Pairwise Instance Relation Augmentation for Long-tailed Multi-label Text Classification

TL;DR

This paper introduces PIRAN, a novel data augmentation method for long-tailed multi-label text classification that generates diverse instances for underrepresented tail labels, significantly improving classification performance.

Contribution

The paper proposes PIRAN, a relation-based augmentation network that balances tail and head labels by generating high-level feature instances with regularizers to ensure diversity and consistency.

Findings

PIRAN outperforms state-of-the-art methods on three benchmark datasets.

It significantly improves tail label classification accuracy.

The approach effectively balances label distribution in long-tailed datasets.

Abstract

Multi-label text classification (MLTC) is one of the key tasks in natural language processing. It aims to assign multiple target labels to one document. Due to the uneven popularity of labels, the number of documents per label follows a long-tailed distribution in most cases. It is much more challenging to learn classifiers for data-scarce tail labels than for data-rich head labels. The main reason is that head labels usually have sufficient information, e.g., a large intra-class diversity, while tail labels do not. In response, we propose a Pairwise Instance Relation Augmentation Network (PIRAN) to augment tailed-label documents for balancing tail labels and head labels. PIRAN consists of a relation collector and an instance generator. The former aims to extract the document pairwise relations from head labels. Taking these relations as perturbations, the latter tries to generate new document instances in high-level feature space around the limited given tailed-label instances. Meanwhile, two regularizers (diversity and consistency) are designed to constrain the generation process. The consistency-regularizer encourages the variance of tail labels to be close to head labels and further balances the whole datasets. And diversity-regularizer makes sure the generated instances have diversity and avoids generating redundant instances. Extensive experimental results on three benchmark datasets demonstrate that PIRAN consistently outperforms the SOTA methods, and dramatically improves the performance of tail labels.