Roshan Prakash Rane

ICON decomposition reveals the true reliance of deep models on concepts, debunking misleading correlations that traditional methods often overlook.

A novel deep shape regression model reveals covariate effects in hippocampal outlines that align with established neuroimaging findings, showcasing its potential in health applications.

SSL embeddings, typically superior for speech tasks, surprisingly lose their edge to hand-crafted acoustic features when classifying mild cognitive impairment from speech, challenging assumptions about representation learning in clinical applications.