search menu icon-carat-right cmu-wordmark

Towards Security Defect Prediction with AI

October 2018 Poster
Nathan M. VanHoudnos

This poster describes research comparing a state-of-the-art AI system to existing static analysis approaches for defect prediction.

Publisher:

Software Engineering Institute

Abstract

In this project, the SEI investigated the limits of the current state-of-the-art AI system for detecting buffer overflows and compared it with current static analysis tools. Researchers also developed a code generator, sa-bAbI, capable of producing an arbitrarily large number of code samples of controlled complexity.