Metraipolis › archive › dc_fa34a620f51c
Latent Decision Flow Rewards for Clinical Guideline Reasoning — arXiv:2610.38709
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.38709
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-31T06:12:53.812Z
- SHA-256 of the archived text
5c280e1aa29ff54a90a45c32a99ead30b501fce0e2841756f68a8a8578583a3a- Stable URL of this page
https://metraipolis.com/documento/dc_fa34a620f51c
How to cite
Latent Decision Flow Rewards for Clinical Guideline Reasoning — arXiv:2610.38709 — arXiv (en), https://arxiv.org/abs/2610.38709 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fa34a620f51c (Metraipolis archive dc_fa34a620f51c, sha256 5c280e1aa29ff54a90a45c32a99ead30b501fce0e2841756f68a8a8578583a3a)
BibTeX
@misc{dcfa34a620f51c,
title = {Latent Decision Flow Rewards for Clinical Guideline Reasoning — arXiv:2610.38709},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.38709},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fa34a620f51c; sha256 5c280e1aa29ff54a90a45c32a99ead30b501fce0e2841756f68a8a8578583a3a},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.38709. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Clinical practice guidelines encode evidence-based recommendations as conditional decision flows, which makes them a natural source of process supervision for training reasoning models. Two properties of clinical reasoning make this supervision difficult to use. A decision flow consistent with a guideline does not by itself determine the correct answer for a given patient, which we call the path--answer gap, and several valid decision flows can support the same clinical decision, which we call path multiplicity. Current training approaches address parts of this problem. Answer rewards supervise the final decision, and process rewards computed against a reference flow supervise one valid decision flow per case. Supervised fine-tuning on annotated flows teaches guideline-consistent reasoning but constrains the model to one trajectory per case and can reduce answer accuracy. We introduce GUIDER, a method that jointly optimizes a decision flow controller and a reasoning policy. The controller learns to generate plausible decision flows from guideline-derived supervision, and the policy learns to generate a decision flow and a final answer for each patient case. GUIDER computes the process reward for the policy over several controller-generated flows, which accounts for path multiplicity, and gates this reward on answer correctness, which addresses the path--answer gap. Joint training allows the process supervision to evolve with the policy. On MedGUIDE, GUIDER with Qwen3-14B achieves 82.5 weighted accuracy and 73.5 guideline adherence, 3.3 and 3.1 points above the strongest clinical reinforcement learning baseline. With Qwen3-14B and Gemma4-12B backbones, GUIDER also improves over CRPO, MRPO, and FaithMed on the out-of-domain benchmarks AMEGA-LLM, RealMedQA, and CancerGUIDE. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.38709 — 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 2026-10-31T06:07:08.673Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).