Question-Aware Sentence Gating Networks for Question and Answering

TL;DR

This paper introduces question-aware sentence gating networks that enhance question answering by integrating sentence-level information into word representations, leading to improved accuracy across multiple datasets.

Contribution

The paper presents a novel model that incorporates sentence-level information into word encoding for QA, improving performance over existing methods.

Findings

Consistent accuracy improvements on various QA datasets.

Effective integration of sentence-level info into word representations.

Wide applicability to neural network-based QA models.

Abstract

Machine comprehension question answering, which finds an answer to the question given a passage, involves high-level reasoning processes of understanding and tracking the relevant contents across various semantic units such as words, phrases, and sentences in a document. This paper proposes the novel question-aware sentence gating networks that directly incorporate the sentence-level information into word-level encoding processes. To this end, our model first learns question-aware sentence representations and then dynamically combines them with word-level representations, resulting in semantically meaningful word representations for QA tasks. Experimental results demonstrate that our approach consistently improves the accuracy over existing baseline approaches on various QA datasets and bears the wide applicability to other neural network-based QA models.