Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2605.11400 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910210939944960 |
|---|---|
| author | Bai, Hayes Luo, Yinyi Wang, Wenwen Wen, Qingsong Wang, Jindong |
| author_facet | Bai, Hayes Luo, Yinyi Wang, Wenwen Wen, Qingsong Wang, Jindong |
| contents | Unified multimodal models (UMMs) aim to integrate understanding and generation within a single architecture. However, it remains underexplored how to effectively coordinate these two capabilities for more effective and efficient reasoning. Existing coordination approaches either perform coupling during training, without explicit inference-time coordination, or impose a fixed coordination pattern for all inputs. In this work, we show that multimodal tasks exhibit substantial coordination-path diversity: different inputs favor different coordination paths. This suggests that exploiting such diversity is key to improving performance. We propose UniPath, a framework for adaptively modeling and exploiting coordination-path diversity. Instead of enforcing a single coordination pattern, we represent task solving as the selection and execution of a path, ranging from direct answering to textual inference, visual-thought construction, and hypothesis-based exploration. We construct role-aligned trajectories to train a path-conditioned executor and introduce a lightweight planner mechanism to enable input-dependent path selection. Experiments show that leveraging coordination-path diversity improves performance over fixed coordination strategies while providing interpretable intermediate behaviors. The code is available at:https://github.com/AIFrontierLab/TorchUMM/tree/main/src/umm/post_training/unipath. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_11400 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | UniPath: Adaptive Coordination of Understanding and Generation for Unified Multimodal Reasoning Bai, Hayes Luo, Yinyi Wang, Wenwen Wen, Qingsong Wang, Jindong Multimedia Unified multimodal models (UMMs) aim to integrate understanding and generation within a single architecture. However, it remains underexplored how to effectively coordinate these two capabilities for more effective and efficient reasoning. Existing coordination approaches either perform coupling during training, without explicit inference-time coordination, or impose a fixed coordination pattern for all inputs. In this work, we show that multimodal tasks exhibit substantial coordination-path diversity: different inputs favor different coordination paths. This suggests that exploiting such diversity is key to improving performance. We propose UniPath, a framework for adaptively modeling and exploiting coordination-path diversity. Instead of enforcing a single coordination pattern, we represent task solving as the selection and execution of a path, ranging from direct answering to textual inference, visual-thought construction, and hypothesis-based exploration. We construct role-aligned trajectories to train a path-conditioned executor and introduce a lightweight planner mechanism to enable input-dependent path selection. Experiments show that leveraging coordination-path diversity improves performance over fixed coordination strategies while providing interpretable intermediate behaviors. The code is available at:https://github.com/AIFrontierLab/TorchUMM/tree/main/src/umm/post_training/unipath. |
| title | UniPath: Adaptive Coordination of Understanding and Generation for Unified Multimodal Reasoning |
| topic | Multimedia |
| url | https://arxiv.org/abs/2605.11400 |