Online Defense Strategies for Reinforcement Learning Against Adaptive Reward Poisoning.
Andi NikaAdish SinglaGoran RadanovicPublished in: AISTATS (2023)
Keyphrases
- reinforcement learning
- adaptive strategies
- adaptive control
- online learning
- function approximation
- learning algorithm
- state space
- learning agents
- exploration strategy
- optimal policy
- learning capabilities
- balancing exploration and exploitation
- reward function
- transfer learning
- machine learning
- reinforcement learning algorithms
- learning agent
- model free
- initially unknown
- real time
- multi agent reinforcement learning
- partially observable environments
- neural network
- temporal difference
- control strategies
- optimal control
- network security
- dynamic programming
- multi agent
- long run
- markov decision processes
- learning process