Syntax-based Deep Matching of Short Texts

TL;DR

This paper introduces DeepMatch_tree, a novel deep learning approach that leverages dependency tree patterns for improved short-text matching, demonstrating superior performance on social media response tasks.

Contribution

The paper presents a new dependency tree-based pattern mining and deep neural network framework for short-text matching, outperforming existing models.

Findings

DeepMatch_tree outperforms competitors on social media response matching

Dependency tree patterns enhance matching accuracy

Sparse neural network structure improves learning efficiency

Abstract

Many tasks in natural language processing, ranging from machine translation to question answering, can be reduced to the problem of matching two sentences or more generally two short texts. We propose a new approach to the problem, called Deep Match Tree (DeepMatch$_{tree}$), under a general setting. The approach consists of two components, 1) a mining algorithm to discover patterns for matching two short-texts, defined in the product space of dependency trees, and 2) a deep neural network for matching short texts using the mined patterns, as well as a learning algorithm to build the network having a sparse structure. We test our algorithm on the problem of matching a tweet and a response in social media, a hard matching problem proposed in [Wang et al., 2013], and show that DeepMatch$_{tree}$ can outperform a number of competitor models including one without using dependency trees and one based on word-embedding, all with large margins