基于图神经网络的软件源码漏洞检测方法

Software source code vulnerability detection method based on graph neural network

常规的漏洞检测方法通过激活函数的方式,标识出存在漏洞的代码,少有对代码中的词义进行分析,导致检测效果不佳。因此,设计了基于图神经网络的软件源码漏洞检测方法。提取软件源码漏洞特征,针对漏洞特征进行词义分析。利用图神经网络构建软件源码漏洞检测模型,对源码漏洞做二次检测,去掉重读的源码,得到正确的代码。采用对比实验,验证了该方法的检测效果更佳。

The conventional methods of vulnerability detection identify the code with vulnerability by means of activation function,and seldom analyze the meaning of the words in the code,which results in poor detection results.Therefore,a software source code vulnerability detection method based on graph neural network is designed.The software source code vulnerability features are extracted and their word meaning is analyzed.The software source code vulnerability detection model is constructed using graph neural network,and the source code vulnerability is detected twice to remove the reread source code and get the correct code.Comparison experiments are used to verify that the method has better detection results.