Özet:

Abstract The issue of the distribution of fake news and misinformation on social media platforms is a rising global concern, which has affected India as well. This research paper introduces an innovative method for identifying and detecting fake news in India using a multimodal adversarial network. The approach presented in this study leverages both text and image characteristics to encompass the multimodal aspects of fake news. Adversarial training is employed to learn robust and discriminative features/characteristics that enable differentiating authentic news from fabricated news. Evaluation of the proposed method is conducted on an Indian fake news events dataset and achieves a high accuracy and  F1-score of 0.89 and 0.90 respectively. The experiment results indicate that the proposed multimodal adversarial network approach is effective in detecting fake news in the Indian context and thus helpful in mitigating the dissemination of misinformation.

Anahtar Kelimeler: