Dov Gordon, Assistant Professor, Computer Science, is studying secure computation.
Secure computation allows multiple parties to compute on a distributed data set, without revealing anything other than the result of the computation. In theory, it offers the possibility of leveraging user data in all the ways we do today, without sacrificing user privacy.
Over the last decade, the performance of secure computation has been so improved that it has nearly reached theoretical limits, yet data sizes and the number of data sources continue to grow, unabated.
For this project, Gordon will complete three tasks. First, he will look at the problem of computing on large volumes of data, among only a few computing parties. Second, he will explore computing that involves a large number of parties. Third, he will consider the problem of storing data in a dynamic, distributed data plane, and interacting with the distributed data to perform computations satisfying varying security and privacy requirements.
In this project, Gordon is building on previous research he conducted examining privacy, efficiency and the associated tradeoffs. In this work, he proposes to further explore such tradeoffs, developing new relaxations, and deepening our understanding of recently proposed relaxations, with the aim of ensuring the relevance of secure computation in the modern data environment.
Gordon received $514,202 from the National Science Foundation for this work. Funding began in February 2020 and will conclude in late January 2025.