|
Work item:
|
Y.RA-FML
|
|
Subject/title:
|
Requirements and reference architecture of IoT and smart city & community service based on federated machine learning
|
|
Status:
|
Under study [Issued from previous study period]
|
|
Approval process:
|
AAP
|
|
Type of work item:
|
Recommendation
|
|
Version:
|
New
|
|
Equivalent number:
|
-
|
|
Timing:
|
2023-Q1 (Medium priority)
|
|
Liaison:
|
IEEE; ITU-T SG13; ITU-T SG16; ITU-T SG17; ITU-T FG-ML5G; ISO/IEC JTC1/SC42; GSMA
|
|
Supporting members:
|
CICT, ZTE, China Unicom, China Mobile, China Telecom, MIIT China
|
|
Summary:
|
The widespread popularity of data-driven services and applications transforms the IoT and Smart City & Community (SC&C) system from a traditional data collecting and transportation network into a more holistic architecture with AI-native data processing and service delivery capability. One of the key challenges in designing an AI-based architecture for large IoT and SC&C networking systems is to implement distributed data processing and learning across a large number of decentralized datasets that can be owned or managed by different entities such as cities, communities, buildings, devices, government and business entities. Federated machine learning (FML) is an emerging distributed AI framework that enables collaborative machine learning (ML) and model construction across decentralized datasets. It offers a viable solution for data-driven data learning and synthesis across a wide variety of entities across large SC&C networking systems. The main purpose of this recommendation is to provide a feasible and standardized solution for the IoT and SC&C relevant services and applications to use and deploy FML-enabled collaborative data learning across distributed and decentralized data sources.
|
|
Comment:
|
-
|
|
Reference(s):
|
|
|
Historic references:
|
|
Contact(s):
|
|
| ITU-T A.5 justification(s): |
|
|
|
|
First registration in the WP:
2020-07-21 13:53:52
|
|
Last update:
2022-08-02 16:44:16
|
