Multimodal model improves image generation by learning from itself
UniEvo-VL: An On-policy Self-Distillation Training Recipe for Multimodal Model Self-improvement
Artificial IntelligenceComputer Vision and Pattern Recognition
Summary
Some AI models can both understand and create content using images and text together. The authors developed a new method where the model teaches itself by comparing its initial answers to feedback it generates on its own answers. This approach helps the model get better at generating images without needing an external teacher or guide. Testing showed that this self-teaching method improved performance on image generation tasks.
What this means in practice
- •For ai developers: Improve multimodal models’ image generation by integrating self-feedback during runtime without extra labeled data.
- •For content creation platforms: Enhance visual content generation tools by adopting models that self-correct and improve quality through internal feedback.$Commercial implications: Enables selling advanced AI image generation services with higher quality and less reliance on external supervision.
Authors
Fang Wu, Da Xing, Yanjie Huang, Junxi Wang, Ji Wang, Hejia Geng, Guancheng Wan, Bowen Zuo, Xiaomin Li, Shixiang Tang, Xinyu Xiang, Zehong Wang, Shiyi Du, Peng Xia, Shuangjia Zheng, Yining Hong, Li Erran Li, Jure Leskovec, Yejin Choi
Abstract
Modern multimodal models bring generation and understanding into a single unified system, which enables them to provide and learn from their own feedback. Motivated by this unified capacity, we introduce UniEvo-VL, a self-evolving framework for multimodal models to learn from this constructive self-correction feedback during test-time compute. Instead of relying on a separate, often larger, teacher, we leverage their self-critiques as privileged information and ask a single multimodal model to act as both teacher and student with different contexts. The student only sees the vanilla question, while the teacher conditions on the privileged critique. Then training minimizes the per-state divergence between their denoising diffusion distributions over the student's own sampling trajectories. Experiments demonstrate that UniEvo-VL improves the image generation capabilities of multimodal models, while maintaining their sensitivity to additional reflection information. Specifically, we build on top of the open-source Qwen-image-2512 and observe a significant performance gain from 0.747 to 0.808 on GenEval and from 32.97 to 35.53 on GenEval2 Soft-TIFA. Moreover, attempts with more powerful external critics (e.g., GPT5.6-Luna) show that multimodal models with strong judge capabilities can anticipate a higher self-evolving ceiling. Last but not least, mixed text-rendering outcomes show that our self-improvements may not be uniform across different tasks. Our study aims to shed light on the current hot recursive self-improvement research line to enhance the user experience when using multimodal models without external supervision or guidance.