Abstract:Code Language Models (codeLMs) and Graph Neural Networks (GNNs) are widely used in code vulnerability detection. However, a critical yet often overlooked issue is that GNNs primarily rely on aggregating information from adjacent nodes, limiting structural information transfer to single-layer updates. In code graphs, nodes and relationships typically require cross-layer information propagation to fully capture complex program logic and potential vulnerability patterns. Furthermore, while some studies utilize codeLMs to supplement GNNs with code semantic information, existing integration methods have not fully explored the potential of their collaborative effects. To address these challenges, we introduce Vul-LMGNNs that integrates pre-trained CodeLMs with GNNs, leveraging knowledge distillation to facilitate cross-layer propagation of both code semantic knowledge and structural information. Specifically, Vul-LMGNNs utilizes Code Property Graphs (CPGs) to incorporate code syntax, control flow, and data dependencies, while employing gated GNNs to extract structural information in the CPG. To achieve cross-layer information transmission, we implement an online knowledge distillation (KD) program that enables a single student GNN to acquire structural information extracted from a simultaneously trained counterpart through an alternating training procedure. Additionally, we leverage pre-trained CodeLMs to extract semantic features from code sequences. Finally, we propose an "implicit-explicit" joint training framework to better leverage the strengths of both CodeLMs and GNNs. In the implicit phase, we utilize CodeLMs to initialize the node embeddings of each student GNN. Through online knowledge distillation, we facilitate the propagation of both code semantics and structural information across layers. In the explicit phase, we perform linear interpolation between the CodeLM and the distilled GNN to learn a late fusion model. The proposed method, evaluated across four real-world vulnerability datasets, demonstrated superior performance compared to 17 state-of-the-art approaches. Our source code can be accessed via GitHub: https: //github.com/Vul-LMGNN/vul-LMGGNN.

VulLibGen: Generating Names of Vulnerability-Affected Packages via a Large Language Model

MegaVul: A C/C++ Vulnerability Dataset with Comprehensive Code Representations.

Categorizing and Predicting Invalid Vulnerabilities on Common Vulnerabilities and Exposures

VulDetectBench: Evaluating the Deep Capability of Vulnerability Detection with Large Language Models

Identifying Vulnerable Third-Party Libraries from Textual Descriptions of Vulnerabilities and Libraries

Identifying Vulnerable Third-Party Java Libraries from Textual Descriptions of Vulnerabilities and Libraries

Enhancing Large Language Models for Secure Code Generation: A Dataset-driven Study on Vulnerability Mitigation

Vul-LMGNNs: Fusing Language Models and Online-Distilled Graph Neural Networks for Code Vulnerability Detection

Vul-RAG: Enhancing LLM-based Vulnerability Detection via Knowledge-level RAG

How Far Have We Gone in Vulnerability Detection Using Large Language Models

Harnessing Large Language Models for Software Vulnerability Detection: A Comprehensive Benchmarking Study

Towards Effectively Detecting and Explaining Vulnerabilities Using Large Language Models

Transformer-Based Language Models for Software Vulnerability Detection

VulnLLMEval: A Framework for Evaluating Large Language Models in Software Vulnerability Detection and Patching

Combining Software Metrics and Text Features for Vulnerable File Prediction

Detecting software vulnerabilities using Language Models

An Empirical Study of Automated Vulnerability Localization with Large Language Models

VULNERLIZER: Cross-analysis Between Vulnerabilities and Software Libraries

Is Your AI-Generated Code Really Safe? Evaluating Large Language Models on Secure Code Generation with CodeSecEval

VulZoo: A Comprehensive Vulnerability Intelligence Dataset

ChatNVD: Advancing Cybersecurity Vulnerability Assessment with Large Language Models