Exploiting Global and Local Hierarchies for Hierarchical Text Classification

Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 4030 - 4039 December 7-11, 2022 ©2022 Association for Computational Linguistics Exploiting Global and Local Hierarchies for Hierarchical Text Classification Ting Jiang1, Deqing Wang1,4,∗, Leilei Sun1, Zhongzhi Chen2, Fuzhen Zhuang1,3,4, Qinghong Yang2 1SKLSDE, School of Computer, Beihang University, Beijing, China 2School of Software, Beihang University, Beijing, China 3Institute of Artificial Intelligence, Beihang University, Beijing, China 4Zhongguancun Laboratory, Beijing, China {royokong, dqwang, leileisun, jongjyh, zhuangfuzhen, yangqh}@buaa.edu.cn Abstract Hierarchical text classification aims to leverage label hierarchy in multi-label text classification. Existing methods encode label hierarchy in a global view, where label hierarchy is treated as the static hierarchical structure containing all labels. Since global hierarchy is static and irrelevant to text samples, it makes these meth- ods hard to exploit hierarchical information. Contrary to global hierarchy, local hierarchy as a structured labels hierarchy corresponding to each text sample. It is dynamic and rele- vant to text samples, which is ignored in pre- vious methods. To exploit global and local hi- erarchies, we propose Hierarchy-guided BERT with Global and Local hierarchies (HBGL), which utilizes the large-scale parameters and prior language knowledge of BERT to model both global and local hierarchies. Moreover, HBGL avoids the intentional fusion of seman- tic and hierarchical modules by directly mod- eling semantic and hierarchical information with BERT. Compared with the state-of-the-art method HGCLR, our method achieves signifi- cant improvement on three benchmark datasets. Our code is available at http://github. com/kongds/HBGL. 1 Introduction Hierarchical text classification (HTC) focuses on assigning one or more labels from the label hier- archy to a text sample (Sun and Lim, 2001). As a special case of multi-label text classification, HTC has various applications such as news categoriza- tion (Kowsari et al., 2017) and scientific paper clas- sification (Lewis et al., 2004b). The methods in HTC aim to improve prediction accuracy by mod- eling the large-scale, imbalanced, and structured label hierarchy (Mao et al., 2019a). To model the label hierarchy, recent meth- ods (Zhou et al., 2020; Chen et al., 2021; Wang et al., 2022) view hierarchy as a directed acyclic ∗ Corresponding Author. graph and model label hierarchy based on graph encoders. However, the input of graph encoders is static, considering that all HTC text samples share the same hierarchical structure, which leads graph encoders to model the same graph redundantly. To solve this problem, Wang et al. (2022) directly dis- cards the graph encoder during prediction, but this method still suffers from the same problem during training. Moreover, since the target labels corre- sponding to each text sample could be either a single-path or a multi-path in HTC (Zhou et al., 2020), recent methods only consider the graph of the entire label hierarchy and ignore the subgraph corresponding to each text sample. This subgraph can contain structured label co-occurrence infor- mation. For instance, a news report about France travel is labeled “European” under the parent label “World” and “France” under a different parent label “Travel Destinations”. There is a …

Content truncated to 500 words