Title: The Cryptographer’s Symphony: Composing Efficient and Secure Computation
Date: Monday, April 7, 2025
Time: 11:00am - 12:30pm (ET)
Location: Coda C0908 "Home Park"
(Meeting ID: 216 539 361 674, Passcode: VZ3uj7AG)
Stanislav Peceny
Ph.D. Candidate in the School of Cybersecurity and Privacy
Georgia Institute of Technology
Dissertation Defense Committee:
Prof. Vladimir Kolesnikov (Advisor, College of Computing, Georgia Institute of Technology)
Prof. Alexandra Boldyreva (College of Computing, Georgia Institute of Technology)
Prof. David Heath (School of Computing and Data Science, UIUC)
Prof. Wenke Lee (College of Computing, Georgia Institute of Technology)
Prof. Vassilis Zikas (College of Computing, Georgia Institute of Technology)
Abstract:
In this work, I improve concrete efficiency of secure multiparty computation (MPC) by (1) designing new protocols that address key bottlenecks of generic MPC primitives and techniques, such as the preprocessing of correlated randomness, improved handling of complex control flow and expensive memory access, and (2) creating special-purpose protocols for commonly used functionalities, such as merging, sorting, and permuting that can be integrated into generic MPC for enhanced performance.
More lyrically, I envision secure computation as a symphony, an intricate composition where every element must harmonize to create a seamless performance. Just as a symphony relies on each instrument in an orchestra, secure computation depends on efficient generic primitives, each playing a crucial role in the overall arrangement. For the symphony to be harmonious, every instrument must produce clear, well-tuned notes. Likewise, in secure computation, these foundational primitives must be optimized for efficiency. My work focuses on improving pseudorandom correlation generators and random access, ensuring these fundamental components are efficient.
Yet, individual excellence alone does not create a masterpiece. A symphony is not merely a collection of instruments playing in isolation. It thrives on synchronization, where each component contributes to a cohesive whole. Similarly, secure computation requires generic techniques that integrate these primitives into an efficient framework. In this area, I work on improving conditional branching, ensuring that secure computation can make decisions efficiently without compromising privacy.
Within an orchestra, instruments are grouped into sections, each with a distinct role, yet all must be carefully tuned and coordinated to achieve harmony. In secure computation, these sections correspond to special-purpose protocols, built from generic primitives but tailored for specific tasks. Just as a strings section enhances a full orchestra, these specialized protocols improve the efficiency and practicality of secure computation. My work in this area includes optimizing secure merge, sort, and permutation, designing these fundamental building blocks to seamlessly integrate within larger secure computation frameworks.
By advancing both the foundational primitives and the techniques that orchestrate them, through generic MPC alongside finely tuned special-purpose MPC, we can achieve a symphony of secure computation that is both powerful and elegant. For this reason, I name my thesis: The Cryptographer’s Symphony.