[12-20] Proving that Programs are Differentially Private" "Quantitative information flow: when is one leak worse than another?"

文章来源:  |  发布时间:2019-12-20  |  【打印】 【关闭

  

Time:  

10:00 a.m.-1200 a.m.  2019-12-20 

Venue: 

Lecture room (334),  Building 5, SKLCS, Institute of Software, CAS 

Talk 1 

Proving that Programs are Differentially Private 

Speaker: 

Prof. Annabelle McIver, University of New South Wales & Data61 Sydney, Australia 

Abstract:   

In this talk we extend recent work in Quantitative Information Flow (QIF) to provide tools for the analysis of programs that aim to implement differentially private mechanisms. We demonstrate how differential privacy can be expressed using loss functions, and how to use this idea in conjunction with a QIF-enabled program semantics to verify differentially private guarantees. Finally we describe how to use this approach experimentally using Kuifje, a recently developed tool for analysing information-flow properties of programs. 

Talk 2 

Quantitative information flow: when is one leak worse than another? 

Speaker: 

Prof. Carroll Morgan,  University of New South Wales & Data61 Sydney, Australia 

Abstract:   

Cybersecurity is typically concerned with limiting the escape of private information from computer systems. In its original form, this was applied as an absolute criterion, that is "No information must flow." More recently, however, the criterion has been weakened to "No more than a certain amount of information must flow." 

But how do you measure "the amount" of information that is flowing? Recent work (the last 10 years) has shown that the traditional Shannon Entropy measure of information flow is not the answer. This talk explains why it is not the answer,  how we can move to more general measures, what they are, and what are the good properties that they have.

The topic is related to the upcoming monograph "The Science of Quantitative Information Flow" (Springer), due in early 2020.