InterAgent: Physics-based Multi-agent Command Execution via Diffusion on Interaction Graphs

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Bin, Zhang, Ruichi, Liang, Han, Zhang, Jingyan, Zhang, Juze, Chen, Xin, Xu, Lan, Yu, Jingyi, Wang, Jingya
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909957461377024
author Li, Bin
Zhang, Ruichi
Liang, Han
Zhang, Jingyan
Zhang, Juze
Chen, Xin
Xu, Lan
Yu, Jingyi
Wang, Jingya
author_facet Li, Bin
Zhang, Ruichi
Liang, Han
Zhang, Jingyan
Zhang, Juze
Chen, Xin
Xu, Lan
Yu, Jingyi
Wang, Jingya
contents Humanoid agents are expected to emulate the complex coordination inherent in human social behaviors. However, existing methods are largely confined to single-agent scenarios, overlooking the physically plausible interplay essential for multi-agent interactions. To bridge this gap, we propose InterAgent, the first end-to-end framework for text-driven physics-based multi-agent humanoid control. At its core, we introduce an autoregressive diffusion transformer equipped with multi-stream blocks, which decouples proprioception, exteroception, and action to mitigate cross-modal interference while enabling synergistic coordination. We further propose a novel interaction graph exteroception representation that explicitly captures fine-grained joint-to-joint spatial dependencies to facilitate network learning. Additionally, within it we devise a sparse edge-based attention mechanism that dynamically prunes redundant connections and emphasizes critical inter-agent spatial relations, thereby enhancing the robustness of interaction modeling. Extensive experiments demonstrate that InterAgent consistently outperforms multiple strong baselines, achieving state-of-the-art performance. It enables producing coherent, physically plausible, and semantically faithful multi-agent behaviors from only text prompts. Our code and data will be released to facilitate future research.
format Preprint
id arxiv_https___arxiv_org_abs_2512_07410
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle InterAgent: Physics-based Multi-agent Command Execution via Diffusion on Interaction Graphs
Li, Bin
Zhang, Ruichi
Liang, Han
Zhang, Jingyan
Zhang, Juze
Chen, Xin
Xu, Lan
Yu, Jingyi
Wang, Jingya
Computer Vision and Pattern Recognition
Humanoid agents are expected to emulate the complex coordination inherent in human social behaviors. However, existing methods are largely confined to single-agent scenarios, overlooking the physically plausible interplay essential for multi-agent interactions. To bridge this gap, we propose InterAgent, the first end-to-end framework for text-driven physics-based multi-agent humanoid control. At its core, we introduce an autoregressive diffusion transformer equipped with multi-stream blocks, which decouples proprioception, exteroception, and action to mitigate cross-modal interference while enabling synergistic coordination. We further propose a novel interaction graph exteroception representation that explicitly captures fine-grained joint-to-joint spatial dependencies to facilitate network learning. Additionally, within it we devise a sparse edge-based attention mechanism that dynamically prunes redundant connections and emphasizes critical inter-agent spatial relations, thereby enhancing the robustness of interaction modeling. Extensive experiments demonstrate that InterAgent consistently outperforms multiple strong baselines, achieving state-of-the-art performance. It enables producing coherent, physically plausible, and semantically faithful multi-agent behaviors from only text prompts. Our code and data will be released to facilitate future research.
title InterAgent: Physics-based Multi-agent Command Execution via Diffusion on Interaction Graphs
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.07410