A new approach to software vulnerability detection based on CPG analysis

Detecting source code vulnerabilities is an essential issue today. In this paper, to improve the efficiency of detecting in software written C/C++, we propose use a combination Deep Graph Convolutional Neural Network (DGCNN) and property graph (CPG). Specifically, 3 main proposed phases research method include: phase 1: building feature profiles code. At step, suggest using analysis techniques such as Word2vec, one hot encoding standardize analyze code; 2: extracting features based on profiles. Accordingly, at phase, model extract 3: classifying extracted 2 find normal containing security vulnerabilities. Some scenarios for comparing evaluating study compared with other approaches have taken show superior effectiveness our approach. Besides, result proves that paper not only correct reasonable, but it also opens up new approach task