| title | LoopVL: Recurrent Visual Intelligence |
|---|---|
| authors | Zhe Qian, Ziyang Gong, Zhongxing Xu, Hehan Li, Zhonghua Wang, Fei Luo, Mingxuan Wang, Xue Yang, Shiwei liu, Yanbiao Ma, Junchi Yan, Jungong Han |
| date | 2026-09-29 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2609.38426 |
| affiliation | Renmin University of China |
| upvotes | 47 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| abstract | We introduce LoopVL to study whether Loop Transformers can be effectively extended to vision- language models. LoopVL combines Module-Loop and Model-Loop computation to iteratively update a unified vision-language state through shared modules. We train LoopVL from scratch through language pre-training, multimodal training, and post-training. LoopVL outperforms a range of similarly sized and larger non-recurrent models on multimodal understanding and visual reasoning benchmarks. We also observe Visual Aha Moments in LoopVL, characterized by pronounced shifts in visual attention across loops. LoopVL provides practical evidence for recurrent vision-language modeling and offers an intuitive perspective on how shared parameters can support deeper multimodal computation over continuously evolving visual-language states. |
cite
@misc{qian2026loopvl,
title={LoopVL: Recurrent Visual Intelligence},
author={Zhe Qian and Ziyang Gong and Zhongxing Xu and Hehan Li and Zhonghua Wang and Fei Luo and Mingxuan Wang and Xue Yang and Shiwei liu and Yanbiao Ma and Junchi Yan and Jungong Han},
year={2026},
eprint={2609.38426},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2609.38426}
}