NFV异常检测方法、装置、设备、存储介质和程序产品
By constructing a collaborative architecture between detection agents and annotation agents, and utilizing reinforcement learning and large language models for multimodal data processing, the problem of low efficiency and accuracy in NFV anomaly detection is solved, achieving self-iterative optimization and efficient fault location.
Patent Information
- Application Number
- CN202610414140.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-03-31
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies for NFV anomaly detection have low efficiency and accuracy, are difficult to adapt to dynamic log characteristics, and cannot effectively correlate multi-source information, making it difficult to locate the root cause of complex faults.
A dual-agent architecture is constructed, in which a detection agent and a labeling agent collaborate. Through a reinforcement learning policy network and a large language model, multimodal data features are extracted and fused to generate a state vector. The detection agent is automatically updated when a fault cannot be identified. Generative adversarial networks are used to generate synthetic log samples for enhanced training, thereby achieving self-iterative optimization.
It improves the efficiency and accuracy of NFV anomaly detection, reduces the workload of manual annotation, and enables the system to continuously adapt to dynamic network environments, thereby improving the accuracy and efficiency of root cause localization for complex faults.
Smart Images
Figure CN122420188A_ABST