Differentially Private Analysis of Graphs

Presented on November 7, 2016 at 11:00 a.m. in the Klaus Advanced Computing Building, Room 1116ESofya Raskhodnikova is an associate professor of Computer Science and Engineering at Penn State. Her research interests include sublinear-time algorithms, private data analysis, approximation algorithms, and randomized algorithms.Runtime: 55:16 minutesMany types of data can be represented as graphs, where nodes correspond to individuals and edges capture relationships between them. Examples include datasets capturing “friendships” in an online social network, financial transactions, email communication, doctor-patient relationships, and romantic ties. On one hand, such datasets contain sensitive information about individuals. On the other hand, global information that can be gleaned from their analysis can provide significant benefits to society. Several naive attempts at anonymizing sensitive data by stripping obvious identifying information resulted in spectacular failures. In this talk, we discuss algorithms for analyzing network data that satisfy a rigorous notion of privacy called node differential privacy. We present several techniques for designing node differentially private algorithms, based on combinatorial analysis, network flow, and linear and convex programming. Based on joint work with A. Smith (FOCS 2016) and with S. Kasiwisvanathan, K. Nissim, A. Smith (TCC 2013