Feedback-directed fuzzing (a.k.a. coverage-guided or grey-box fuzzing) algorithms and their efficient implementation have shown themselves to be a force in the field of automated testing and bug discovery. These algorithms do great at exploring search spaces through mutation, when each evaluation run is cheap. Large language models have emerged as a force in the generation of both natural language and code, showing promise for search tasks over these spaces. This keynote will cover projects that show the power of feedback-directed-fuzzing on its own, and how large language models can be used most effectively in conjunction with and complement to automated testing techniques.

Caroline Lemieux is an Assistant Professor at the University of British Columbia. Her research interests centre around improving the correctness and reliability of software systems by developing automated methods for engineering tasks such as testing, debugging, and comprehension. Her work on fuzz testing has been awarded an ACM SIGSOFT Distinguished Paper Award, Distinguished Artifact Award, Tool Demonstration Award, and Best Paper Award (Industry Track). Before UBC, she was a postdoctoral researcher at Microsoft Research NYC, studying the use of deep learning for automated testing. She completed her PhD at UC Berkeley, advised by Koushik Sen; there, she was the recipient of a Berkeley Fellowship for Graduate Study, and a Google PhD Fellowship in Programming Technologies and Software Engineering. She received her B.Sc. in Computer Science and Mathematics at the University of British Columbia.