Strange Loop

Your Secrets are Safe with Julia: A Compiler for Secure Computation

Homomorphic Encryption (HE) provides the ability to compute functions over encrypted data without the need to decrypt. By computing on the encrypted values, HE is able to ensure that private data stays private. Traditionally, HE has been a tool that only cryptography experts have had the expertise to use. In this talk we will explore Ramparts, a compiler that translates a subset of the Julia programming language into a form suitable for computation under HE. Because the computation model for HE is different than traditional compute models we required a strange kind of compiler. Ramparts provides a proof of concept for the idea that data scientists, and programmers generally, should not have to compromise on the tools they use in order to take advantage of homomorphic encryption.

Jason Dagit

Jason Dagit

Galois, Inc

I joined Galois in 2009 and have worked on a wide range of projects ranging over identity management, mobile security, static analysis, fully homomorphic encryption, and control flow integrity. I have a keen interest in programming languages and making theoretical things practical. I hold an M.S. in Computer Science from Oregon State University and served on the committee between 2011 and 2014. I enjoy spending time with my two lovely cats, studying computer architectures from the 80s, and participating in speedgaming classic NES and SNES era games.