Metraipolis › archive › dc_7674d6ee5ce7
ConRub-Med: Reinforcement Learning with Consensus Rubrics for Open-Ended Medical Question Answering — arXiv:2608.10996
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.10996
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T06:14:24.802Z
- SHA-256 of the archived text
0b63fca332d652cf86a9ba089b18ac71aea1b007078400809bea79d4b987c9fb- Stable URL of this page
https://metraipolis.com/documento/dc_7674d6ee5ce7
How to cite
ConRub-Med: Reinforcement Learning with Consensus Rubrics for Open-Ended Medical Question Answering — arXiv:2608.10996 — arXiv (en), https://arxiv.org/abs/2608.10996 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7674d6ee5ce7 (Metraipolis archive dc_7674d6ee5ce7, sha256 0b63fca332d652cf86a9ba089b18ac71aea1b007078400809bea79d4b987c9fb)
BibTeX
@misc{dc7674d6ee5ce7,
title = {ConRub-Med: Reinforcement Learning with Consensus Rubrics for Open-Ended Medical Question Answering — arXiv:2608.10996},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.10996},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7674d6ee5ce7; sha256 0b63fca332d652cf86a9ba089b18ac71aea1b007078400809bea79d4b987c9fb},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.10996. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Reinforcement learning with verifiable rewards has been especially effective in mathematics and coding, where answers can be checked automatically. Many open-ended medical questions lack comparably cheap outcome verifiers: responses may be partly correct, incomplete, or contain clinically consequential errors. Rubrics written or validated by physicians offer strong clinical grounding, but involving experts in every instance is costly. Model-generated rubrics make this supervision scalable. We introduce ConRub-Med to preserve useful distinctions as rubric feedback moves from construction to policy optimization. For each prompt, three heterogeneous language models propose atomic criteria independently; a separate model reviews them, retaining only criteria with semantic support from all three generators. Three-State scoring distinguishes correct coverage, missing information, and incorrect claims. Errors receive negative rather than zero credit. When every response in a complete Group Relative Policy Optimization (GRPO) group receives the same final reward, a pairwise judge provides sequence advantages only if both candidate orders agree, without changing the scalar rewards. Groups without ties use vanilla GRPO. In a blinded study matched by question, two medical experts rate panels from the full pipeline as more clinically relevant than panels produced by one generator. Across the evaluated open models, ConRub-Med ranks first on six of nine benchmarks and achieves the highest medical and generalization averages. Using the resulting rubric dataset of 5,166 prompts, it scores $38.98 \pm 1.04$ (mean $\pm$ SD) on HealthBench-Hard, compared with InfiMed-ORBIT's 33.60 with 8,000 samples and 37.30 with 28,000. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-12 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2608.10996 — 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 2026-10-19T06:08:50.337Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-12?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).