Structure-Enhanced Prompt Learning for Graph-Based Code Vulnerability Detection

Recent advances in prompt learning have opened new avenues for enhancing natural language understanding in domain-specific tasks, including code vulnerability detection. Motivated by the limitations of conventional binary classification methods in capturing complex code semantics, we propose a novel...

Full description

Saved in:
Bibliographic Details
Main Authors: Wei Chang, Chunyang Ye, Hui Zhou
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/15/11/6128
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850129052717809664
author Wei Chang
Chunyang Ye
Hui Zhou
author_facet Wei Chang
Chunyang Ye
Hui Zhou
author_sort Wei Chang
collection DOAJ
description Recent advances in prompt learning have opened new avenues for enhancing natural language understanding in domain-specific tasks, including code vulnerability detection. Motivated by the limitations of conventional binary classification methods in capturing complex code semantics, we propose a novel framework that integrates a two-stage prompt optimization mechanism with hierarchical representation learning. Our approach leverages graphon theory to generate task-adaptive, structurally enriched prompts by encoding both contextual and graphical information into trainable vector representations. To further enhance representational capacity, we incorporate the pretrained model CodeBERTScore, a syntax-aware encoder, and Graph Neural Networks, enabling comprehensive modeling of both local syntactic features and global structural dependencies. Experimental results on three public datasets—FFmpeg+Qemu, SVulD and Reveal—demonstrate that our method performs competitively across all benchmarks, achieving accuracy rates of 64.40%, 83.44% and 90.69%, respectively. These results underscore the effectiveness of combining prompt-based learning with graph-based structural modeling, offering a more accurate and robust solution for automated vulnerability detection.
format Article
id doaj-art-196d3c5930744ab998f9527bd44d9779
institution OA Journals
issn 2076-3417
language English
publishDate 2025-05-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-196d3c5930744ab998f9527bd44d97792025-08-20T02:33:07ZengMDPI AGApplied Sciences2076-34172025-05-011511612810.3390/app15116128Structure-Enhanced Prompt Learning for Graph-Based Code Vulnerability DetectionWei Chang0Chunyang Ye1Hui Zhou2School of Cybersapce Security, Hainan University, Haikou 570228, ChinaSchool of Computer Science and Technology, Hainan University, Haikou 570228, ChinaSchool of Computer Science and Technology, Hainan University, Haikou 570228, ChinaRecent advances in prompt learning have opened new avenues for enhancing natural language understanding in domain-specific tasks, including code vulnerability detection. Motivated by the limitations of conventional binary classification methods in capturing complex code semantics, we propose a novel framework that integrates a two-stage prompt optimization mechanism with hierarchical representation learning. Our approach leverages graphon theory to generate task-adaptive, structurally enriched prompts by encoding both contextual and graphical information into trainable vector representations. To further enhance representational capacity, we incorporate the pretrained model CodeBERTScore, a syntax-aware encoder, and Graph Neural Networks, enabling comprehensive modeling of both local syntactic features and global structural dependencies. Experimental results on three public datasets—FFmpeg+Qemu, SVulD and Reveal—demonstrate that our method performs competitively across all benchmarks, achieving accuracy rates of 64.40%, 83.44% and 90.69%, respectively. These results underscore the effectiveness of combining prompt-based learning with graph-based structural modeling, offering a more accurate and robust solution for automated vulnerability detection.https://www.mdpi.com/2076-3417/15/11/6128vulnerability detectionsyntax awareprompt learninggraphoncode property graph
spellingShingle Wei Chang
Chunyang Ye
Hui Zhou
Structure-Enhanced Prompt Learning for Graph-Based Code Vulnerability Detection
Applied Sciences
vulnerability detection
syntax aware
prompt learning
graphon
code property graph
title Structure-Enhanced Prompt Learning for Graph-Based Code Vulnerability Detection
title_full Structure-Enhanced Prompt Learning for Graph-Based Code Vulnerability Detection
title_fullStr Structure-Enhanced Prompt Learning for Graph-Based Code Vulnerability Detection
title_full_unstemmed Structure-Enhanced Prompt Learning for Graph-Based Code Vulnerability Detection
title_short Structure-Enhanced Prompt Learning for Graph-Based Code Vulnerability Detection
title_sort structure enhanced prompt learning for graph based code vulnerability detection
topic vulnerability detection
syntax aware
prompt learning
graphon
code property graph
url https://www.mdpi.com/2076-3417/15/11/6128
work_keys_str_mv AT weichang structureenhancedpromptlearningforgraphbasedcodevulnerabilitydetection
AT chunyangye structureenhancedpromptlearningforgraphbasedcodevulnerabilitydetection
AT huizhou structureenhancedpromptlearningforgraphbasedcodevulnerabilitydetection