Graph Fusion Network for Text Classification
Yong Dai,Linjun Shou,Ming Gong,Xiaolin Xia,Zhao Kang,Zenglin Xu,Daxin Jiang
DOI: https://doi.org/10.1016/j.knosys.2021.107659
IF: 8.139
2022-01-01
Knowledge-Based Systems
Abstract:Text classification is an important and classical problem in natural language processing. Recently, Graph Neural Networks (GNNs) have been widely applied in text classification and achieved outstanding performance. Despite the success of GNNs on text classification, existing methods are still limited in two main aspects. On the one hand, transductive methods cannot easily adapt to new documents. Since transductive methods incorporate all documents into their text graph, they need to reconstruct the whole graph and retrain their system from scratch when new documents come. However, this is not applicable to real-world situations. On the other hand, many state-of-the-art algorithms ignore the quality of text graphs, which may lead to sub-optimal performance. To address these problems, we propose a Graph Fusion Network (GFN), which can overcome these limitations and boost text classification performance. In detail, in the graph construction stage, we build homogeneous text graphs with word nodes, which makes the learning system capable of making inference on new documents without rebuilding the whole text graph. Then, we propose to transform external knowledge into structural information and integrate different views of text graphs to capture more structural information. In the graph reasoning stage, we divide the process into three steps: graph learning, graph convolution, and graph fusion. In the graph learning step, we adopt a graph learning layer to further adapt text graphs. In the graph fusion step, we design a multi-head fusion module to integrate different opinions. Experimental results on five benchmarks demonstrate the superiority of our proposed method.