Multi-hop Reading Comprehension across Multiple Documents by Reasoning over Heterogeneous Graphs
Ming Tu, Guangtao Wang, Jing Huang, Yun Tang, Xiaodong He, Bowen Zhou JD AI Research
Abstract
Multi-hop reading comprehension (RC) across documents poses new challenge over single-document RC because it requires reasoning over multiple documents to reach the final answer. In this paper, we propose a new model to tackle the multi-hop RC problem. We introduce a heterogeneous graph with different types of nodes and edges, which is named as Heterogeneous Document-Entity (HDE) graph. The advantage of HDE graph is that it contains different granularity levels of information including candidates, documents and entities in specific document contexts. Our proposed model can do reasoning over the HDE graph with nodes representation initialized with co-attention and self-attention based context encoders. We employ Graph Neural Networks (GNN) based message passing algorithms to accumulate evidences on the proposed HDE graph. Evaluated on the blind test set of the Qangaroo WikiHop data set, our HDE graph based single model delivers competitive result, and the ensemble model achieves the state-of-the-art performance.
中文速览
多跳阅读理解（multi-hop reading comprehension）要求模型跨多篇文档进行推理才能找到答案，远比单文档问答难得多。作者提出了一种名为"异构文档-实体图"（Heterogeneous Document-Entity graph，HDE图）的方法，将候选答案、文档和实体作为不同类型的节点纳入同一个图中，并通过协同注意力和自注意力机制为各节点初始化查询感知的表示，再利用图神经网络（Graph Neural Network，GNN）在图上做消息传递来逐步累积跨文档的推理证据。在WikiHop公开测试集上，单模型超越了此前发表的最优结果，集成模型登顶排行榜第一，并以微弱优势超过了人类水平，且全程无需使用ELMo等预训练上下文嵌入。这项工作表明，将多粒度信息统一建模进异构图并结合GNN推理，是解决多文档多跳阅读理解的有效路径。
原文 arXiv:1905.07374;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1905.07374v2