Is Graph Structure Necessary for Multi-hop Question Answering?
Nan Shao†, Yiming Cui‡†, Ting Liu‡, Shijin Wang†§, Guoping Hu† †State Key Laboratory of Cognitive Intelligence, iFLYTEK Research, China ‡Research Center for Social Computing and Information Retrieval (SCIR), Harbin Institute of Technology, Harbin, China §iFLYTEK AI Research (Hebei), Langfang, China
Abstract
Recently, attempting to model texts as graph structure and introducing graph neural networks to deal with it has become a trend in many NLP research areas. In this paper, we investigate whether the graph structure is necessary for multi-hop question answering. Our analysis is centered on HotpotQA. We construct a strong baseline model to establish that, with the proper use of pre-trained models, graph structure may not be necessary for multi-hop question answering. We point out that both graph structure and adjacency matrix are task-related prior knowledge, and graph-attention can be considered as a special case of self-attention. Experiments and visualized analysis demonstrate that graph-attention or the entire graph structure can be replaced by self-attention or Transformers.
中文速览
多跳问答（multi-hop question answering）任务长期依赖图神经网络（GNN）来建模实体间关系，但这种图结构到底有多大必要性，一直缺乏系统验证。研究者以HotpotQA为测试床，在重新实现经典图网络模型DFGN的基础上发现：只要将预训练模型（RoBERTa）以微调方式使用，直接去掉整个图结构对最终指标几乎没有影响；反之，若将预训练模型冻结参数作为特征提取器，图结构才能带来约4%的显著提升。理论上，图注意力（graph-attention）本质上是自注意力（self-attention）的一个特例——邻接矩阵不过是人工定义的先验知识，而自注意力完全能自行学习到实体间的对应模式；仅堆叠两层Transformer就能达到与复杂图网络相当的效果，注意力可视化也进一步揭示预训练模型已自发学会实体间、属性与实体间、共指消解等多种推理模式。这一结论提醒未来将图结构引入NLP任务的研究者，需要更严格地论证图结构的必要性与优越性，而不能简单将其视为"标配"。
原文 arXiv:2004.03096;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2004.03096v2