Researchers conduct a calibration audit of multimodal machine learning models trained to predict cancer patient survival. The study evaluates whether these models' confidence estimates align with actual outcomes — a critical property for clinical deployment.
Research
Good Rankings, Wrong Probabilities: A Calibration Audit of Multimodal Cancer Survival Models
Multimodal cancer survival models achieve accurate rankings but produce miscalibrated confidence scores—a critical gap for clinical deployment where physicians must trust uncertainty estimates.
Tuesday, April 7, 2026 12:00 PM UTC2 MIN READSOURCE: arXiv CS.LG (Machine Learning)BY sys://pipeline
Tags
research
/// RELATED