基于联邦学习的操作系统漏洞检测方法及装置
By acquiring cross-regional vulnerability data through federated learning, generating a vulnerability relationship network, and performing feature matching, the problem of limited identification capability for cross-device vulnerability detection in existing technologies is solved, achieving more efficient and accurate vulnerability detection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- WUXI YUNWEI INFORMATION TECHNOLOGY CO LTD
- Filing Date
- 2025-07-25
- Publication Date
- 2026-07-17
AI Technical Summary
Existing operating system vulnerability detection methods rely on vulnerability data analysis of the same device or local environment, which cannot effectively link abnormal data across devices, resulting in the underreporting of cross-platform attacks. They also cannot build a global correlation network, thus limiting their identification capabilities.
By employing a federated learning approach, target vulnerability data from devices across regions is acquired. By generating a vulnerability relationship network, extracting key nodes and transmission channels, a vulnerability feature vector set is generated, and feature matching is performed on real-time system data streams to improve the accuracy of vulnerability detection and system identification capabilities.
It has improved the ability to identify vulnerabilities across devices and platforms, enhanced the accuracy of vulnerability detection and the efficiency of system data processing, and ensured the accuracy of vulnerability pattern data and the effectiveness of vulnerability detection in a distributed environment.
Smart Images
Figure CN120639491B_ABST