{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,4]],"date-time":"2026-08-04T05:43:58Z","timestamp":1785822238496,"version":"3.56.0"},"reference-count":75,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/legalcode"}],"funder":[{"name":"Jilin Province Science and Technology Innovation Platform Construction Project","award":["YDZJ202302CXJD027"],"award-info":[{"award-number":["YDZJ202302CXJD027"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Access"],"published-print":{"date-parts":[[2026]]},"DOI":"10.1109\/access.2026.3716267","type":"journal-article","created":{"date-parts":[[2026,7,23]],"date-time":"2026-07-23T19:10:22Z","timestamp":1784833822000},"page":"115480-115505","source":"Crossref","is-referenced-by-count":0,"title":["Severity Matters: Risk-Calibrated Reinforcement Fine-Tuning for Clinically Aligned Medical Vision\u2013Language Models"],"prefix":"10.1109","volume":"14","author":[{"ORCID":"https:\/\/orcid.org\/0009-0001-0639-3947","authenticated-orcid":false,"given":"Yu","family":"Su","sequence":"first","affiliation":[{"name":"Changchun University of Science and Technology","place":["Changchun, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Peng","family":"Wang","sequence":"additional","affiliation":[{"name":"Changchun University of Science and Technology","place":["Changchun, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-024-03302-1"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.52202\/068431-2011"},{"key":"ref3","article-title":"Toward clinically explainable AI for medical diagnosis: A foundation model with human-compatible reasoning via reinforcement learning","author":"Lin","year":"2025","journal-title":"arXiv:2509.03906"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-032-04981-0_32"},{"key":"ref5","article-title":"Med-r1: Reinforcement learning for generalizable medical reasoning in vision-language models","author":"Lai","year":"2025","journal-title":"arXiv:2503.13939"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1109\/tnn.1998.712192"},{"key":"ref7","article-title":"CheXalign: Preference fine-tuning in chest X-ray interpretation models without human feedback","author":"Hein","year":"2024","journal-title":"arXiv:2410.07025"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1145\/3746027.3754507"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.ijcnlp-short.26"},{"key":"ref10","article-title":"RadVLM: A multitask conversational vision-language model for radiology","author":"Deperrois","year":"2025","journal-title":"arXiv:2502.03333"},{"key":"ref11","article-title":"MedGemma technical report","volume-title":"arXiv:2507.05201","author":"Sellergren","year":"2025"},{"key":"ref12","article-title":"MAIRA-2: Grounded radiology report generation","volume-title":"arXiv:2406.04449","author":"Bannur","year":"2024"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1016\/j.patter.2023.100802"},{"key":"ref14","article-title":"CheXNet: Radiologist-level pneumonia detection on chest X-rays with deep learning","author":"Rajpurkar","year":"2017","journal-title":"arXiv:1711.05225"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v33i01.3301590"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.emnlp-main.460"},{"key":"ref17","article-title":"MMedPO: Aligning medical vision-language models with clinical-aware multimodal preference optimization","volume-title":"Proc. 42nd Int. Conf. Mach. Learn. (ICML)","author":"Zhu"},{"key":"ref18","first-page":"1468","article-title":"Policy gradient for coherent risk measures","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"28","author":"Tamar"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1016\/j.enganabound.2023.10.027"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.32604\/cmes.2022.022088"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.32604\/cmc.2024.050736"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1142\/S0219876221420111"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.acl-long.459"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1016\/j.artmed.2023.102633"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2022.102510"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.01897"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.01954"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.acl-long.451"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.00325"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.01112"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.00718"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1109\/WACV57701.2024.00781"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2024.3421599"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v38i3.28038"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-73001-6_11"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2024.3398350"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v38i17.29826"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-16452-1_54"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1109\/JBHI.2024.3422168"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2023.3343391"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.acl-long.514"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.1109\/TMI.2024.3372638"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-025-09422-z"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.52202\/075280-1240"},{"key":"ref45","first-page":"249","article-title":"Clinically accurate chest X-ray report generation","volume-title":"Proc. 4th Mach. Learn. Healthcare Conf. (MLHC)","volume":"106","author":"Liu"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.findings-emnlp.319"},{"key":"ref47","article-title":"RadHiera: Semantic hierarchical reinforcement learning for medical report generation","volume-title":"arXiv:2511.10065","author":"Du","year":"2025"},{"key":"ref48","first-page":"449","article-title":"A distributional perspective on reinforcement learning","volume-title":"Proc. 34th Int. Conf. Mach. Learn.","author":"Bellemare"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v32i1.11791"},{"key":"ref50","first-page":"2151","article-title":"SelectiveNet: A deep neural network with an integrated reject option","volume-title":"Proc. 36th Int. Conf. Mach. Learn.","author":"Geifman"},{"key":"ref51","first-page":"1321","article-title":"On calibration of modern neural networks","volume-title":"Proc. 34th Int. Conf. Mach. Learn.","author":"Guo"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.5555\/1642194.1642224"},{"key":"ref53","first-page":"1437","article-title":"A comprehensive survey on safe reinforcement learning","volume":"16","author":"Garc\u00eda","year":"2015","journal-title":"J. Mach. Learn. Res."},{"key":"ref54","first-page":"311","article-title":"BLEU: A method for automatic evaluation of machine translation","volume-title":"Proc. 40th Annu. meeting Assoc. Comput. Linguistics","author":"Papineni"},{"key":"ref55","first-page":"74","article-title":"Rouge: A package for automatic evaluation of summaries","volume-title":"Proc. Text Summarization Branches Out","author":"Lin"},{"key":"ref56","article-title":"BERTScore: Evaluating text generation with BERT","volume-title":"Proc. Int. Conf. Learn. Represent.","author":"Zhang"},{"key":"ref57","article-title":"RadGraph: Extracting clinical entities and relations from radiology reports","volume-title":"arXiv:2106.14463","author":"Jain","year":"2021"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-acl.765"},{"key":"ref59","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.emnlp-main.117"},{"key":"ref60","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-emnlp.21"},{"key":"ref61","volume-title":"Constrained Markov Decision Processes","author":"Altman","year":"1999"},{"key":"ref62","first-page":"22","article-title":"Constrained policy optimization","volume-title":"Proc. 34th Int. Conf. Mach. Learn.","author":"Achiam"},{"key":"ref63","article-title":"Policy gradients with variance related risk criteria","volume-title":"Proc. 29th Int. Conf. Mach. Learn.","author":"Di Castro"},{"key":"ref64","doi-asserted-by":"publisher","DOI":"10.2307\/2282952"},{"key":"ref65","article-title":"DeepSeekMath: Pushing the limits of mathematical reasoning in open language models","author":"Shao","year":"2024","journal-title":"arXiv:2402.03300"},{"key":"ref66","article-title":"DAPO: An open-source LLM reinforcement learning system at scale","author":"Yu","year":"2025","journal-title":"arXiv:2503.14476"},{"key":"ref67","doi-asserted-by":"publisher","DOI":"10.1038\/s41597-019-0322-0"},{"key":"ref68","doi-asserted-by":"publisher","DOI":"10.1093\/jamia\/ocv080"},{"key":"ref69","article-title":"LoRA: Low-rank adaptation of large language models","author":"Hu","year":"2021","journal-title":"arXiv:2106.09685"},{"key":"ref70","article-title":"A vision-language foundation model to enhance efficiency of chest X-ray interpretation","author":"Chen","year":"2024","journal-title":"arXiv:2401.12208"},{"key":"ref71","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-025-58344-x"},{"key":"ref72","article-title":"Qwen3-VL technical report","volume-title":"arXiv:2511.21631","author":"Bai","year":"2025"},{"key":"ref73","first-page":"10835","article-title":"Scaling laws for reward model overoptimization","volume-title":"Proc. 40th Int. Conf. Mach. Learn.","author":"Gao"},{"key":"ref74","doi-asserted-by":"publisher","DOI":"10.52202\/068431-0687"},{"key":"ref75","article-title":"Language models (Mostly) know what they know","author":"Kadavath","year":"2022","journal-title":"arXiv:2207.05221"}],"container-title":["IEEE Access"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/6287639\/11323511\/11622461.pdf?arnumber=11622461","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,4]],"date-time":"2026-08-04T04:47:11Z","timestamp":1785818831000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11622461\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":75,"URL":"https:\/\/doi.org\/10.1109\/access.2026.3716267","relation":{},"ISSN":["2169-3536"],"issn-type":[{"value":"2169-3536","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026]]}}}