Generative AI-based malicious code analysis method and system

• Main Authors: XUE SUJIN, HU MIAOXIN, ZHENG JIANMIN
• Format: Patent
• Language: Chinese; English
• Published: 08.09.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention provides a malicious code analysis method and system based on generative AI.The method comprises the following steps that malicious code data are collected and preprocessed, a malicious code sample is obtained, and the malicious code sample comprises a binary file and a source code; performing feature extraction on the binary file by adopting a static analysis method to obtain a first feature vector, and performing feature extraction on the source code by adopting TF-IDF to obtain a second feature vector; a generative AI model is constructed, the first feature vector and the second feature vector are utilized to train the generative AI model, and the generative AI model comprises an embedded layer, two LSTM layers, an input repetition layer and a full connection layer; evaluating the trained generative AI model, inputting the malicious code to be tested into the generative AI model for analysis, and finally obtaining a prediction result. The method has the advantages of high efficiency, accuracy