Sentence Similarity Learning Method based on Attention Hybrid Model
Yue Wang, Xiaoqiang Di, Jinqing Li, Huamin Yang, Lin Bi · Journal of Physics Conference Series · 2018
Sentence similarity learning is a vital task in Natural Language Processing (NLP) such as document summarization and question answering. In this paper, we propose a method to compute semantic similarity between sentences which is based on the attention hybrid model. Our method utilizes Bidirectional Long Short-Term Memory Networks (BLSTM) and Convolutional Neural Networks (CNN) to extract the semantic features of a sentence. And it learns the representation of each sentence with word-level attention. Then the attentive representations are concatenated and fed into the output layer to compute the score of sentences similarity. Finally, the public datasets of the Quora is used to test the proposed method and experiment results show that our method is effective and outperforms other methods.