Label-Specific Dual Graph Neural Network for Multi-Label Text Classification

Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, pages 3855–3864 August 1–6, 2021. ©2021 Association for Computational Linguistics 3855 Label-Specific Dual Graph Neural Network for Multi-Label Text Classification Qianwen Ma1,2, Chunyuan Yuan1,2, Wei Zhou1* and Songlin Hu1,2 1 Institute of Information Engineering, Chinese Academy of Sciences 2 School of Cyber Security, University of Chinese Academy of Sciences {maqianwen,yuanchunyuan,zhouwei,husonglin}@iie.ac.cn Abstract Multi-label text classification is one of the fun- damental tasks in natural language process- ing. Previous studies have difficulties to dis- tinguish similar labels well because they learn the same document representations for differ- ent labels, that is they do not explicitly ex- tract label-specific semantic components from documents. Moreover, they do not fully ex- plore the high-order interactions among these semantic components, which is very helpful to predict tail labels. In this paper, we propose a novel label-specific dual graph neural net- work (LDGN), which incorporates category information to learn label-specific components from documents, and employs dual Graph Convolution Network (GCN) to model com- plete and adaptive interactions among these components based on the statistical label co- occurrence and dynamic reconstruction graph in a joint way. Experimental results on three benchmark datasets demonstrate that LDGN significantly outperforms the state-of-the-art models, and also achieves better performance with respect to tail labels. 1 Introduction Automatically labeling multiple labels of docu- ments is a fundamental and practical task in nat- ural language processing. Recently, with the growth of data scale, multi-label text classifica- tion(MLTC) has attracted more attention, since it is often applied to many fields such as sentiment analysis (Liu and Chen, 2015; Li et al., 2016), emotion recognition (Wang et al., 2016; Jabreel and Moreno, 2019), web page tagging (Jain et al., 2016) and so on. However, the number of labels and documents and the complex relations of labels render it an unsolved and challenging task. Existing studies for multi-label text classifica- tion mainly focus on learning enhanced document *Corresponding Author representation (Liu et al., 2017) and modeling la- bel dependency (Zhang et al., 2018; Yang et al., 2018; Tsai and Lee, 2019) to improve classifica- tion performance. Although they have explored the informative words in text content, or consid- ered the label structure and label semantics to cap- ture label correlations, these models cannot dis- tinguish similar labels well (e.g., the categories Prices vs Consumer Prices in Reuters News). The main reason is that most of them neglect the semantic connections between labels and in- put documents and they learn the same document representations for different labels, which cannot issue the label similarity problem. More specif- ically, they do not explicitly consider the corre- sponding semantic parts of each label in the docu- ment. Recently, some studies (You et al., 2019; Xiao et al., 2019; Du et al., 2019) have used attention mechanism to explore the above semantic connec- tions, and learn a label-specific document repre- sentation for classification. These methods have obtained promising results in …

Content truncated to 500 words