Journal of Northeastern University(Natural Science) ›› 2024, Vol. 45 ›› Issue (1): 18-25.DOI: 10.12068/j.issn.1005-3026.2024.01.003

• Information & Control • Previous Articles     Next Articles

MMCSC: A Cross-Modal Approach to Fake News Detection

Yue ZHAO1, Kun HAO1, Jing ZHAO2,3, Jun-chang XIN2,3   

  1. 1.School of Medicine & Biological Information Engineering,Northeastern University,Shenyang 110169,China
    2.School of Computer Science & Engineering,Northeastern University,Shenyang 110169,China
    3.Key Laboratory of Big Data Management and Analytics,Liaoning Province,Shenyang 110169,China. Corresponding author: XIN Jun-chang,E-mail: xinjunchang@mail. neu. edu. cn
  • Received:2022-07-28 Online:2024-01-15 Published:2024-04-02

Abstract:

Current fake news detection methods based on news content do not take into account the higher-level semantic correlation of different modalities, and lack information that can be used to judge news, thus lacking effective use of social network information for news with important distinguishing features. Address to this problem, a fake news detection method based on news content is proposed. By extracting high-level semantic features of multi-modal news such as text, images and videos, the high-level semantic information of different modalities is analyzed, and the cross-modal topic consistency and cross-modal emotional consistency are designed. On this basis, a fake news detection model MMCSC (multi-modal feature content semantic consistency) is designed with cross-modal content semantic consistency. Experiments show that the proposed MMCSC has better detection effect than the traditional method.

Key words: fake news detection, content semantic consistency, cross-modal topic consistency, cross-modal emotional consistency

CLC Number: