An artificial intelligence-based network attack tracing and defense method and system

By constructing a dynamic evolution graph structure and a generative adversarial network, the problems of multi-source data fusion and proactive defense are solved, enabling highly accurate network attack tracing and defense, and forming a self-evolving intelligent defense system.

CN122437688APending Publication Date: 2026-07-21ASPIRE TECH (SHENZHEN) LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
ASPIRE TECH (SHENZHEN) LTD
Filing Date
2026-04-20
Publication Date
2026-07-21

AI Technical Summary

Technical Problem

Existing network attack attribution technologies are unable to effectively integrate multi-source heterogeneous data, lack attribution capabilities when facing proxy and encrypted traffic, and lack proactive defense capabilities, making them unable to cope with attacks on unknown vulnerabilities.

Method used

We construct a dynamic evolution graph structure based on graph neural networks, generate deceptive network assets through generative adversarial networks to actively lure attackers, and improve the accuracy of attribution tracing and defense capabilities through incremental learning.

Benefits of technology

It achieves a source location accuracy rate of up to 96.5% in complex network environments, can predict attacker behavior in real time, and forms a self-evolving intelligent defense system that dynamically adapts to new attack methods.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122437688A_ABST
    Figure CN122437688A_ABST
Patent Text Reader

Abstract

The application discloses a network attack tracing and defense method and system based on artificial intelligence and belongs to the technical field of network security. The method comprises the following steps: acquiring multi-modal network security data and performing pretreatment; based on the pretreated data, a dynamic evolution graph structure representing network attack behaviors is constructed, time-space features of the graph structure are extracted through a graph neural network, attack paths are reconstructed, and attack behaviors are predicted; based on the reconstruction result of the attack paths and the prediction result of the attack behaviors, a deceptive network asset matched with the characteristics of an attacker is generated; the deceptive network asset is deployed into a real network to lure the attacker and collect interactive data of the attacker; and the interactive data is fed back to the graph neural network for incremental learning, and model parameters of attack path reconstruction and attack behavior prediction are updated. The application realizes intelligent perception, accurate tracing and active defense of network attacks by constructing a dynamic evolution attack knowledge graph and actively luring attackers by using a generative adversarial network.
Need to check novelty before this filing date? Find Prior Art