[4Pin1] インタラクティブ(2)

[4Pin1-26] Support for Searching for Articles to be Verified for Fact Checking

〇Kaori Uchiyama¹, Kaito Suzuki¹, Tsubasa Tagami¹, Kazuaki Hanawa¹, Kentaro Inui^1,2, Atsushi Komiya³, Atsuo Fujimura³, Akinori Machino, Hitofumi Yanai⁴, Ryo Yamashita⁴ (1. Tohoku University, 2. RIKEN Center for Advanced Intelligence Project, 3. SmartNews, Inc., 4. Watchdog for Accuracy inNews-reporting, Japan)

Keywords:Natural Language Processing

False information have become a social problem and it is necessary to verify huge information such as news articles on the Internet. In this paper, we constructed a support system to efficiently verify the authenticity of news articles. This system estimates news articles that describe incorrect information based on posts that refer to them on SNS. We found that it can be expected to improve efficiency of searching for articles that require human verification by using the created system.