Title: Towards Trustworthy Modern Systems Using Data Science
Speaker: Neil Gong, Ph.D. Candidate, University of California Berkeley
Abstract: Modern systems such as social medias, mobile computing platforms, and wearable devices bring new benefits to almost all aspects of our lives. However, they are also plagued by both conventional and emerging threats to security and privacy. In this talk, I will demonstrate that data science, including machine learning, network science, and natural language processing, can play an important role towards the design of trustworthy modern systems.
First, I will present my work on leveraging network science to design secure account recovery methods. Specifically, companies such as Facebook and Microsoft have recently proposed Social Authentication, a social network based approach, to recover lost user accounts. My work provides the first systematic study about the security of Social Authentication, and my results can guide the design of more secure Social Authentication. Second, I will discuss the framework called SybilBelief that I developed to detect fake (Sybil) accounts in online social networks. The SybilBelief framework models a social network as a pairwise Markov Random Fields and uses Loopy Belief Propagation to infer the labels (i.e., benign or fake) of accounts. SybilBelief demonstrates how probabilistic graphical model techniques can be applied to security problems. Third, I will demonstrate that diverse private information, including user attributes, relationships between users, author identity of text contents, and user interests, can be inferred from public data with high accuracies via big data analytics. This line of research shows how data science can be used as new privacy attacks. Finally, I will describe my future research agenda where I plan to apply data science to other important security and privacy problems as well as bring security and privacy to data science.
Speaker Bio: Neil Zhenqiang Gong is a Ph.D candidate in the Department of Electrical Engineering and Computer Science at the University of California, Berkeley. He is broadly interested in security, privacy, and their intersection with data science. In particular, he recently focuses on security and privacy issues in social media and mobile computing. His research has been widely covered by popular media such as WIRED, NPR, Slashdot, etc.