C/C++ vulnerability static detection method based on neural network and deep learning

Abstract

The invention relates to a C / C++ vulnerability static detection method based on a neural network and deep learning, and belongs to the field of information safety. The method comprises the following steps: performing data cleaning and code slicing on a to-be-detected source code; converting the code slice into an input vector of the CVDF-LZW code; converting the input vector into equal-length input, and performing normalization processing; inputting the keyword information of the normalized vector into 14 neurons, wherein the output of the normalized vector is a 14-dimensional vector; inputting the normalized vectors into different neural networks, and outputting corresponding operation dimension vectors in the vulnerability feature vectors; integrating the 14-dimensional vector and the operation dimension vector into a 35-dimensional vulnerability feature vector; extracting associated neurons from the vulnerability feature vectors in a non-full connection mode, obtaining a six-dimensional output vector through linear function conversion, and converting corresponding vulnerability type feature values into corresponding probability values through softmax multi-classification layers.According to the method, high-precision detection can be realized for various vulnerabilities, and the efficiency and the adaptability are improved.

Description

technical field [0001] The invention belongs to the field of information security, and in particular relates to a static detection method for C / C++ vulnerabilities based on neural network and deep learning. Background technique [0002] At present, software security has become one of the hot topics that people pay more and more attention to. According to the CSEC2017 model, knowledge domains are divided into six aspects: data security, software security, system security, personnel security, organizational security, and social security, with software security among them. With the explosive growth of software, the research on software security vulnerabilities is facing severe challenges. In the field of source code-oriented vulnerability static detection, the traditional manual detection method is inefficient and cannot effectively deal with the ever-increasing software vulnerabilities in various forms, and the recognition accuracy depends on the technical level of software s...

AI Technical Summary

Problems solved by technology

[0005] The technical problem to be solved by the present invention is how to provide a static detection method for C / C++ vulnerabilities based on neural network and deep learning, so as to solve the problems of low efficiency, poor adaptability, slow detection speed and inability to detect differences in morphological characteristics of existing vulnerability detection methods Larger vulnerabilities and other issues

Images