Metraipolis › archive › dc_c5d8cc673eb2

DR Pixel: Page-Native Deep Research with Vision-Language Models — arXiv:2611.03097

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.03097
License (as archived)
CC0-1.0
Archived
2026-11-03T18:24:43.738Z
SHA-256 of the archived text
44dc5ab7c6708def120a3a76921399b19f39a1b3aa9b2df6dafb3abc17c701b6
Stable URL of this page
https://metraipolis.com/documento/dc_c5d8cc673eb2

How to cite

DR Pixel: Page-Native Deep Research with Vision-Language Models — arXiv:2611.03097 — arXiv (en), https://arxiv.org/abs/2611.03097 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c5d8cc673eb2 (Metraipolis archive dc_c5d8cc673eb2, sha256 44dc5ab7c6708def120a3a76921399b19f39a1b3aa9b2df6dafb3abc17c701b6)
BibTeX
@misc{dcc5d8cc673eb2,
  title = {DR Pixel: Page-Native Deep Research with Vision-Language Models — arXiv:2611.03097},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.03097},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c5d8cc673eb2; sha256 44dc5ab7c6708def120a3a76921399b19f39a1b3aa9b2df6dafb3abc17c701b6},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.03097. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Complex documents such as scientific papers contain rich information in tables, figures, and equations; however, existing deep research (DR) agents typically reduce papers to OCR-extracted text. We instead ask whether vision-language models (VLMs) can enable page-native deep research: directly retrieving and reasoning over rendered document page images. We introduce DR Pixel, the first open-source end-to-end multimodal DR agent that uses rendered page images as its sole retrieval and reasoning unit, operating over a corpus of 1.06M pages. To evaluate this setting, we construct DRPixelBench, a benchmark on scientific papers requiring reasoning over text, figures, tables, and equations. Surprisingly, page-native DR already works well without training, outperforming OCR-based baselines across state-of-the-art models like GPT-5.4, Gemini-3.5-Flash, and Qwen3.6-27B, while requiring fewer tool calls. We next turn to DR training. A key challenge is that teacher trajectories for agentic supervised finetuning are noisy, mixing critical reasoning and action steps with redundant or failed attempts. To address this, we introduce counterfactual credit reweighting, a fine-grained credit assignment method that weights individual reasoning and action steps by their counterfactual contribution to the final answer. Our training improves retrieval recall by about 15 points and final accuracy by over 10 points on DRPixelBench, and likewise outperforms the untrained baseline on four existing benchmarks. On out-of-distribution ViDoRe-V3, DR Pixel comes within 1.8 points of an oracle that is given the gold pages directly. Our results suggest that the conventional OCR-then-retrieve pipeline is not only unnecessary but suboptimal: directly retrieving and reasoning over visual document pages makes complex DR both more effective and more efficient.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Computation and Language (cs.CL); Information Retrieval (cs.IR).
Pagina del lavoro: https://arxiv.org/abs/2611.03097 — 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 2026-11-03T18:10:08.058Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).