Unknown network threat attack and defense method and system based on gene evolution

CN116760576BActive Publication Date: 2026-02-27WUHAN UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202310606083.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-05-23
Publication Date
2026-02-27
Estimated Expiration
2043-05-23

AI Technical Summary

Technical Problem

Existing technologies are vulnerable to unknown threats in cyber-physical system network intrusion detection, and existing adversarial sample generation methods are computationally complex and difficult to apply to real-time detectors.

Method used

We employ an improved genetic algorithm and generative adversarial network framework to design an adversarial sample generation method based on gene evolution. This method generates simulated diverse attack traffic data to attack black-box models and improves detection performance through adversarial training.

Benefits of technology

The generated attack samples can effectively induce intrusion detectors to misclassify, while improving the ability to resist attacks and defend against them, thus enhancing the robustness of network intrusion detection.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116760576B_ABST
    Figure CN116760576B_ABST
Patent Text Reader

Abstract

The application discloses a kind of unknown network threat attack and defense method, system based on gene evolution, it is related to a kind of method and corresponding defense mechanism for deep learning model attack, belong to artificial intelligence information security technical field.The method in it uses network intrusion detection dataset, designs new fitness calculation method, generates the adversarial sample that is closer to real data, i.e.unknown network threat attack sample using improved genetic algorithm, attacks network intrusion detection model under black box state, and proposes defense measures based on adversarial generation network architecture model for the spatial distribution features of these unknown network threat attack samples.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Artificial intelligence optimization system and method adopting adversarial resistance training

    CN109766991A

  • Network intrusion detection method and system based on adversarial auto-encoder, and medium

    CN115913643A