CooT: Learning to Coordinate In-Context with Coordination Transformers

Huai-Chih Wang, Hsiang-Chun Chuang, Hsi-Chun Cheng, Dai-Jie Wu, Shao-Hua Sun
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:125587-125625, 2026.

Abstract

Effective coordination among unfamiliar partners remains a major challenge in multi-agent systems. Existing approaches, such as population-based methods, improve robustness through diversity but often lack mechanisms for efficient adaptation beyond the training distribution. Fine-tuning is also impractical for few-shot learning because it requires a large number of interactions for meaningful improvement. To address these limitations, we propose Coordination Transformers (CooT), a framework that leverages in-context learning (ICL) for real-time partner adaptation. Unlike prior ICL approaches that focus on task generalization, CooT is designed to generalize across diverse partner behaviors. Trained on trajectories from behavior-preferring agents, it learns to align actions with partner intentions purely through observation. We evaluate CooT on two challenging multi-agent benchmarks: Overcooked and Google Research Football. Results show that CooT consistently outperforms population-based methods, gradient-based fine-tuning, and Meta-RL baselines, achieving stable and rapid adaptation without parameter updates. Human evaluations also identify CooT as a preferred collaborator, and our ablations confirm its ability to adapt quickly to new partners and remain stable under sudden partner changes, making it reliable for real-world human-AI collaboration.

Cite this Paper


BibTeX
@InProceedings{pmlr-v306-wang26ag, title = {{C}oo{T}: Learning to Coordinate In-Context with Coordination Transformers}, author = {Wang, Huai-Chih and Chuang, Hsiang-Chun and Cheng, Hsi-Chun and Wu, Dai-Jie and Sun, Shao-Hua}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {125587--125625}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/wang26ag/wang26ag.pdf}, url = {https://proceedings.mlr.press/v306/wang26ag.html}, abstract = {Effective coordination among unfamiliar partners remains a major challenge in multi-agent systems. Existing approaches, such as population-based methods, improve robustness through diversity but often lack mechanisms for efficient adaptation beyond the training distribution. Fine-tuning is also impractical for few-shot learning because it requires a large number of interactions for meaningful improvement. To address these limitations, we propose Coordination Transformers (CooT), a framework that leverages in-context learning (ICL) for real-time partner adaptation. Unlike prior ICL approaches that focus on task generalization, CooT is designed to generalize across diverse partner behaviors. Trained on trajectories from behavior-preferring agents, it learns to align actions with partner intentions purely through observation. We evaluate CooT on two challenging multi-agent benchmarks: Overcooked and Google Research Football. Results show that CooT consistently outperforms population-based methods, gradient-based fine-tuning, and Meta-RL baselines, achieving stable and rapid adaptation without parameter updates. Human evaluations also identify CooT as a preferred collaborator, and our ablations confirm its ability to adapt quickly to new partners and remain stable under sudden partner changes, making it reliable for real-world human-AI collaboration.} }
Endnote
%0 Conference Paper %T CooT: Learning to Coordinate In-Context with Coordination Transformers %A Huai-Chih Wang %A Hsiang-Chun Chuang %A Hsi-Chun Cheng %A Dai-Jie Wu %A Shao-Hua Sun %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-wang26ag %I PMLR %P 125587--125625 %U https://proceedings.mlr.press/v306/wang26ag.html %V 306 %X Effective coordination among unfamiliar partners remains a major challenge in multi-agent systems. Existing approaches, such as population-based methods, improve robustness through diversity but often lack mechanisms for efficient adaptation beyond the training distribution. Fine-tuning is also impractical for few-shot learning because it requires a large number of interactions for meaningful improvement. To address these limitations, we propose Coordination Transformers (CooT), a framework that leverages in-context learning (ICL) for real-time partner adaptation. Unlike prior ICL approaches that focus on task generalization, CooT is designed to generalize across diverse partner behaviors. Trained on trajectories from behavior-preferring agents, it learns to align actions with partner intentions purely through observation. We evaluate CooT on two challenging multi-agent benchmarks: Overcooked and Google Research Football. Results show that CooT consistently outperforms population-based methods, gradient-based fine-tuning, and Meta-RL baselines, achieving stable and rapid adaptation without parameter updates. Human evaluations also identify CooT as a preferred collaborator, and our ablations confirm its ability to adapt quickly to new partners and remain stable under sudden partner changes, making it reliable for real-world human-AI collaboration.
APA
Wang, H., Chuang, H., Cheng, H., Wu, D. & Sun, S.. (2026). CooT: Learning to Coordinate In-Context with Coordination Transformers. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:125587-125625 Available from https://proceedings.mlr.press/v306/wang26ag.html.

Related Material