Non-Interference Preserving and Optimising Compilation with Hyperproperty Simulations
ACM SIGPLAN via YouTube
Learn AI, Data Science & Business — Earn Certificates That Get You Hired
Learn the Skills Netflix, Meta, and Capital One Actually Hire For
Overview
Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off
One annual plan covers every course and certificate on Coursera. 40% off for a limited time.
Get Full Access
Watch this conference talk from PriSC 2025 where Julian Rosemann, Sebastian Hack, and Deepak Garg present their research on hyperproperty simulations, a novel framework for secure compilation that preserves security policies during the compilation process. Learn how this approach goes beyond classical compiler correctness by addressing indistinguishability-based properties like non-interference that are characterized by multiple traces rather than individual ones. Discover how the researchers overcome limitations of existing secure compilation approaches by modeling the preservation of arbitrary k-hyperproperties, allowing for more flexible compiler optimizations while maintaining security guarantees. The presentation demonstrates the framework's expressiveness through a leakage-based non-interference-preserving dead code elimination pass, verified in the Coq proof assistant. This 26-minute talk was presented at the PriSC 2025 workshop sponsored by ACM SIGPLAN.
Syllabus
[PriSC'25] Non-Interference Preserving and Optimising Compilation with Hyperproperty Simulations
Taught by
ACM SIGPLAN