A Compare-Aggregate Model for Matching Text Sequences

TL;DR

This paper introduces a compare-aggregate framework for text sequence matching in NLP tasks, utilizing word-level comparison and CNN-based aggregation, with findings that simple comparison functions can outperform complex neural methods.

Contribution

The paper proposes a novel compare-aggregate model that effectively combines word-level matching with CNN aggregation, highlighting the effectiveness of simple comparison functions.

Findings

Simple element-wise comparison functions outperform neural tensor networks.

The model achieves strong results across four datasets.

CNN-based aggregation effectively captures matching information.

Abstract

Many NLP tasks including machine comprehension, answer selection and text entailment require the comparison between sequences. Matching the important units between sequences is a key to solve these problems. In this paper, we present a general "compare-aggregate" framework that performs word-level matching followed by aggregation using Convolutional Neural Networks. We particularly focus on the different comparison functions we can use to match two vectors. We use four different datasets to evaluate the model. We find that some simple comparison functions based on element-wise operations can work better than standard neural network and neural tensor network.