Federal learning training acceleration method for heterogeneous scene

Abstract

The invention discloses a federated learning training acceleration method for a heterogeneous scene. The federated learning training acceleration method for the heterogeneous scene comprises the following steps: S1, distributing a training task to a server and a client; and S2, running a client algorithm and a server algorithm according to the training task to obtain a client running result and a server running result. According to the federated learning training acceleration method for the heterogeneous scene provided by the invention, the problem of low synchronization efficiency in existing federated learning can be solved.

Description

technical field [0001] The invention relates to the technical field of machine learning, in particular to a federated learning training acceleration method for heterogeneous scenarios. Background technique [0002] Driven by resource and privacy considerations, for decades we have witnessed the rise of federated learning (FL) in the era of big data. Federated learning as a distributed paradigm such as figure 1 As shown, traditional centralized systems have gradually been replaced to enable artificial intelligence (AI) at the edge of the network. In federated learning, each client uses its own collected data to train its local model without sharing raw data with other clients. Client federations with the same interest can join together to derive a shared model by periodically synchronizing their local parameters under the coordination of a central server. However, due to the heterogeneity and dynamics of the edge environment, federated learning may encounter the problem of...

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to provide a federated learning training acceleration method for heterogeneous scenarios to solve the problem of low synchronization efficiency in existing federated learning

Images