Artificial Intelligence

Constraint-based Differential Privacy: Private Data Release for Complex Tasks

Friday, March 23, 2018, 11:00am - 12:00pm PDTiCal
6th floor large conference room
This event is open to the public.
AI Seminar
Ferdinando Fioretto, Universityof of Michigan

Advances in artificial intelligence and data science have allowed the development of products that leverage individuals data to provide valuable services. However, the use of this massive quantity of personal information raises fundamental privacy concerns. Differential Privacy (DP) has emerged as the de-facto standard to addresses the sensitivity of such information and can be used to release privacy-preserving datasets. However, when these private datasets are used as inputs to complex machine learning or optimization tasks, they may produce results that are fundamentally different from those obtained on the original dataset.

In this talk, we will focus on the problem of releasing private datasets for complex data analysis tasks.  We will introduce the notion of Constrained-Based Differential Privacy (CBDP) which allow us to cast the data release problem to an optimization problem whose goal is to preserve the salient features of the original dataset. Finally, we will discuss two applications of CBDP for large socio-technical systems related to the optimization of operations in energy networks and transportation systems.



Ferdinando (Nando) Fioretto is a postdoctoral researcher at the University of Michigan. His research focuses on artificial intelligence, data privacy, and optimization. Nando has published in several top-ranked artificial intelligence journals and conferences and has served the program committee of various artificial intelligence conferences, including AAAI, IJCAI, AAMAS, and CP. He is the recipient of a best student paper award (CMSB, 2013), a most visionary paper award (AAMAS workshop series, 2017), and a best AI dissertation award (AI*IA, 2017).

« Return to Events