Rajveer Pall

TrustShift: Shift Type, Not Shift Magnitude, Determines Machine-Learning Failure Modes - A Cross-Domain Audit Protocol and a Failure Taxonomy

Deployment evaluation implicitly assumes that a larger distribution shift means greater risk. Across four domains, we find this assumption false. We introduce TrustShift, a single pre-registered protocol measuring discrimination, calibration, and subgroup reliability, and apply i...

Privacy-Preserving Federated Learning for Diabetes Risk Prediction Across Demographically Heterogeneous Hospital Nodes

Background and Objective. Diabetes risk models achieve strong discrimination on centralised cohorts, but centralisation is constrained by privacy regulation, and models trained at a single site risk demographic bias and miscalibration. We evaluate whether federated learning (FL)...