| title | Evidence-RL: Towards Evidence-intensive Visual Reasoning |
|---|---|
| authors | Haojie Huang, Xinlei Yu, Chengming Xu, Zhangquan Chen, Cheng Yang, Qingdong He, Yu Yang, Jiangning Zhang, Xiaobin Hu |
| date | 2026-08-08 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2608.08021 |
| upvotes | 14 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| abstract | Vision-Language Models (VLMs) should answer from concrete image evidence rather than language priors, dataset shortcuts, or irrelevant visual context. Existing perception-aware post-training methods encourage image use through global perturbations or attention proxies, but they do not test whether a sampled answer causally depends on the local evidence that supports it. We propose Counterfactual Evidence Disentanglement (CED), a training-time evidence audit for VLM grounding. For each response, CED neutralizes an object-centric Evidence Region and compares the resulting support drop against matched non-evidence Regions. We combine this signal with answer correctness inside GRPO, rewarding correct answers that rely on the evidence path rather than shortcut or nuisance paths. CED uses weak object-level proposals, requires no question-specific evidence annotations, and adds no inference-time overhead. Across nine public benchmarks and four backbones, CED outperforms prior RL-based post-training methods, with targeted analyses verifying its object-centric signal. |
cite
@misc{huang2026evidence,
title={Evidence-RL: Towards Evidence-intensive Visual Reasoning},
author={Haojie Huang and Xinlei Yu and Chengming Xu and Zhangquan Chen and Cheng Yang and Qingdong He and Yu Yang and Jiangning Zhang and Xiaobin Hu},
year={2026},
eprint={2608.08021},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2608.08021}
}