IC Colloquium: Privacy at Scale
By: Alexandra Henzinger - MIT
IC Faculty candidate
Abstract
Today, we have no choice but to entrust digital services with our data: our queries, clicks, prompts, and much more. In doing so, we leave behind a digital footprint that reveals where we go, what we care about, and who we are. Users pay the price for this lack of privacy; our data is exposed to thousands of parties, resold for profit, trained on, or stolen by attackers.
This talk argues that digital services need not come at the cost of our privacy. As a step towards this goal, I will describe the design of a search engine that learns nothing about what its users are searching for. Building and scaling such a privacy-protecting system requires changes across the stack: from system architecture, to algorithms, to cryptographic operations. The result is the first system that can privately search over hundreds of millions of webpages in seconds, and whose core techniques have seen adoption in Apple’s iOS (“Enhanced Visual Search”).
Bio
Alexandra Henzinger is a Ph.D. student in the Parallel and Distributed Operating Systems group at MIT, advised by Henry Corrigan-Gibbs. Her research builds computer systems that deliver privacy and security at scale. Alexandra has received an NSF GRFP fellowship, an MIT EECS Great Educators fellowship, and was selected as a 2024 EECS Rising Star. Her work has been invited to the Journal of Cryptology.
More information
IC Faculty candidate
Abstract
Today, we have no choice but to entrust digital services with our data: our queries, clicks, prompts, and much more. In doing so, we leave behind a digital footprint that reveals where we go, what we care about, and who we are. Users pay the price for this lack of privacy; our data is exposed to thousands of parties, resold for profit, trained on, or stolen by attackers.
This talk argues that digital services need not come at the cost of our privacy. As a step towards this goal, I will describe the design of a search engine that learns nothing about what its users are searching for. Building and scaling such a privacy-protecting system requires changes across the stack: from system architecture, to algorithms, to cryptographic operations. The result is the first system that can privately search over hundreds of millions of webpages in seconds, and whose core techniques have seen adoption in Apple’s iOS (“Enhanced Visual Search”).
Bio
Alexandra Henzinger is a Ph.D. student in the Parallel and Distributed Operating Systems group at MIT, advised by Henry Corrigan-Gibbs. Her research builds computer systems that deliver privacy and security at scale. Alexandra has received an NSF GRFP fellowship, an MIT EECS Great Educators fellowship, and was selected as a 2024 EECS Rising Star. Her work has been invited to the Journal of Cryptology.
More information
Practical information
- General public
- Free
Contact
- Host: Alessandro Chiesa