The problem of Fake News is arguably one of the most serious challenges facing the media industry and a threat to democratic societies worldwide. By exploiting the current infrastructure of social media platforms where contents can be created and disseminated to a large audience with a little to zero cost, and the click-through-rate based revenue model of the media ecosystem, the problem has reached to an unprecedented level. To mitigate such fake news spread, researchers from multiple disciplines have proposed various strategies, envisioned automated and semi-automated checking systems and recommended policy changes across the media ecosystem. In this tutorial, we analyze the current state of the Fake News problem, its variations, and summarize the recent research findings on this emerging and timely topic. Specifically, we focus on the efforts from the Artificial Intelligence (AI) and Database (DB) communities to detect fake news and curb its spread and show their strengths and limitations. Finally, we present a set of research questions and suggest how they can be tackled using the combination of both AI and DB approaches. Date: Saturday, February 3, 2018 Time: 9 AM - 1 PM
- Definition and Problem Formulation
- Typology of Fake News
- Computational Questions
- Discussion on DB vs. AI Approaches
- Online Fake News Detection Challenge
- ML Based Solutions
- Feature Categories
- Available Databases
- Supervised and Unsupervised Approaches
- What is Lacking Now?
Naeemul Hassan is an assistant professor in the computer and information science department at the University of Mississippi. He has interest in research areas related to Database, Data Mining, and Natural Language Processing. He received a Ph.D. in Computer Science from the University of Texas at Arlington.
Dongwon Lee is an associate professor in the College of Information Sciences and Technology (a.k.a. iSchool) of The Pennsylvania State University, USA. He researches broadly in Data Science, in particular, on the management of and mining in data in diverse forms including structured records, text, multimedia, social media, and Web. He is also interested in applying the human computation framework to solve data science problems, and detecting/curbing challenging online frauds using machine learning techniques.
