Safety Aware Reinforcement Learning by Identifying Comprehensible Constraints in Expert Demonstrations.
Leopold MüllerLars BöckingMichael FärberPublished in: SafeAI@AAAI (2022)
Keyphrases
- reinforcement learning
- function approximation
- global constraints
- expert knowledge
- real time
- neural network
- database
- learning algorithm
- linear constraints
- model free
- black box
- autonomous learning
- constraint programming
- learning problems
- domain experts
- optimal policy
- semi supervised
- state space
- learning process
- machine learning