Guest Speaker: Stefan Feuerriegel
All dates for this event occur in the past.
Thursday, October 24th, 2019
480 Dreese Labs
2015 Neil Avenue, Columbus, Ohio 43210
Neural Learning of Misinformation Cascades from Propagation Dynamics
Misinformation in social media poses a significant threat to society and, hence, scalable strategies for an early detection of online cascades with misinformation are in dire need. In this work, we propose a novel tree-structured neural network for learning propagation dynamics in Twitter cascades. The network builds upon a Tree-LSTM that is carefully engineered to the structure of online information cascades. The resulting Tree-LSTM overcomes inherent limitations from feature engineering, since it learns propagation features along the complete cascade. Its effectiveness is demonstrated based on a classification task where the veracity of 2,156 Twitter cascades is predicted. We improve the detection performance in terms of AUC over the status quo with cascade features by 4.1%.
Bio: Stefan Feuerriegel heads the Chair of Management Information Systems at ETH Zurich as an assistant professor (tenure track). His group aims at developing data-driven innovations for business analytics. Key research themes related to, for instance, interactive digital platforms, clickstream analytics, and healthcare analytics. Stefan obtained his PhD from the University of Freiburg, Germany. He further collaborated with various institutions during his research visits, including the University of New South Wales (UNSW) in Sydney, the National Institute of Informatics (NII) in Tokyo, the University of Texas at Austin, and Carnegie Mellon University (CMU) in Pittsburgh.
Host: Huan Sun