Adversarial Attack against Deep Reinforcement Learning with Static Reward Impact Map.
Patrick P. K. ChanYaxuan WangDaniel S. YeungPublished in: AsiaCCS (2020)
Keyphrases
- reinforcement learning
- multi agent
- eligibility traces
- function approximation
- reinforcement learning algorithms
- reinforcement learning methods
- state space
- optimal policy
- reward function
- partially observable environments
- temporal difference
- model free
- maximum a posteriori
- learning agent
- function approximators
- state action
- supervised learning
- countermeasures
- markov decision processes
- image reconstruction
- real robot
- intrusion detection
- bandit problems
- dynamic programming