Recommendation ITU-T Y.3186 (04/2024) Requirements and framework for distributed joint learning to enable machine learning in future networks including IMT-2020
Summary
History
FOREWORD
Table of Contents
1 Scope
2 References
3 Definitions
     3.1 Terms defined elsewhere
     3.2 Terms defined in this Recommendation
4 Abbreviations and acronyms
5 Conventions
6 Overview
7 Scenarios of DJL in future networks including IMT-2020
     7.1 Scenarios from the perspective of the data source of participating DJL nodes
          7.1.1 The IMT-2020 network enables DJL with the use of network data
          7.1.2 The IMT-2020 network enables DJL with the use of network data and external data
     7.2 Scenarios from the perspective of data distribution in DJL
8 Requirements of DJL in future networks including IMT-2020
     8.1 JL task management
     8.2 DJL node management
     8.3 DJL resource management
     8.4 DJL model management
     8.5 DJL data management
     8.6 DJL interfaces
     8.7 DJL connectivity
     8.8 DJL reliability management
     8.9 DJL security and privacy protection
9 Framework of DJL in future networks including IMT-2020
     9.1 Architectural components of DJL
          9.1.1 Training services function
          9.1.2 Data services function
          9.1.3 ML consumer
          9.1.4 MLFO services function
          9.1.5 Model repository
     9.2 Architectural framework
     9.3 Deployment options of the DJL components
          9.3.1 DJL nodes deployed on a machine learning overlay
          9.3.2 DJL nodes deployed on UEs and AFs
     9.4 DJL data transmission optimization
          9.4.1 DJL data transmission optimization based on network control in the ML underlay network
          9.4.2 DJL data transmission optimization based on policy control in the IMT-2020 network
10 Flow diagram for DJL
11 Security considerations
Bibliography