Metraipolis › archive › dc_7df71e24dc12
MedAgent-R1: Faithfulness-Aware Reinforcement Learning for Evidence-Grounded Medical Reasoning — arXiv:2608.30676
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.30676
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T07:44:57.953Z
- SHA-256 of the archived text
ec5fbfc7b49066818bdfad2265486048e309d4b4606c90024a8dc076c74bea51- Stable URL of this page
https://metraipolis.com/documento/dc_7df71e24dc12
How to cite
MedAgent-R1: Faithfulness-Aware Reinforcement Learning for Evidence-Grounded Medical Reasoning — arXiv:2608.30676 — arXiv (en), https://arxiv.org/abs/2608.30676 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7df71e24dc12 (Metraipolis archive dc_7df71e24dc12, sha256 ec5fbfc7b49066818bdfad2265486048e309d4b4606c90024a8dc076c74bea51)
BibTeX
@misc{dc7df71e24dc12,
title = {MedAgent-R1: Faithfulness-Aware Reinforcement Learning for Evidence-Grounded Medical Reasoning — arXiv:2608.30676},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.30676},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7df71e24dc12; sha256 ec5fbfc7b49066818bdfad2265486048e309d4b4606c90024a8dc076c74bea51},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.30676. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
When medical AI systems hallucinate clinical reasoning, the consequences extend beyond incorrect answers: fabricated justifications that superficially reference retrieved evidence can mislead clinicians into unsafe treatment decisions. Medical reasoning agents must therefore produce not only correct answers but also faithful justifications that clinicians can verify against cited evidence. We identify a systematic failure mode in RL-trained retrieval agents: outcome-only rewards improve accuracy while degrading faithfulness, a phenomenon we term confident hallucination. The agent learns to answer from parametric memory and backfill plausible but unsupported justifications; citation fabrication rates rise from 16.5% to 31.8% even as accuracy improves by 5 points over the supervised baseline. We address this with a faithfulness-gated reward design: accuracy credit is conditioned on evidence grounding via a hard gate, complemented by retrieval validity and conciseness signals that close exploitation paths unique to agentic retrieval. The resulting system, MedAgent-R1, reduces citation fabrication from 31.8% to 4.7% and raises evidence completeness from 58.7 to 82.6 while maintaining 75.1% accuracy, with 13.2-point gains on HealthBench Safety. Under the same agentic retrieval setup, MedAgent-R1 outscores GPT-4o on faithfulness-specific dimensions (Factual Support 4.55 vs. 4.25; Overclaiming 4.40 vs. 4.15) while remaining below GPT-4o in overall accuracy, suggesting that explicit faithfulness training yields evidence-grounding gains not achieved by scaling alone. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-01 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2608.30676 — 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 2026-10-17T07:28:26.619Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).