Sentence Similarity Learning by Lexical Decomposition and Composition

TL;DR

This paper introduces a novel sentence similarity model that considers both similarities and dissimilarities by decomposing lexical semantics, leading to improved performance on answer selection and paraphrase tasks.

Contribution

The work presents a new approach that decomposes words into similar and dissimilar components and employs a CNN to better capture semantic relations.

Findings

Achieves state-of-the-art results on answer sentence selection.

Obtains competitive results on paraphrase identification.

Demonstrates the effectiveness of lexical decomposition in similarity learning.

Abstract

Most conventional sentence similarity methods only focus on similar parts of two input sentences, and simply ignore the dissimilar parts, which usually give us some clues and semantic meanings about the sentences. In this work, we propose a model to take into account both the similarities and dissimilarities by decomposing and composing lexical semantics over sentences. The model represents each word as a vector, and calculates a semantic matching vector for each word based on all words in the other sentence. Then, each word vector is decomposed into a similar component and a dissimilar component based on the semantic matching vector. After this, a two-channel CNN model is employed to capture features by composing the similar and dissimilar components. Finally, a similarity score is estimated over the composed feature vectors. Experimental results show that our model gets the state-of-the-art performance on the answer sentence selection task, and achieves a comparable result on the paraphrase identification task.