Software Systems
CONCORD: Clone-Aware Contrastive Learning for Source Code improves existing pre-trained models to learn better representations that consequently become more efficient in both identifying semantically equivalent programs and differentiating buggy from non-buggy code.
Spoq significantly reduces the time and manual effort needed to verify software systems.
Professor Steven Bellovin remembers his mentor Frederick P. Brooks, Jr., in a blog post about the computer scientist and author of The Mythical Man-Month.
About
The Software Systems group pursues fundamental research in all aspects of the design, implementation, analysis, verification, and evaluation of software systems.
The group conducts research with systems at all scales, from handheld devices to cloud computing data centers. They take an experimental systems approach in building real systems to investigate new research ideas, create tools to enable developers to quickly and correctly build complex systems, and teach students how to do it.
Software Systems Laboratory 
Programming Systems Laboratory 
Languages and Compilers Group 
Database Research Group 
Add to GMail