Defense Against Reward Poisoning Attacks in Reinforcement Learning

Understand

We study defense strategies against reward poisoning attacks in reinforcement learning.

• As a threat model, we consider attacks that minimally alter rewards to make the attacker's target policy uniquely optimal under the poisoned rewards, with the optimality gap specified by an attack parameter.
• Our goal is to design agents that are robust against such attacks in terms of the worst-case utility w.r.t.
• the true, unpoisoned, rewards while computing their policies under the poisoned rewards.