Improved Private Information Retrieval Schemes from Matching Vectors and Derivatives
Institute for Advanced Study via YouTube
Google AI Professional Certificate - Learn AI Skills That Get You Hired
Future-Proof Your Career: AI Manager Masterclass
Overview
Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off
One annual plan covers every course and certificate on Coursera. 40% off for a limited time.
Get Full Access
Explore a Computer Science/Discrete Mathematics Seminar presentation on "Improved Private Information Retrieval Schemes from Matching Vectors and Derivatives" delivered by Swastik Kopparty from the University of Toronto. Delve into the evolution of Private Information Retrieval (PIR) methods, which allow users to access database entries without revealing which information they're retrieving. Learn about significant breakthroughs in the field and discover a new, simplified PIR scheme for t=2 servers that offers superpolynomial improvements over previous approaches. The presentation covers how this scheme achieves 3-server PIR communication of exp((log n)^(1/3)), a substantial improvement over Efremenko's previous exp((log n)^(1/2)) approach. This 79-minute seminar, presented at the Institute for Advanced Study, represents joint work with Fatemeh Ghasemi and Madhu Sudan.
Syllabus
10:30am|Simonyi Hall 101 and Remote Access
Taught by
Institute for Advanced Study