CRCS Lunch Seminar
Date: Monday, March 24, 2014
Time: 1:00pm – 2:30pm
Place: 60 Oxford St., Room 330
Speaker: David Xiao, Harvard CRCS and CNRS, LIAFA at Université Paris 7
Progress in information technology continues to accelerate, exemplified by the recent and rapid adoption of smartphones. This progress has opened the possibility of collecting massive amounts of data, tantalizing both researchers and entrepreneurs with the possibility of gaining new insights into people's behaviors. However, this promise stands in tension with the need to protect the privacy of the individuals generating this data.
In this talk we will explore how to balance these concerns using ideas from game theory and differential privacy. By formally modeling both incentives and privacy loss, we are able to rigorously study what kinds of trade-offs are achievable.
In particular, we will discuss two recent lines of work exploring such trade-offs. In the first, payments are used to compensate individuals for their loss in privacy. For example, this can model the case of a marketing or polling agency that pays individuals in order to participate in surveys. In the second line of work, we consider settings where the individuals have a stake in the outcome of the data analysis. For example, this can model a government agency who needs individuals' data in order to make policy decisions.
In both settings, we use differential privacy as a key concept to quantify the privacy loss incurred by players. We will present constructions of mechanisms achieving meaningful trade-offs between incentives and privacy, while also satisfying desirable properties such as truthfulness and individual rationality. We will also discuss negative results showing that in some cases incentives and privacy may be incompatible.
David Xiao is a research scientist at the CNRS and is based at the LIAFA, a computer science lab at Université Paris 7. He is currently a visiting scholar at CRCS. His research interests lie at the intersection of cryptography / privacy and computational complexity. He received his Ph.D in computer science from Princeton University in 2009.