Weakly-supervised Text Classification Based on Keyword Graph

TL;DR

This paper introduces ClassKG, a weakly-supervised text classification framework that leverages keyword correlations via GNNs, improving pseudo-labeling and classification accuracy.

Contribution

It proposes a novel keyword graph approach with GNNs and self-supervised pretraining to exploit keyword correlations for better weakly-supervised text classification.

Findings

Outperforms existing methods on various datasets

Effectively models keyword correlations with GNNs

Improves pseudo-label quality through self-supervised pretraining

Abstract

Weakly-supervised text classification has received much attention in recent years for it can alleviate the heavy burden of annotating massive data. Among them, keyword-driven methods are the mainstream where user-provided keywords are exploited to generate pseudo-labels for unlabeled texts. However, existing methods treat keywords independently, thus ignore the correlation among them, which should be useful if properly exploited. In this paper, we propose a novel framework called ClassKG to explore keyword-keyword correlation on keyword graph by GNN. Our framework is an iterative process. In each iteration, we first construct a keyword graph, so the task of assigning pseudo labels is transformed to annotating keyword subgraphs. To improve the annotation quality, we introduce a self-supervised task to pretrain a subgraph annotator, and then finetune it. With the pseudo labels generated by the subgraph annotator, we then train a text classifier to classify the unlabeled texts. Finally, we re-extract keywords from the classified texts. Extensive experiments on both long-text and short-text datasets show that our method substantially outperforms the existing ones