|
Work item:
|
F.FLC-Heter-Arch
|
|
Subject/title:
|
Functional architecture of federated learning controller based on heterogeneous resources
|
|
Status:
|
Under study [Issued from previous study period]
|
|
Approval process:
|
AAP
|
|
Type of work item:
|
Recommendation
|
|
Version:
|
New
|
|
Equivalent number:
|
-
|
|
Timing:
|
2026 (Medium priority)
|
|
Liaison:
|
ITU-T SG2, ITU-T SG13, ITU-T SG17
|
|
Supporting members:
|
China Telecommunications Corporation, Inspur, Institute of Computing Technology Chinese Academy of Sciences, State Grid Corporation of China, Beijing University of Posts and Telecommunications
|
|
Summary:
|
This Recommendation focuses on the functional architecture of federated learning controller based on heterogeneous resources, which enables real-time acquisition and evaluation of heterogeneous resource information for federated learning tasks. It achieves fine-grained control of training processes such as node selection, task splitting, network topology, data interaction, and parameter aggregation, in order to improve the efficiency and performance of model training in heterogeneous resource scenarios.
|
|
Comment:
|
-
|
|
Reference(s):
|
|
|
Historic references:
|
|
Contact(s):
|
|
| ITU-T A.5 justification(s): |
|
|
|
|
First registration in the WP:
2024-06-20 16:40:26
|
|
Last update:
2025-03-05 14:19:23
|
