1 paper
Xingyuan Li, Mengyue Wu
Detecting medical conditions from speech acoustics is fundamentally a weakly-supervised learning problem: a single, often noisy, session-level label must be linked to nuanced patte…