Multi-hop Graph Convolutional Network with High-order Chebyshev Approximation for Text Reasoning
This work addresses a specific bottleneck in graph-based NLP models, offering an incremental improvement for researchers in natural language processing.
The paper tackles the problem of missing important non-consecutive dependencies in single-hop graph reasoning for NLP tasks by proposing a multi-hop graph convolutional network with high-order Chebyshev approximation, achieving verified efficacy on four transductive and inductive NLP tasks.
Graph convolutional network (GCN) has become popular in various natural language processing (NLP) tasks with its superiority in long-term and non-consecutive word interactions. However, existing single-hop graph reasoning in GCN may miss some important non-consecutive dependencies. In this study, we define the spectral graph convolutional network with the high-order dynamic Chebyshev approximation (HDGCN), which augments the multi-hop graph reasoning by fusing messages aggregated from direct and long-term dependencies into one convolutional layer. To alleviate the over-smoothing in high-order Chebyshev approximation, a multi-vote-based cross-attention (MVCAttn) with linear computation complexity is also proposed. The empirical results on four transductive and inductive NLP tasks and the ablation study verify the efficacy of the proposed model. Our source code is available at https://github.com/MathIsAll/HDGCN-pytorch.