Unbiased risk estimators can mislead: A case study of learning with complementary labels
Journal
Proceedings of Machine Learning Research
Journal Volume
PartF168147-3
Pages
1907-1916
ISSN
26403498
Date Issued
2020
Author(s)
Abstract
In weakly supervised learning, unbiased risk estimator (URE) is a powerful tool for training classifiers when training and test data are drawn from different distributions. Nevertheless, UREs lead to overfitting in many problem settings when the models are complex like deep networks. In this paper, we investigate reasons for such overfitting by studying a weakly supervised problem called learning with complementary labels. We argue the quality of gradient estimation matters more in risk minimization. Theoretically, we show that a URE gives an unbiased gradient estimator (UGE). Practically, however, UGEs may suffer from huge variance, which causes empirical gradients to be usually far away from true gradients during minimization. To this end, we propose a novel surrogate complementary loss (SCL) framework that trades zero bias with reduced variance and makes empirical gradients more aligned with true gradients in the direction. Thanks to this characteristic, SCL successfully mitigates the overfitting issue and improves URE-based methods.
Event(s)
37th International Conference on Machine Learning, ICML 2020
Subjects
Risk perception
Supervised learning
Different distributions
Gradient estimation
Gradient estimator
Overfitting
Risk minimization
Unbiased risk estimator
Weakly supervised learning
Zero bias
Learning systems
Publisher
ML Research Press
Type
conference paper