Partial AUC Maximization from Positive-unlabeled Data

Name: Partial AUC Maximization from Positive-unlabeled Data
Uploaded: 2026-10-04T08:35:11.755090+00:00
Duration: 681 s
Description: The method maximizes partial Area Under the ROC Curve (pAUC) using only positive and marginal densities from positive-unlabeled data, avoiding the need for labeled negative samples. It reformulates pAUC using these densities to create an empirical estimator that allows classifier training without requiring negative labels.

The method maximizes partial Area Under the ROC Curve (pAUC) using only positive and marginal densities from positive-unlabeled data, avoiding the need for labeled negative samples. It reformulates pAUC using these densities to create an empirical estimator that allows classifier training without requiring negative labels.

Open the full episode