SoK: Adversarial Machine Learning Attacks and Defences in Multi-Agent Reinforcement Learning

Maxwell Standen, Junae Kim, Claudia Szabo
1/11/2023
cs.LGcs.AIcs.CR

No official code found. Best implementation: asokraju/Adv-MARL (25 stars, last commit 2023-02-01).

Abstract

Multi-Agent Reinforcement Learning (MARL) is vulnerable to Adversarial Machine Learning (AML) attacks and needs adequate defences before it can be used in real world applications. We have conducted a survey into the use of execution-time AML attacks against MARL and the defences against those attacks. We surveyed related work in the application of AML in Deep Reinforcement Learning (DRL) and Multi-Agent Learning (MAL) to inform our analysis of AML for MARL. We propose a novel perspective to understand the manner of perpetrating an AML attack, by defining Attack Vectors. We develop two new frameworks to address a gap in current modelling frameworks, focusing on the means and tempo of an AML attack against MARL, and identify knowledge gaps and future avenues of research.

View on arXivView PDF

Code Implementations(1)

asokraju/Adv-MARLCommunity73%

Adversarial attacks in consensus-based multi-agent reinforcement learning

255Shell, PythonSep 4, 20203 years ago
gridworldmultiagent-reinforcement-learningreinforcement-learning

Similar Papers

Efficient Adversarial Attacks on Online Multi-agent Reinforcement Learning

90% match2023

Enhancing the Robustness of QMIX against State-adversarial Attacks

89% match2023

Constrained Black-Box Attacks Against Multi-Agent Reinforcement Learning

89% match1 repo32025

Robust Deep Reinforcement Learning Through Adversarial Attacks and Training : A Survey

88% match2024

Attacking Cooperative Multi-Agent Reinforcement Learning by Adversarial Minority Influence

88% match1 repo172023

Adversarial Reinforcement Learning for Large Language Model Agent Safety

86% match2025

Cite this paper

@article{standen2023sok,
  title  = {SoK: Adversarial Machine Learning Attacks and Defences in Multi-Agent Reinforcement Learning},
  author = {Maxwell Standen and Junae Kim and Claudia Szabo},
  year   = {2023},
  eprint = {2301.04299},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2301.04299v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2301.04299.svg)](https://codeofpaper.com/paper/2301.04299)

Discussion