Self-adaptive distributed support vector regression method for resisting Byzantine fault

By employing an adaptive distributed support vector regression method, utilizing a smooth objective function and MOM aggregation mechanism, the robustness and asymptotic efficiency issues of distributed regression methods under Byzantine faults are addressed, achieving efficient prediction and stability in complex environments.

CN122087769APending Publication Date: 2026-05-26QINGDAO UNIV
0 Cites 0 Cited by

Patent Information

Application Number
CN202610205386.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-02-12
Publication Date
2026-05-26

AI Technical Summary

Technical Problem

Existing distributed support vector regression methods suffer from performance degradation under Byzantine fault conditions, making it difficult to achieve robust estimations, and traditional improved methods have low asymptotic efficiency.

Method used

An adaptive distributed support vector regression method is adopted. By introducing the smoothing concept to approximate the objective function, and combining the MOM aggregation mechanism and robust gradient estimation, a robust global gradient update mechanism is designed to suppress the influence of abnormal machines and improve the asymptotic efficiency of global gradient estimation.

Benefits of technology

In the presence of Byzantine faults, it improves the robustness and prediction accuracy of distributed regression methods, making it applicable to a variety of real-world distributed systems. It shows significant advantages, especially in scenarios with potentially malicious machines or unreliable devices, and possesses good scalability and computational efficiency.

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

Abstract

The invention discloses an adaptive distributed support vector regression method for resisting a Byzantine fault, and the method comprises the steps: carrying out the smoothing of an objective function of an SVR model, and constructing an empirical risk minimization objective function of the SVR; enabling t to be equal to 1, and enabling the main machine to obtain initial parameter estimation on local data of the main machine; in the t-th iteration process, MOM estimation is obtained based on an MOM aggregation method and serves as global gradient estimation; then obtaining a steady global gradient on the basis of global gradient estimation; after the master machine obtains a steady global gradient, global parameters are updated by adopting a gradient descent method; and finally, judging whether a convergence condition is met or not, if not, repeating the steps, and otherwise, obtaining global optimal parameter estimation. According to the method, a robust aggregation mechanism is introduced on the basis of the MOM method, so that the influence of an abnormal machine can be effectively inhibited, the stability and robustness of the system are remarkably improved while the prediction precision is guaranteed, and the method can be widely applied to various actual distributed systems.
Need to check novelty before this filing date? Find Prior Art