Hierarchical Taxonomy-Aware and Attentional Graph Capsule RCNNs for Large-Scale Multi-Label Text Classification
| Autor: | Qiran Gong, Bo Li, Lifang He, Hao Peng, Jianxin Li, Senzhang Wang, Philip S. Yu, Renyu Yang, Lihong Wang |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
Normalization (statistics)
Computer science Matrix representation 02 engineering and technology Semantics computer.software_genre Text mining 020204 information systems 0202 electrical engineering electronic engineering information engineering Multi-label classification Contextual image classification business.industry Deep learning Graph theory Graph Medoid Computer Science Applications Computational Theory and Mathematics Embedding Graph (abstract data type) Artificial intelligence business Feature learning computer Natural language processing Information Systems |
| Zdroj: | IEEE Transactions on Knowledge and Data Engineering. 33:2505-2519 |
| ISSN: | 2326-3865 1041-4347 |
| Popis: | CNNs, RNNs, GCNs, and CapsNets have shown significant insights in representation learning and are widely used in various text mining tasks such as large-scale multi-label text classification. Most existing deep models for multi-label text classification consider either the non-consecutive and long-distance semantics or the sequential semantics. However, how to coherently take them into account is still far from studied. In addition, most existing methods treat output labels as independent medoids, ignoring the hierarchical relationships among them, which leads to a substantial loss of useful semantic information. In this paper, we propose a novel hierarchical taxonomy-aware and attentional graph capsule recurrent CNNs framework for large-scale multi-label text classification. Specifically, we first propose to model each document as a word order preserved graph-of-words and normalize it as a corresponding word matrix representation preserving both the non-consecutive, long-distance and local sequential semantics. Then the word matrix is input to the proposed attentional graph capsule recurrent CNNs for effectively learning the semantic features. To leverage the hierarchical relations among the class labels, we propose a hierarchical taxonomy embedding method to learn their representations, and define a novel weighted margin loss by incorporating the label representation similarity. Extensive evaluations on three datasets show that our model significantly improves the performance of large-scale multi-label text classification by comparing with state-of-the-art approaches. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|