Acknowledging the Unknown for Multi-label Learning with Single Positive Labels

TL;DR

This paper introduces a novel approach for multi-label learning with only one positive label per image, treating unannotated labels as unknowns rather than negatives, and employs entropy-maximization and pseudo-labeling techniques to improve performance.

Contribution

It proposes a new framework for single positive multi-label learning that acknowledges unknown labels, using entropy-maximization loss and asymmetric pseudo-labeling for better supervision.

Findings

Significantly outperforms existing methods on four benchmarks.

Achieves state-of-the-art results in single positive multi-label learning.

Demonstrates the effectiveness of treating unannotated labels as unknowns.

Abstract

Due to the difficulty of collecting exhaustive multi-label annotations, multi-label datasets often contain partial labels. We consider an extreme of this weakly supervised learning problem, called single positive multi-label learning (SPML), where each multi-label training image has only one positive label. Traditionally, all unannotated labels are assumed as negative labels in SPML, which introduces false negative labels and causes model training to be dominated by assumed negative labels. In this work, we choose to treat all unannotated labels from an alternative perspective, i.e. acknowledging they are unknown. Hence, we propose entropy-maximization (EM) loss to attain a special gradient regime for providing proper supervision signals. Moreover, we propose asymmetric pseudo-labeling (APL), which adopts asymmetric-tolerance strategies and a self-paced procedure, to cooperate with EM loss and then provide more precise supervision. Experiments show that our method significantly improves performance and achieves state-of-the-art results on all four benchmarks. Code is available at https://github.com/Correr-Zhou/SPML-AckTheUnknown.