Enhancing Security Vulnerability Detection in Source Code

Description

In this thesis, we refine the Juliet Test Suite for C/C++ to improve software vulnerability detection.
The dataset has been reorganized, separating each test case into individual 'bad' function and 'good' function
test cases, diverging from the original setup where one 'good' function encompassed multiple good functions.
This restructuring is anticipated to assist in better understanding the capability of Graph Neural Networks (GNNs)
for vulnerability detection. The experiments conducted with this revised dataset aim to provide insights into
GNNs' performance in handling specific and detailed test cases.