SCIR-QA at SemEval-2017 Task 3: CNN Model Based on Similar and Dissimilar Information between Keywords for Question Similarity

Proceedings of the 11th International Workshop on Semantic Evaluations (SemEval-2017), pages 305–309, Vancouver, Canada, August 3 - 4, 2017. c⃝2017 Association for Computational Linguistics SCIR-QA at SemEval-2017 Task 3: CNN Model Based on Similar and Dissimilar Information between Keywords for Question Similarity Le Qi, Yu Zhang, Ting Liu Research Center for Social Computing and Information Retrieval Harbin Institute of Technology lqi,zhangyu,tliu@ir.hit.edu.cn Abstract We describe a method of calculating the similarity between questions in communi- ty QA. Questions in cQA are usually very long and there are a lot of useless infor- mation about calculating the similarity be- tween questions. Therefore, we imple- ment a CNN model based on similar and dissimilar information on questions key- words. We extract the keywords of ques- tions, and then model the similar and dis- similar information between the keyword- s, and use the CNN model to calculate the similarity. 1 Introduction We participate in SemEval-2017 Task 3 Subtask B (Nakov et al., 2017) on Community Question An- swering. In this task, we are given a question from community forum (named original question) and 10 related questions. We need to re-rank the relat- ed questions according to their similarity between the origin question. Both the original question and the related ques- tion have question subject and question body. The subject is short. The body is long and contains a lot of useless information. In our system, we try to use keywords to replace questions to locate more important information on the question, so we use a keyword extraction algorithm that combines syn- tactic information to get more accurate keywords. Then we use a CNN model based on similar and dissimilar information between questions to cal- culate the similarity of questions. The model can make good use of similar information and dissimi- lar information between questions to get better re- sults. The paper is organized as follows: Section 2 in- troduces our system. Section 3 introduces the ex- periment. And in section 4, there are the conclu- sions. 2 Model In this section we describe our system in detail. In Section 2.1 we show how we extract keyword- s from the subject and body, and then in Section 2.2 we describe how to construct the CNN mod- el based on similar and dissimilar information on question keywords. 2.1 Keyword extraction First, we cut the question subject and question body. Then, we extract keywords from each sub- sentence. We combine all the extracted keywords together as a result. We use an unsupervised keyword extraction method based on dependency analysis. The method uses syntactic dependency relations be- tween words as clues. For the given question, we not only use the statistical information and word vector information, but also construct the depen- dency graph to calculate the correlation intensity between words, and then construct the weighted graph according to the dependency degree, and use the TextRank algorithm (Mihalcea and Ta- rau, 2004) to iterate to calculate the word im- portance score. The main steps include prepro- cessing, the construction …

Content truncated to 500 words