Increasing Transparency of Reinforcement Learning using Shielding for Human Preferences and Explanations.
Georgios AngelopoulosLuigi MangiacapraAlessandra RossiClaudia Di NapoliSilvia RossiPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- multi agent
- human subjects
- decision making
- function approximation
- markov decision processes
- transfer learning
- hidden markov models
- preference elicitation
- user preferences
- optimal policy
- computational models
- knowledge base
- artificial intelligence
- causal models
- learning classifier systems
- reinforcement learning algorithms
- multi agent reinforcement learning
- behavioural cloning