Active Learning for Risk-Sensitive Inverse Reinforcement Learning.
Rui ChenWenshuo WangZirui ZhaoDing ZhaoPublished in: CoRR (2019)
Keyphrases
- risk sensitive
- inverse reinforcement learning
- active learning
- utility function
- preference elicitation
- control policies
- markov decision processes
- reward function
- optimal control
- model free
- temporal difference
- supervised learning
- learning algorithm
- optimality criterion
- markov decision problems
- machine learning
- dynamic programming
- decision theoretic
- reinforcement learning
- optimal policy
- semi supervised
- decision problems
- decision theory
- reinforcement learning algorithms
- partially observable
- decision makers
- finite state
- multi objective
- theoretical framework
- learning process
- training set
- state space
- probability distribution