Breadth First Reasoning Graph for Multi-hop Question Answering

Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 5810–5821 June 6–11, 2021. ©2021 Association for Computational Linguistics 5810 Breadth First Reasoning Graph for Multi-hop Question Answering Yongjie Huang1 and Meng Yang1,2,∗ 1School of Computer Science and Engineering, Sun Yat-sen University / Guangzhou, China 2Key Laboratory of Machine Intelligence and Advanced Computing, Sun Yat-sen University, Ministry of Education / China huangyj229@mail2.sysu.edu.cn, yangm6@mail.sysu.edu.cn Abstract Recently Graph Neural Network (GNN) has been used as a promising tool in multi-hop question answering task. However, the un- necessary updations and simple edge construc- tions prevent an accurate answer span extrac- tion in a more direct and interpretable way. In this paper, we propose a novel model of Breadth First Reasoning Graph (BFR-Graph), which presents a new message passing way that better conforms to the reasoning process. In BFR-Graph, the reasoning message is re- quired to start from the question node and pass to the next sentences node hop by hop until all the edges have been passed, which can effec- tively prevent each node from over-smoothing or being updated multiple times unnecessarily. To introduce more semantics, we also define the reasoning graph as a weighted graph with considering the number of co-occurrence enti- ties and the distance between sentences. Then we present a more direct and interpretable way to aggregate scores from different levels of granularity based on the GNN. On Hot- potQA leaderboard, the proposed BFR-Graph achieves state-of-the-art on answer span pre- diction. 1 Introduction Typical Question Answering (QA) or Reading Comprehension (RC) task aims at exploring a de- sired answer through a single evidence document or paragraph. Recently, a more challenging multi-hop QA task, where we need to reason over multiple paragraphs to find the answer, is gradually catching attention. One example from HotpotQA dataset (Yang et al., 2018) is shown in Fig. 1. One method for achieving multi-hop QA is to concatenate all the paragraphs together and treat it as a typical single-hop QA task (Yang et al., 2018), then existing QA techniques can be applied. Al- though multi-hop QA can be solved to some extent, *Corresponding author. Figure 1: One example from HotpotQA dataset. “s1”, “s2”, ... denote the sentences in paragraphs. The model needs to find the answer and supporting sentences by reasoning over multiple sentences and paragraphs. It’s obvious that the reasoning is in a ordered process from the question to “s1”, “s2” and finally to “s4”. this method lacks interpretation of the reasoning process from one hop to the next hop. Graph Neural Networks (GNN) is a natural way to represent the solving procedure of multi-hop QA. For instance, nodes in GNN represent sen- tences/entities in the paragraphs, and from the up- dation through edges we can get interactive mes- sage between them, which is similar to the process of reasoning. Thus, a more reasonable method is to construct GNN to simulate the reasoning process among multiple paragraphs (Ding et al., 2019; Qiu et al., 2019; …

Content truncated to 500 words