Neyman-Pearson multiclass classification under label noise via empirical likelihood
Qiong Zhang · 2026
Read Neyman-Pearson multiclass classification under label noise via empirical likelihood by Qiong Zhang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-21623.txt.
Introduction
Zhang and co-authors extend Neyman-Pearson classification to training data with noisy labels. This framework caps class-specific error rates, which matters when errors cost very different amounts. Older methods assume clean labels, and noise can make the target values impossible to recover. A density ratio model with exponential tilting restores recovery, they show. Their method then estimates class shares, posteriors, and the noise process from noisy data alone. No map of label errors is needed in advance. Tests came close to oracle scores.
Catalog record
| CREATOR | Qiong Zhang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 172426 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.21623