Synthetic Curriculum Reinforces Compositional Text-to-Image Generation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911283041796096 |
|---|---|
| author | Wang, Shijian Fu, Runhao Zhao, Siyi Zhan, Qingqin Wang, Xingjian Jin, Jiarui Lu, Yuan Wu, Hanqian Chen, Cunjian |
| author_facet | Wang, Shijian Fu, Runhao Zhao, Siyi Zhan, Qingqin Wang, Xingjian Jin, Jiarui Lu, Yuan Wu, Hanqian Chen, Cunjian |
| contents | Text-to-Image (T2I) generation has long been an open problem, with compositional synthesis remaining particularly challenging. This task requires accurate rendering of complex scenes containing multiple objects that exhibit diverse attributes as well as intricate spatial and semantic relationships, demanding both precise object placement and coherent inter-object interactions. In this paper, we propose a novel compositional curriculum reinforcement learning framework named CompGen that addresses compositional weakness in existing T2I models. Specifically, we leverage scene graphs to establish a novel difficulty criterion for compositional ability and develop a corresponding adaptive Markov Chain Monte Carlo graph sampling algorithm. This difficulty-aware approach enables the synthesis of training curriculum data that progressively optimize T2I models through reinforcement learning. We integrate our curriculum learning approach into Group Relative Policy Optimization (GRPO) and investigate different curriculum scheduling strategies. Our experiments reveal that CompGen exhibits distinct scaling curves under different curriculum scheduling strategies, with easy-to-hard and Gaussian sampling strategies yielding superior scaling performance compared to random sampling. Extensive experiments demonstrate that CompGen significantly enhances compositional generation capabilities for both diffusion-based and auto-regressive T2I models, highlighting its effectiveness in improving the compositional T2I generation systems. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_18378 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Synthetic Curriculum Reinforces Compositional Text-to-Image Generation Wang, Shijian Fu, Runhao Zhao, Siyi Zhan, Qingqin Wang, Xingjian Jin, Jiarui Lu, Yuan Wu, Hanqian Chen, Cunjian Computer Vision and Pattern Recognition Text-to-Image (T2I) generation has long been an open problem, with compositional synthesis remaining particularly challenging. This task requires accurate rendering of complex scenes containing multiple objects that exhibit diverse attributes as well as intricate spatial and semantic relationships, demanding both precise object placement and coherent inter-object interactions. In this paper, we propose a novel compositional curriculum reinforcement learning framework named CompGen that addresses compositional weakness in existing T2I models. Specifically, we leverage scene graphs to establish a novel difficulty criterion for compositional ability and develop a corresponding adaptive Markov Chain Monte Carlo graph sampling algorithm. This difficulty-aware approach enables the synthesis of training curriculum data that progressively optimize T2I models through reinforcement learning. We integrate our curriculum learning approach into Group Relative Policy Optimization (GRPO) and investigate different curriculum scheduling strategies. Our experiments reveal that CompGen exhibits distinct scaling curves under different curriculum scheduling strategies, with easy-to-hard and Gaussian sampling strategies yielding superior scaling performance compared to random sampling. Extensive experiments demonstrate that CompGen significantly enhances compositional generation capabilities for both diffusion-based and auto-regressive T2I models, highlighting its effectiveness in improving the compositional T2I generation systems. |
| title | Synthetic Curriculum Reinforces Compositional Text-to-Image Generation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2511.18378 |