Private Information Retrieval: Improvements and Limitations
Kevin Yeo, Google and Columbia University
November 23, 2023 2:00pm, in DC 1302 and Zoom
Private information retrieval (PIR) is a very promising cryptographic tool that enables privacy-preserving data querying that has endless implications to real-world applications. Unfortunately, PIR’s high cost remains a hindrance in widespread adoption. In this talk, I cover three PIR topics. First, I will motivate the importance of PIR by walking through a real-world use case at Google deployed today. Next, I will present recent improvements to the concrete efficiency for keyword and batch PIR. Finally, I will discuss recent developments in PIR with preprocessing and explain some limitations and lower bounds to the approach.
Kevin Yeo is a research engineering manager at Google NYC in the Private Computing group as well as a PhD student at Columbia University in the CS theory group. His research interests include cryptography, security, privacy and data structures spanning from understanding theoretical limits to building large-scale systems with strong privacy guarantees. His work has received two USENIX Security distinguished paper awards (2019 & 2023). Kevin's research, including Password Checkup and Private Set Membership, has been deployed to billions of users across Android and Chrome and Kevin's work has also influenced IETF standards for blind signatures.