Seminario impartido por Dr. Cuneyt Akcora.

SUMMARY. In this talk, we will describe our work in augmenting network connections by using information from user interactions with applications in Trust, Anomaly Detection, Privacy and Link Prediction. In these studies we used social network data primarily from Facebook, Twitter, Orkut, and scientific network data from DBLP and the ACM digital library. We will outline the Homophily and Heterophily theories and explain how we can quantify online user behavior in terms of these theories. Our research indicates that most users choose to live in a closed bubble of Homophily, and this tendency can be used against their privacy. Throughout the talk we will describe the state of the art in privacy research and explain how governments and organizations pry data out of a social network.

BIOGRAPHY. Cuneyt was a Fulbright Scholar at the State University of New York at Buffalo, where he worked on public opinion mining and sentiment analysis on microblogs. In 2014, he completed his Ph.D. at the Università Degli Studi Dell'Insubria in Italy, under the supervision of Elena Ferrari and Barbara Carminati. His Ph.D. studies focused on analyzing privacy effects of user interactions on online social networks. He visited the Social Computing Group (Francesco Bonchi) at Yahoo! Research Barcelona (2011), Data Security and Privacy lab (Murat Kantarcioglu) at the University of Texas at Dallas (2012) and the Data Analytics Group (Paolo Papotti) of Qatar Computing Research Institute (2014). Before starting his Post-Doctoral studies at the University of Texas at Dallas in 2016, he worked at the Intelligent Search Lab at Huawei Research Turkey.