Estimation of the Label-Noise Transition Matrix with Performance Guarantees via Selective Classification
Abstract
Modern machine learning depends heavily on massive datasets, but obtaining high-quality annotations at scale is often expensive. As a result, learning from noisily-labeled data has become common, making accurate estimation of the label-noise transition matrix crucial. However, existing transition matrix estimators rely on the fragile estimation of class-posteriors and do not provide finite-sample performance guarantees. In this work, we propose a novel methodology to estimate the transition matrix based on one-sided selective classification. This approach bypasses class-posterior estimation, provides finite-sample performance guarantees, and leverages flexible learning methods for binary classification. Moreover, we introduce effective algorithms to implement the proposed methodology and provide their refined finite-sample performance bounds.
Related
- Robust dimension-free estimation of simple random tensors: optimal guarantees under heavy tails and adversarial contamination
- Mutual Information Estimation via Triangular Transport Maps: Finite Sample Guarantees and Sample Allocation
- Neural Estimation of Summary Statistics for Simulation-Based Inference with Non-Asymptotic Guarantees
- Statistical Guarantees for Covariance Estimation with Adversarial Corruption under Constraints
- Differential Privacy Guarantees in Small Area Estimation
BibTeX
@misc{juan2026estimation,
title = {{Estimation of the Label-Noise Transition Matrix with Performance Guarantees via Selective Classification}},
author = {Xabier de Juan and Santiago Mazuelas and Yilun Zhu and Clayton Scott},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.39829},
eprint = {2609.39829},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.39829},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Estimation of the Label-Noise Transition Matrix with Performance Guarantees via Selective Classification
AU - Xabier de Juan
AU - Santiago Mazuelas
AU - Yilun Zhu
AU - Clayton Scott
PY - 2026
DA - 2026/09/30
PB - arXiv
DO - 10.48550/arxiv.2609.39829
UR - https://arxiv.org/abs/2609.39829
AB - Modern machine learning depends heavily on massive datasets, but obtaining high-quality annotations at scale is often expensive. As a result, learning from noisily-labeled data has become common, making accurate estimation of the label-noise transition matrix crucial. However, existing transition matrix estimators rely on the fragile estimation of class-posteriors and do not provide finite-sample performance guarantees. In this work, we propose a novel methodology to estimate the transition matrix based on one-sided selective classification. This approach bypasses class-posterior estimation, provides finite-sample performance guarantees, and leverages flexible learning methods for binary classification. Moreover, we introduce effective algorithms to implement the proposed methodology and provide their refined finite-sample performance bounds.
ER -