作者: Houfeng Wang , Xiaodong Zhang , Shuming Ma , Hao Wang , Xu Sun
DOI:
关键词:
摘要: Most question answering (QA) systems are based on raw text and structured knowledge graph. However, corpora hard for QA system to understand, graph needs intensive manual work, while it is relatively easy obtain semi-structured tables from many sources directly, or build them automatically. In this paper, we an end-to-end answer multiple choice questions with as its knowledge. Our answers queries by two steps. First, finds the most similar tables. Then measures relevance between each candidate table cells, choose related cell source of answer. The evaluated TabMCQ dataset, gets a huge improvement compared state art.