There Is No Turning Back: A Self-Supervised Approach for Reversibility-Aware Reinforcement Learning.
Nathan GrinsztajnJohan FerretOlivier PietquinPhilippe PreuxMatthieu GeistPublished in: NeurIPS (2021)
Keyphrases
- reinforcement learning
- markov chain
- function approximation
- state space
- reinforcement learning algorithms
- model free
- markov processes
- learning algorithm
- learning process
- markov decision processes
- dynamic programming
- direct policy search
- control problems
- temporal difference
- learning problems
- search algorithm
- decision making
- artificial intelligence
- neural network
- robot control
- database
- multi agent reinforcement learning
- evolutionary learning
- autonomous learning
- continuous state
- reinforcement learning methods
- real time
- action space
- optimal policy
- machine learning
- information systems
- control system