M$^2$S-Net: Multi-Modal Similarity Metric Learning based Deep Convolutional Network for Answer Selection

TL;DR

This paper introduces M$^2$S-Net, a deep convolutional network leveraging multi-modal similarity metric learning for answer selection, outperforming previous models on the TREC-QA benchmark.

Contribution

The paper presents a novel end-to-end deep learning framework that models interdependence between sentence pairs using multi-modal similarity metrics.

Findings

Surpasses previous state-of-the-art on TREC-QA in MAP and MRR.

Demonstrates effectiveness of multi-modal similarity learning in answer selection.

Outperforms models that only capture individual sentence semantics.

Abstract

Recent works using artificial neural networks based on distributed word representation greatly boost performance on various natural language processing tasks, especially the answer selection problem. Nevertheless, most of the previous works used deep learning methods (like LSTM-RNN, CNN, etc.) only to capture semantic representation of each sentence separately, without considering the interdependence between each other. In this paper, we propose a novel end-to-end learning framework which constitutes deep convolutional neural network based on multi-modal similarity metric learning (M$^2$S-Net) on pairwise tokens. The proposed model demonstrates its performance by surpassing previous state-of-the-art systems on the answer selection benchmark, i.e., TREC-QA dataset, in both MAP and MRR metrics.