February 20, 2018 –
Title: Machine Learning Meets Applications: Discovering Precursors for use in Event Modeling and Forecasting
Modeling and forecasting large-scale societal events such as civil unrest, disease outbreaks, and turmoil in economic markets is a key problem of interest to social scientists and policy makers. Forecasting algorithms are expected not only to make accurate predictions, but also to provide insights into causative attributes that influence an event's evolution. In this talk I will introduce a framework for discovering event precursors using the machine learning paradigm known as multi-instance learning. Using large-scale distributed representations of news articles and multi-task learning, I will demonstrate how this framework can provide clues into the spatio-temporal progression of events. I will discuss how this approach sheds insight into several issues of interest to policy makers, such as: what types of precursors distinguish violent events from non-violent events, are events influenced more by local versus national happenings, and how significantly can evolving circumstances alter the probability of events of interest? Results will be shown for data from a range of countries in Latin America, Asia, and the Middle East.
Yue Ning is a Ph.D. candidate in the Computer Science department at Virginia Tech. She has also worked as a research intern at Yahoo Research for two summers. Her research interests include applied machine learning, knowledge discovery from data, and AI applications. Her Ph.D. dissertation is focused on modeling and understanding event propagation in large-scale news and social media datasets. Yue's other research projects have included personal recommender systems, information extraction in deep learning, and topical analysis across social media and news media. Her work has been published in SIGKDD, CIKM, SDM, ACM RecSys, and AAAI. She has won student travel awards from 2015 CRA-Women Graduate Cohort Workshop, KDD 2016, and the 2017 ACM Capital Region Celebration of Women in Computing.