| title | LogT: Logically Think with Images for Visual Search |
|---|---|
| authors | Yanjun Fu, Quanzeng You, Jiadong Guo, Yujie Lu, Sai Charitha Akula, Haiping Wu, Miao Liu, Sanghamitra Dutta, Lu Yuan |
| date | 2026-11-14 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2611.19200 |
| affiliation | AI at Meta |
| upvotes | 13 (Hugging Face Daily Papers, 2026-11-21) |
| links | abs pdf hf |
| cite | |
| related | Towards Physics of Multimodal Pretraining: Knowledge Flow, Modality Synergy, Early Unification, and Recipes TUA-Bench: A Benchmark for General-Purpose Terminal-Use Agents VLM3: Vision Language Models Are Native 3D Learners Realiz3D: 3D Generation Made Photorealistic via Domain-Aware Learning MIMESIS: Learning User Simulators as Training Environments for Interactive Agents |
| abstract | Thinking with images empowers vision-language models (VLMs) to use an external Zoom-in tool for fine-grained visual search. Recent works employ supervised fine-tuning (SFT) and reinforcement learning (RL) with tool-specific designs to enable such capabilities. However, achieving logical reasoning alongside accurate and efficient tool use remains a significant challenge. To address this, we shift the focus away from increasing RL complexity and instead identify SFT data quality as a key bottleneck. We then propose LogT (short for Logically Think with Images), a fully automated SFT data engine that explicitly enforces logic-chain-guided reasoning. LogT generates an SFT dataset featuring challenging, visually grounded questions and structured tool-use traces with human-like logic. We validate our approach by fine-tuning two base models, Qwen2.5-VL-7B and InternVL3.5-8B, on the dataset and evaluating them across seven benchmark test sets. We show that LogT-SFT based on Qwen2.5-VL-7B outperforms the state-of-the-art trained with both SFT and RL, despite using 3.1times less data and >120times fewer GPU hours. Furthermore, applying a simple RL recipe without tool-specific design on LogT-SFT yields a model that surpasses prior methods with complex RL setups by a substantial 9.59-point margin. In addition to strong performance gains, our model exhibits higher tool-use accuracy and efficiency, as well as more structured and logically consistent reasoning traces, directly validating the effectiveness of our approach. |
bibtex
@misc{fu2026logt,
title={LogT: Logically Think with Images for Visual Search},
author={Yanjun Fu and Quanzeng You and Jiadong Guo and Yujie Lu and Sai Charitha Akula and Haiping Wu and Miao Liu and Sanghamitra Dutta and Lu Yuan},
year={2026},
eprint={2611.19200},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2611.19200}
}