Lifecycle management using ML model identification and ML function identification

Abstract

A method includes requesting to use machine learning ML features based on using an ML model applicable to the ML features; receiving an activation message relating to at least one of: activation of one ML model available and applicable to the ML feature, the one ML model associated with at least one function, or activation of one of the at least one function associated with the one ML model available and applicable to the ML feature, wherein the activation message indicates that one ML model and / or one function are / is activated; starting to use an ML model based on the received activation message; and reporting at least one of an ML model performance key performance indicator (KPI) of one ML model, or a functional performance (KPI) of one function associated with one ML model.

Description

Technical Field

[0001] The present disclosure relates to a method and apparatus for identifying lifecycle management (LCM) using machine learning (ML) model identification and ML functionality. Background Art

[0002] The following description of the background technology may include insights, discoveries, understandings, or disclosures of at least some examples of embodiments of the present disclosure, or disclosed connections not known to the relevant prior art, but provided by the present disclosure. Some of these contributions of the present disclosure may be specifically noted below, while other contributions of the present disclosure will be clear from the relevant context.

[0003] In the past few years, there has been an increasing expansion of communication networks, for example wire-based communication networks such as Integrated Services Digital Network (ISDN), Digital Subscriber Line (DSL), or wireless communication networks such as cellular third generation (3G) such as cdma2000 (Code Division Multiple Access) systems, Universal Mobile Telecommunications System (UMTS), fourth generation (4G) communication networks or enhanced communication networks based on, for example, Long Term Evolution (LTE) or Long Term Evolution-Advanced (LTE-A), fifth generation (5G) communication networks, sixth generation (6G) communication networks, cellular second generation communication networks such as Global System for Mobile Communications (GSM), General Packet Radio System (GPRS), Enhanced Data Rates for Global Evolution (EDGE) or other wireless communication systems such as Wireless Local Area Networks (WLAN), Bluetooth or Worldwide Interoperability for Microwave Access (WiMAX), which has occurred all over the world. Various organizations (such as the European Telecommunications Standards Institute (ETSI), the 3rd Generation Partnership Project (3GPP), Telecommunications and Internet Convergence Services and Protocols for Advanced Networks (TISPAN), the International Telecommunication Union (ITU), the 3rd Generation Partnership Project 2 (3GPP2), the Internet Engineering Task Force (IETF), the IEEE (Institute of Electrical and Electronics Engineers), the WiMAX Forum, etc.) are developing standards or specifications for telecommunication networks and access environments.

[0004] In this context, within the machine learning paradigm, model lifecycle management (LCM) encompasses data processing, training, deployment, and monitoring. LCM ensures the validity and robustness of models for complex tasks. Although machine learning solutions have been applied to the 5G NR air interface, LCM interactions and signaling have yet to be explored. As agreed upon by RAN1#110, the different LCM components are as follows.

[0005]

[0006] To ensure interoperability between models in different LCM components and different nodes (UE, NW), 3GPP has proposed two different LCM-based identifications: a) LCM based on model identification and b) LCM based on function identification. In RAN1#111, the working assumptions and agreements for model identification and function identification are defined. Figure 1 and Figure 2 The diagrams show the model ID-based LCM and function (ID)-based LCM proposed by Nokia in 3GPP RAN1#112. In addition to the LCM of AI / ML models, 3GPP also identified three use cases as AI / ML applications. Figure 3 An example of how the two LCM mechanisms can be applied to one of these use cases is shown. Rel-18 NR air interface use cases include: CSI feedback enhancement, beam management, and positioning enhancement.

[0007]

[0008]

[0009] In addition, RAN1 also reaches an agreement on UE-NW cooperation based on signaling and model transmission on the air interface, as described below.

[0010]

[0011] In general, the LCM process ensures the continuous integration and validity of AI / ML models throughout their lifecycle. Although LCM of models is well defined in the field of machine learning, the architectural framework is still under discussion in the 3GPP NR air interface. During the discussion, two different LCM mechanisms were identified: model-based LCM mechanism and function-based LCM mechanism. The former allows the continuous development of models and ensures smooth integration with other underlying operations including models. The latter allows the continuous integration of ML functions in UE and NW communications. However, it has not yet been found.

[0012] a) How to organize ML models and ML functions to ensure smooth collaboration between (multiple) UEs and NW?

[0013] b) How to define the interaction between model-based LCM and feature-based LCM?

[0014] c) How can these two LCM mechanisms be used to ensure different levels of collaboration (level x, level y, and level z)?

[0015] d) How to provide a scalable architecture to accommodate the increasing number of ML-enabled features?

[0016] Figure 3The example of the definition of an ML-enabled feature in

[0045] illustrates only one of several different ways to organize and define the use of ML models and ML functions within the overall ML feature.

[0017] For example, another approach is to define ML capabilities such that one capability ID corresponds to one ML feature (use case) and multiple ML features are configured for different sub-use cases, scenarios (across Figure 3 The columns in ) define different ML models (identified by ID).

[0018] Another approach is when there is a one-to-one mapping between ML features and ML models, potentially both can be identified with a single ID ( Figure 3 One ML model in each column of ).

[0019] By combining the above methods, further variations are possible.

[0020] These options lead to increased complexity both in the actual implementation solution and in the UE specifications (UE capabilities, features, testability, requirements, etc.), and ultimately to increased LCM process complexity.

[0021] Therefore, improvements are needed. In particular, LCM using ML model identification and ML feature identification is needed.

[0022] As disclosed herein, LCM using ML model identification and ML function identification can address these issues.

[0023] Therefore, it is an object of the present disclosure to improve upon the prior art.

[0024] The abbreviations used in this manual have the following meanings:

[0025] 2G: Second Generation

[0026] 3G: Third Generation

[0027] 3GPP: Third Generation Partnership Project

[0028] 3GGP2: 3rd Generation Partnership Project 2

[0029] 4G: Fourth Generation

[0030] 5G: Fifth Generation

[0031] 6G: Sixth Generation

[0032] AI: Artificial Intelligence

[0033] AP: Access Point

[0034] BS: Base Station

[0035] CDMA: Code Division Multiple Access

[0036] CSI: Channel State Information

[0037] DSL: Digital Subscriber Line

[0038] EDGE: Enhanced Data rates for Global Evolution

[0039] EEPROM: Electrically Erasable Programmable Read-Only Memory

[0040] eNB: Evolved Node B

[0041] ETSI: European Telecommunications Standards Institute

[0042] FID: Function ID

[0043] gNB: Next Generation Node B

[0044] GPRS: General Packet Radio System

[0045] GSM: Global System for Mobile Communications

[0046] ID: Identification

[0047] IEEE: Institute of Electrical and Electronics Engineers

[0048] ISDN: Integrated Services Digital Network

[0049] ITU: International Telecommunication Union

[0050] KPI: Key Performance Indicator

[0051] LCM: Lifecycle Management

[0052] LTE: Long Term Evolution

[0053] LTE-A: Long Term Evolution Advanced

[0054] MANET: Mobile Ad Hoc Network

[0055] MAC CE: MAC Control Element

[0056] MID: Model ID

[0057] ML: Machine Learning

[0058] NB: Node B

[0059] NW: Network

[0060] RAM: Random Access Memory

[0061] RAN: Radio Access Network

[0062] ROM: Read-Only Memory

[0063] TISPAN: Telecom and Internet Convergence Services and Protocols for Advanced Networks

[0064] UE: User Equipment

[0065] UMTS: Universal Mobile Telecommunications System

[0066] UUID: Universally Unique Identifier

[0067] UWB: Ultra Wideband

[0068] WCDMA: Wideband Code Division Multiple Access

[0069] WiMAX: Worldwide Interoperability for Microwave Access

[0070] WLAN: Wireless Local Area Network Summary of the Invention

[0071] Various examples of embodiments of the present disclosure are intended to improve the prior art. Therefore, at least some examples of embodiments of the present disclosure are intended to address at least part of the above-mentioned difficulties and / or problems and disadvantages.

[0072] Various aspects of examples of embodiments of the present disclosure are set out in the accompanying claims and relate to methods, apparatus, and computer program products related to LCM using ML model identification and ML function identification.

[0073] This object is achieved by the method, the device and the non-transitory storage medium specified in the appended claims. Advantageous further developments are listed in the respective dependent claims.

[0074] The use of ML model identification and ML function identification to enable LCM according to any of the aspects of the appended claims allows solving at least some of the above identified / derived problems and disadvantages.

[0075] Thus, improvements are achieved by methods, apparatus, and computer program products for enabling LCM using ML model identification and ML function identification.

[0076] In more detail, the present specification discloses LCM using ML model identification and ML function identification, which is superior to proprietary solutions because, in addition to other advantageous technical effects, it also achieves reduced complexity of UE specifications and actual implementation solutions, as well as reduced complexity of LCM procedures.

[0077] Other advantages will become apparent from the detailed description that follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Some embodiments of the present disclosure are described below, by way of example only, with reference to the accompanying drawings, in which:

[0079] Figure 1 The LCM module based on model identification for ML model training and inference process [3GPP RAN1#112] is shown;

[0080] Figure 2 Shows the LCM module based on functional identification for ML functions and ML reasoning mode procedures [3GPP RAN1#112];

[0081] Figure 3 An example of ML-enabled feature definition and the use of function IDs and model IDs and their related information is shown;

[0082] Figure 4 Different diagrams showing a function (FID) or a set of functions (FID1, FID2, ...) to be associated with a model (MID) or a set of models (MID1, MID2);

[0083] Figure 5 (parts 1 / 2 and 2 / 2) shows Framework C: the call flow of function (ID)-based LCM and ML model ID-based LCM using function and / or ML model switching in collaboration level z;

[0084] Figure 6 (part 1 / 2 and part 2 / 2) shows Framework B: Call flow for function (ID)-based LCM monitoring and switching at the UE using Framework B in collaboration level y;

[0085] Figure 7 (part 1 / 2 and part 2 / 2) shows Framework B: Call flow for function (ID)-based LCM monitoring and switching at the UE using Framework B in collaboration level z;

[0086] Figure 8 shows a flowchart illustrating steps corresponding to methods according to various examples of the embodiments;

[0087] Figure 9 shows a flowchart illustrating steps corresponding to methods according to various examples of the embodiments;

[0088] Figure 10 shows a block diagram illustrating apparatuses according to various examples of embodiments; and

[0089] Figure 11 Block diagrams illustrating apparatuses according to various examples of embodiments are shown. DETAILED DESCRIPTION

[0090] Basically, in order to correctly establish and process communications between two or more endpoints (e.g., communication stations or elements or functions, such as terminal devices, user equipment (UE) or other communication network elements, databases, servers, hosts, etc.), one or more network elements or functions (e.g., virtualized network functions) may be involved, such as communication network control elements or functions, for example, access network elements, such as access points (APs), radio base stations (BSs), relay stations, eNBs, gNBs, etc.; and core network elements or functions, such as control nodes, support nodes, service nodes, gateways, user plane functions, access and mobility functions, etc., which may belong to one communication network system or different communication network systems.

[0091] In the following, different exemplary embodiments will be described using a communication network architecture based on the 3GPP standard, such as 5G / NR (or 6G / NR), as an example of an exemplary communication network to which the embodiments may be applied, without limiting the embodiments to such an architecture. It will be clear to those skilled in the art that the embodiments may also be applied to other types of communication networks in which mobile communication principles are integrated, such as 4G and / or LTE (or 5G, 6G or another "XG"), such as Wi-Fi, Worldwide Interoperability for Microwave Access (WiMAX), Personal Communications Service (PCS), Wideband Code Division Multiple Access (WCDMA), systems using ultra-wideband (UWB) technology, mobile ad hoc networks (MANET), wired access, etc. In addition, without loss of generality, some examples of the embodiments are described with respect to mobile communication networks, but the principles of the present disclosure can be extended and applied to any other type of communication network, such as a wired communication network or a data center network.

[0092] The following examples and embodiments should be understood as illustrative examples only. Although the specification may refer to "an," "one," or "some" examples or embodiments in several places, this does not necessarily mean that each such reference relates to the same (multiple) examples or (multiple) embodiments, nor does it necessarily mean that the feature applies only to a single example or embodiment. Individual features of different embodiments may also be combined to provide further embodiments. Furthermore, terms such as "comprising" and "including" should be understood as not limiting the described embodiments to consist only of those features that have been mentioned; such examples and embodiments may also include features, structures, units, modules, etc. that are not specifically mentioned.

[0093] The basic system architecture of a (long-range) communication network including a mobile communication system (in which some examples of the embodiments are applicable) may include the architecture of one or more communication networks, including (multiple) radio access network subsystems and (multiple) core networks. Such an architecture may include one or more communication network control elements or functions, access network elements, radio access network elements, access service network gateways or base transceiver stations, such as base stations (BS), access points (AP), NodeBs (NBs), eNBs or gNBs, distributed or centralized units (CUs), which control the corresponding coverage area or (multiple) cells, and one or more communication stations communicating therewith, such as communication elements or functions, such as user equipment (e.g., customer equipment), mobile devices or terminal devices, such as UEs, or another device with similar functionality, such as modem chipsets, chips, modules, etc., which may also be part of a station, element, function or application capable of communication, such as a UE, an element or function that can be used for a machine-to-machine communication architecture, or as a separate element attached to an element, function or application capable of communication, etc., capable of communicating via one or more channels via one or more communication beams to send several types of data in multiple access domains. In addition, it may include (core) network elements or network functions ((core) network control elements or network functions, (core) network management elements or network functions) such as gateway network elements / functions, mobility management entities, mobile switching centers, servers, databases, etc.

[0094] The general functionality and interconnection of the described elements and functions (which also depends on the actual network type) are known to those skilled in the art and are described in the corresponding specifications, so their detailed description is omitted here. However, it should be noted that in addition to those described in detail below, several additional network elements and signaling links may be used to communicate with elements, functions or applications (such as communication endpoints), communication network control elements (such as servers, gateways, radio network controllers) and other elements of the same or other communication networks.

[0095] The communication network architecture considered in the examples of the embodiments may also be able to communicate with other networks, such as the public switched telephone network or the Internet. The communication network may also be able to support cloud services for virtual network elements or their functions, wherein it should be noted that the virtual network portion of the telecommunications network may also be provided by non-cloud resources, such as an internal network, etc. It should be understood that the network elements and / or corresponding functions of the access system, core network, etc. may be implemented by using any node, host, server, access node or entity, etc. suitable for such purpose. In general, the network functions may be implemented as network elements on dedicated hardware, software instances running on dedicated hardware, or virtualized functions instantiated on an appropriate platform (e.g., cloud infrastructure).

[0096] In addition, network elements (such as communication elements, such as UE, mobile devices, terminal devices), control elements or functions (such as access network elements, such as base stations (BS), eNB / gNB, radio network controllers), core network control elements or functions (such as gateway elements) or other network elements or functions as described herein, (core) network management elements or functions and any other elements, functions or applications may be implemented by software, for example by a computer program product for a computer and / or by hardware. In order to perform its corresponding processing, the corresponding device, node, function or network element used may include several parts, modules, units, components, etc. (not shown), which are required for control, processing and / or communication / signaling functions. Such components, modules, units and components may include, for example, one or more processors or processor units, including one or more processing sections for executing instructions and / or programs and / or for processing data, storage or memory units or components for storing instructions, programs and / or data for use as a work area for the processor or processing section (e.g., ROM, RAM, EEPROM, etc.), input or interface components for inputting data and instructions by software (e.g., floppy disk, CD-ROM, EEPROM, etc.), user interfaces for providing monitoring and operation possibilities to the user (e.g., screen, keyboard, etc.), other interfaces or components for establishing links and / or connections under the control of the processor unit or section (e.g., wired and wireless interface components, radio interface components including, for example, antenna units, components for forming a radio communication section), etc., wherein the corresponding components forming the interface (such as the radio communication section) may also be located at a remote site (e.g., a radio head or radio station, etc.). It should be noted that in this specification, a processing section should not be regarded as meaning only a physical part of one or more processors, but may also be regarded as a logical division of the indicated processing tasks performed by one or more processors.

[0097] It should be understood that, according to some examples, a so-called "liquid" or flexible network concept may be employed, wherein the operations and functions of a network element, network function, or another entity of the network may be performed in a flexible manner in different entities or functions, such as nodes, hosts, or servers. In other words, the "division of labor" between the involved network elements, functions, or entities may vary from case to case.

[0098] Now refer to Figure 4 , Figure 4 Different diagrams of a function (FID) or a set of functions (FID1 , FID2 , . . . ) to be associated with a model (MID) or a set of models (MID1 , MID2 ) are shown according to various examples of embodiments.

[0099] exist Figure 4 , frame A can be considered to represent prior art. Figure 4 , it is assumed that the UE provides the gNB with at least information about supported capabilities (at the UE) and optionally also about supported ML model(s) (at the UE). Where the capabilities and ML model(s) are associated with a certain ML-enabled feature, e.g. Figure 3 shown.

[0100] ML models can be identified by a model ID and optionally supplemented by associated meta information. Functions can be identified by a function ID and optionally supplemented by associated meta information.

[0101] According to at least some examples of the embodiments, further reference is made to Figure 4 , there can be several frameworks to associate (multiple) FIDs and (multiple) MIDs. For example, Figure 4 Three different frameworks are shown. In framework A 410, a function (ID) is linked / associated with a set of ML models (ID). These associated ML models support the same function. In framework B 420, a given ML model (ID) is linked / associated with a set of different functions (ID). In this case, the associated functions are supported by the same ML model. In framework C 430, the ML model (ID) is independent of the function (ID). In this framework, any combination of frameworks A 410 and B 420 is supported.

[0102] With the framework disclosed in this paper, it is our goal to alleviate the aforementioned complexity issues and provide a scalable structure that can adapt to an increasing number of ML-enabled features (version 19, 6G) and underlying ML models. Figure 4 Frames B 420 and C 430 summarize the two types of solutions proposed in this specification.

[0103] This specification discloses Figure 4 Two possible associations between the LCM process signaling of the ML model configurations corresponding to frames B 420 and C 430 in FIG. The goal is to enable efficient collaboration and signaling between (multiple) UEs and the NW throughout the ML-enabled feature process. Some key concepts are highlighted below.

[0104] For frame B 420 or C 430, according to various examples of embodiments:

[0105] -The first node (e.g., UE) may provide, to the second node (e.g., gNB), information about the association between each of its (multiple) ML models and the functions provided by these ML models.

[0106] The first node may receive an activation message from the second node for activating one of the ML models, wherein the activation may include at least a configuration of a desired functional output KPI to be monitored by the second node.

[0107] -After receiving the ML model activation, the first node may start the corresponding ML model management process (model of relevant functions (monitoring, activation, deactivation, switching)).

[0108] - The first node may report the configuration function output KPI to the second node.

[0109] - The second node can start the appropriate ML function management process (model (monitoring, activation, deactivation, switching)) based on the function output KPI received from the first node.

[0110] In this specification, the following terms are introduced:

[0111] FID (Function ID): During the function identification process, a function or a group of related functions is uniquely identified by an ID (called an FID). Assuming the FID is a bit string, it can be a set of functions for a use case (such as activation, deactivation, or switching). However, the format of the FID is beyond the scope of this specification. An example format is further shown below.

[0112] MID (Model ID): During the model identification process, a model is uniquely identified by an ID (called the MID). This ID allows the model to be uniquely identified throughout its lifecycle. The format of this MID is also outside the scope of this specification. An example format is shown further below.

[0113] MID-LCM (MID-based LCM): The lifecycle management of ML models is done by

[0114] MID identifier.

[0115] FID-LCM (FID-based LCM): The lifecycle management of functions is identified by FID.

[0116] Furthermore, in accordance with at least some examples of embodiments, the following will be considered.

[0117] In Framework B 420, model-based LCM and function-based LCM are interdependent, where a given ML model is associated with a different (set of) functions. Interoperability between model LCM and function LCM requires that their configurations maintain a certain consistency.

[0118] In this framework B 420 according to various examples of embodiments, it is assumed that the UE has a set (or multiple) of ML models, and the NW requires the UE to use / activate a specific function. The function LCM is controlled by the NW, while the ML model LCM can be controlled by the UE, or alternatively by the NW. When the ML model LCM and the function LCM operate in different nodes, their configurations still need to be consistent to make them compatible.

[0119] In Framework C 430, the model-based LCM and the function-based LCM are completely independent. Therefore, in order to ensure effective interoperability between the two LCMs, strict consistency between the model and the function is required (configuration, monitoring, etc.).

[0120] The calling flow of the framework C 430 according to various examples of the embodiment.

[0121] Referring now to FIG5 (part 1 / 2 and part 2 / 2; connected at A-A' and B-B'), FIG5 illustrates Framework C: Call flow of function (ID)-based and ML model ID-based LCMs utilizing function and / or ML model switching in collaboration level z.

[0122] A detailed example call flow is given in Figure 5. According to various examples of embodiments, it is assumed that the UE has functionality and requires a model from the NW. Both the model and functionality will be controlled by the NW. However, in both cases, LCM operations are independent of each other, although synchronization and compatibility of operations need to be carefully handled and combined using configuration and signaling.

[0123] Referring to Figure 5, the calling process steps are as follows:

[0124] Step 0 (not shown): UE 500 sends a signal to NW (e.g., represented by gNB 510).

[0125] Indicates that it requires the use of specific ML features (e.g., beam management use case).

[0126] Step 1 (only for level z cooperation): If the UE 500 does not have a suitable model for the ML feature, the NW transmits the preferred model with the MID and associated meta information.

[0127] Step 2: The UE 500 determines the ML function association of the received model.

[0128] Option 1

[0129] Step 3: If the UE 500 accepts the ML function, the UE 500 will send an ACK (eg, confirmation message) to the NW in order to use the received model.

[0130] Step 4: UE 500 will configure its capabilities for the model and exchange its capability configuration to NW. After it is ready, it will request NW to enable ML features.

[0131] Step 5: The NW will enable / activate the ML function.

[0132] Step 6: UE 500 will start using the model.

[0133] Step 7-8: Since the model-based LCM is controlled by the NW, the UE 500 will send periodic, aperiodic / triggered MID performance reports to the NW. If the NW decides to switch the functionality based on the reported performance, the NW signals to switch the model.

[0134] Step 9: If handover is required, the UE 500 may decide to continue or fall back from step 1. Otherwise, proceed to step 10.

[0135] Step 10-12: Since the function-based LCM is controlled by the NW, the UE 500 will send periodic, aperiodic / triggered function reports to the NW. If the NW decides to switch the function based on the reported performance, the NW signals to switch the function.

[0136] Option 2

[0137] Step 13: The UE 500 rejects the association of its ML function with the received ML model.

[0138] Step 14: NW selects another model and continues with step 1 or back.

[0139] Hereinafter, reference is made to the calling flow of the framework B 420 according to various examples of the embodiments.

[0140] Furthermore, in accordance with at least some examples of embodiments, the following will be considered.

[0141] In frame B 420, a model is associated / linked to one or more functions. An example could be an intermediate KPI generated from the model. The following process may occur.

[0142] - Framework B 420 can be implemented in UE or NW.

[0143] - If Framework B 420 is implemented in NW, both Function (ID) based LCM and Model (ID) based LCM will be completely controlled by NW.

[0144] If frame B 420 is in the UE, three possible interactions can occur depending on the level of cooperation:

[0145] ○ During x-level collaboration, the functional and model LCMs are transparent to the NW.

[0146] ○ During level y collaboration, capabilities need to be indicated to achieve the requirements, configurations, and KPIs associated with a specific use case. In this case, since capabilities (sets of capabilities) are linked to models, the (multiple) models need to be

[0147] ■Register to NW,

[0148] ■Activated by NW,

[0149] Then the function (function set)

[0150] ■Activated by NW

[0151] ■Through NW's LCM management

[0152] o During level z, the UE may not have the model(s) required for a specific use case,

[0153] And request(s) the specified model(s) from NW. In this case,

[0154] ■Model LCM is maintained by NW,

[0155] ■ Function LCM is maintained by NW.

[0156] Figures 6 and 7 illustrate the appropriate invocation flow for Framework B, assuming that the model LCM is maintained by the UE, but the function-based LCM is maintained by the gNB. Since the gNB has more knowledge of the surrounding environment, it can help the UE effectively use the model through function-based LCM.

[0157] Referring now to FIG. 6 (part 1 / 2 and part 2 / 2; connected at AA' and BB'), FIG. 6 illustrates Framework B: a call flow for function (ID)-based LCM monitoring and switching at the UE using Framework B in collaboration level y.

[0158] According to various examples of the embodiment, the calling process of collaboration level y in framework B 420 is shown in FIG6 , and the steps are as follows:

[0159] Step 1: The UE 600 selects a model for ML-enabled features (e.g., beam management use case) and indicates and registers the model meta-information (identified by MID) and capabilities (or capability sets) associated with the model to the NW (e.g., represented by the gNB 610).

[0160] Information (identified by FID(s)).

[0161] Step 2: The NW determines the meta-information of the indicated function (function set) FID that enables the features of the ML.

[0162] Step 3-4: NW sends ACK signals (eg, confirmation messages) of (multiple) FIDs.

[0163] The UE 600 configures the FID option for the model.

[0164] Step 5: NW sends FID Activation via MAC CE.

[0165] Step 6-7: UE 600 starts using the model with the activated FID(s).

[0166] Model-related LCM operations such as model (monitoring, updating) are performed by the UE and are transparent to the NW.

[0167] Step 8: The UE 600 sends the performance KPI report of the FID(s) to the NW periodically / aperiodically / event-triggered.

[0168] Steps 9-10: The NW evaluates the performance of the FID and if the performance is not satisfactory, the NW may send a FID deactivation via MAC CE. The NW may also send a new FID activation via MAC CE.

[0169] Step 11: If there is a function switch in the UE 600, the operation of step 6 continues or rolls back.

[0170] Referring now to FIG. 7 (part 1 / 2 and part 2 / 2; connected at AA' and BB'), FIG. 7 illustrates Framework B: a call flow for function (ID) based LCM monitoring and switching at the UE using Framework B in collaboration level z.

[0171] According to at least some examples of the embodiments, the calling process of collaboration level z in frame B 420 is shown in FIG. 7 , and the steps are as follows:

[0172] Step 0 (not shown): The UE 700 indicates to the NW (e.g., represented by the gNB 710) the features (e.g., beam management use case) that enable ML.

[0173] Step 1: NW transmits the model and the functions (or function sets) associated with the model. The meta information of the model is identified by the MID, and the function (or function set) information is identified by the (multiple) FIDs associated with the model.

[0174] Step 2: UE 700 determines the functionality (functionality set) of the indicated ML-enabled feature

[0175] Meta information of FID(s).

[0176] Step 3-4: UE 700 sends an ACK signal (e.g., confirmation message) for both the MID and the FID(s). This signal ensures that UE 700 is able to use the model and that UE 700 supports the functional configuration indicated in the FID(s). UE 700 configures the FID options for the model.

[0177] Step 5-6: NW sends FID activation and MID activation signals to UE via MAC CE.

[0178] Step 6-7: UE 700 starts using the model with the active FID(s). Model-related LCM operations such as model (monitoring, updating) and function-related LCM operations such as function monitoring, handover are performed by the NW.

[0179] Step 8: UE 700 sends performance KPI reports of both MID and FID(s) to NW periodically / aperiodically / event-triggered.

[0180] Steps 9-10: The NW evaluates the performance of the FID and if the performance is not satisfactory, the NW may send a FID deactivation via MAC CE. The NW may also send a new FID activation via MAC CE.

[0181] Step 11: If there is a function switch in the UE 700, the operation of step 7 continues or rolls back.

[0182] Step 12: Since the NW is also monitoring the model, it may need to update the model. In this case, the NW can trigger a model switch.

[0183] Steps 13-14: The NW first sends an FID deactivation signal via the MAC CE, and then sends an MID deactivation signal via the MAC CE.

[0184] Step 15: If the UE 700 receives a model switching indication, it needs to use another model. If the model is not available in the UE 700, then step 1 is continued.

[0185] Examples of elements in the FID and MID are provided for completeness, but the content is not limited thereto.

[0186] The FID may contain at least one of the following:

[0187] a) Function ID: a label / tag / UNDI that uniquely identifies the function and combination of the following items b)-g).

[0188] b) (Multiple) Additional IDs / Labels / Tags:

[0189] a. If model identification is applied, each additional ID / label corresponds to a model ID.

[0190] b. Otherwise, each additional ID / tag corresponds to a model to be monitored, which can identify performance changes of the functions supported by the model.

[0191] c) Applicable scenarios / configurations / parameters / conditions for model function enablement: including system and intermediate KPIs for function-based LCM purposes.

[0192] d) Input data types / sources and preparation / preprocessing: including an indication of any latency-sensitive ML-specific data processing, e.g., as an indication of the expected latency budget for such operations

[0193] e)(multiple) non-ML operations (optional): Indicates any non-ML operations / algorithms involved in the model functionality, e.g., as an indication of the expected latency budget for such operations

[0194] f) Output Data and Post-Processing (optional): Includes an indication of any latency-sensitive ML-specific output post-processing, e.g., as an indication of the expected latency budget for such operations.

[0195] Show

[0196] g) Enable and (partially) control b)-f) operations and corresponding function-based LCM specific control signaling configuration(s)

[0197] The MID can contain at least one of the following:

[0198] a) Model ID: a label / tag / UNDI that can uniquely identify an ML model (implementation version, etc.) within an ML-enabled feature (across several potential capabilities) or only within a specified capability, and

[0199] b) Information related to AI / ML models

[0200] i. When using proprietary ML model formats, relevant information may include:

[0201] Potential additional meta-information required for function-based LCM

[0202] ii. When using the Open ML Model Format, relevant information may include:

[0203] Model input data (size, features) and preparation / preprocessing, including an indication of any feature extraction, feature selection, or any other latency-sensitive ML-specific data processing, e.g., as an indication of the expected latency budget for such operations

[0204] Model output data (size, features) and post-processing (optional), including an indication of any latency-sensitive ML-specific output post-processing, e.g., as an indication of the expected latency budget for such operations

[0205] · Specific control signaling configuration(s) that enable and (partially) control i)-ii) operations and corresponding LCM based on model ID

[0206] Hereinafter, other examples of the embodiments will be described with respect to the above-mentioned methods and / or apparatuses.

[0207] Now refer to Figure 8 , Figure 8 Flowcharts illustrating steps corresponding to methods according to various examples of embodiments are shown. Figure 8 The method steps shown may represent the above reference Figure 4 to at least a portion of the method / processing steps described in Figure 7. In addition, Figure 8 The illustrated method may be applied to the UE 500 , 600 and / or 700 described above with reference to FIG. 5 to FIG. 7 .

[0208] In particular, according to Figure 8 , in S810 , the method includes requesting to use the ML feature based on using the ML model applicable to the ML feature.

[0209] It should be noted that such a request may indicate a need to use a specific ML feature as described above with reference to FIG5 to FIG7 (e.g., steps 0 and / or 1). Thus, such an ML feature may be, for example, a beam management use case. Furthermore, the ML model may indicate the transmission / reception and / or selection of a (preferred) ML model as described above with reference to FIG5 to FIG7 (e.g., step 1). The ML model may be selected by the ML model described above with reference to FIG5 to FIG7 (e.g., step 1). Figure 4 The MID representation and / or identification described in relation to FIG. 7 .

[0210] Furthermore, at S820, the method includes receiving an activation message related to at least one of the following:

[0211] - activation of an ML model that is applicable to and suitable for ML features, an ML model associated with at least one feature, or

[0212] - Activation of one of at least one function associated with an ML model that is applicable and adapted for use with the ML feature.

[0213] The activation message indicates that an ML model and / or a feature is activated.

[0214] It should be noted that the reception of such an activation message may represent at least a portion of the enabling / activation described above with reference to FIG5 to FIG7 (e.g., steps 5 and 6). Thus, the activation message may be provided by an access network element, such as the gNB 510, 610, and / or 710 described above with reference to FIG5 to FIG7. Furthermore, such an ML model and / or such a function to be activated may represent the activation of the ML model and / or such a function as described above with reference to FIG5 to FIG7. Figure 4 7 (e.g., steps 5 and 6) and / or selected (preferred) ML models and (ML) functions. In addition, the term "available" may be understood as meaning that the ML models and / or functions are available and ready for use at a device (e.g., an endpoint terminal), which may be represented by the UE 500, 600, and / or 700 as described above with reference to FIG5 to FIG7. In addition, the expression "at least one function" may, for example, include the above-mentioned references to Figure 4 The FID1 and FID2 and / or the (ML) functions described above with reference to Figures 5 to 7. Therefore, an ML model may correspond to the (ML) functions described above with reference to Figures 5 to 7. Figure 4 To the MID described in FIG. 7 .

[0215] Furthermore, at S830 , the method includes starting to use an ML model based on the received activation message.

[0216] It should be noted that such initiation may represent at least a portion of the initiation described above with reference to Figures 5 to 7 (eg, steps 6 and 7).

[0217] Furthermore, in S840 , the method includes reporting at least one of the following: an ML model performance key performance indicator KPI of an ML model, or a function performance KPI of a function associated with an ML model.

[0218] It should be noted that such reporting may represent at least part of the reporting / sending of (performance) reports described above with reference to Figures 5 to 7, such as steps 7 and 10 in Figure 5, step 8 in Figure 6 and step 8 in Figure 7.

[0219] Furthermore, according to at least some examples of embodiments, the method may further include configuring one of at least one functions provided by an ML model; and providing a function configuration resulting from the configuration. The function configuration indicates at least one of a function performance KPI to be reported or an ML model performance KPI to be reported. The method may further include requesting at least one of: activation of an ML model or activation of a function; and receiving an activation message in response to the request.

[0220] It should be noted that this configuration may represent the configuration described above with reference to step 4 of Figure 5. Furthermore, this request may represent the request described above with reference to step 4 of Figure 5.

[0221] Furthermore, according to various examples of embodiments, the method may further include receiving an ML model applicable to the ML feature; and determining a functional association of the received ML model, the functional association indicating an association between the received ML model and at least one available function. If the determined functional association is accepted, the method may further include providing a confirmation message indicating that the received ML model is available. If the determined functional association is rejected, the method may further include providing a rejection message indicating that the received ML model is unavailable.

[0222] It should be noted that such reception may represent the reception described above with reference to step 1 of Figure 5. Furthermore, such determination may represent the determination described above with reference to step 2 of Figure 5. Such acceptance may represent the acceptance described above with reference to step 3 of Figure 5. Such rejection may represent the rejection described above with reference to step 13 of Figure 5.

[0223] Furthermore, according to various examples of embodiments, the method may further include: if the determined functional association is accepted, accepting the configuration as one of the at least one functions associated with the received ML model; and providing a functional configuration resulting from the configuration. The functional configuration indicates at least one of a functional performance KPI to be reported or an ML model performance KPI to be reported. The method may further include requesting at least one of: activation of the received ML model representing an ML model, or activation of one of the at least one functions associated with the received ML model, the received ML model representing an ML model. The method may further include receiving an activation message in response to the request.

[0224] It should be noted that this configuration may represent the configuration described above with reference to step 4 of Figure 5. Furthermore, this request may represent the request described above with reference to step 4 of Figure 5.

[0225] Optionally, according to at least some examples of the embodiments, the method may further include at least one of the following: receiving signaling for switching an ML model in response to the report; and deciding whether to use another ML model that can be used and applied to the ML feature, wait for receipt of a new ML model applied to the ML feature, or fall back, or receiving signaling for switching one function; and configuring another function based on the determined function association.

[0226] It should be noted that such reception may represent the reception described above with reference to FIG5, steps 7 to 8 and steps 10 to 12. Furthermore, such decision may represent the decision described above with reference to FIG5, step 9.

[0227] Furthermore, according to various examples of embodiments, the method may further include selecting an ML model that is applicable to and usable for the ML feature; indicating the selected ML model; and registering information related to at least the selected ML model and at least one function associated with the selected ML model. The method may further include receiving a confirmation signal for at least one function associated with the selected ML model; configuring one of the at least one functions associated with the selected ML model; wherein the configuration includes configuring at least one of a function performance KPI to be reported or an ML model performance KPI to be reported; and receiving an activation message, wherein the selected ML model indicates one ML model.

[0228] It should be noted that such selection, indication, and registration may represent the selection, indication, and registration described above with reference to step 1 of Figure 6. Furthermore, such reception may represent the reception described above with reference to steps 3 to 5 of Figure 6. Furthermore, such configuration may represent the configuration described above with reference to steps 3 to 4 of Figure 6.

[0229] Furthermore, according to at least some examples of embodiments, the method may further include controlling lifecycle management (LCM) operations associated with the selected ML model.

[0230] It should be noted that this control may represent the control described above with reference to steps 6 to 7 of FIG. 6 .

[0231] In addition, according to various examples of the embodiment, the method may further include: receiving deactivation of the function associated with the selected ML model in response to the report, or receiving deactivation of the function associated with the selected ML model in response to the report and receiving activation of another function of the at least one function associated with the selected ML model.

[0232] It should be noted that such receiving may represent the receiving described above with reference to steps 9 to 10 of FIG. 6 .

[0233] Furthermore, according to various examples of embodiments, the method may further include receiving an ML model applicable to an ML feature and at least one function associated with the received ML model; and determining information about the received at least one function related to the ML feature. The method may further include providing a confirmation signal for both the received ML model and the received at least one function, wherein the confirmation signal indicates that both the received ML model and the received at least one function are available. The method may further include configuring one of the received at least one function; wherein the configuration includes configuring at least one of a function performance KPI to be reported or an ML model performance KPI to be reported; and receiving an activation message, wherein the received ML model represents an ML model, and wherein the received one function represents a function.

[0234] It should be noted that this receiving may represent the receiving described above with reference to step 1 of FIG. 7 . Furthermore, this determining may represent the determining described above with reference to step 2 of FIG. 7 . Furthermore, this providing may represent the providing described above with reference to steps 3 and 4 of FIG. 7 . Furthermore, this configuring may represent the configuring described above with reference to steps 3 and 4 of FIG. 7 . Furthermore, this receiving may represent the receiving described above with reference to step 5 of FIG. 7 .

[0235] Optionally, according to at least some examples of the embodiments, the method may further include: receiving deactivation of the received function in response to the report, or receiving deactivation of the received function in response to the report and receiving activation of another function of the at least one received function.

[0236] It should be noted that such receiving may represent the receiving described above with reference to steps 9 to 10 of FIG. 7 .

[0237] In addition, according to various examples of embodiments, the method may further include: receiving, in response to the report, an indication that an update of the received ML model is required; receiving deactivation of the received function; receiving deactivation of the received ML model; and performing one of the following: using another ML model that can be used and applied to the ML feature, or waiting for receipt of a new ML model applicable to the ML feature and receipt of at least one function associated with the new ML model.

[0238] It should be noted that such receiving may represent the receiving described above with reference to steps 12 to 14 of Figure 7. Furthermore, such executing may represent the executing described above with reference to step 15 of Figure 7.

[0239] Furthermore, according to various examples of embodiments, the method may further include providing association information regarding an association between at least one available ML model and at least one function provided by the at least one ML model. Thus, if such association information is obtained at / for an endpoint terminal and provided, for example, to an access network element (e.g., a gNB), the access network element (i.e., the network) may be able to identify (multiple) ML models and associate at least one available (i.e., ready-to-use) function at such endpoint terminal. Furthermore, the network may use such association information to determine whether to perform an action at such endpoint terminal, such as updating at least one of the available ML models or available functions, or, for example, transferring / providing additional ML models and / or functions to be available at the endpoint terminal, if a specific ML feature is required.

[0240] In addition, according to various examples of embodiments, the method may further include controlling LCM operations (regarding ML models and / or functions) may include at least one of: monitoring ML models and / or functions, activating ML models and / or functions, deactivating ML models and / or functions, or switching ML models and / or functions.

[0241] The above solution enables LCM using ML model identification and ML function identification. Therefore, the above solution has the advantage that it enables LCM using ML model identification and ML function identification in a more efficient and / or more secure and / or more robust and / or fault-resistant and / or flexible and / or with reduced complexity.

[0242] Now refer to Figure 9 , Figure 9 Flowcharts illustrating steps corresponding to methods according to various examples of embodiments are shown. Figure 9 The method steps shown may represent the above reference Figure 4 to at least a portion of the method / processing steps described in Figure 7. In addition, Figure 9 The illustrated approach may be applied to access network elements, such as gNB 510, 610, and / or 710 described above with reference to Figures 5 to 7.

[0243] It should be noted that Figure 8 and Figure 9 Similar terms / expressions used in the present invention may be similarly understood, and thus repeated explanations thereof may be omitted.

[0244] Specifically, according to Figure 9 Regarding an ML feature that needs to be used based on use of an ML model applicable to machine learning ML features, in S910, the method includes providing an activation message related to at least one of the following: activation of an ML model that can be used and applied to the ML feature, association of an ML model with at least one function, or activation of one of at least one functions associated with an ML model that can be used and applied to the ML feature.

[0245] It should be noted that such provision may refer to the provision / transmission described above with reference to Figures 5 to 7 (e.g., steps 5 to 6). In addition, the term "available" may be understood to mean that the ML model and / or function is available and ready for use at a device (e.g., an endpoint terminal), which may be represented by the UE 500, 600, and / or 700 described above with reference to Figures 5 to 7.

[0246] Furthermore, in S920 , the method includes activating an ML model and / or a function.

[0247] It should be noted that such activation may represent the enabling / activation described above with reference to FIG. 5 to FIG. 7 (eg, steps 5 to 6 ).

[0248] Furthermore, at S930 , the method includes receiving a report indicating at least one of: an ML model performance key performance indicator (KPI) of an ML model, or a function performance KPI of a function associated with an ML model.

[0249] It should be noted that such receiving may represent at least a portion of the receiving described above with reference to FIG. 5 to FIG. 7 (eg, steps 7 to 8 and 10 to 12 ).

[0250] Furthermore, at S940 , the method includes executing an action related to an ML model and / or a function based on the received report.

[0251] It should be noted that such execution action may represent at least a portion of such execution action derived from the above reference to FIG5 to FIG7 (eg, steps 7 to 14), such as the action to be executed by the NW as described in FIG5 to FIG7

[0252] In addition, according to at least some examples of the embodiments, the method may also include: receiving a functional configuration indicating at least one of a functional performance KPI included in a report to be received, or an ML model performance KPI included in a report to be received; receiving a request regarding at least one of: activation of an ML model, or activation of a function; and providing an activation message in response to the received request.

[0253] It should be noted that such receiving may represent the receiving described above with reference to step 4 of Figure 5. Furthermore, such providing may represent the providing described above with reference to step 4 of Figure 5.

[0254] In addition, according to various examples of embodiments, the method may further include: providing an ML model applicable to the ML feature; and if the association between the provided ML model and at least one available function is accepted, receiving a confirmation message indicating that the provided ML model is available, wherein if the association is rejected, receiving a rejection message indicating that the provided ML model is not available.

[0255] It should be noted that such providing may represent the providing as described above with reference to step 1 of Figure 5. Furthermore, such receiving may represent the receiving as described above with reference to steps 3 and 13 of Figure 5.

[0256] In addition, according to various examples of the embodiment, the method may further include: if the association is accepted, receiving a function configuration generated by the configuration of one of the at least one functions accepted as associated with the provided ML model, wherein the function configuration indicates a function performance KPI included in the report to be received, or at least one of the ML model performance KPIs included in the report to be received; receiving a request regarding at least one of the following: activation of the provided ML model representing an ML model, or activation of one of the at least one functions accepted as associated with the provided ML model, the provided ML model representing a function associated with an ML model; and providing an activation message in response to the received request.

[0257] It should be noted that such receiving may represent the receiving described above with reference to step 4 of FIG. 5 .

[0258] Optionally, according to at least some examples of the embodiments, the method may further include at least one of the following:

[0259] -Control lifecycle management (LCM) operations associated with an ML model, or

[0260] Control LCM operations related to a function;

[0261] - The execution action may further include at least one of the following:

[0262] Based on the controlled LCM operation associated with an ML model, determine whether to switch a function; if it is determined that a function is to be switched, provide signaling for switching an ML model, or

[0263] Based on the controlled LCM operation related to a function, it is determined whether a function is to be switched; if it is determined that a function is to be switched, signaling for switching a function is provided.

[0264] It should be noted that this LCM control may represent the LCM control described above with reference to FIG5, steps 7 to 8 and steps 10 to 12. Furthermore, this decision may represent the decision described above with reference to FIG5, steps 7 to 8.

[0265] In addition, according to various examples of embodiments, the method may further receive an indication related to a selected ML model that is available and applicable to the ML feature; receive a registration of information related to at least the selected ML model and at least one function associated with the selected ML model; determine that the registration is received; based on the determination, provide a confirmation signal of at least one function associated with the selected ML model; and provide an activation message, wherein the selected ML model represents an ML model.

[0266] It should be noted that such receiving may represent the receiving described above with reference to step 1 of Figure 6. Furthermore, such determining may represent the determining described above with reference to step 2 of Figure 6. Furthermore, such providing may represent the providing described above with reference to steps 3 to 5 of Figure 6.

[0267] In addition, according to at least some examples of the embodiments, the performing action may further include at least one of: providing deactivation of a function associated with the selected ML model, or providing deactivation of a function associated with the selected ML model and providing activation of another function of at least one function associated with the selected ML model.

[0268] It should be noted that this provision may represent the provision described above with reference to steps 9 to 10 of FIG. 6 .

[0269] In addition, according to various examples of embodiments, the method may further include: providing an ML model applicable to the ML feature, and at least one function associated with the provided ML model; receiving confirmation signals for both the provided ML model and the provided at least one function, wherein the confirmation signals indicate that both the provided ML model and the provided at least one function are available; and providing an activation message, wherein the provided ML model represents one ML model, and wherein the provided one function represents one function.

[0270] It should be noted that such providing and receiving may represent the providing and receiving described above with reference to steps 1 to 5 of FIG. 7 .

[0271] In addition, according to various examples of embodiments, the method may further include: controlling LCM operations related at least to the provided ML model and one provided function; and wherein the performing action includes at least one of the following: providing deactivation of the provided function, or providing deactivation of the provided function and providing activation of another function of the at least one provided function based on the controlled LCM operations.

[0272] It should be noted that this provision may represent the provision described above with reference to steps 9 to 10 of FIG. 7 .

[0273] Optionally, according to at least some examples of the embodiments, the method may further: control LCM operations related at least to a provided ML model and a provided function; and wherein the performing actions include at least one of the following: triggering switching of an ML model by providing an indication that an update of the provided ML model is required based on the controlled LCM operations; providing deactivation of the provided function; and providing deactivation of the provided ML model.

[0274] It should be noted that such triggering and providing may represent the triggering and providing described above with reference to steps 12 to 14 of FIG. 7 .

[0275] In addition, according to various examples of the embodiment, the method may further include receiving Figure 8 Thus, if such association information is obtained at / for an endpoint terminal and received, for example, to an access network element (e.g., a gNB), the method may further include identifying (a plurality of) ML models and associating at least one available (i.e., ready-to-use) function at such endpoint terminal. Furthermore, correspondingly, the method may further include using such association information to determine whether to perform an action at such endpoint terminal, such as updating at least one of the available ML models or available functions, or transferring / providing additional ML models and / or functions to be available at the endpoint terminal, such as when a specific ML feature is required.

[0276] Therefore, the method may further include determining that no ML model and / or function is available for using the specific ML feature based on the received association information; and providing an appropriate ML model and / or function for using the specific ML feature.

[0277] In addition, according to various examples of embodiments, the method may further include controlling LCM operations (regarding ML models and / or functions) may include at least one of: monitoring ML models and / or functions, activating ML models and / or functions, deactivating ML models and / or functions, or switching ML models and / or functions.

[0278] The above solution allows the use of ML model identifiers and ML function identifiers for LCM. Therefore, the above solution has the advantage that it can use ML model identifiers and ML function identifiers to enable LCM in a more efficient and / or more secure and / or more robust and / or fault-resistant and / or flexible and / or with reduced complexity.

[0279] Now refer to Figure 10 , Figure 10 Block diagrams illustrating apparatuses according to various examples of embodiments are shown.

[0280] Specifically, Figure 10 A block diagram illustrating an apparatus 1000 according to various examples of embodiments is shown. The apparatus 1000 may represent an endpoint terminal, such as the UE described above with reference to FIG5 to FIG7 , which may participate in LCM using an ML model identifier and an ML function identifier. Furthermore, even when referring to an endpoint terminal, the endpoint terminal may be another device or function with a similar task, such as a chipset, chip, module, application, etc., which may also be part of a network element or connected to a network element as a separate element. It should be understood that each block and any combination thereof may be implemented by various means or combinations thereof, such as hardware, software, firmware, one or more processors, and / or circuit systems.

[0281] Figure 10 The illustrated apparatus 1000 may include a processing circuit system, a processing function, a control unit, or a processor 1010, such as a CPU or the like, which is adapted to enable LCM using an ML model identifier and an ML function identifier. The processor 1010 may include one or more processing portions or functions dedicated to a specific process as described below, or the process may be run within a single processor or processing function. The portion for performing such a specific process may also be provided as a discrete component, or provided within one or more other processors, processing functions, or processing portions, such as within a physical processor (e.g., a CPU) or within one or more physical or virtual entities. Reference numerals 1031 and 1032 represent input / output (I / O) units or functions (interfaces) connected to the processor or processing function 1010. The I / O units 1031 and 1032 may be combined units including communication devices toward several entities / elements, or may include a distributed structure with multiple different interfaces for different entities / elements. Reference numeral 1020 represents a memory that may be used, for example, to store data and programs to be executed by the processor or processing function 1010 and / or serve as working memory for the processor or processing function 1010. It should be noted that the memory 1020 may be implemented by using one or more memory portions of the same or different types of memory, but may also represent external memory, such as an external database provided on a cloud server.

[0282] The processor or processing function 1010 is configured to perform processing related to the above-mentioned processing. In particular, the processor or processing circuit system or function 1010 includes one or more of the following sub-parts. Sub-part 1011 is a necessary part that can be used as a part for using the ML model. Part 1011 can be configured according to Figure 8 The S810 of the embodiment of the present invention performs processing. In addition, the sub-section 1012 is a receiving section, which can be used as a section for receiving an activation message. Section 1012 can be configured according to Figure 8 S820 performs processing. In addition, sub-section 1013 is a starting section, which can be used as a section for starting to use the ML model. Section 1013 can be configured according to Figure 8 S830 performs processing. In addition, sub-section 1014 is a reporting section, which can be used as a section for reporting KPIs. Section 1014 can be configured according to Figure 8 The processing of S840 is executed.

[0283] According to various examples of the embodiment, for Figure 10 The device 1000 may also consider the following.

[0284] According to various examples of embodiments, the apparatus 1000 may include at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus 1000 to at least:

[0285] Requesting to use ML features based on using an ML model suitable for machine learning ML features;

[0286] Receiving an activation message related to at least one of the following:

[0287] activation of an ML model available and applicable to ML features at apparatus 1000, an ML model associated with at least one function, or

[0288] activation of one of at least one functionality associated with one of the ML models available and applicable to the ML feature at the device,

[0289] wherein the activation message indicates that an ML model and / or a function is activated;

[0290] Based on the received activation message, starting to use an ML model; and

[0291] Report at least one of the following:

[0292] ML model performance key performance indicator (KPI) for an ML model, or

[0293] Feature performance KPI for a feature associated with an ML model.

[0294] According to various examples of embodiments, the apparatus 1000 may also be caused to configure one of at least one functions provided by an ML model; provide a function configuration resulting from the configuration, wherein the function configuration indicates at least one of a function performance KPI to be reported or an ML model performance KPI to be reported; request at least one of: activation of an ML model or activation of a function; and receive an activation message in response to the request.

[0295] According to various examples of embodiments, the device 1000 may also be caused to receive an ML model applicable to an ML feature; determine a functional association of the received ML model, the functional association indicating an association between the received ML model and at least one available function; and if the device 1000 is caused to determine the functional association results in the device 1000 accepting the determined functional association, provide a confirmation message indicating that the received ML model is available, wherein if the result of the device 1000 being caused to determine the functional association results in the device 1000 rejecting the determined functional association, provide a rejection message indicating that the received ML model is not available.

[0296] According to various examples of embodiments, if the device 1000 accepts the determined function association, the device 1000 may also be caused to configure a function that is accepted as one of the at least one functions associated with the received ML model; provide a function configuration generated by the configuration, wherein the function configuration indicates a function performance KPI to be reported, or at least one of the ML model performance KPIs to be reported; request at least one of the following: activation of the received ML model representing an ML model, or activation of one of the at least one functions accepted as associated with the received ML model, the received ML model representing a function associated with the ML model; and receive an activation message in response to the request.

[0297] According to various examples of embodiments, the device 1000 may also be caused to receive at least one of signaling for switching one ML model in response to the report; and decide whether to use another ML model available and applicable to the ML feature at the device, wait for receipt of a new ML model applicable to the ML feature, or fall back, or receive signaling for switching one function; and configure another function based on the determined functional association.

[0298] According to various examples of embodiments, the apparatus 1000 may also be caused to select an ML model that is available and applicable for ML features at the apparatus; indicate the selected ML model; and register information related to at least the selected ML model and at least one function associated with the selected ML model; receive a confirmation signal of at least one function associated with the selected ML model; configure one of the at least one functions associated with the selected ML model; wherein the apparatus 1000 is caused to configure one function including configuring at least one of: a function performance KPI to be reported, or an ML model performance KPI to be reported; and receive an activation message, wherein the selected ML model represents one ML model.

[0299] According to various examples of embodiments, the apparatus 1000 may also be caused to control lifecycle management (LCM) operations associated with the selected ML model.

[0300] According to various examples of embodiments, in response to the apparatus 1000 being caused to report, the apparatus 1000 may also be caused to receive deactivation of a function associated with the selected ML model, or receive deactivation of a function associated with the selected ML model and receive activation of another function of the at least one function associated with the selected ML model.

[0301] According to various examples of embodiments, the device 1000 may also be caused to receive an ML model applicable to an ML feature, and at least one function associated with the received ML model; determine information of the received at least one function related to the ML feature; provide confirmation signals for both the received ML model and the received at least one function, wherein the confirmation signal indicates that both the received ML model and the received at least one function are available at the device; configure one of the received at least one function; wherein the device is caused to perform configuration including configuration including at least one of a function performance KPI to be reported or an ML model performance KPI to be reported; and receive an activation message, wherein the received ML model represents an ML model, and wherein the received one function represents a function.

[0302] According to various examples of embodiments, in response to the device 1000 being caused to report, the device 1000 may also be caused to receive the deactivation of the received function, or receive the deactivation of the received function and receive the activation of another function of the received at least one function.

[0303] According to various examples of embodiments, in response to the apparatus 1000 being caused to report, the apparatus 1000 may also be caused to receive an indication that an update of the received ML model is required; receive a deactivation of the received function; receive a deactivation of the received ML model; and perform one of the following: use another ML model available and applicable to the ML feature at the apparatus, or wait for receipt of a new ML model applicable to the ML feature and receipt of at least one function associated with the new ML model.

[0304] Now refer to Figure 11 , Figure 11 Block diagrams illustrating apparatuses according to various examples of embodiments are shown.

[0305] Specifically, Figure 11 A block diagram illustrating an apparatus according to various examples of embodiments is shown. The apparatus may represent an access network element, such as the gNB described above with reference to Figures 5 to 7 , which may participate in LCM using an ML model identifier and an ML function identifier. Furthermore, even when reference is made to an access network element, the access network element may be another device or function with similar tasks, such as a chipset, chip, module, application, etc., which may also be part of a network element or connected to the network element as a separate element. It should be understood that each block and any combination thereof may be implemented by various means or combinations thereof, such as hardware, software, firmware, one or more processors, and / or circuit systems.

[0306] Figure 11The illustrated apparatus 1100 may include a processing circuit system, a processing function, a control unit, or a processor 1110, such as a CPU, etc., which is suitable for enabling LCM using an ML model identifier and an ML function identifier. The processor 1110 may include one or more processing portions or functions dedicated to a specific process as described below, or the process may be run within a single processor or processing function. The portion for performing such a specific process may also be provided as a discrete component, or provided within one or more other processors, processing functions, or processing portions, such as within a physical processor (e.g., a CPU) or within one or more physical or virtual entities. Reference numerals 1131 and 1132 represent input / output (I / O) units or functions (interfaces) connected to the processor or processing function 1110. The I / O units 1131 and 1132 may be combined units including communication devices toward several entities / elements, or may include a distributed structure with multiple different interfaces for different entities / elements. Reference numeral 1120 represents a memory that may be used, for example, to store data and programs to be executed by the processor or processing function 1110 and / or serve as working memory for the processor or processing function 1110. It should be noted that the memory 1120 may be implemented by using one or more memory portions of the same or different types of memory, but may also represent external memory, such as an external database provided on a cloud server.

[0307] The processor or processing function 1110 is configured to perform processing related to the above-mentioned processing. In particular, the processor or processing circuit system or function 1110 includes one or more of the following sub-parts. Sub-part 1111 is a providing part, which can be used as a part for providing activation messages. Part 1111 can be configured according to Figure 9 S910 performs processing. In addition, sub-section 1112 is an activation section, which can be used as a section for activating ML models and / or functions. Section 1112 can be configured according to Figure 9 S920 performs processing. In addition, sub-section 1113 is a receiving section, which can be used as a section for receiving reports. Section 1113 can be configured according to Figure 9 S930 performs processing. In addition, sub-section 1114 is an execution action section, which can be used as a section for executing actions related to ML models and / or functions. Section 1114 can be configured according to Figure 9 S940 executes the processing.

[0308] According to various examples of the embodiment, for Figure 11 The device 1100 may also consider the following.

[0309] According to various examples of embodiments, the apparatus 1100 may include at least one processor; and at least one memory storing instructions, which, when executed by the at least one processor, cause the apparatus 1100 to at least:

[0310] With respect to an ML feature required for use by another device based on use of an ML model applicable to the machine learning ML feature, providing an activation message related to at least one of the following: activation of an ML model available and applicable to the ML feature at the other device, activation of an ML model associated with at least one function, or activation of a function in at least one function associated with an ML model available and applicable to the ML feature at the other device; activation of an ML model and / or a function; a report indicating at least one of the following: an ML model performance key performance indicator (KPI) of an ML model, or a function performance KPI of a function associated with an ML model; and performing an action related to an ML model and / or a function based on the received report.

[0311] According to various examples of embodiments, the device 1100 may also be caused to: receive a functional configuration indicating at least one of a functional performance KPI included in a report to be received, or an ML model performance KPI included in a report to be received; receive a request regarding at least one of: activation of an ML model, or activation of a function; and provide an activation message in response to the received request.

[0312] According to various examples of embodiments, the device 1100 may also be caused to; provide an ML model applicable to an ML feature; and if the association between the provided ML model and at least one available function is accepted, receive a confirmation message indicating that the provided ML model is available at the other device, wherein if the association is rejected, receive a rejection message indicating that the provided ML model is not available at the other device.

[0313] According to various examples of embodiments, the apparatus 1100 may further be caused to: if the association is accepted, receive a function configuration generated by the configuration of one of the at least one functions accepted as associated with the provided ML model, wherein the function configuration indicates a function performance KPI included in the report to be received, or at least one of the ML model performance KPIs included in the report to be received; receive a request regarding at least one of: activation of the provided ML model representing an ML model, or activation of one of the at least one functions accepted as associated with the provided ML model, the provided ML model representing a function associated with the ML model; and provide an activation message in response to the received request.

[0314] According to various examples of embodiments, the apparatus 1100 may further be caused to at least one of: control a lifecycle management LCM operation associated with an ML model, or control an LCM operation associated with a function;

[0315] The device 1100 being caused to perform an action may also include the device 1100 being caused to perform at least one of the following: determining whether to switch a function based on a controlled LCM operation related to an ML model; if it is determined that a function is to be switched, providing signaling for switching an ML model; determining whether to switch a function based on a controlled LCM operation related to a function; if it is determined that a function is to be switched, providing signaling for switching a function.

[0316] According to various examples of embodiments, the device 1100 may also be caused to: receive an indication related to a selected ML model that is available and applicable for the ML feature at the other device; receive a registration of information related to at least the selected ML model and at least one function associated with the selected ML model; determine information that the registration is received; provide a confirmation signal of at least one function associated with the selected ML model based on the determined information; and provide an activation message, wherein the selected ML model represents one ML model.

[0317] According to various examples of the embodiment, the device 1100 being caused to perform an action may also include the device 1100 being caused to perform at least one of the following: providing deactivation of a function associated with the selected ML model, or providing deactivation of a function associated with the selected ML model and providing activation of another function of the at least one function associated with the selected ML model.

[0318] According to various examples of embodiments, the device 1100 may further be caused to: provide an ML model applicable to an ML feature, and at least one function associated with the provided ML model; receive confirmation signals for both the provided ML model and the provided at least one function, wherein the confirmation signal indicates that both the provided ML model and the provided at least one function are available at the other device; and provide an activation message, wherein the provided ML model represents one ML model, and wherein the provided one function represents one function.

[0319] According to various examples of embodiments, the apparatus 1100 may also be caused to control LCM operations related to at least the provided ML model and the provided one function; and

[0320] Wherein the apparatus 1100 is caused to perform an action may further include the apparatus 1100 being caused to perform at least one of: providing deactivation of a provided function, or providing deactivation of a provided function and providing activation of another function of the at least one provided function based on the controlled LCM operation.

[0321] According to various examples of embodiments, the apparatus 1100 may also be caused to control LCM operations related to at least the provided ML model and the provided one function; and

[0322] The apparatus 1100 being caused to perform an action may further include the apparatus 1100 being caused to perform at least one of: triggering switching of an ML model by providing an indication that an update of a provided ML model is required based on the controlled LCM operation; providing deactivation of a provided function; and providing deactivation of a provided ML model.

[0323] It should be noted that, as mentioned above Figure 10 and Figure 11 The apparatuses 1000 and 1100 may include other / additional sub-parts, which may allow the apparatuses 1000 and 1000 to perform the functions described above with reference to FIG5 to FIG7 and / or Figure 8 and Figure 9 The method / method steps described.

[0324] In addition, according to various examples of embodiments, a computer program product for a computer can be provided, comprising a software code portion, which, when the above-mentioned product is run on a computer, is used to perform the steps of any one of the appended claims 1 to 9 or any one of the appended claims 10 to 15, wherein optionally, the computer program product includes a computer-readable medium storing the above-mentioned software code portion, and / or the computer program product can be directly loaded into the internal memory of the computer and / or can be sent via a network through at least one of upload, download and push processes.

[0325] It should be understood

[0326] -The access technology used to transmit services between entities in the communication network can be any suitable existing or future technology, such as WLAN (wireless local area network), WiMAX (worldwide interoperability for microwave access), LTE, LTE-A, 5G, Bluetooth, infrared, etc.; in addition, the embodiments can also apply wired technology, such as IP-based access technology, such as a wired network or a fixed line.

[0327] -Embodiments suitable for being implemented as software code or portion thereof and run using a processor or processing functionality are independent of the software code and may be specified using any known or future developed programming language, such as a high-level programming language, such as object-oriented C, C, C++, C#, Java, Python, Javascript, other scripting languages, etc., or a low-level programming language, such as machine language or assembly language.

[0328] -The implementation of the embodiments is hardware independent and can be implemented using any known or future developed hardware technology or any mixture of these technologies, such as a microprocessor or CPU (Central Processing Unit), MOS (Metal Oxide Semiconductor), CMOS (Complementary MOS), BiMOS (Bipolar MOS), BiCMOS (Bipolar CMOS), ECL (Emitter Coupled Logic) and / or TTL (Transistor Transistor Logic).

[0329] - embodiments may be implemented as individual devices, apparatuses, units, components or functions, or in a distributed manner, e.g., one or more processors or processing functions may be used or shared in a process, or one or more processing portions may be used and shared in a process, where one physical processor or more than one physical processor may be used to implement one or more processing portions dedicated to a particular process being described,

[0330] - A device may be formed from or include a semiconductor chip, a chipset, or a combination thereof

[0331] (Hardware) module to achieve;

[0332] - The embodiments may also be implemented as any combination of hardware and software, such as ASIC (Application Specific IC (Integrated Circuit)) components, FPGA (Field Programmable Gate Array) or CPLD (Complex Programmable Logic Device) components or DSP (Digital Signal Processor) components.

[0333] The embodiments may also be implemented as a computer program product including a computer usable medium having computer readable program code embodied therein, the computer readable program code being adapted to execute the processes described in the embodiments, wherein the computer usable medium may be a non-transitory medium.

[0334] Although the present disclosure has been described herein before with reference to specific embodiments, the present disclosure is not limited thereto and various modifications may be made.

Claims

1. A method comprising: Requesting ( S810 ) to use the ML feature based on using an ML model suitable for machine learning the ML feature; An activation message is received (S820) related to at least one of the following: activation of an ML model applicable to and adapted for the ML feature, the ML model being associated with at least one function, or activation of one of at least one function associated with one of the ML models available and applicable to the ML feature, wherein the activation message indicates that the one ML model and / or the one function is activated; Based on the received activation message, starting (S830) using the one ML model; and Report (S840) at least one of the following: the ML model performance key performance indicator (KPI) of the ML model, or A function performance KPI of the one function associated with the one ML model.

2. The method according to claim 1, further comprising: configuring the one of the at least one function provided by the one ML model; providing a functional configuration resulting from said configuration, wherein the functional configuration indicates at least one of the functional performance KPI to be reported, or the ML model performance KPI to be reported; Request at least one of the following: the activation of the one ML model, or said activation of said one function; and The activation message is received in response to the request.

3. The method according to claim 1, further comprising: receiving an ML model applicable to the ML feature; determining a functional association of the received ML model, the functional association representing an association between the received ML model and at least one available function; as well as If the determined functional association is accepted, providing a confirmation message indicating that the received ML model is available, wherein If the determined functional association is rejected, a rejection message indicating that the received ML model is unusable is provided.

4. The method according to claim 3, further comprising: If the determined functional association is accepted, configuration is accepted as one of the at least one functions associated with the received ML model; providing a functional configuration resulting from said configuration, wherein the functional configuration indicates at least one of the functional performance KPI to be reported, or the ML model performance KPI to be reported; Request at least one of the following: the received activation of the ML model representing the one ML model, or being accepted as the activation of the one of the at least one functionality associated with the received ML model, the received ML model representing the one functionality associated with the one ML model; and The activation message is received in response to the request.

5. The method according to any one of claims 1 to 4, further comprising at least one of the following: in response to the report, receiving a signaling for switching the one ML model; and deciding whether to use another ML model available and applicable to the ML feature, wait for receipt of a new ML model applicable to the ML feature, or fall back, or receiving signaling for switching the one function; and configuring another function based on the determined functional association, or receiving a deactivation of the functionality associated with the selected ML model, or Deactivation of the function associated with the selected ML model is received, and activation of another function of the at least one function associated with the selected ML model is received.

6. The method according to claim 1, further comprising: receiving an ML model applicable to the ML feature and at least one function associated with the received ML model; determining information of the received at least one function related to the ML feature; providing confirmation signals for both the received ML model and the received at least one function, wherein the confirmation signal indicates that both the received ML model and the received at least one function are available; configuring a function of the received at least one function; wherein the configuration includes: configuration of at least one of the functional performance KPI to be reported, or the ML model performance KPI to be reported; as well as The activation message is received, wherein the received ML model represents the one ML model, and wherein the received one function represents the one function.

7. The method according to claim 1 or 6, further comprising: In response to the report, receiving a deactivation of the received function, or In response to the report, a deactivation for the received function is received, and an activation of another one of the received at least one function is received.

8. The method according to any one of claims 1, 6 or 7, further comprising: receiving, in response to the report, an indication that an update of the received ML model is required; receiving a deactivation for the received function; receiving a deactivation for the received ML model; and Do one of the following: Use another ML model that is applicable and suitable for the ML features, or Waiting for receipt of a new ML model applicable to the ML feature and receipt of at least one function associated with the new ML model.

9. A method comprising: Regarding the ML feature that needs to be used based on use of the ML model applicable to the machine learning ML feature, providing (S910) an activation message related to at least one of the following: activation of an ML model applicable to and adapted for the ML feature, the ML model being associated with at least one function, or activation of one of at least one functions associated with an ML model that is applicable and adapted for use with the ML feature; activating ( S920 ) the one ML model and / or the one function; Receive (S930) a report indicating at least one of the following: the ML model performance key performance indicator (KPI) of the ML model, or a function performance KPI of the one function associated with the one ML model; as well as Based on the received report, an action related to the one ML model and / or the one function is performed ( S940 ).

10. The method according to claim 9, further comprising: receiving a function configuration indicating at least one of the function performance KPI included in the report to be received or the ML model performance KPI included in the report to be received; Receive a request for at least one of the following: the activation of the one ML model, or said activation of said one function; and The activation message is provided in response to the received request.

11. The method according to claim 9, further comprising: Providing an ML model applicable to the ML features; as well as If the association between the provided ML model and at least one available function is accepted, receiving a confirmation message indicating that the provided ML model is available, wherein If the association is rejected, a rejection message is received indicating that the provided ML model is unusable.

12. The method according to claim 11, further comprising: If the association is accepted, then receiving a function configuration resulting from a configuration accepted as one of the at least one functions associated with the provided ML model, wherein the function configuration indicates at least one of the function performance KPI included in the report to be received or the ML model performance KPI included in the report to be received; Receive a request for at least one of the following: representing the activation of the provided ML model of the one ML model, or being accepted as the activation of the one of the at least one functionality associated with the provided ML model, the provided ML model representing the one functionality associated with the one ML model; and providing said activation message in response to said received request, The method further comprises at least one of the following: Control lifecycle management (LCM) operations associated with the one ML model, or controlling LCM operations associated with the one function; The execution action includes at least one of the following: determining whether to switch the one function based on the controlled LCM operation associated with the one ML model; and If it is decided to switch the one function, signaling for switching the one ML model is provided, or determining whether to switch the one function based on the controlled LCM operation associated with the one function; and If it is decided to switch the one function, signaling for switching the one function is provided.

13. The method according to claim 9, further comprising: providing an ML model applicable to the ML feature, and at least one function associated with the provided ML model; receiving confirmation signals for both the provided ML model and the provided at least one function, wherein the confirmation signal indicates that both the provided ML model and the provided at least one function are available; and The activation message is provided, wherein the provided ML model represents the one ML model, and wherein the provided one function represents the one function.

14. The method according to claim 9 or 13, further comprising: controlling LCM operations associated with at least the provided ML model and the provided one function; and The execution action includes at least one of the following: based on the controlled LCM operation, Provide for deactivation of the functionality provided, or Deactivation of the provided function is provided, and activation of another one of the provided at least one function is provided.

15. The method according to any one of claims 9, 13 or 14, further comprising: controlling LCM operations associated with at least the provided ML model and the provided one function; and The execution action includes at least one of the following: based on the controlled LCM operation, triggering switching of the one ML model by providing an indication that an update of the provided ML model is required; Provide for deactivation of the said functionality provided; as well as Provides deactivation of the provided ML model.

16. An apparatus (1000; 1100), comprising: at least one processor; as well as At least one memory stores instructions that, when executed by the at least one processor, cause the apparatus (1000) to at least perform the method according to any one of claims 1 to 8, or cause the apparatus (1100) to at least perform the method according to any one of claims 9 to 15.

Images