Secure Computation is a powerful concept from cryptography that enables collaboration in the absence of trust. Despite its great potential for solving practical problems in collaborative situations, it has not yet been widely adopted in practice. Indeed, until recently there were few practical implementations of secure computation protocols, and even recent implementations were forced to restrict themselves to weak forms of security for the sake of efficiency. This is because the dominant paradigm for achieving strong security, since the invention of the first such protocols, has relied on zero-knowledge proofs, and yields protocols that are too inefficient even for simple computations.
We are developing radically different new architectures for efficient secure computation protocols that bypass the need for such zero-knowledge proofs. Our architectures are based on a novel principled approach to developing new secure computation protocols, with consequences to the theory and practice of modern cryptography. Our research will identify new (partial) security properties inherent in simple protocols, and study how such properties can add up to strong security guarantees through carefully developed methods for composing protocols.
Secure multiparty computation is an idea whose time has come, as evidenced by the several projects around the world engaged in translating the theoretical results to practical implementations, and the number of research projects outside cryptography that seek to exploit it. Widespread availability and deployment of secure computation protocols has the potential to be a disruptive technology, enabling new avenues of cooperation in areas with sensitive information. Apart from the technological impact, we strive to bring the advances in cryptography to a broader computer science audience, by integrating our research into graduate and undergraduate education, and outreach efforts.
