Thursday, March 23 2023
11:30 - 13:00

Alladi Ramakrishnan Hall

Robustness in Cryptography

Sruthi Sekar

UC Berkeley

This talk will be in hybrid mode with the speaker joining us online. Please find below the zoom link:

Join Zoom Meeting
zoom.us/j/96200820854

Meeting ID: 962 0082 0854
Passcode: 115103


Standard security definitions and adversarial models in several widely-used cryptographic schemes like encryption and authentication break under a notorious class of attacks called *side-channel attacks. *These give additional information about a secret to the adversary than what is intended by the actual model*. *Two such class of attacks include *leakage attacks *(which deal with passive adversaries gaining extra information by observing something like power consumption, or acoustics, etc) and *tampering attacks *(which deal with active adversaries gaining extra information by actively tampering the system through heating or cutting wires, etc).

Several lines of research have focused on making existing schemes *robust *against side-channel attacks. In this talk, I will present our contributions on making existing cryptographic primitives resilient to leakage and tampering attacks in an efficiency-preserving way. Some such primitives that I will talk about include non-malleable codes, zero-knowledge proofs (that offer protection against the active tampering attacks), and leakage-resilient secret sharing, big-key encryptions(that offer protection against leakage attacks).

There are two crucial sides to building robust schemes. One is the theoretical side that involves defining the correct mathematical security definition, understanding the capacity of the model, and designing asymptotically optimal protocols. The other is to deploy the solutions in practice and overcome the challenges in attaining concrete efficiency. Through this talk, I will present both these facets of my research. The techniques that are needed to build these systems, while at the heart are cryptographic, require a blend of two other important areas -- pseudorandomness and coding theory. I will highlight these techniques through the course of my talk.



Download as iCalendar

Done