A Search-Enhanced Path Mining and Ranking Method for Cross-lingual Knowledge Base Question Answering

Knowledge Base Question Answering (KBQA) is a popular research direction in knowledge graphs, and most of KBQA work focuses on natural language parsing. In order to enhance the application of cross-lingual knowledge, the 16th China Conference on Knowledge Graph and Semantic Computing (CCKS2022) has released a cross-lingual knowledge base question answering (CKBQA) task. This paper presents the submission of our team (HW-TSC) to the CKBQA task, in which we propose a search-enhanced path mining and ranking method. The method divides the process of CKBQA into four parts: question classification, principal entity extraction, search-enhanced candidate path mining and candidate path ranking. Finally, both the preliminary and final evaluation results prove the effectiveness of our method.