One Risk to Rule Them All: A Risk-Sensitive Perspective on Model-Based Offline Reinforcement Learning.
Marc RigterBruno LacerdaNick HawesPublished in: NeurIPS (2023)
Keyphrases
- risk sensitive
- model free
- reinforcement learning
- markov decision processes
- optimal control
- reinforcement learning algorithms
- function approximation
- temporal difference
- risk neutral
- policy iteration
- markov decision problems
- state space
- utility function
- learning algorithm
- average reward
- markov decision chains
- transfer learning
- optimal policy
- dynamic programming
- multi agent
- machine learning
- reward function
- linear programming
- finite horizon
- control policies
- optimal solution
- real time