Misinformation is often edited and repeated across multiple platforms, websites, languages, and formats (audio, video, image, text). In this talk, Dr. Scott A. Hale (Oxford Internet Institute and Meedan) and Ashkan Kazemi (PhD candidate, University of Michigan) will detail initiatives to crowdsource misleading content through "tiplines" on messaging platforms like WhatsApp and Telegram as well as state-of-the-art natural language processing approaches to group messages making the same claims.
While much effort focuses on large, high-resource languages and unencrypted platforms, Tiplines offer the opportunity for users of end-to-end encrypted platforms to share potentially misleading content with fact-checking organizations and check whether that content has been fact-checked. This talk will show how knowledge distillation can be used to create machine learning models for claim matching that perform well on low-resource as well as high-resource languages.
Meedan, a non-profit building digital tools for global journalism and translation, has built Check, an open-source web-based service to make it easy for fact-checking organizations to run tiplines on a variety of platforms. The research detailed in this talk has been integrated into Check and is used by over a dozen fact-checking organizations across the globe. Meedan is currently working to create better infrastructure to allow academics, practitioners, and community leaders to collaborate more easily on misinformation response.
About the speakers
Dr. Scott A. Hale is an Associate Professor and Senior Research Fellow at the Oxford Internet Institute of the University of Oxford, Director of Research at Meedan, and a Fellow at the Alan Turing Institute. His cross-disciplinary research focuses on advancing equitable access to quality information. Scott develops and applies new techniques in the computational sciences to social science questions and puts the results into practice with industry and policy partners. He is particularly interested in multilingual natural language processing, computational sociolinguistics, mobilization/collective action, agenda setting, and misinformation and has a strong track record in building tools and teaching programs that enable wider access to new methods and forms of data.
Ashkan Kazemi is a 4th year PhD candidate in the Language and Information Technology (LIT) lab at the Computer Science and Engineering department at University of Michigan. He works with his advisor Rada Mihalcea on natural language processing and applications in studying social media, and broadly computational social science. His PhD focus is on using NLP to develop methods for understanding and responding to misinformation, within both English and non-English speaking communities. Ashkan is an organizer for "ACL Year-Round Mentorship", a program that provides mentorship to students interested in NLP research from around the world and has served in the program committee of EMNLP, ACL and NAACL conferences in the past.