Network attack data automatic generation method and system based on attack principle

Abstract

The invention provides a network attack data automatic generation method and system based on an attack principle, and the method comprises the steps: building an attack principle model according to the public information of a network attack, wherein the network attack public information comprises network attack analysis code data and a network attack public report; establishing a data generation system based on an attack principle according to the attack principle model; and training a network security defense model through the data generated by the data generation system based on the attack principle. According to the invention, vivid and effective target data can be generated in a large-batch and customized mode through editing of the configuration file, the data requirements of the network security defense AI model in the training and testing stages can be met, and according to the generation system designed and achieved through the method, a network security defense AI model user can edit the configuration file simply, and therefore, a large amount of target attack data can be generated in a customized manner for model training and testing.

Description

technical field [0001] The invention relates to the technical field of network security, in particular to a method and system for automatically generating network attack data based on attack principles. Background technique [0002] In recent years, the rapid development of artificial intelligence technology has been widely used in various fields. Similarly, artificial intelligence technology can help network security defense, which can effectively enhance the defense's potential. At present, there are already some network security products that apply artificial intelligence technology, and their security defense capabilities and industry competitiveness have been significantly improved. As we all know, traditional artificial intelligence models (AI models) such as machine learning, neural network, and deep learning not only require relevant theoretical foundations, but also cannot do without the support of data sets. It is worth noting that the prerequisite for the AI ​​m...

AI Technical Summary

Problems solved by technology

[0005] The present invention provides a method and system for automatically generating network attack data based on attack principles. The method can customize and generate a large amount of realistic network attack data with controllable and adjustable completeness according to the needs of AI model users. These data are used to meet The AI ​​model needs the quality and quantity of training and testing data; finally, the real and available attack data is used as the verification data of the trained model, so as to solve the shortage of available data and insufficient completeness in the current AI model training process. question

Images