Robustifying Reinforcement Learning Agents via Action Space Adversarial Training.
Kai Liang TanYasaman EsfandiariXian Yeow Lee AakankshaSoumik SarkarPublished in: ACC (2020)
Keyphrases
- action space
- reinforcement learning agents
- reinforcement learning
- state space
- state abstraction
- markov decision processes
- multi agent
- supervised learning
- dynamic environments
- transfer learning
- real valued
- training set
- feature selection
- optimal policy
- situation calculus
- action selection
- single agent
- stochastic processes
- search algorithm
- decision making