Sentence Similarity Computation Based on Deep Learning

Abstract

Computing the semantic similarity of two sentences is a task central to automated understanding of natural languages. In data mining Technology, sentence similarity computation is a key problem. As we know a sentence can be presented by many kinds of style, if we want to describe what a sentence means we should dip into the semantic level and consider about the dependency structure. In this paper, we propose a novel method for calculate the sentence similarity score between two sentences based on deep learning. The sentence similarity score is computed using semantic knowledge notably the synonymy relationships and syntactico-semantic knowledge especially the semantic class and thematic role. Furthermore, we show that the use of recurrent and recursive neural networks can provide a 16% to 70% improvement over baseline models.