Salvato in:
| Autori principali: | , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2510.25682 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866917256323137536 |
|---|---|
| author | Zheng, Jiani Teng, Zhiyang Qiu, Kunpeng Li, Xiangtai Wang, Anran Tian, Yu Tian, Ye Wang, Haochen Wang, Zhuochen |
| author_facet | Zheng, Jiani Teng, Zhiyang Qiu, Kunpeng Li, Xiangtai Wang, Anran Tian, Yu Tian, Ye Wang, Haochen Wang, Zhuochen |
| contents | Unified Vision-Language Models (UVLMs) perform both understanding and generation within a single architecture. Since these models rely on heterogeneous data and supervision, balancing both generation and understanding in reinforcement learning (RL) is challenging. To address this challenge, we propose PairUni, a unified framework that reorganizes data into understanding-generation (UG) pairs and aligns optimization accordingly. Specifically, we construct a unified paired dataset by synthesizing aligned instances via cross-modal semantic completion and retrieving semantically related samples. These paired structures expose cross-task semantic correspondences and support consistent policy learning. To leverage this structure, we present PairGRPO, a pair-aware variant based on Group Relative Policy Optimization. It assigns a similarity score to each pair to modulate the advantage, strengthening learning from well-aligned examples and reducing task interference. Extensive experiments across diverse UVLM architectures (Autoregressive and Discrete Diffusion) and scales (1B to 14B) demonstrate that PairUni yields consistent improvements over strong baselines. Notably, our method also demonstrates strong generalization by improving performance on image editing tasks without using any editing-specific data. Codes are available at https://github.com/Haochen-Wang409/PairUni. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_25682 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | PairUni: Pairwise Training for Unified Multimodal Language Models Zheng, Jiani Teng, Zhiyang Qiu, Kunpeng Li, Xiangtai Wang, Anran Tian, Yu Tian, Ye Wang, Haochen Wang, Zhuochen Computation and Language Unified Vision-Language Models (UVLMs) perform both understanding and generation within a single architecture. Since these models rely on heterogeneous data and supervision, balancing both generation and understanding in reinforcement learning (RL) is challenging. To address this challenge, we propose PairUni, a unified framework that reorganizes data into understanding-generation (UG) pairs and aligns optimization accordingly. Specifically, we construct a unified paired dataset by synthesizing aligned instances via cross-modal semantic completion and retrieving semantically related samples. These paired structures expose cross-task semantic correspondences and support consistent policy learning. To leverage this structure, we present PairGRPO, a pair-aware variant based on Group Relative Policy Optimization. It assigns a similarity score to each pair to modulate the advantage, strengthening learning from well-aligned examples and reducing task interference. Extensive experiments across diverse UVLM architectures (Autoregressive and Discrete Diffusion) and scales (1B to 14B) demonstrate that PairUni yields consistent improvements over strong baselines. Notably, our method also demonstrates strong generalization by improving performance on image editing tasks without using any editing-specific data. Codes are available at https://github.com/Haochen-Wang409/PairUni. |
| title | PairUni: Pairwise Training for Unified Multimodal Language Models |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2510.25682 |