Human-Feedback Shield Synthesis for Perceived Safety in Deep Reinforcement Learning.
Daniel MartaChristian PekGaspar Isaac MelsiónJana TumovaIolanda LeitePublished in: IEEE Robotics Autom. Lett. (2022)
Keyphrases
- reinforcement learning
- function approximation
- multi agent
- relevance feedback
- temporal difference
- state space
- optimal policy
- human body
- human interaction
- reinforcement learning algorithms
- user satisfaction
- user engagement
- program synthesis
- reinforcement learning methods
- action space
- perceived usefulness
- optimal control
- machine learning
- human subjects
- human experts
- information systems