Deep Fusion LSTMs for Text Semantic Matching

Abstract

Recently, there is rising interest in modelling the interactions of text pair with deep neural networks. In this paper, we propose a model of deep fusion LSTMs (DF-LSTMs) to model the strong interaction of text pair in a recursive matching way. Specifically, DF-LSTMs consist of two interdependent LSTMs, each of which models a sequence under the in-fluence of another. We also use exter-nal memory to increase the capacity of LSTMs, thereby possibly capturing more complicated matching patterns. Experiments on two very large datasets demonstrate the efficacy of our proposed architecture. Furthermore, we present an elaborate qualitative analysis of our models, giving an intuitive understanding how our model worked.