Reward Poisoning Attack Against Offline Reinforcement Learning.
Yinglun XuRohan GumasteGagandeep SinghPublished in: CoRR (2024)
Keyphrases
- reinforcement learning
- state space
- function approximation
- eligibility traces
- reinforcement learning algorithms
- learning algorithm
- real time
- dynamic programming
- temporal difference
- model free
- partially observable environments
- supervised learning
- markov decision processes
- function approximators
- reward function
- action selection
- optimal control
- optimal policy
- total reward
- learning capabilities
- malicious users
- policy gradient
- countermeasures
- dos attacks
- policy iteration
- learning problems
- mobile robot
- machine learning
- neural network