Towards Building an Automated Fact-Checking System

Proceedings of the 2017 ACM International Conference on Management of Data(2017)

引用 3|浏览6
ABSTRACTFact-checking has become a prominent issue in today's society, yet a system that can fact-check automatically has yet to surface. In this paper I explore the hurdles of implementing such a system and assess the limitations of currently available technologies. The results of this research show that there is still much work to be done in parsing natural language inputs and mapping them to queries that can be used to extract information from knowledge bases. In particular I explore the use of a question generation tool as a means of mapping claims to natural language queries. I observe that, while moderately successfully, current tools still require substantial improvements to be useful as part of an end-to-end fact-checking system. However, the scaffolding for tackling this problem exists, and ultimately I worked to understand where the shortcomings lie and their nature.
AI 理解论文
Chat Paper