NLP for More Realistic Fact-Checking

Explore cutting-edge research on Natural Language Processing (NLP) for realistic fact-checking in this 47-minute conference talk by Prof. Iryna Gurevych, presented as part of the INSAIT Tech Series. Delve into the challenges of misinformation in today's information society, including the COVID-19 infodemic and the increasing use of hallucinating large language models. Discover how automated fact-checking (AFC) research aims to assist humans in countering misinformation. Learn about the gap between NLP-based AFC approaches and real-world requirements, and understand the efforts to bridge this divide. Explore new datasets developed to investigate naturally occurring ambiguities when comparing realistic claims with evidence. Gain insights into the specific problem of scientific information being twisted by misinformation and the ongoing work to create novel datasets and models for reconstructing fallacious arguments. Benefit from the expertise of Prof. Gurevych, a distinguished Computer Science professor, director of the Ubiquitous Knowledge Processing Lab at TU Darmstadt, and current president of the Association for Computational Linguistics, as she shares her research on machine learning for large-scale language understanding and text semantics.