[TOC]

  1. Title: Efficient Reward Poisoning Attacks on Online Deep Reinforcement Learning
  2. Author: Yinglun Xu et. al.
  3. Publish Year: 30 May 2022
  4. Review Date: Tue, Dec 27, 2022

Summary of paper

Motivation

Contribution

limitation

Some key terms

LPE (learned policy evasion) attack

Uniformly at random (UR) attack

Potential future work

we can get some insight about the theoretical analysis of their attack methods based on certain assumptions on the efficiency of the DRL algorithm which yields several insightful implications