Adversarial attacks in consensus-based multi-agent reinforcement learning
This work addresses security vulnerabilities in cooperative distributed MARL systems, which is an incremental but important contribution for researchers and practitioners in multi-agent systems.
The paper investigates adversarial attacks on consensus-based multi-agent reinforcement learning (MARL) algorithms, demonstrating that an adversarial agent can manipulate all other agents to optimize its own objective, revealing the fragility of these standard algorithms.
Recently, many cooperative distributed multi-agent reinforcement learning (MARL) algorithms have been proposed in the literature. In this work, we study the effect of adversarial attacks on a network that employs a consensus-based MARL algorithm. We show that an adversarial agent can persuade all the other agents in the network to implement policies that optimize an objective that it desires. In this sense, the standard consensus-based MARL algorithms are fragile to attacks.