High-reliability multi-mode anomaly detection method and system for micro-service system

Through multimodal data learning and probabilistic potential coding technology, combined with adaptive fusion methods, the problems of low accuracy and poor reliability of abnormal detection in microservice systems are solved, and efficient multimodal data integration and abnormal detection effects are achieved.

CN119739591AActive Publication Date: 2025-04-01ANHUI UNIV
5 Cites 0 Cited by

Patent Information

Application Number
CN202411799393.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-09
Publication Date
2025-04-01
Estimated Expiration
2044-12-09

AI Technical Summary

Technical Problem

The existing microservice system abnormality detection technology has low accuracy and poor reliability, making it difficult to effectively integrate multimodal data, resulting in noise and information conflicts, affecting the learning and detection effects of abnormal characteristics.

Method used

The discriminant representation of the state of the multimodal data is used to learn the system state, and the characteristic representation of each modal data is learned through label data, and the fusion weight of each modal data is allocated through an adaptive fusion method, combining the fusion weight and the characteristic representation of the modal data for abnormal detection.

Benefits of technology

It greatly improves the accuracy and reliability of abnormal detection of microservice systems, adapts to noise scenarios, enhances the robustness of the model, and can effectively integrate multimodal data to reduce the impact of noise and information conflicts.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to a high-reliability multi-mode anomaly detection method and system for a micro-service system. The multi-modal anomaly detection method for the micro-service system comprises the following steps: collecting various modal data in the micro-service system, respectively encoding the various modal data to obtain feature representations, and respectively sampling and reconstructing each modal feature representation by using Gaussian distribution representation so as to understand noisy multi-modal content. After modal features are obtained, in order to measure different contributions of modal contents, a fusion weight of each piece of modal data is distributed through a self-adaptive fusion method, and finally, the fusion weights and feature representations of the modal data are combined and input into an anomaly detector to judge whether the micro-service system is abnormal or not. Therefore, the accuracy and reliability of anomaly detection of the micro-service system are greatly improved.
Need to check novelty before this filing date? Find Prior Art