Attacking c-MARL More Effectively: A Data Driven Approach
This work addresses the security vulnerability of c-MARL systems, which is crucial for applications like autonomous vehicles or robotics, but it is incremental as it builds on existing adversarial attack methods in MARL.
The paper tackles the problem of evaluating the robustness of cooperative multi-agent reinforcement learning (c-MARL) agents against adversarial attacks by proposing a model-based approach called c-MBA, which crafts stronger state perturbations to lower team rewards and introduces victim-agent selection and data-driven targeted failure states, outperforming baselines in all tested environments.
In recent years, a proliferation of methods were developed for cooperative multi-agent reinforcement learning (c-MARL). However, the robustness of c-MARL agents against adversarial attacks has been rarely explored. In this paper, we propose to evaluate the robustness of c-MARL agents via a model-based approach, named c-MBA. Our proposed formulation can craft much stronger adversarial state perturbations of c-MARL agents to lower total team rewards than existing model-free approaches. In addition, we propose the first victim-agent selection strategy and the first data-driven approach to define targeted failure states where each of them allows us to develop even stronger adversarial attack without the expert knowledge to the underlying environment. Our numerical experiments on two representative MARL benchmarks illustrate the advantage of our approach over other baselines: our model-based attack consistently outperforms other baselines in all tested environments.