As artificial intelligence systems move toward clinical deployment, ensuring reliable prediction behavior is fundamental for safety-critical decision-making tasks. One proposed safeguard is selective prediction, where models can defer uncertain predictions to human experts for review. In this work, we empirically evaluate the reliability of uncertainty-based selective prediction in multilabel clinical condition classification using multimodal ICU data. Across a range of state-of-the-art unimodal and multimodal models, we find that selective prediction can substantially degrade performance despite strong standard evaluation metrics. This failure is driven by severe class-dependent miscalibration, whereby models assign high uncertainty to correct predictions and low uncertainty to incorrect ones, particularly for underrepresented clinical conditions. Our results show that commonly used aggregate metrics can obscure these effects, limiting their ability to assess selective prediction behavior in this setting. Taken together, our findings characterize a task-specific failure mode of selective prediction in multimodal clinical condition classification and highlight the need for calibration-aware evaluation to provide strong guarantees of safety and robustness in clinical AI.
Attention-Based Synthetic Data Generation for Calibration-Enhanced Survival Analysis: A Case Study for Chronic Kidney Disease Using Electronic Health Records
Centre for Big Data Research in Health (CBDRH)(健康大数据研究中心);The University of New South Wales(新南威尔士大学)
Calibrating Bayesian Learning via Regularization, Confidence Minimization, and Selective Inference
King’s Communications, Learning & Information Processing (KCLIP) lab(国王通讯、学习与信息处理(KCLIP)实验室);Centre for Intelligent Information Processing Systems (CIIPS)(智能信息处理系统中心);Department of Engineering, King’s College London(伦敦国王学院工程系)
Interpretable and Calibrated Classification of Clinical Data Using Supervised Feature Binarization
Worcester Polytechnic Institute(伍斯特理工学院);Centro de Vacunación e Investigación (CEVAXIN)(疫苗接种与研究中心(CEVAXIN));Universidad Tecnológica de Panamá(巴拿马技术大学);Massachusetts Institute of Technology(麻省理工学院);Broad Institute of MIT and Harvard(麻省理工学院和哈佛大学布罗德研究所);McGill University(麦吉尔大学);The Montreal Neurological Hospital-Institute(蒙特利尔神经学医院研究所)
Partial Causal Structure Learning for Valid Selective Conformal Inference under Interventions
Department of Biostatistics, Vanderbilt University Medical Center(生物统计学系,范德比尔特大学医学中心);Department of Informatics, King’s College London(信息学系,伦敦国王学院);Department of Biostatistics, MD Anderson Cancer Center(生物统计学系,MD安德森癌症中心)
A Hybrid Machine Learning Approach for Synthetic Data Generation with Post Hoc Calibration for Clinical Tabular Datasets
organization= Department of Electrical \& Electronic Engineering, Bangladesh University of Business;organization= Department of Electrical Engineering;Computer Science, University of Michigan , country= United States
Tipping the Balance: Impact of Class Imbalance Correction on the Performance of Clinical Risk Prediction Models
Aalborg University(奥尔堡大学);Novo Nordisk A/S(诺和诺德公司);Bowling Green State University(鲍灵格林州立大学);Heidelberg University(海德堡大学);HMS Analytical Software GmbH(HMS分析软件有限公司);DKFZ Hector Cancer Institute(德国癌症研究中心赫克托癌症研究所);Gødstrup Hospital(戈兹楚普医院)
ReliableNet: A Chance-Constrained Approach to Trustworthy Classification in Deep Learning
School of Agriculture and Science, University of KwaZulu-Natal(夸祖鲁-纳塔尔大学农业与科学学院);AIMS Research and Innovation Centre, African Institute for Mathematical Sciences(非洲数学科学研究所AIMS研究与创新中心);African Institute for Mathematical Sciences(非洲数学科学研究所)
Selective Learning: Towards Robust Calibration with Dynamic Regularization
College of Intelligence and Computing, Tianjin University(天津大学智能与计算学部);Rutgers University(罗格斯大学);A*STAR Centre for Frontier AI Research (CFAR)(新加坡科技研究局前沿人工智能研究中心)
Department of Computer Science, New Jersey Institute of Technology(计算机科学系,新泽西理工学院);Department of Mathematical Sciences, New Jersey Institute of Technology(数学科学系,新泽西理工学院)
University College London(伦敦大学学院);NIHR University College London Hospitals Biomedical Research Centre(英国国家健康研究所伦敦大学学院医院生物医学研究中心);McGill University(麦吉尔大学)
Scalable Clinical Data Infrastructure and Comparative ML Evaluation for Hospitalisation Risk Prediction in Elderly Patients with Multiple Long-Term Conditions using CPRD
School of Information, University of Sheffield(谢菲尔德大学信息学院);School of Medicine, Faculty of Medicine and Health, University of Leeds(利兹大学医学与健康学院医学院);Division of Informatics, University of Manchester(曼彻斯特大学信息学系);Academic Unit for Ageing & Stroke Research, University of Leeds(利兹大学衰老与中风研究学术单元);Institute of Population Health, University of Liverpool(利物浦大学人口健康研究所)
A causal viewpoint on prediction model performance under changes in case-mix: discrimination and calibration respond differently for prognosis and diagnosis predictions
Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht University, The Netherlands, Department of Data Science and Biostatistics(朱利叶斯健康科学与初级保健中心,乌得勒支大学医学中心,乌得勒支大学,荷兰,数据科学与生物统计学系)
Data Alchemy: Mitigating Cross-Site Model Variability Through Test Time Data Calibration
Children’s National Hospital, Washington, DC, USA(儿童医院国家医院);Universidad Politécnica de Madrid, Madrid, Spain(马德里理工大学);Universitat Pompeu Fabra, Barcelona, Spain(庞培法布拉大学);Nvidia Corporation, Santa Clara, CA, USA(英伟达公司);George Washington University, Washington, DC, USA(乔治华盛顿大学)
Imputation-free transformer learning enables robust Alzheimer's disease prediction and calibrated uncertainty quantification across heterogeneous clinical cohorts
Platform of Bioinformatics(生物信息平台);Lausanne University Hospital(洛桑大学医院);Faculty of Biology and Medicine(生物学与医学学院);University of Lausanne(洛桑大学);Leenaards Memory Centre(Leenaards记忆中心);Department of Clinical Neurosciences(临床神经科学系)