Reward Delay Attacks on Deep Reinforcement Learning.
Anindya SarkarJiarui FengYevgeniy VorobeychikChristopher D. GillNing ZhangPublished in: CoRR (2022)
Keyphrases
- reinforcement learning
- reinforcement learning algorithms
- function approximation
- eligibility traces
- state space
- countermeasures
- model free
- machine learning
- reward function
- markov decision processes
- malicious users
- multi agent
- learning algorithm
- total reward
- security protocols
- temporal difference
- terrorist attacks
- traffic analysis
- security threats
- partially observable
- transfer learning
- watermarking scheme
- optimal control
- supervised learning
- dynamical systems
- security mechanisms
- dynamic programming
- learning agent
- average reward
- reinforcement learning methods
- digital images
- malicious attacks
- neural network