Traditional vulnerability detection methods

The gap

Evidence profile

Sourced from the future work and stated research gap of the source papers, classified as general, drawn from work published between 2024 and 2026, spanning 2 journals. Those papers have been cited 86 times in total.

Supporting evidence — 3 representative gaps

• Large Language Model for Vulnerability Detection and Repair: Literature Review and the Road Ahead (2024) · ACM Transactions on Software Engineering and Methodology · cited 86× · doi

  The use of Large Language Models (LLMs) for vulnerability detection and repair has been garnering increasing attention. This paper presents a systematic literature review of 58 primary studies on LLMs for vulnerability detection and repair. This review begins by analyzing the types of LLMs used in primary studies, shedding light on researchers’ preferences for different LLMs. Subsequently, we categorized a variety of techniques for adapting LLMs. Through our analysis, this review also identifies the limitations in this field and proposes a research roadmap outlining promising avenues for future exploration. In the future, we plan to broaden this literature review by incorporating additional vulnerability-related tasks, such as vulnerability localization and vulnerability assessment.

  generalfuture workevidence 5/5
  Keywords: https database llms vulnerability review detection repair literature primary future national foundation arxiv ieee sciencedirect
• APPLICATION OF LARGE LANGUAGE MODELS IN SMART CONTRACT VULNERABILITY DETECTION (2026) · Journal of Computer Science and Electrical Engineering · doi

  Traditional vulnerability detection methods have limitations such as high false positive rates and poor adaptability to complex logical vulnerabilities. Large Language Models have exceptional code understanding and reasoning capabilities, but their application in smart contract vulnerability detection is still in its early stages.

  generalstated research gapevidence 5/5
  Keywords: traditional vulnerability detection methods have limitations high false
• VFDelta: A Framework for Detecting Silent Vulnerability Fixes by Enhancing Code Change Learning (2026) · ACM Transactions on Software Engineering and Methodology · doi

  Existing methods for detecting vulnerability fixes do not effectively highlight nuanced differences in code changes. Previous approaches fine-tune code embedding models and classification models separately, limiting overall effectiveness. There is a need for a framework that can capture fine-grain changes in code and improve the detection of vulnerability fixes.

  generalstated research gapevidence 5/5
  Keywords: existing methods detecting vulnerability fixes effectively highlight nuanced

Questions about this gap