Human-Computer Interaction in Multimodal Fake News Detection

In the era of digital information, the problem of fake news has become one of the most serious challenges for society. In response to this, researchers from various fields are focusing on developing advanced methods for detecting fake news. Multimodal fake news detection requires advanced algorithms and technologies that can analyze and integrate diverse data. Using a multimodal approach that integrates different data sources, such as text, image, and audio, allows for a more comprehensive analysis and identification of fake news. In the context of Human-Computer Interaction (HCI), it is crucial that these technologies are not only effective, but also intuitive and easy for users to use. Studies show that interactive interfaces that visualize the fake news detection process can significantly improve users' ability to identify disinformation. The paper presents the proposal of the method for evaluating the quality of diagrams visualizing advanced multimodal models for fake news detection that facilitate understanding and verification of information. The proposed method includes nine metrics divided into two categories: qualitative metrics, which relate to the subjective assessment of the model's visualization, and quantitative metrics, which assess whether the diagram includes the key elements of the model architecture.