Uni-Synergy: Bridging Understanding and Generation for Personalized Reasoning via Co-operative Reinforcement Learning
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866918494506844160 |
|---|---|
| author | Shen, Zijun Yang, Sihan An, Ruichuan Guo, Ziyu Liang, Hao Lu, Ming Zhang, Renrui Zhang, Wentao |
| author_facet | Shen, Zijun Yang, Sihan An, Ruichuan Guo, Ziyu Liang, Hao Lu, Ming Zhang, Renrui Zhang, Wentao |
| contents | Unified Multimodal Models (UMMs) excel in general tasks but struggle to bridge the gap between personalized understanding and generation. Prior works largely rely on implicit token-level alignment via supervised fine-tuning, which fails to fully capture the potential synergy between comprehension and creation. In this work, we propose Sync-R1, an end-to-end reinforcement learning framework that jointly optimizes personalized understanding and generation within a single, explicit reasoning loop. Through this unified feedback process, Sync-R1 enables personalized comprehension to guide content creation, while the resulting generation quality reciprocally refines understanding within an integrated reward landscape. To efficiently orchestrate this dual-task synergy, we introduce Sync-GRPO, a reinforcement learning method utilizing an ensemble reward system. Furthermore, we propose Dynamic Group Scaling (DGS), which adaptively filters low-potential trajectories to reduce gradient variance and accelerate convergence. To better reflect real-world complexity, we introduce UnifyBench++, featuring denser textual descriptions and richer user contexts. Experimental results demonstrate that Sync-R1 achieves state-of-the-art performance, showcasing superior cross-task reasoning and robust personalization without requiring complex cold-start procedures. The code and the UnifyBench++ dataset will be released at: https://github.com/arctanxarc/UniCTokens. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_10445 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Uni-Synergy: Bridging Understanding and Generation for Personalized Reasoning via Co-operative Reinforcement Learning Shen, Zijun Yang, Sihan An, Ruichuan Guo, Ziyu Liang, Hao Lu, Ming Zhang, Renrui Zhang, Wentao Computer Vision and Pattern Recognition Unified Multimodal Models (UMMs) excel in general tasks but struggle to bridge the gap between personalized understanding and generation. Prior works largely rely on implicit token-level alignment via supervised fine-tuning, which fails to fully capture the potential synergy between comprehension and creation. In this work, we propose Sync-R1, an end-to-end reinforcement learning framework that jointly optimizes personalized understanding and generation within a single, explicit reasoning loop. Through this unified feedback process, Sync-R1 enables personalized comprehension to guide content creation, while the resulting generation quality reciprocally refines understanding within an integrated reward landscape. To efficiently orchestrate this dual-task synergy, we introduce Sync-GRPO, a reinforcement learning method utilizing an ensemble reward system. Furthermore, we propose Dynamic Group Scaling (DGS), which adaptively filters low-potential trajectories to reduce gradient variance and accelerate convergence. To better reflect real-world complexity, we introduce UnifyBench++, featuring denser textual descriptions and richer user contexts. Experimental results demonstrate that Sync-R1 achieves state-of-the-art performance, showcasing superior cross-task reasoning and robust personalization without requiring complex cold-start procedures. The code and the UnifyBench++ dataset will be released at: https://github.com/arctanxarc/UniCTokens. |
| title | Uni-Synergy: Bridging Understanding and Generation for Personalized Reasoning via Co-operative Reinforcement Learning |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.10445 |