Towards Zero-Shot Coordination between Teams of Agents: The N-XPlay Framework
This addresses the challenge of enabling autonomous agents to coordinate effectively in complex real-world multi-team systems, representing an incremental advancement over existing two-agent methods.
The paper tackled the problem of zero-shot coordination in multi-agent systems by extending it to multi-team settings, introducing N-player Overcooked as a benchmark and proposing the N-XPlay framework, which showed improved ability to balance intra-team and inter-team coordination compared to Self-Play in scenarios with up to five agents.
Zero-shot coordination (ZSC) -- the ability to collaborate with unfamiliar partners -- is essential to making autonomous agents effective teammates. Existing ZSC methods evaluate coordination capabilities between two agents who have not previously interacted. However, these scenarios do not reflect the complexity of real-world multi-agent systems, where coordination often involves a hierarchy of sub-groups and interactions between teams of agents, known as Multi-Team Systems (MTS). To address this gap, we first introduce N-player Overcooked, an N-agent extension of the popular two-agent ZSC benchmark, enabling evaluation of ZSC in N-agent scenarios. We then propose N-XPlay for ZSC in N-agent, multi-team settings. Comparison against Self-Play across two-, three- and five-player Overcooked scenarios, where agents are split between an ``ego-team'' and a group of unseen collaborators shows that agents trained with N-XPlay are better able to simultaneously balance ``intra-team'' and ``inter-team'' coordination than agents trained with SP.