Coarse-grain Fine-grain Coattention Network for Multi-evidence Question Answering

Abstract

End-to-end neural models have made significant progress in question answering, however recent studies show that these models implicitly assume that the answer and evidence appear close together in a single document. In this work, we propose the Coarse-grain Fine-grain Coattention Network (CFC), a new question answering model that combines information from evidence across multiple documents. The CFC consists of a coarse-grain module that interprets documents with respect to the query then finds a relevant answer, and a fine-grain module which scores each candidate answer by comparing its occurrences across all of the documents with the query. We design these modules using hierarchies of coattention and self-attention, which learn to emphasize different parts of the input. On the Qangaroo WikiHop multi-evidence question answering task, the CFC obtains a new state-of-the-art result of 70.6% on the blind test set, outperforming the previous best by 3% accuracy despite not using pretrained contextual encoders.

arXiv commentICLR 2019; 9 pages, 7 figures

References in corpus (8)

• Teaching Machines to Read and Comprehend
• Bidirectional Attention Flow for Machine Comprehension
• End-To-End Memory Networks
• A Sentence Compression Based Framework to Query-Focused Multi-Document Summarization
• Exploring Graph-structured Passage Representation for Multi-hop Reading Comprehension with Graph Neural Networks
• Global-Locally Self-Attentive Dialogue State Tracker
• Dynamic Self-Attention : Computing Attention over Words Dynamically for Sentence Embedding
• Efficient and Robust Question Answering from Minimal Context over Documents

Cited by in corpus (13)

• Multi-hop Question Answering via Reasoning Chains
• Hierarchical Graph Network for Multi-hop Question Answering
• Multi-hop Reading Comprehension across Multiple Documents by Reasoning over Heterogeneous Graphs
• Select, Answer and Explain: Interpretable Multi-hop Reading Comprehension over Multiple Documents
• Answering Complex Open-Domain Questions with Multi-Hop Dense Retrieval
• Do Multi-hop Readers Dream of Reasoning Chains?
• Graph Sequential Network for Reasoning over Sequences
• LUKE-Graph: A Transformer-based Approach with Gated Relational Graph Attention for Cloze-style Reading Comprehension
• NLProlog: Reasoning with Weak Unification for Question Answering in Natural Language
• Multi-grained Evidence Inference for Multi-choice Reading Comprehension
• ClueReader: Heterogeneous Graph Attention Network for Multi-hop Machine Reading Comprehension
• Complex Knowledge Base Question Answering: A Survey
• Fine-tuning Multi-hop Question Answering with Hierarchical Graph Network