Detecting Corrupted Labels Without Training a Model to Predict.

Zhaowei Zhu, Zihao Dong, Yang Liu

Browse the full ICML paper archive.