Thesis Oral Defense - Elisaweta Masserova

— 5:45pm

Location:
In Person and Virtual - ET - Traffic21 Classroom, Gates Hillman 6501 and Zoom

Speaker:
ELISAWETA MASSEROVA, Ph.D. Candidate, Computer Science Department, Carnegie Mellon University
https://www.andrew.cmu.edu/user/elisawem/

Distributed Cryptography as a Service

Today’s world is undeniably data-driven. The explosion of the Internet has generated vast volumes of data, and the advent of machine learning has unlocked captivating applications that thrive on this data. In such a world, it is evident that the ability to store, transmit, and process data securely is paramount. Distributing trust is one of the fundamental cryptographic principles that enable such security, and it is at the core of key cryptographic tools such as multi-party computation (MPC) and randomness generation. As the demand for secure and reliable cryptographic solutions grows, there is increasing interest in offering such distributed protocols as a service. 

These services are typically expected to run continuously for long periods of time, requiring significant resource commitments from all participating parties. One approach to mitigate this issue is to design distributed cryptographic protocols that are stateless. With such protocols, parties can contribute to the execution of a distributed cryptographic protocol by participating only for a short time, without committing to a long-term computation. 

In this work, we study such mostly stateless protocols. We start by introducing a blockchain-based MPC protocol which does not require parties to be online at the same time and requires no interaction between the participants. We construct this protocol in the blockchain model and under the assumption of what we call Conditional Storage and Retrieval (CSaR) systems. In our next step, we eliminate the CSaR requirement and design a stateless MPC protocol without relying on this assumption. 

More concretely, we focus on the recently introduced You Only Speak Once (YOSO) paradigm. In this model participating parties are allowed to send only a single message; i.e., they speak only once. We improve the state of the art in YOSO MPC by designing a protocol with better communication complexity than the currently known solutions. Then, we focus on improving the efficiency of special-purpose YOSO MPC protocols. Specifically, we consider the task of distributed randomness generation, and design a suite of protocols, each balancing different trade-offs in terms of underlying assumptions, efficiency, and corruption threshold.

Thesis Committee 

Bryan Parno (Co-chair)
Vipul Goyal (Co-chair)
Elaine Shi
Antigoni Polychroniadou (J.P. Morgan AI Research)
Tal Rabin (University of Pennsylvania  / Amazon Web Services)

In Person and Zoom Participation.  See announcement.

Event Website:
https://csd.cmu.edu/calendar/thesis-oral-defense-elisaweta-masserova