Combining Graph-Based Learning With Automated Data Collection for Code Vulnerability Detection

This paper presents FUNDED (Flow-sensitive vUl-Nerability coDE Detection), a novel learning framework for building vulnerability detection models. Funded leverages the advances in graph neural networks (GNNs) to develop graph-based method capture and reason about program's control, data, call dependencies. Unlike prior work that treats program as sequential sequence or an untyped graph, learns operates on representation of source code, which individual statements are connected other through relational edges. By capturing syntax, semantics flows, finds better code downstream software task. To provide sufficient training data build effective deep model, we combine probabilistic statistical assessments automatically gather high-quality samples from open-source projects. provides many real-life vulnerable complement limited available standard databases. We apply identify vulnerabilities at function level code. evaluate large real-world datasets with programs written C, Java, Swift Php, compare it against six state-of-the-art Experimental results show significantly outperforms alternative approaches across evaluation settings.