首页|'Checkpoint Averaging To Mitigate And/Or Eliminate Catastrophic@@Forgetting Of Ma chine Learning Model(S) In Decentralized Learning@@Thereof' in Patent Application Approval Process (USPTO@@20240386318)
'Checkpoint Averaging To Mitigate And/Or Eliminate Catastrophic@@Forgetting Of Ma chine Learning Model(S) In Decentralized Learning@@Thereof' in Patent Application Approval Process (USPTO@@20240386318)
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
This patent application has not been assigned to a company or institution.The following quote was obtained by the news editors from the background informa tion supplied bythe inventors: “Decentralized learning of machine learning (ML) model(s) is an increasingly popular MLtechnique for updating ML model(s) due t o various privacy considerations. In one common implementationof decentralized learning, an on-device ML model is stored locally on a client device of a user, and a globalML model, that is a cloud-based counterpart of the on-device ML mod el, is stored remotely at a remotesystem (e.g., a server or cluster of servers) . During a given round of decentralized learning, the clientdevice can check-in to a population of client devices that will be utilized in the given round of d ecentralizedlearning, download a global ML model or weights thereof from the re mote system (e.g., to be utilized asthe on-device ML model), generate an update for the weight of the global ML model based on processinginstance(s) of client data locally at the client device and using the on-device ML model, and upload theupdate for the weight of the global ML model back to the remote system and w ithout transmitting theinstance(s) of the client device. The remote system can utilize the update received from the client device,and additional updates gener ated in a similar manner at additional client devices and that are receivedfrom the additional client devices, to update the weights of the global ML model.
NETL
NSTL
