CAMA: Exploring Collusive Adversarial Attacks in c-MARL

TL;DR

This paper introduces a unified framework called CAMA for collusive adversarial attacks in cooperative multi-agent reinforcement learning, proposing three novel attack modes and analyzing their effectiveness and stealthiness.

Contribution

The paper presents the first design of three collusive attack modes in c-MARL, along with a unified framework and theoretical analysis of their impact and stealth.

Findings

Collusive attacks significantly enhance attack effectiveness.

The attacks maintain high stealthiness and stability.

Experimental results confirm the additive adversarial synergy.

Abstract

Cooperative multi-agent reinforcement learning (c-MARL) has been widely deployed in real-world applications, such as social robots, embodied intelligence, UAV swarms, etc. Nevertheless, many adversarial attacks still exist to threaten various c-MARL systems. At present, the studies mainly focus on single-adversary perturbation attacks and white-box adversarial attacks that manipulate agents' internal observations or actions. To address these limitations, we in this paper attempt to study collusive adversarial attacks through strategically organizing a set of malicious agents into three collusive attack modes: Collective Malicious Agents, Disguised Malicious Agents, and Spied Malicious Agents. Three novelties are involved: i) three collusive adversarial attacks are creatively proposed for the first time, and a unified framework CAMA for policy-level collusive attacks is designed; ii) the attack effectiveness is theoretically analyzed from the perspectives of disruptiveness, stealthiness, and attack cost; and iii) the three collusive adversarial attacks are technically realized through agent's observation information fusion, attack-trigger control. Finally, multi-facet experiments on four SMAC II maps are performed, and experimental results showcase the three collusive attacks have an additive adversarial synergy, strengthening attack outcome while maintaining high stealthiness and stability over long horizons. Our work fills the gap for collusive adversarial learning in c-MARL.