一种基于反汇编语义信息的可执行软件溢出漏洞预测方法
By designing overflow vulnerability behavior patterns from disassembled semantic information, utilizing the transformation features of convolutional neural networks and temporal convolutional networks, and combining them with a multilayer perceptron model, the problem of insufficient generalization ability in overflow vulnerability detection in existing technologies is solved, achieving accurate prediction and efficient detection of overflow vulnerabilities in executable software.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- YANSHAN UNIV
- Filing Date
- 2023-12-15
- Publication Date
- 2026-07-17
AI Technical Summary
Existing vulnerability detection methods are unable to comprehensively detect overflow vulnerabilities in executable software or fail under complexity and encryption measures, and the diversity of features after disassembly leads to insufficient generalization ability.
The design incorporates overflow vulnerability behavior patterns from disassembled semantic information. By converting binary and assembly code features using convolutional neural networks and temporal convolutional networks, and combining this with a multilayer perceptron model, overflow vulnerability features are accurately extracted and input to improve prediction generalization capabilities.
It achieves accurate prediction of executable software overflow vulnerabilities, improves the generalization ability and applicability of prediction, and reduces the consumption of computing resources.
Smart Images
Figure CN117874762B_ABST