Personalized collaborative learning method and device based on gradual freezing of parameters

Abstract

The invention provides a personalized collaborative learning method and device based on gradual freezing of parameters, and the method comprises the steps: receiving a global model, sent by a central server, of the current communication round at the beginning of each communication round, splicing the global model with a local model of the previous communication round according to a mask matrix, obtaining a local initial model of the communication round; determining the number of training rounds of the communication round according to the variable parameters, and training a local initial model of the communication round according to the number of the training rounds; and sending the local model of the current communication round after the training is completed, and updating the mask matrix according to the freezing parameter corresponding to the local model of the current communication round after the training is completed. According to the method, the number of communication rounds required by collaborative learning can be reduced, and the communication overhead in a model training process is reduced.

Description

technical field [0001] The invention relates to the technical field of edge computing, in particular to a personalized collaborative learning method and device based on gradual freezing of parameters. Background technique [0002] With the advent of the era of the Internet of Everything, more and more edge devices are connected to the network. The data generated by various edge devices and the corresponding computing, storage and communication requirements are growing rapidly. Scenarios such as malicious traffic detection, industrial Internet, smart home, autonomous driving, and the Internet of Vehicles all require agile and fast transmission, calculation, storage, and decision-making of large amounts of data generated in a short period of time. Computing paradigms to handle business needs in many complex scenarios. For edge computing scenarios, the existing cloud computing model has problems such as weak real-time performance, insufficient bandwidth, high energy consumpti...

AI Technical Summary

Problems solved by technology

[0006] To this end, the present invention proposes a personalized collaborative learning method and device based on gradual freezing of parameters, which is used to solve the communication efficiency and data non-independent and identical distribution problems of collaborative learning in edge intelligence scenarios, and reduces the communication required for collaborative learning. Number of rounds, reducing communication overhead during training

Images