ESA: Example Sieve Approach for Multi-Positive and Unlabeled Learning

TL;DR

This paper introduces ESA, a novel example sieve method for multi-positive and unlabeled learning that improves classifier training by selecting examples based on Certain Loss values, addressing risk shift issues.

Contribution

The paper proposes ESA, a new example selection approach using Certain Loss to mitigate risk shift in MPU learning, with proven optimal convergence and superior experimental performance.

Findings

ESA outperforms previous methods on real-world datasets.

The risk estimator in ESA is consistent and converges optimally.

Experimental results demonstrate improved classifier accuracy.

Abstract

Learning from Multi-Positive and Unlabeled (MPU) data has gradually attracted significant attention from practical applications. Unfortunately, the risk of MPU also suffer from the shift of minimum risk, particularly when the models are very flexible as shown in Fig.\ref{moti}. In this paper, to alleviate the shifting of minimum risk problem, we propose an Example Sieve Approach (ESA) to select examples for training a multi-class classifier. Specifically, we sieve out some examples by utilizing the Certain Loss (CL) value of each example in the training stage and analyze the consistency of the proposed risk estimator. Besides, we show that the estimation error of proposed ESA obtains the optimal parametric convergence rate. Extensive experiments on various real-world datasets show the proposed approach outperforms previous methods.