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
This conference talk presents a novel approach to automatically detect type unsoundness in libraries that use unsafe features internally. Learn how researchers from Imperial College London leverage under-approximate reasoning and separation logic to identify memory safety bugs in code. The presentation demonstrates how incorrectness logic can be used to refute type assignments, providing a scalable method for bug detection in systems with ownership type systems. The talk was delivered at the Theory and Practice of Static Analysis workshop (TPSA'25) in January 2025, sponsored by ACM SIGPLAN.
