Hohei Chan

Hohei Chan

MS student at SCUT and intern with the Kling Team at Kuaishou. Research: reinforcement learning, agents, and multimodal generation.

PADiff: Predictive and Adaptive Diffusion Policies for Ad Hoc Teamwork

Published in Proceedings of the AAAI Conference on Artificial Intelligence, 2026

In offline multi-agent collaboration scenarios, we propose a diffusion-based decision framework that enables the modeling of various collaboration modes for unknown teammate agents.

Related links:

Bib PDF

BibTeX

@inproceedings{chan2026padiff,
  title={PADiff: Predictive and Adaptive Diffusion Policies for Ad Hoc Teamwork},
  author={Chan, Hohei and Zhang, Xinzhi and Xiang, Antao and Zhang, Weinan and Zhao, Mengchen},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={40},
  number={24},
  pages={19943--19951},
  year={2026}
}