Towards Imbalanced Large Scale Multi-label Classification with Partially Annotated Labels

TL;DR

This paper proposes a novel approach for large-scale multi-label classification with partial labels, addressing label imbalance through pseudo-labeling, a new loss function, and dynamic training, achieving superior results on benchmark datasets.

Contribution

It introduces a pseudo-labeling technique, a new imbalance-aware loss function, and a dynamic training scheme for effective learning with partial labels in large label spaces.

Findings

Outperforms state-of-the-art methods on multiple datasets.

In some cases, surpasses full-label training performance.

Effectively mitigates label imbalance and missing labels issues.

Abstract

Multi-label classification is a widely encountered problem in daily life, where an instance can be associated with multiple classes. In theory, this is a supervised learning method that requires a large amount of labeling. However, annotating data is time-consuming and may be infeasible for huge labeling spaces. In addition, label imbalance can limit the performance of multi-label classifiers, especially when some labels are missing. Therefore, it is meaningful to study how to train neural networks using partial labels. In this work, we address the issue of label imbalance and investigate how to train classifiers using partial labels in large labeling spaces. First, we introduce the pseudo-labeling technique, which allows commonly adopted networks to be applied in partially labeled settings without the need for additional complex structures. Then, we propose a novel loss function that leverages statistical information from existing datasets to effectively alleviate the label imbalance problem. In addition, we design a dynamic training scheme to reduce the dimension of the labeling space and further mitigate the imbalance. Finally, we conduct extensive experiments on some publicly available multi-label datasets such as COCO, NUS-WIDE, CUB, and Open Images to demonstrate the effectiveness of the proposed approach. The results show that our approach outperforms several state-of-the-art methods, and surprisingly, in some partial labeling settings, our approach even exceeds the methods trained with full labels.