Information Surgery: Detecting and Fixing Misinformation Using AI

Explore the cutting-edge approach to combating AI-generated misinformation in this 36-minute talk by Professor Heng Ji from the University of Illinois at Urbana-Champaign. Delve into the innovative "Information Surgeon" model, which utilizes state-of-the-art multimedia joint knowledge extraction techniques to analyze fine-grained event, entity, and relation elements. Learn how this probabilistic graphical neural network model fuses multiple indicators to detect misinformation with high accuracy and explainability. Discover the novel graph-to-text generation approach for creating training data through knowledge element manipulation. Gain insights into the challenges of document-level fake news detection and the importance of evaluating the consistency of information across various data modalities. Understand how this research extends traditional information extraction to future event prediction and cross-lingual, cross-media analysis, potentially revolutionizing the way we consume and verify news media worldwide.