Seminar: Privacy-Preserving and Functional Information Systems
PhD Candidate, University of South Florida
Monday, March 30, 2020
10:00am - 11:00am
Information systems generally involve storage and analytics of large-scale data, many of which may be highly sensitive (e.g., personal/medical information, financial transactions). Thus, it is critical to ensure that these systems not only provide essential functionalities in a large scale efficiently but also achieve a high level of security against cyber threats. However, there are critical research challenges in enabling the security and privacy of such information systems while, at the same time, preserving their original functionalities and efficiency (e.g., search, update, analytics). Specifically, although standard encryption provides data confidentiality, it prevents data queries and analytics. Although some functional encryption techniques can offer encrypted data processing and computation, they have been shown to be highly costly and even leak the access patterns, which may invalidate the usability and security of real-life applications.
In this talk, I will present a new series of privacy-enhancing technologies for critical cyber-infrastructures. To overcome the efficiency and privacy leakage challenges of the state-of-the-art, my solutions feature innovative designs in the distributed computation model. I have developed novel distributed oblivious search and access protocols by harnessing efficient multi-party computation protocols and special data structure paradigms together, which avoids restricted system model and costly operations (e.g, homomorphic encryption), thereby offering both high efficiency and security against malicious adversaries simultaneously. The theoretical and experimental analyses have demonstrated that my proposed frameworks offer significant efficiency and security advantages over the state-of-the-art. Finally, I will conclude my talk by outlining some of my future research directions to achieve advanced privacy-preserving and functional data processing platforms such as oblivious distributed file systems and privacy-preserving machine learning.
Thang Hoang is a Ph.D. candidate in the Department of Computer Science and Engineering at the University of South Florida. He received an M.S. degree in Computer Science from Chonnam National University (South Korea) in 2014, and a B.S. degree in Computer Science from the University of Science, VNU-HCMC (Vietnam) in 2010. His research interests include applied cryptography, data privacy, privacy-enhancing technologies, and mobile security. He has published over 20 research articles in respected security conferences and journals along with several patents. He has won two best paper awards and also served as a regular program committee member in a security workshop (CoSDEO).