Refactoring programs to improve the performance of deep learning for vulnerability detection

This poster was used for a poster presentation about refactoring programs as a method of data augmentation. It earned Honorable Mention at ISU Research Day 2022.

Intuitively, when we refactor programs, we can produce a program that looks quite different syntactically because many tokens are changed, but is identical semantically because the behavior does not change. We applied this technique and found that it is a useful and general method for improving the performance of deep learning-based vulnerability detection.