Secure multi-party computation is a subfield of cryptography that focuses on allowing parties to jointly compute a function while keeping their inputs private from each other. It originated in the late 1970s with the development of mental poker, a cryptographic method for simulating game playing without a trusted third party. By the late 1980s, researchers had published papers demonstrating how to securely compute any function in this setting.
UC Berkeley
Fall 2022
This graduate seminar focuses on the development of secure systems built from decentralized trust, including end-to-end encryption systems and secure collaborative learning. It requires a solid introduction to cryptography and systems. Topics include blockchain, smart contracts, and zero-knowledge proofs, among others.
No concepts data
+ 19 more concepts