首页|Researchers Submit Patent Application, 'Decentralized Learning Of Machine Learni ng Model(S) Through Utilization Of Stale Updates( S) Received From Straggler Comp uting Device(S)', for Approval (USPTO 20240095582)
Researchers Submit Patent Application, 'Decentralized Learning Of Machine Learni ng Model(S) Through Utilization Of Stale Updates( S) Received From Straggler Comp uting Device(S)', for Approval (USPTO 20240095582)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
No assignee for this patent application has been made.News editors obtained the following quote from the background information suppli ed by the inventors:“Decentralized learning of machine learning (ML) model(s) i s an increasingly popular ML techniquefor updating ML model(s) due to various p rivacy considerations. In one common implementation ofdecentralized learning, a n on-device ML model is stored locally on a client device of a user, and a globa lML model, that is a cloud-based counterpart of the on-device ML model, is stor ed remotely at a remotesystem (e.g., a server or cluster of servers). During a given round of decentralized learning, the clientdevice, using the on-device ML model, can process an instance of client data detected at the client deviceto generate predicted output, and can generate an update for the global ML model ba sed on processingthe instance of client data. Further, the client device can tr ansmit the update to the remote system. Theremote system can utilize the update received from the client device, and additional updates generated ina similar manner at additional client devices and that are received from the additional cl ient devices, toupdate global weight(s) of the global ML model. The remote syst em can transmit the updated globalML model (or updated global weight(s) of the updated global ML model), to the client device and theadditional client devices . The client device and the additional client devices can then replace the respective on-device ML models with the updated global ML model (or replace respectiv e on-device weight(s)of the respective on-device ML models with the updated glo bal weight(s) of the global ML model), therebyupdating the respective on-device ML models.
NETL
NSTL
