CoMA: Compositional Human Motion Generation with Multi-modal Agents
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (11) · 9206-9214
DOI: 10.1609/aaai.v40i11.37878 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Shanlin Sun und 7 weiteren Autor:innen. Die Arbeit wurde bisher 6-mal zitiert (Crossref).
Abstract (Originalsprache)
3D human motion generation has seen substantial advancement in recent years. While state-of-the-art approaches have improved performance significantly, they still struggle with complex and detailed motions unseen in training data, largely due to the scarcity of motion datasets and the prohibitive cost of generating new training examples. To address these challenges, we introduce CoMA, an agent-based solution for complex human motion generation, editing, and comprehension. CoMA leverages multiple collaborative agents powered by large language and vision models, alongside a mask transformer-based motion generator featuring body part-specific encoders and codebooks for fine-grained control. Our framework enables generation of both short and long motion sequences with detailed instructions, text-guided motion editing, and self-correction for improved quality. Evaluations on the HumanML3D dataset demonstrate competitive performance against state-of-the-art methods. Additionally, we create a set of context-rich, compositional, and long text prompts, where user studies show our method significantly outperforms existing approaches.
Zitieren
APA
Sun, S., Xu, J., Araujo, G. D., Zhou, S., Zhang, H., Huang, Z., You, C., & Xie, X. (2026). CoMA: Compositional Human Motion Generation with Multi-modal Agents. Proceedings of the AAAI Conference on Artificial Intelligence, 40(11), 9206-9214. https://doi.org/10.1609/aaai.v40i11.37878
Vancouver
Sun S, Xu J, Araujo GD, Zhou S, Zhang H, Huang Z, et al. CoMA: Compositional Human Motion Generation with Multi-modal Agents. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(11):9206-9214. doi:10.1609/aaai.v40i11.37878
BibTeX
@article{Sun2026,
author = {Sun, Shanlin and Xu, Jiaqi and Araujo, Gabriel De and Zhou, Shenghan and Zhang, Hanwen and Huang, Ziheng and You, Chenyu and Xie, Xiaohui},
title = {CoMA: Compositional Human Motion Generation with Multi-modal Agents},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {11},
pages = {9206--9214},
doi = {10.1609/aaai.v40i11.37878}
}
EUVIMED