Kihong Heo, Mukund Raghothaman (soon to join USC CS), Xujie Si, and Mayur Naik are the recipients of the Distinguished Paper Award at the Conference on Programming Language Design and Implementation (PLDI 2019) for their paper "Continuously Reasoning about Programs using Differential Bayesian Inference". Their abstract could be found at https://www.youtube.com/watch?v=TebtiCG4HCA
Software developers are struggling to manage the rapidly growing size and complexity of the systems they develop, and are increasingly looking to sophisticated automatic techniques to help them improve the quality of their code. Static analysis is one such popular technique, and promises to automatically find bugs in software before it is even run. Recent years have witnessed the widespread adoption of static analysis tools by industry, from aerospace and safety-critical medical applications to security-conscious cloud service providers. Unfortunately, the thoroughness of static analyzers is a double-edged sword, and developers are often overwhelmed by the number of warnings produced by tools. In their paper, Dr. Naik, with his postdocs Kihong Heo and Mukund Raghothaman, and his Ph.D.c Xujie Si, developed an AI-based system to prioritize these alarms based on how relevant they are to the developers’ immediate changes. Their system, Drake, will be more effective at identifying vulnerabilities, and promises to eliminate defects as soon as they are introduced into the code-base. Our PRECISE team plans on deploying Drake on commercial software hosting services such as GitHub, in the coming months.