Shaping Zero-Shot Coordination via State Blocking
About
Zero-shot coordination (ZSC) aims to enable agents to cooperate with independently trained partners without prior interaction, a key requirement for real-world multi-agent systems and human-AI collaboration. Existing approaches have largely emphasized increasing partner diversity during training, yet such strategies often fall short of achieving reliable generalization to unseen partners. We introduce State-Blocked Coordination (SBC), a simple yet effective framework that improves ZSC by inducing diverse interaction scenarios without direct environment modification. Specifically, SBC generates a family of virtual environments through state blocking, allowing agents to experience a wide range of suboptimal partner policies. Across multiple benchmarks, SBC demonstrates superior performance in zero-shot coordination, including strong generalization to human partners.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Coordination | Overcooked Cramped Room layout v1 | SP257 | 14 | |
| Coordination | Overcooked Counter Circuit layout v1 | SP246 | 7 | |
| Coordination | Overcooked Coordination Ring layout v1 | SP333 | 7 | |
| Zero-shot Coordination | Overcooked Counter Circuit layout v1 | SP Score246 | 7 | |
| Zero-shot Coordination | Overcooked Coordination Ring layout v1 | SP333 | 7 | |
| Coordination | Overcooked Forced Coordination layout v1 | SP193 | 7 | |
| Zero-shot Coordination | Overcooked Forced Coordination layout v1 | SP193 | 7 | |
| Zero-shot Coordination | Overcooked Asymmetric Advantages layout v1 | SP500 | 7 | |
| Zero-shot Coordination | Multi-Destination Spread | SP982 | 6 |