多源先验门控增强威胁入侵检测与关联分析方法及系统
By constructing a multi-source prior knowledge system and a knowledge-enhanced gating adapter, the problem of insufficient accuracy and robustness of the detection model in the computing power network hub is solved, and efficient detection of similar attack variants and real-time response in resource-constrained environments are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUIZHOU UNIV
- Filing Date
- 2026-06-16
- Publication Date
- 2026-07-17
AI Technical Summary
In computing power network hubs, existing detection models struggle to effectively distinguish between complex heterogeneous traffic, similar attack variants, class imbalance, and edge real-time detection. Furthermore, they lack a unified expression of multi-source prior knowledge and a dynamic enhancement mechanism, resulting in insufficient detection accuracy and robustness.
A multi-source prior knowledge system is constructed, including structural priors, category conditional statistical priors, and global feature importance priors. Traffic features are dynamically modulated through a knowledge-enhanced gating adapter, and online detection is performed in conjunction with a lightweight temporal detection model to achieve dynamic enhancement or suppression of features.
It improves the ability to detect similar attack variants, reduces false positive and false negative rates, is suitable for deployment in resource-constrained environments, and maintains millisecond-level detection latency.
Smart Images
Figure CN122419989A_ABST