Samplable Anonymous Aggregation for Private Federated Data Analysis
Kunal Talwar
Abstract:
In this talk, I will revisit the problem of designing scalable protocols for private statistics and private federated learning when each device holds its private data. Locally differentially private algorithms require little trust but are provably limited in their utility. Centrally differentially private algorithms can allow significantly better utility but require a trusted curator. This gap has led to significant interest in the design and implementation of simple cryptographic primitives, that can allow central-like utility guarantees without having to trust a central server.
I will discuss a new primitive that allows for efficient implementation of several commonly used algorithms, and allows for privacy accounting that is close to that in the central setting without requiring the strong trust assumptions it entails. I will discuss a system architecture that implements the new primitive. Time permitting, I will also talk about Euclidean norm verification in this setting, and discuss some open questions.
Based on these joint works with several wonderful collaborators: https://arxiv.org/abs/2307.15017 and https://arxiv.org/abs/2311.10237
Bio:
Kunal Talwar is a Research Scientist at Apple, leading a research group focusing on foundations of ML and Private Data Analysis. His research interests span various aspects of Computer Science including Differential Privacy, Machine Learning, Algorithms and Data Structures. Prior to joining Apple, he worked at Microsoft Research in Silicon Valley from 2004 to 2014, and at Google Brain from 2014 to 2019. He has made major contributions to Differential Privacy, Metric Embeddings and Discrepancy Theory. His work has been recognized by the Casper Bowden Privacy Enhancing Technologies award in 2009, ICBS Frontiers of Science Award in 2024, and best paper awards at ICLR 2017 and FORC 2022.