Seminar: Foundations of Practical Privacy: A Multifaceted Way to Think About Privacy and Utility

ABSTRACT

Owing, on the one hand to the popularity and diversity of social media platforms, and to sophisticated end-user information collection infrastructure becoming an integral part of all major online retailers, and on the other to the artificial-intelligence supported advancements in data mining and classification, the notion of privacy has been attracting a lot of recent interest from the academia and industry alike. Many recent successful de-anonymization or private-information inference endeavors have further demonstrated that traditional notions of privacy, achieved by allowing each user to control their own information disclosure, is nothing more than a dream of the past. This is mainly because regular users cannot fathom the amount of private information that can be extracted by machines from seemingly safe and unrelated disclosures. This is the one area in which machines proved to be better than humans -- and they are better beyond most people's imagination. To allow private individuals, at different levels of knowledge and maturity, a fighting chance in the grab for their private information, automated mechanisms, from account-management tools to database privacy protections, have already started to be deployed and will no doubt filter most of our information before disclosure in the future, according to a multitude of criteria established by laws, contracts and personal preferences. It is therefore essential to understand what privacy means, how it is valued and how its definition and worth evolve over time and over different contexts, as well as how privacy protection mechanisms can be employed. In this talk, we touch upon some of the practical issues arising in privacy, from the multiuser problem to time-based privacy-centered investment strategies, and from proper metrics of privacy in realistic scenarios, to meaningful measures of the impact of privacy loss.

George T. Amariucai
Associate Professor, Kansas State University

George Amariucai is an Associate Professor of Computer Science at Kansas State University. His research interests lie in the area of cyber security, and focus on a broad set of topics, from theoretical and practical privacy to manipulation in social networks, and from community resilience to the security of cyber-physical systems. He received his PhD degree from Louisiana State University in 2009, and subsequently worked as an adjunct faculty for the Department of Electrical and Computer Engineering at Iowa State University, before joining Kansas State University in 2017.