Webinar: AI Empowered Metadata Research And Information Analysis In Uncertain Environments (DCMI)

This webinar presents studies on AI empowered metadata research and information analysis in uncertain environments. Metadata research in uncertain environments involves various aspects of events, information, and people. The traditional metadata elements include the time, location, and stakeholders of an event, title, keywords, author and length of information as well as age, gender, and occupation of people, etc. In addition to traditional metadata elements, AI can also dynamically generate implicit metadata, such as semantic features and credibility of textual and visual information, severity levels of events, as well as needs, attentions, opinions, and influence of people, etc. Accurately identifying or utilizing AI techniques to generate metadata elements for various objects can facilitate various information analysis tasks empowered by AI in uncertain environments and improve the performance of AI models, such as classification and severity assessment of disaster losses, crisis information generation capability assessment of large language models, profiling the roles of various social media platforms in public events, archetypes of influential users in social question-answering sites, and prediction of users’ attention shift.