Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2604.13540 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911594699554816 |
|---|---|
| author | Jiang, Yibo Wu, Tao Jiang, Rui Lu, Yehao Cai, Chaoxiang Qin, Zequn Li, Xi |
| author_facet | Jiang, Yibo Wu, Tao Jiang, Rui Lu, Yehao Cai, Chaoxiang Qin, Zequn Li, Xi |
| contents | Unified Multimodal Models (UMMs) aim to integrate visual understanding and generation within a single structure. However, these models exhibit a notable capability mismatch, where their understanding capability significantly outperforms their generation. This mismatch indicates that the model's rich internal knowledge, while effective for understanding tasks, remains underactivated during generation. To address this, we draw inspiration from the human ``Thinking-While-Drawing'' paradigm, where humans continuously reflect to activate their knowledge and rectify intermediate results. In this paper, we propose UniRect-CoT, a training-free unified rectification chain-of-thought framework. Our approach unlocks the ``free lunch'' hidden in the UMM's powerful inherent understanding to continuously reflect, activating its internal knowledge and rectifying intermediate results during generation.We regard the diffusion denoising process in UMMs as an intrinsic visual reasoning process and align the intermediate results with the target instruction understood by the model, serving as a self-supervisory signal to rectify UMM generation.Extensive experiments demonstrate that UniRect-CoT can be easily integrated into existing UMMs, significantly enhancing generation quality across diverse complex tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_13540 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Free Lunch for Unified Multimodal Models: Enhancing Generation via Reflective Rectification with Inherent Understanding Jiang, Yibo Wu, Tao Jiang, Rui Lu, Yehao Cai, Chaoxiang Qin, Zequn Li, Xi Computer Vision and Pattern Recognition Artificial Intelligence Unified Multimodal Models (UMMs) aim to integrate visual understanding and generation within a single structure. However, these models exhibit a notable capability mismatch, where their understanding capability significantly outperforms their generation. This mismatch indicates that the model's rich internal knowledge, while effective for understanding tasks, remains underactivated during generation. To address this, we draw inspiration from the human ``Thinking-While-Drawing'' paradigm, where humans continuously reflect to activate their knowledge and rectify intermediate results. In this paper, we propose UniRect-CoT, a training-free unified rectification chain-of-thought framework. Our approach unlocks the ``free lunch'' hidden in the UMM's powerful inherent understanding to continuously reflect, activating its internal knowledge and rectifying intermediate results during generation.We regard the diffusion denoising process in UMMs as an intrinsic visual reasoning process and align the intermediate results with the target instruction understood by the model, serving as a self-supervisory signal to rectify UMM generation.Extensive experiments demonstrate that UniRect-CoT can be easily integrated into existing UMMs, significantly enhancing generation quality across diverse complex tasks. |
| title | Free Lunch for Unified Multimodal Models: Enhancing Generation via Reflective Rectification with Inherent Understanding |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2604.13540 |