HSM-QA: Question Answering System Based on Hierarchical Semantic Matching

In recent years, Question Answering (QA) systems have gained popularity as a means of acquiring knowledge. However, the prevalent approach matching question-answer pairs still suffers from low precision and efficiency due to inherent ambiguity natural language descriptions. To address these issues, we propose novel QA based on hierarchical semantic matching, termed HSM-QA. Specifically, HSM-QA is decomposed into two main steps, i.e., query-question query-answer matchings, respectively. For Siamese network applied calculate similarity between pairs, which recalls most similar questions their corresponding answers candidates. terms adopt idea pairwise algorithm single-stream structure relevance query answer, best-matching candidates are ranked returned. After training, steps combined an efficient scheme for different languages, e.g ., English Chinese. Furthermore, lack Chinese datasets, collect massive amount text data social media generate new dataset via pre-trained model. Extensive experiments conducted six datasets validate our The experimental results demonstrate superior performance method than set compared methods.