EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

SV-GEN: Synergizing LLM-Empowered Variable Semantics and Graph Transformers for Vulnerability Detection

Zhaohui Liu, Haocheng Yang, Wenjie Xie

Future Internet · 27.04.2026 · 18 (5) · 236

DOI: 10.3390/fi18050236 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Future Internet, online erschienen am 27.04.2026. Verfasst von Zhaohui Liu und 2 weiteren Autor:innen. Die Arbeit nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Deep-learning-based vulnerability detection has made substantial progress, but two limitations remain prominent. Sequence-based methods linearize source code and thus weaken the explicit modeling of control-flow and data-flow dependencies. Graph-based methods preserve program structure, yet conventional graph neural networks still have difficulty capturing long-range interactions in large code property graphs (CPGs). In addition, standard CPGs usually lack explicit variable semantics and security-critical node roles, which limits their ability to represent vulnerability-relevant program behavior. To address these issues, we propose SV-GEN, a vulnerability detection framework that combines large-language-model-driven semantic enhancement with hybrid sequence-graph learning. The novelty of SV-GEN lies in introducing a semantically enriched code property graph, termed Sem-CPG, which augments conventional CPGs with variable semantic roles and security-oriented node labels, and in coupling this representation with an adaptive fusion mechanism over structural and sequential views. Specifically, we use a large language model as an external semantic annotator to assign variable roles and identify source, sink, and sanitizer nodes, and then encode the resulting Sem-CPG with a Graph Transformer while modeling the code sequence with GraphCodeBERT. A learnable gating module is further used to adaptively fuse the graph-level and sequence-level representations for final prediction. Experiments on Devign, ReVeal, and DiverseVul show that SV-GEN achieves competitive or superior overall performance across benchmarks, with particularly strong improvements on the large and highly imbalanced DiverseVul dataset.

Zitieren

APA

Liu, Z., Yang, H., & Xie, W. (2026). SV-GEN: Synergizing LLM-Empowered Variable Semantics and Graph Transformers for Vulnerability Detection. Future Internet, 18(5), 236. https://doi.org/10.3390/fi18050236

Vancouver

Liu Z, Yang H, Xie W. SV-GEN: Synergizing LLM-Empowered Variable Semantics and Graph Transformers for Vulnerability Detection. Future Internet. 2026;18(5):236. doi:10.3390/fi18050236

BibTeX
@article{Liu2026,
  author = {Liu, Zhaohui and Yang, Haocheng and Xie, Wenjie},
  title = {SV-GEN: Synergizing LLM-Empowered Variable Semantics and Graph Transformers for Vulnerability Detection},
  journal = {Future Internet},
  year = {2026},
  volume = {18},
  number = {5},
  pages = {236},
  doi = {10.3390/fi18050236}
}
Kostenlos starten →