Please note: This talk will take place in DC 1304 and online.
Alex Ozdemir, PhD student
Computer Science, Stanford University
We introduce collaborative zk-SNARKs: zero-knowledge proofs about secret data distributed among different parties. These generalize conventional zk-SNARKs, which apply to secret data held by one party. Since collaborative zk-SNARKs do not require secrets to be centralized to create a proof, they might allow for distributed systems with better privacy properties.
We define this new primitive, describe an approach for constructing it, give concrete constructions, implement, and evaluate.
We find a surprising fact: despite the fact that creating a collaborative proof requires a secure multi-party protocol, it can often be created in nearly the same time as a conventional (single-prover) proof.
This work was presented at Usenix Security in 2022.
Bio: Alex Ozdemir is a PhD student at Stanford advised by Dan Boneh and Clark Barrett. He studies the full stack of programmable cryptography using ideas from cryptography, programming languages, and formal methods. Programmable cryptosystems (such as zero-knowledge proofs, multi-party computation, homomorphic encryption, ...) are tools for complex systems that are private and secure — without cryptographic expertise.
To attend this talk in person, please go to DC 1304. You can also attend on Zoom.