Breadth First Reasoning Graph for Multi-hop Question Answering

Abstract

Recently Graph Neural Network (GNN) has been used as a promising tool in multi-hop question answering task. However, the unnecessary updations and simple edge constructions prevent an accurate answer span extraction 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 required 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 effectively 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 entities 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 prediction.

Multi-Hop Knowledge Graph Question Answer Method Based on Relation Knowledge Enhancement

“…Zhou et al [14] presented the interpretable reasoning network (INR). The INR model utilizes an interpretable, hop-by-hop reasoning process to answer the question.…”

A Survey on Multi-hop Question Answering and Generation

“…There has been a significant research in the recent years to solve the task. A variety of methods model the task as performing inference over static or dynamic graphs to find the reasoning paths [14,15,30,34,40,61,113,128,142,179,181]. A number of works have also attempted to decompose the multi-hop questions into single hop questions or generate follow-up questions based on the retrieved information [14,95,102,138,181].…”

“…An LSTM is applied to fuse the output of 2 graphs. • Huang and Yang [61] form a sentence graph and add an edge between sentences 𝑠 𝑖 and 𝑠 𝑗 with the weight 𝑤 𝑖 𝑗 given by:…”

“…• Pointer network for reasoning chain prediction: Chen et al [18] encode the question and obtain query dependent paragraph encoding by using BERT. Sentence encodings are extracted from the paragraph encoding similar to Huang and Yang [61]. Alternative baselines are a) BERT-Sent that obtains query aware encoding of each sentence by using BERT and b) BiDAF-Para that uses BiDAF [127] to encode paragraphs.…”

“…• Chen et al [18] concatenate question and all sentences of all the retrieved reasoning chains and feed to BERT to perform the 4 tasks. • Huang and Yang [61] propose to use the output of the GNN model. Sentence scores are computed using MLPs on the sentence node representations.…”