Expertise-Oriented Explainable Question Routing

Question routing aims at routing questions to the most suitable expert with relevant expertise for answering, which is a fundamental issue in Community Question Answering (CQA) websites. Most existing question routing methods usually learn representation of the expert’s interest based on his/her historical answered questions, which will be used to match the target question. However, they always ignore the modeling of expert’s ability to answer questions, and in fact, precisely modeling both expert answering interest and expertise is crucial to the question routing. In this paper, we design a novel Expertise-oriented Modeling explainable Question Routing (EMQR) model based on a multi-task learning framework. In our approach, we propose to learn expert representation by fully capturing the expert’s ability and interest from his/her historical answered questions and the corresponding received vote scores respectively. Furthermore, based on the representations of expert and target question, a multi-task learning model is adopted to predict the most suitable expert and his/her potential vote score, which could provide the intuitive explanation that why routes the question to the expert. Experimental results on six real-world CQA datasets demonstrate the superiority of EMQR, which significantly outperforms existing state-of-the-art methods.