Metraipolis › archive › dc_a68fd15e905b
On-Policy Visual Evidence Distillation — arXiv:2609.36838
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.36838
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T16:26:25.109Z
- SHA-256 of the archived text
8b7e60f44f65508b60ef5e60ca8fe7f3479f48787847bdf0da4a7ec7a3698c0b- Stable URL of this page
https://metraipolis.com/documento/dc_a68fd15e905b
How to cite
On-Policy Visual Evidence Distillation — arXiv:2609.36838 — arXiv (en), https://arxiv.org/abs/2609.36838 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a68fd15e905b (Metraipolis archive dc_a68fd15e905b, sha256 8b7e60f44f65508b60ef5e60ca8fe7f3479f48787847bdf0da4a7ec7a3698c0b)
BibTeX
@misc{dca68fd15e905b,
title = {On-Policy Visual Evidence Distillation — arXiv:2609.36838},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.36838},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a68fd15e905b; sha256 8b7e60f44f65508b60ef5e60ca8fe7f3479f48787847bdf0da4a7ec7a3698c0b},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.36838. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Visual agents solve problems by interleaving reasoning with image operations, and on-policy distillation (OPD) provides guidance from a strong teacher on student-generated interaction trajectories. However, image operations change the evidence available for subsequent reasoning, so local errors in evidence acquisition (Acquire), reading (Read), or answer grounding (Ground) can propagate through the trajectory and lead to incorrect answers. Existing multimodal OPD methods primarily construct or contrast auxiliary views of the original image to strengthen supervision, without explicitly modeling the connections between student actions, resulting observations, and subsequent reasoning. This limits their ability to provide corrections tailored to different failure stages. We introduce Reflection on Visual Evidence (ReVuE), an on-policy distillation method for visual agents. ReVuE compares multiple student-generated trajectories for the same query, summarizes the observed visual evidence, and diagnoses the first failure across the Acquire, Read, and Ground stages. The resulting reflections provide training-time context for the teacher. We group and reweight token-level distillation losses according to how strongly these reflections affect the teacher's predictions. This design translates trajectory-level evidence diagnosis into targeted token-level supervision, guiding students to improve their visual evidence acquisition and reasoning. Across 11 benchmarks spanning the Qwen2.5-VL and InternVL3.5 model families, ReVuE outperforms all evaluated OPD baselines in weighted-average scores for perception, mathematical reasoning, and general tasks. ReVuE also reduces redundancy in reasoning and tool calls while improving tool-call accuracy and task accuracy. Code is available at this https URL Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.AI). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI); Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2609.36838 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-07T15:54:53.894Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).