinforcement learning
- learning tasks
- neural network
- learning systems
- machine learning
- language acquisition
- online learning
- human experts
- human behavior
- human subjects
- motor skills
- connectionist networks
- multi attribute
- learning scenarios
- learning experience
- supervised learning
- prior knowledge
- multi agent systems