An Attention Mechanism for Answer Selection Using a Combined Global and Local View

TL;DR

This paper introduces a novel attention mechanism for neural question answering that combines global and local views, improving answer selection performance on the InsuranceQA dataset.

Contribution

It presents a new attention mechanism that incorporates both global answer embeddings and local input features, advancing neural question answering methods.

Findings

Outperforms state-of-the-art on InsuranceQA

Visualizes attention focus areas

Analyzes parameter effects on performance

Abstract

We propose a new attention mechanism for neural based question answering, which depends on varying granularities of the input. Previous work focused on augmenting recurrent neural networks with simple attention mechanisms which are a function of the similarity between a question embedding and an answer embeddings across time. We extend this by making the attention mechanism dependent on a global embedding of the answer attained using a separate network. We evaluate our system on InsuranceQA, a large question answering dataset. Our model outperforms current state-of-the-art results on InsuranceQA. Further, we visualize which sections of text our attention mechanism focuses on, and explore its performance across different parameter settings.