Content for TS 28.105 Word version: 18.3.0
6 AI/ML management use cases and requirements
6.1 General
6.2a ML training phase
6.3 AI/ML emulation phase
6.4 ML entity deployment phase
6.5 AI/ML inference phase
...
...
6 AI/ML management use cases and requirements p. 17
6.1 General p. 17
6.2 Void
6.2a ML training phase |R18| p. 18
6.2a.1 ML training p. 18
6.2a.1.1 Description p. 18
6.2a.1.2 Use cases p. 18
6.2a.1.2.1 ML training requested by consumer p. 18
6.2a.1.2.2 ML training initiated by producer p. 19
6.2a.1.2.3 ML entity selection p. 19
6.2a.1.2.4 Managing ML training processes p. 19
6.2a.1.2.5 Handling errors in data and ML decisions p. 20
6.2a.1.2.6 ML entity joint training p. 20
6.2a.1.2.7 ML entity validation performance reporting p. 21
6.2a.1.2.8 Training data effectiveness reporting p. 21
6.2a.1.3 Requirements for ML training p. 21
6.2a.2 Performance management for ML training and testing p. 23
6.2a.2.1 Description p. 23
6.2a.2.2 Use cases p. 23
6.2a.2.2.1 Performance indicator selection for ML training and testing p. 23
6.2a.2.2.2 ML entity performance indicators query and selection for ML training and testing p. 24
6.2a.2.2.3 MnS consumer policy-based selection of ML entity performance indicators for ML training and testing p. 24
6.2a.2.3 Requirements for ML training and testing performance management p. 25
6.2a.3 ML testing p. 25
6.3 AI/ML emulation phase |R18| p. 26
6.3.1 Description p. 26
6.3.2 Use cases p. 26
6.3.3 Requirements for Managing AI/ML Inference emulation p. 27
6.4 ML entity deployment phase |R18| p. 27
6.4.1 ML entity loading p. 27
6.5 AI/ML inference phase |R18| p. 28
6.5.1 AI/ML inference performance management p. 28
6.5.1.1 Description p. 28
6.5.1.2 Use cases p. 29
6.5.1.2.1 AI/ML inference performance evaluation p. 29
6.5.1.2.2 AI/ML performance measurements selection based on MnS consumer policy p. 29
6.5.1.3 Requirements for AI/ML inference performance management p. 29
6.5.2 AI/ML update control p. 30
6.5.2.1 Description p. 30
6.5.2.2 Use cases p. 30
6.5.2.2.1 Availability of new capabilities or ML entities p. 30
6.5.2.2.2 Triggering ML entity update p. 30
6.5.2.3 Requirements for AIML update control p. 31
6.5.3 AI/ML inference capabilities management p. 31
6.5.3.1 Description p. 31
6.5.3.2 Use cases p. 32
6.5.3.2.1 Identifying capabilities of ML entities p. 32
6.5.3.2.2 Mapping of the capabilities of ML entities p. 32
6.5.3.3 Requirements for AI/ML inference capabilities management p. 32
6.5.4 AI/ML inference capability configuration management p. 33
6.5.4.1 Description p. 33
6.5.4.2 Use cases p. 33
6.5.4.2.1 Managing NG-RAN AI/ML-based distributed Network Energy Saving p. 33
6.5.4.2.2 Managing NG-RAN AI/ML-based distributed Mobility Optimization p. 33
6.5.4.2.3 Managing NG-RAN AI/ML-based distributed Load Balancing p. 33
6.5.4.3 Requirements for AI/ML inference management p. 34
6.5.5 Executing AI/ML Inference p. 34
