Xuefeng Du

Examining how in-distribution labels impact out-of-distribution detection in machine learning.

A new method integrates human input to enhance OOD learning for machine learning models.

A new framework improves detection of false outputs in language models using unlabeled data.