|
Work item:
|
Y.Suppl.UC-NRS-DLT (Y.2345 series)
|
|
Subject/title:
|
Use cases of network resource sharing based on distributed ledger technology for supporting large-scale deep learning models
|
|
Status:
|
[Carried to next study period]
|
|
Approval process:
|
Agreement
|
|
Type of work item:
|
Supplement
|
|
Version:
|
New
|
|
Equivalent number:
|
-
|
|
Timing:
|
-
|
|
Liaison:
|
ITU-T SG16, ISO/IEC JTC1 SC42, 3GPP TSG SA, ETSI ENI
|
|
Supporting members:
|
China Telecom, Huawei, China Unicom
|
|
Summary:
|
In recent years, there has been rapid development in deep learning models. However, the training and deployment of large-scale deep learning models pose significant challenges as they require extensive network resources, which are only available to a select few organizations. Large-scale deep learning models refers to those models with billions of parameters, exceeding the capacity of a single device. Through the decentralized scheme, the training process of large-scale deep learning models can be significantly accelerated and the cost can be reduced. To fulfill the potential of decentralization for the training of large-scale deep learning models, it is necessary to supplement the framework of NRS-DLT to supporting large-scale deep learning models.
This new Supplement aims to provide the general considerations, use cases and network expectations of NRS-DLT for large-scale deep learning models.
|
|
Comment:
|
-
|
|
Reference(s):
|
|
|
Historic references:
|
|
Contact(s):
|
|
| ITU-T A.5 justification(s): |
|
|
|
|
First registration in the WP:
2024-03-22 09:20:41
|
|
Last update:
2024-09-20 15:00:32
|
