Medical Diagnosis
Overview
Medical diagnosis studies how to combine symptoms, records, laboratory tests, medical images, physiological signals, and molecular data to identify disease and support clinical decisions. Methods range from conventional machine learning and deep models such as CNNs and graph networks to multimodal foundation models, retrieval-augmented language models, and agents that ask questions or request tests as evidence accumulates. Systems are evaluated on tasks including image and signal classification, electronic-record prediction, and conversational diagnosis, using resources such as MIMIC and ADNI alongside disease-specific and external-site cohorts. A central challenge is that strong benchmark results may not transfer to real clinics, where data are incomplete or shifted and models must provide calibrated, evidence-grounded recommendations without worsening bias or unsafe decisions.
Latest papers
263 papers26 papers in the last four weeks, down 13% on the four weeks before. 0.2% of all new papers.