GB2603469A

GB2603469A - Network analytics model training

Info

Publication number: GB2603469A
Application number: GB2101220.8A
Authority: GB
Inventors: Gutierrez Estevez David
Original assignee: Samsung Electronics Co Ltd
Current assignee: Samsung Electronics Co Ltd
Priority date: 2021-01-28
Filing date: 2021-01-28
Publication date: 2022-08-10
Also published as: WO2022164225A1; GB202101220D0

Abstract

A first network entity providing a model training service to a second network entity, comprising: receiving, from the second network entity, a service request message comprising one or more of: an identification of one or more requested trained models, and an identification of one or more analytics associated with the one or more requested trained models; and in response to receiving the service request message, providing, to the second network entity, output information comprising trained model information requested for each of the one or more requested trained models, wherein the service request message comprises a request for at least one report of the output information. The service request message may comprise a subscription request to the model training service. The output information may comprise a notification corresponding to the subscription request. The identification of one or more requested trained models may comprise model or analytics IDs. A trained model based on new data may be updated by the first network entity. A request to unsubscribe from the model training service may be received from the second network entity. A method for a second network entity, first and second network entities and a network are also claimed.

Description

Intellectual Property Office Application No G132101220.8 RTM Date:14 June 2021 The following terms are registered trade marks and should be read as such wherever they occur in this document: 3 GP P Intellectual Property Office is an operating name of the Patent Office www.gov.uk/ipo

Network Analytics Model Training BACKGROUND

Field

Certain examples of the present disclosure provide methods, apparatus and systems for providing model training for network analytics providers. For example, certain examples of the present disclosure provide methods, apparatus and systems for providing model training for an NWDAF providing network analytics in a 3GPP 5G network.

Description of the Related Art

Herein, the following documents are referenced: [1] 3GPP, TS 23.501 "System Architecture for the 5G System; Stage 2".

[2] 3GPP, TS 23.288: Architecture enhancements for 5G System (5G5) to support network data analytics services, Rel-16 (12-2020).

[3] 3GPP, TR 23.700-91: Study on enablers for network automation for the 5G System (5G5); Phase 2, Rel-17 (12-2020).

Various acronyms, abbreviations and definitions used in the present disclosure are defined at

the end of this description.

There is an ever-increasing desire to improve the performance of communication networks so that user experience can be enhanced without the network operator investing unnecessarily in excessive equipment. In other words, network operators are keen to optimize the performance of their installed fleet of infrastructure. In the past, network optimization was largely a manually-managed process, with skilled operators adjusting network parameters as required. Over time, more automation has been introduced. More recently still, Artificial Intelligence (Al) and Machine Learning (ML) techniques have be employed. In 5th Generation (5G) networks, there are different network structures and protocols which have been employed to enhance user experience. Therefore, there is a desire to make best use of these new structures and protocols to improve network performance and/or user experience.

Al has been identified as a key enabler for end-to-end network automation in 5G in all network domains, including the domains subject to the standardization process of Radio Access Network (RAN), Core Network (ON), and Management System, also known as Operations, Administration and Maintenance (OAM). Hence, standardization and industry bodies are now in the process of developing specification support for data analytics to enable Al models assist with the ever-increasing complex task of autonomously operating and managing the network.

On the RAN side, the pioneering 0-RAN alliance was established in 2018 by leading operators with the vision of developing open specifications for an open and efficient RAN that leverages Al for automating different network functions (N Fs) and reducing operating expenses (OPEX). Furthermore, standardized support for data analytics by 3GPP is particularly advanced already in Rel-16 on the CN side and the control plane. A data analytics framework anchored in the new so-called network data analytics function (NWDAF), located within the 5GC as a network function following the service-based architecture principles of 5GC has been defined with the purpose of enhancing multiple control-plane functionalities of the network. Moreover, on the OAM side a management data analytics service (MDAS) is also being specified by 3GPP to assist in dealing with longer-term management aspects of the network [1]. The joint operation of RAN analytics entities, NWDAF and MDAS is still work in progress within the relevant bodies.

In 3GPP 5GS, a Network Function (NF) is defined (e.g. in 3GPP TS 23.501) as a 3GPP adopted or 3GPP defined processing function in a network, which has defined functional behaviour and 3GPP defined interfaces. NFs in 3GPP 5GC include the NWDAF (as defined in 3GPP TS 23.288). NWDAF represents operator managed network analytics logical function providing network data analytics to a NF. Stage 2 architecture enhancements for 5G5 to support network data analytics services in 5GC are defined in 3GPP TS 23.288 (e.g. V 16.4.0). The NWDAF is part of the architecture specified in 3GPP TS 23.501 (e.g. V 16.5.1). The NWDAF services are used to expose analytics from the NWDAF to the consumer NF. The NWDAF notifies network status analytics information to the NFs that are subscribed to it.

The recently frozen 3GPP Rel-16 has specified the NWDAF framework as coarsely depicted in Figure 1. An analytics service consumer may request a specific type of data analytics to NWDAF, which can be provided by NWDAF in the form of statistics and/or predictions. The analytics consumers defined in Rel-16 are 5GC NFs, Application Functions (AFs), and OAM.

NWDAF then triggers the input data collection by means of an exposure framework defined in [2] where the input data sources may again be 5GC NFs, AFs, and/or OAM. The collected data is then used by NWDAF to perform training and inference, possibly by an Al engine, but the definition of the models is outside the scope of standardization to provide enough flexibility to vendors. This also implies that the Al engine may reside outside NWDAF itself, and the next release of the standard (Rel-17) has already started to study any required interface standardization to enable such NWDAF functional decomposition [3]. However, in the present disclosure, it is assumed the Al engine resides within NWDAF, although the present disclosure is not limited to this case. In any case, the inference results are then fed to the analytics production entity within NWDAF, which delivers the statistics and/or predictions requested by the service consumer.

A number of data analytics types have also been introduced in 3GPP Rel-16, including analytics for network slice and application service experience, NF and network slice load, network performance, UE aspects (communication, mobility, expected and abnormal behaviour), etc., as described in [2]. A network slice is defined (e.g. in 3GPP TS 23.501) as a logical network that provides specific network capabilities and network characteristics.

In addition to the above basic operation, the already ongoing Rel-17 is expanding the Rel-16 NWDAF framework by addressing a number of new use cases and key issues, including the above mentioned NWDAF functional decomposition, the architecture and interaction of multiple NWDAF instances, efficient data collection mechanisms, support for network slice service level agreement (SLA) guarantee, etc. [3].

In 3GPP Rel-17 it has been proposed to decompose NWDAF into two logical functions: * Model Training logical function: An NWDAF containing the Model Training logical function trains ML models and exposes new training services.

* Analytics logical function: An NWDAF containing the Analytics logical function can perform inference, derive analytics information (i.e. derives statistics and/or predictions based on Analytics Consumer request) and expose analytics service.

What is desired is a an architecture for the model training exposure, i.e. new model training services, to be offered by NWDAF.

The above information is presented as background information only to assist with an understanding of the present disclosure. No determination has been made, and no assertion is made, as to whether any of the above might be applicable as prior art with regard to the present invention.

SUMMARY

It is an aim of certain examples of the present disclosure to address, solve and/or mitigate, at least partly, at least one of the problems and/or disadvantages associated with the related art, for example at least one of the problems and/or disadvantages described herein. It is an aim of certain examples of the present disclosure to provide at least one advantage over the related art, for example at least one of the advantages described herein.

The present invention is defined in the independent claims. Advantageous features are defined in the dependent claims.

Other aspects, advantages, and salient features will become apparent to those skilled in the art from the following detailed description, taken in conjunction with the annexed drawings, which disclose examples of the present disclosure.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 illustrates an exemplary NWDAF framework; Figure 2 illustrates an exemplary NWDAF functional decomposition with model training exposure and Figure 3 is a block diagram of an exemplary network entity that may be used in certain examples of the present disclosure.

DETAILED DESCRIPTION

The following description of examples of the present disclosure, with reference to the accompanying drawings, is provided to assist in a comprehensive understanding of the present invention, as defined by the claims. The description includes various specific details to assist in that understanding but these are to be regarded as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the examples described herein can be made without departing from the scope of the invention.

The same or similar components may be designated by the same or similar reference numerals, although they may be illustrated in different drawings.

Detailed descriptions of techniques, structures, constructions, functions or processes known in the art may be omitted for clarity and conciseness, and to avoid obscuring the subject matter of the present invention.

The terms and words used herein are not limited to the bibliographical or standard meanings, but, are merely used to enable a clear and consistent understanding of the invention.

Throughout the description and claims of this specification, the words "comprise", "include" and "contain" and variations of the words, for example "comprising" and "comprises", means "including but not limited to", and is not intended to (and does not) exclude other features, elements, components, integers, steps, processes, operations, functions, characteristics, properties and/or groups thereof.

Throughout the description and claims of this specification, the singular form, for example "a", "an" and "the", encompasses the plural unless the context otherwise requires. For example, reference to "an object" includes reference to one or more of such objects.

Throughout the description and claims of this specification, language in the general form of "X for Y" (where Y is some action, process, operation, function, activity or step and X is some means for carrying out that action, process, operation, function, activity or step) encompasses means X adapted, configured or arranged specifically, but not necessarily exclusively, to do Y. Features, elements, components, integers, steps, processes, operations, functions, characteristics, properties and/or groups thereof described or disclosed in conjunction with a particular aspect, embodiment, example or claim of the present invention are to be understood to be applicable to any other aspect, embodiment, example or claim described herein unless incompatible therewith.

Certain examples of the present disclosure provide methods, apparatus and systems for providing model training for network analytics providers. The following examples are applicable to, and use terminology associated with, 3GPP 5G. For example, certain examples of the present disclosure provide methods, apparatus and systems for providing model training for an NWDAF providing network analytics in a 3GPP 5G network. However, the skilled person will appreciate that the techniques disclosed herein are not limited to these examples or to 3GPP 5G, and may be applied in any suitable system or standard, for example one or more existing and/or future generation wireless communication systems or standards.

For example, the functionality of the various network entities and other features disclosed herein may be applied to corresponding or equivalent entities or features in other communication systems or standards. Corresponding or equivalent entities or features may be regarded as entities or features that perform the same or similar role, function, operation or purpose within the network. For example, the functionality of the NWDAF in the examples below may be applied to any other suitable type of entity providing network analytics.

The skilled person will appreciate that the present invention is not limited to the specific examples disclosed herein. For example: * The techniques disclosed herein are not limited to 3GPP 5G.

* One or more entities in the examples disclosed herein may be replaced with one or more alternative entities performing equivalent or corresponding functions, processes or operations.

* One or more of the messages in the examples disclosed herein may be replaced with one or more alternative messages, signals or other type of information carriers that communicate equivalent or corresponding information.

* One or more further elements, entities and/or messages may be added to the examples disclosed herein.

* One or more non-essential elements, entities and/or messages may be omitted in certain examples.

* The functions, processes or operations of a particular entity in one example may be divided between two or more separate entities in an alternative example.

* The functions, processes or operations of two or more separate entities in one example may be performed by a single entity in an alternative example.

* Information carried by a particular message in one example may be carried by two or more separate messages in an alternative example.

* Information carried by two or more separate messages in one example may be carried by a single message in an alternative example.

* The order in which operations are performed may be modified, if possible, in alternative examples.

* The transmission of information between network entities is not limited to the specific form, type and/or order of messages described in relation to the examples disclosed herein.

Certain examples of the present disclosure may be provided in the form of an apparatus/device/network entity configured to perform one or more defined network functions and/or a method therefor. Certain examples of the present disclosure may be provided in the form of a system (e.g. a network) comprising one or more such apparatuses/devices/network entities, and/or a method therefor.

A network may include a Network Data Analytics Function (NWDAF) entity and one or more other network entities. For example, the network may include an Access and Mobility Management Function (AMF) entity, a Session Management Function (SMF) entity, a Network Slice Selection Function (NSSF) entity, a Network Repository Function (NRF) entity, and an Operation and Maintenance (OAM) entity. The network may include one or more analyfics consumers (for example, one or more of the entities mentioned above and/or one or more other entities) that receive analytics from NWDAF. The network may include one or more training consumers (for example, one or more other NWDAF instances) that receive model training from NWDAF. The skilled person will appreciate that a network may omit one or more of the entities mentioned above and/or may comprise one or more additional entities.

A particular network function can be implemented either as a network element on a dedicated hardware, as a software instance running on a dedicated hardware, or as a virtualised function instantiated on an appropriate platform, e.g. on a cloud infrastructure. An NF service may be defined as a functionality exposed by an NF through a service based interface and consumed by other authorized N Fs.

Certain examples of the present disclosure provide an architecture for the model training exposure, i.e. new model training services, to be offered by NWDAF. Certain examples of the present disclosure provide such an architecture given the following features: * Model Training logical function is not always required to be involved to provide output analytics. In particular, delivery of statistics may not require operation of the Model Training logical function.

* Analytics logical function is not always required to perform inference, depending on analytics requested.

* Model Training logical function and Analytics logical function can be deployed as standalone NWDAFs.

* Model Training logical function may register the supported ML model information to NRF.

Certain examples of the present disclosure provide a set of training services for NWDAF.

Figure 2 illustrates an exemplary architecture featuring exemplary model training exposure services within an NWDAF entity that has been decomposed into Analytics logical function and Model training logical function.

An analytics consumer may be, for example, 5GC NF, AF or OAM. In certain examples, a training consumer may only be another NWDAF instance, or any other suitable type of entity providing network analytics.

The analytics consumer may request either statistics On the past) or predictions On the future), or both, using the parameter of analytics target period. Training services will be only triggered by the Analytics function when the analytics target period is in the future, i.e. when the subscription or request is for predictions. Statistics do not require training services.

The following is a description of the model training exposure as well as a detailed definition of each new model training service in NWDAF. The skilled person will appreciate that the present disclosure is not limited to these specific examples. For example, alternative implementations may use different (e.g. additional, omitted and/or alternative) input parameters and may provide different (e.g. additional, omitted and/or alternative) output information.

The skilled person will appreciate that the term "network entity" used in the present disclosure can also be used to refer to specific functionality within a network entity.

Contents of Model Training Exposure In certain examples, the consumers of the Nnwdaf TrainingSubscription or Nnwdaf_TrainingInfo service operations provided by the NWDAF Training Function, for example described in clause 7 of [2], provide the following input parameters listed below.

A list of Model I D(s): identifies the requested trained model(s).

A list of Analytics ID(s): identifies the analytics associated to the requested trained model(s).

(Only for Nnwdaf_TrainingSubscription) A Notification Target Address (+ Notification Correlation ID), for example as defined in TS 23.502 [3] clause 4.15.1, allowing to correlate notifications received from NWDAF with this subscription.

Training Reporting Information with the following optional parameters: Model type: identifies the type of ML model.

Minimum level of confidence: the expected minimum confidence level that the prediction is correct.

(Only for Nnwdaf_TrainingInfo_Request) Time when training information is needed Of applicable). If the time is reached the consumer does not need to wait for the training information any longer, yet the NWDAF Training Function may send an error response to the consumer.

In certain examples, the NWDAF Training Function provides to the consumer of the Nnwdaf TrainingSubscription or Nnwdaf_TrainingInfo service operations, for example described in clause 7 of [2], the output information listed below: - (Only for Nnwdaf TrainingSubscription) The Notification Correlation Information. -The trained model information requested for each Model ID. NOTE 1: Trained model information can be provided in a variety of forms, e.g a file, an address where the file is located, a data structure, etc. - In addition, the following additional information: Timestamp of trained model generation.

- Validity period, which defines the time period for which the trained model is valid. Probability assertion: confidence in trained model.

Nnwdaf Model Training Services General Services description: this service enables the consumer to request model training and provisioning services.

Nnwdaf Training Subscription service Analyfics function may subscribe to the Training Function's training service so that it receives the updated trained models as new data becomes available and trained models change.

1. General * Service description: Subscribe to NWDAF model training services with specific parameters 2. Nnwdaf_TrainingSubscription_Subscribe service operation * Service operation name: Nnwdaf TrainingSubscription_Subscribe * Description: Subscribe to NWDAF model training services with specific parameters.

* Inputs, Required in certain examples: (Set of) Model ID(s) and their corresponding Analytics ID(s), for example defined in Table 7.1-2 of [2], Notification Target Address (+ Notification Correlation ID).

* Inputs, Optional in certain examples: Subscription Correlation ID (in the case of modification of the training subscription), Training reporting information, model type, minimum level of confidence.

* Outputs, Required in certain examples: When the subscription is accepted: Subscription Correlation ID (required for management of this subscription).

* Outputs, Optional in certain examples: None.

3. Nnwdaf_TrainingSubscripfion_Unsubscribe service operation * Service operation name: Nnwdaf TrainingSubscription_Unsubscribe.

* Description: Unsubscribe from NWDAF training services.

* Inputs, Required in certain examples: Subscription Correlation ID.

* Inputs, Optional in certain examples: None.

* Outputs, Required in certain examples: Operation execution result indication.

* Outputs, Optional in certain examples: None.

4. Nnwdaf_TrainingSubscription_Notify service operation * Service operation name: Nnwdaf TrainingSubscription_Notify * Description: NWDAF notifies the consumer instance that has subscribed to the specific training service.

* Inputs, Required in certain examples: Notification Correlation Information, trained model information.

* Inputs, Optional in certain examples: Timestamp of training, validity period, probability assertion.

* Outputs, Required in certain examples: Operation execution result indication.

* Outputs, Optional in certain examples: None.

Nnwdaf TraininqInfo service 1. General * Service description: This service enables the consumer to request and get models trained by NWDAF 2. Nnwdaf_TrainingInfo_Request service operation * Service operation name: Nnwdaf Training_Request * Description: Request of a trained model by NWDAF with specific parameters.

* Inputs, Required in certain examples: (Set of) Model ID(s) and their corresponding Analytics ID(s) defined in Table 7.1-2 of [2], Notification Target Address (+ Notification Correlation ID).

* Inputs, Optional in certain examples Training reporting information, model type, minimum level of confidence.

* Outputs, Required in certain examples: Trained model information.

* Outputs, Optional in certain examples: Timestamp of training, validity period, probability assertion.

Various examples of the present disclosure allow a Training consumer or Analytics logical function to gain exposure to Model training provided by a Model Training logical function. In certain examples, such exposure may be requested by, and provided to, the Training consumer or Analytics logical function in the form of a subscription (resulting in multiple (e.g. periodic or regular) reports in response to a subscription request) and/or a request (resulting in a single response report).

Certain examples of the present disclosure provide a method, for a first network entity (e.g. an NWDAF or a model training logical function within NWDAF), for providing a model training service to a second network entity (e.g. an Analytics logical function or a different training consumer), the method comprising: receiving, from the second network entity, a service request message comprising one or more of: an identification of one or more requested trained models, and an identification of one or more analytics associated with the one or more requested trained models; and in response to receiving the service request message, providing, to the second network entity, output information comprising trained model information requested for each of the one or more requested trained models. The service request message may comprise a request for a single report of the output information, or a request for (e.g. subscription to) periodic or regular reports of the output information.

Certain examples of the present disclosure provide a method, for a second network entity (e.g. an Analytics logical function or a different training consumer), for receiving a model training service from a first network entity (e.g. an NWDAF or a model training logical function within NWDAF), the method comprising: transmitting, to the first network entity, a service request message comprising one or more of: an identification of one or more requested trained models, and an identification of one or more analytics associated with the one or more requested trained models; and in response to transmitting the service request message, receiving, from the first network entity, output information comprising trained model information requested for each of the one or more requested trained models. The service request message may comprise a request for a single report of the output information, or a request for (e.g. subscription to) periodic or regular reports of the output information.

In certain examples, the service request message comprises a subscription request for subscribing to the model training service, and the output information may comprise a notification corresponding to the subscription request.

In certain examples, the identification of one or more requested trained models may comprise a list of one or more Model IDs, and/or the identification of one or more analytics may comprise a list of one or more Analytics IDs.

In certain examples, the trained model information may be provided in the form of one or more of: a file; information (e.g. an address) specifying the location of a file; and a data structure.

In certain examples, if the service request message is a request for periodic or regular reports of the output information, the service request message may comprise information (e.g. a Notification Target Address and a Notification Correlation ID) for correlating output information (e.g. notifications) provided by the first network entity with the service request message (e.g. subscription).

In certain examples, the service request message may comprises information (e.g. Training Reporting Information) comprising one or more of the following parameters: information (e.g. Model type) for identifying the type of ML model; information (e.g. Minimum level of confidence) indicating an expected minimum confidence level that a prediction is correct; and information indicating a time when training information is needed.

In certain examples, the output information may comprise one or more of Notification Correlation Information; a timestamp of trained model generation; information (e.g. Validity period) indicating a time period for which a trained model is valid; and information (e.g. Probability assertion) indicating a confidence in a trained model.

In certain examples, the method may comprise updating, by the first network entity, a trained model based on new data, and the output information may comprise information based on the updated trained model.

In certain examples, the method may comprise receiving, from the second network entity, a request to unsubscribe from the model training service.

Certain examples of the present disclosure provide a first network entity (e.g. an NWDAF or a model training logical function within NWDAF) configured to perform a method according to any example disclosed herein.

Certain examples of the present disclosure provide a second network entity (e.g. an Analytics logical function or a different training consumer) configured to perform a method according to any example disclosed herein.

Certain examples of the present disclosure provide a network comprising a first network entity according to any example disclosed herein, and a second network entity according to any example disclosed herein.

Figure 3 is a block diagram of an exemplary network entity that may be used in examples of the present disclosure. For example, an analyfics consumer, training consumer, NWDAF, AMF, SMF, NSSF, NRF, OAM and/or other NFs may be provided in the form of the network entity illustrated in Figure 3. The skilled person will appreciate that the network entity illustrated in Figure 3 may be implemented, for example, as a network element on a dedicated hardware, as a software instance running on a dedicated hardware, or as a virtualised function instantiated on an appropriate platform, e.g. on a cloud infrastructure.

The entity 300 comprises a processor (or controller) 301, a transmitter 303 and a receiver 305. The receiver 305 is configured for receiving one or more messages or signals from one or more other network entities. The transmitter 303 is configured for transmitting one or more messages or signals to one or more other network entities. The processor 301 is configured for performing one or more operations and/or functions as described above. For example, the processor 301 may be configured for performing the operations of an analytics consumer, training consumer, NWDAF, AMF, SMF, NSSF, NRF, OAM and/or other NFs.

The techniques described herein may be implemented using any suitably configured apparatus and/or system. Such an apparatus and/or system may be configured to perform a method according to any aspect, embodiment, example or claim disclosed herein. Such an apparatus may comprise one or more elements, for example one or more of receivers, transmitters, transceivers, processors, controllers, modules, units, and the like, each element configured to perform one or more corresponding processes, operations and/or method steps for implementing the techniques described herein. For example, an operation/function of X may be performed by a module configured to perform X (or an X-module). The one or more elements may be implemented in the form of hardware, software, or any combination of hardware and software.

It will be appreciated that examples of the present disclosure may be implemented in the form of hardware, software or any combination of hardware and software. Any such software may be stored in the form of volatile or non-volatile storage, for example a storage device like a ROM, whether erasable or rewritable or not, or in the form of memory such as, for example, RAM, memory chips, device or integrated circuits or on an optically or magnetically readable medium such as, for example, a CD, DVD, magnetic disk or magnetic tape or the like.

It will be appreciated that the storage devices and storage media are embodiments of machine-readable storage that are suitable for storing a program or programs comprising instructions that, when executed, implement certain examples of the present disclosure. Accordingly, certain example provide a program comprising code for implementing a method, apparatus or system according to any example, embodiment, aspect and/or claim disclosed herein, and/or a machine-readable storage storing such a program. Still further, such programs may be conveyed electronically via any medium, for example a communication signal carried over a wired or wireless connection.

While the invention has been shown and described with reference to certain examples, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the scope of the invention, as defined by any appended claims.

Acronyms, Abbreviations and Definitions In the present disclosure, the following acronyms, abbreviations and definitions may be used.

3GPP 3rd Generation Partnership Project 5G 5th Generation 5GC 5G Core Network 5G5 5G System AF Application Function Al Artificial Intelligence AMF Access and Mobility Management Function CN Core Network ID Identifier/Identity MDAS Management Data Analytics Service ML Machine Learning NF Network Function NRF Network Repository Function NWDAF Network Data Analytics Function NSSF Network Slice Selection Function OAM Operation and Maintenance OPEX Operating Expenses RAN Radio Access Network Rel Release SLA Service Level Agreement SMF Session Management Function TR Technical Report

TS Technical Specification

UE User Equipment

Claims

Claims 1. A method, for a first network entity (e.g. an NWDAF or a model training logical function within NVVDAF), for providing a model training service to a second network entity (e.g. an Analytics logical function or a different training consumer), the method comprising: receiving, from the second network entity, a service request message comprising one or more of: an identification of one or more requested trained models, and an identification of one or more analytics associated with the one or more requested trained models; and in response to receiving the service request message, providing, to the second network entity, output information comprising trained model information requested for each of the one or more requested trained models, wherein the service request message comprises a request for a single report of the output information, or a request for (e.g. subscription to) periodic or regular reports of the output information.
2. A method according to claim 1, wherein the service request message comprises a subscription request for subscribing to the model training service, and wherein the output information comprises a notification corresponding to the subscription request.
3. A method according to claim 1 or 2, wherein the identification of one or more requested trained models comprises a list of one or more Model IDs, and/or wherein the identification of one or more analyfics comprises a list of one or more Analytics IDs.
4. A method according to claim 1, 2 or 3, wherein the trained model information is provided in the form of one or more of: a file; information (e.g. an address) specifying the location of a file; and a data structure.
5. A method according to any preceding claim, wherein, if the service request message is a request for periodic or regular reports of the output information, the service request message comprises information (e.g. a Notification Target Address and a Notification Correlation ID) for correlating output information (e.g. notifications) provided by the first network entity with the service request message (e.g. subscription).
6. A method according to any preceding claim, wherein the service request message comprises information (e.g. Training Reporting Information) comprising one or more of the following parameters: information (e.g. Model type) for identifying the type of ML model; information (e.g. Minimum level of confidence) indicating an expected minimum confidence level that a prediction is correct; and information indicating a time when training information is needed.
7. A method according to any preceding claim, wherein the output information comprises one or more of: Notification Correlation Information; a timestamp of trained model generation; information (e.g. Validity period) indicating a time period for which a trained model is valid and information (e.g. Probability assertion) indicating a confidence in a trained model.
8. A method according to any preceding claim, wherein the method comprises updating, by the first network entity, a trained model based on new data, and wherein the output information comprises information based on the updated trained model.
9. A method according to any preceding claim, wherein the method comprises receiving, from the second network entity, a request to unsubscribe from the model training service.
10. A method, for a second network entity (e.g. an Analytics logical function or a different training consumer), for receiving a model training service from a first network entity (e.g. an NWDAF or a model training logical function within NWDAF), the method comprising: transmitting, to the first network entity, a service request message comprising one or more of: an identification of one or more requested trained models, and an identification of one or more analytics associated with the one or more requested trained models; and in response to transmitting the service request message, receiving, from the first network entity, output information comprising trained model information requested for each of the one or more requested trained models, wherein the service request message comprises a request for a single report of the output information, or a request for (e.g. subscription to) periodic or regular reports of the output information.
11. A first network entity (e.g. an NWDAF or a model training logical function within NWDAF) configured to perform a method according to any of claims 1 to 9.
12. A second network entity (e.g. an Analytics logical function or a different training consumer) configured to perform a method according to claim 10.
13. A network comprising a first network entity according to claim 11 and a second network entity according to claim 12.