Registering and discovering external federated learning clients in wireless communication systems

Abstract

The present disclosure provides an application entity in a wireless communication system, comprising: a processor; and a memory coupled with the processor. The processor is configured to cause the application entity to: receive, from at least one other network or application entity of a wireless communication system, a query for acting as a federated learning participant of a particular machine learning enabled analysis event, wherein the wireless communication system includes a core network; determining a federated learning client configuration file for the specific analysis event enabling machine learning, wherein the federated learning client profile comprises the ability of the application entity to act as a federated learning client of at least one data source outside the core network and / or to act as a data source outside the core network for the specific machine learning enabled analysis event; and transmitting a request to register registration information to a service registration entity, the registration information including the federated learning client profile and an identifier of the application entity.

Description

Technical Field

[0001] The subject matter disclosed herein generally relates to the field of implementing registration and discovery of federated learning clients in a wireless communication system. This document defines an application entity, a server entity, and a service registration entity in a wireless communication system. This document also defines methods within the application entity, the server entity, and the service registration entity in the wireless communication system. Background Art

[0002] In the Third Generation Partnership Project (3GPP), fifth-generation (5G) core analytics services are provided by the Network Data Analysis Function (NWDAF). This is described in 3GPP specification TS 23.288. This analytics can collect data from other network functions (NFs), application functions (AFs), or operations and maintenance (OAM) and can be exposed to other NFs, OAMs, and even third parties / AFs to provide statistics and forecasts related to, for example, slice load levels, observed service experience, NF load, network performance, user equipment (UE)-related analytics (mobility, communication), user data congestion, quality of service (QoS) sustainability, data network (DN) performance, and more. Furthermore, in 3GPP SA5 (TS 28.104), Management Data Analysis (MDA) provides data analytics for network management. MDA can be deployed at different levels, such as at the network element level (e.g., gNodeB), domain level (e.g., radio access network (RAN), core network (CN), network slice subnet), or in a centralized manner (e.g., at the public land mobile network (PLMN) level). The goal of MDA is to provide root case analysis on complex problems and optimize network resource allocation (e.g., at the network / domain level, at the slice / slice subnet level).

[0003] Additional analysis capabilities in 3GPP are discussed in 3GPP SA6 (TS 23.436), where the Application Data Analysis Enablement Service (ADAES) is defined for performing application layer and edge / cloud analysis outside the 3GPP domain. ADAES can be implemented as an AF with analysis capabilities and also has interfaces to the UE side (to the ADAE client) and to the OAM.

[0004] ADAES supports analytics (e.g., vertical application layer (VAL) server performance, edge load analysis, location analysis, etc.), which can enable machine learning (ML). TS 23.436 supports different deployment and service models for ADAES. For example, ADAES can be deployed within a PLMN, at an edge computing service provider (ECSP) (e.g., AWS), or at a vertical domain (e.g., BMW). Regarding deployment models, there are three possible scenarios: centralized, distributed, and coordinated. Summary of the Invention

[0005] A key scenario in ML-enabled analytics is federated learning support as part of the NWDAF. Specifically, for ML-enabled analytics, TR 23.700-81 is exploring enhancements to support federated learning within 5G systems. The main functions of the federated learning (FL) architecture include: a 'FL consumer' NWDAF, which is either an NWDAF with an analysis logic function (AnLF) or an NWDAF with a model training logic function (MTLF) for a specific analysis ID; a 'FL server' NWDAF, which is an NWDAF with an MTLF that supports the 'FL aggregation' capability for a specific analysis ID; and a 'FL client' NWDAF, which is an NWDAF with an MTLF that supports the 'FL participant' capability for a specific analysis ID and is selected as an FL client by the 'FL server' NWDAF.

[0006] The current assumption is that the FL client is within the core network (NF, NWDAF); however, it would be beneficial to also consider FL clients belonging to different domains or systems / platforms. By way of example, the FL client can be an AF or UE (behind an AF or UE that supports subscriber-aware Northbound API Access (SNA)). As another example, the FL client can be a Service Enabler Architecture Layer (SEAL) / ADAE server (defined in 3GPP SA6) or a vertical application. As an additional example, the FL client can be a Multi-Access Edge Computing (MEC) service or MEC application (e.g., an RNI service can be enhanced to act as a FL client).

[0007] Currently, in 3GPP SA2, the MTLF determines that a FL is required based on multiple conditions. These include: data not being available from the source (due to privacy issues); target area; and pre-configuration of an analysis ID based on the model requested by the AnLF.

[0008] The benefit of having a cross-domain FL client, and specifically an FL client on the application side, can be to offload some of the processing and communication for FL training from the 5GC in high-load scenarios. Furthermore, it can allow FL training to be performed using more refined data (e.g., vertical application layer, edge / cloud data), which can lead to better predictions. Furthermore, it can allow training on different training data or different regions, which can allow aggregation across different environments (which can also help improve analysis performance). Furthermore, it can allow training on data co-located with the data source, which can improve latency and also avoid sending training data to different domains (which may have some restrictions on sharing raw data, such as from a vertical to a mobile network operator (MNO)). The latter is the main motivation, as it is uncertain whether data producers outside the core network will be willing to send raw data to be trained in the MNO domain, and it would be preferable to train the model in the data producer domain.

[0009] Based on these considerations, a set of heterogeneous FL clients, i.e., vertical FL, is expected, and enhancements are needed to determine how the discovery / selection of FL clients occurs and how ML models are distributed.

[0010] This document discloses a procedure for registering and discovering external federated learning clients in a wireless communication system. The procedure can be implemented by an application entity, a server entity, and a service registration entity in the wireless communication system.

[0011] An application entity in a wireless communication system is provided, comprising: a processor; and a memory coupled to the processor, wherein the processor is configured to cause the application entity to: receive a query from at least one other network or application entity of the wireless communication system for acting as a federated learning participant for a specific machine learning-enabled analysis event, wherein the wireless communication system includes a core network; determine a federated learning client profile for the specific machine learning-enabled analysis event, wherein the federated learning client profile includes the application entity's ability to act as a federated learning client for at least one data source external to the core network and / or act as a data source external to the core network for the specific machine learning-enabled analysis event; and transmit a request for registration information to a service registration entity, the registration information including the federated learning client profile and an identifier of the application entity.

[0012] Further provided is a server entity in a wireless communication system, comprising: a processor; and a memory coupled to the processor, the processor being configured to cause the server entity to: determine a federated learning requirement for a specific machine learning-enabled analysis event, wherein the federated learning requirement uses at least one application entity external to a core network of the wireless communication system; and transmit a request for discovery of information of the at least one application entity to a service registration entity based on the federated learning requirement.

[0013] Further provided is a service registration entity in a wireless communication system, comprising: a processor; and a memory coupled to the processor, wherein the processor is configured to enable the service registration entity to: receive a request for registration information from at least one application entity outside a core network of the wireless communication system, the registration information including the identity of the corresponding application entity and a corresponding federated learning client profile, wherein the federated learning client profile includes the ability of the corresponding application entity to act as a federated learning client of at least one data source outside the core network and / or act as a data source outside the core network for specific machine learning-enabled analysis events; and store the registration information in the memory.

[0014] In addition, a method is provided in an application entity, wherein the application entity is in a wireless communication system, the method comprising: receiving a query from at least one other network or application entity of the wireless communication system for acting as a federated learning participant for a specific machine learning-enabled analysis event, wherein the wireless communication system includes a core mobile network; determining a federated learning client profile for the specific machine learning-enabled analysis event, wherein the federated learning client profile includes the ability of the application entity to act as a federated learning client for at least one data source outside the core network and / or act as a data source outside the core network for the specific machine learning-enabled analysis event; and transmitting a request for registration information to a service registration entity, the registration information including the federated learning client profile and an identifier of the application entity.

[0015] Further provided is a method in a server entity, the server entity being in a wireless communication system, the method comprising: determining a federated learning requirement for a specific machine learning-enabled analysis event, wherein the federated learning requirement uses at least one application entity external to a core network of the wireless communication system; and transmitting a request for discovery of information of the at least one application entity to a service registration entity based on the federated learning requirement.

[0016] A method is further provided in a service registration entity, wherein the service registration entity is in a wireless communication system, the method comprising: receiving a request for registration information from at least one application entity outside a core network of the wireless communication system, the registration information comprising an identity of the corresponding application entity and a corresponding federated learning client profile, wherein the federated learning client profile comprises the capability of the corresponding application entity to act as a federated learning client of at least one data source outside the core network and / or as a data source outside the core network for specific machine learning-enabled analysis events; and storing the registration information. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to illustrate the manner in which the advantages and features of the present disclosure can be obtained, the description of the present disclosure is presented by reference to certain apparatus and methods illustrated in the accompanying drawings. Each of these drawings depicts only certain aspects of the present disclosure and, therefore, is not to be considered as limiting its scope. The drawings may have been simplified for clarity and are not necessarily drawn to scale.

[0018] A method and apparatus for registering and discovering external federated learning clients in a wireless communication system will now be described, by way of example only, with reference to the accompanying drawings, in which:

[0019] Figure 1 Describe an embodiment of a wireless communication system;

[0020] Figure 2 Describe an embodiment of a user equipment device;

[0021] Figure 3 Describe an embodiment of a network node or network entity;

[0022] Figure 4 Describe an embodiment of the Advanced SEAL ADAE architecture;

[0023] Figure 5 Describe an embodiment of a coordinated ADAES deployment model;

[0024] Figure 6 Describe an example of federated learning in ML being deployed;

[0025] Figure 7 An embodiment of a method in an application entity in a wireless communication system is described;

[0026] Figure 8 An embodiment of a method in a server entity in a wireless communication system is described;

[0027] Figure 9 An embodiment of a method in a service registry in a wireless communication system is described;

[0028] Figure 10 An embodiment of a method in an exposure function in a wireless communication system is described;

[0029] Figure 11 Describe an embodiment for registration and discovery of a trusted AF via an NRF; and

[0030] Figure 12 An embodiment for registration and discovery of untrusted AF / UE is described. DETAILED DESCRIPTION

[0031] As will be appreciated by those skilled in the art, aspects of the present disclosure may be embodied as systems, devices, methods, or program products. Thus, the arrangements described herein may be implemented in an entirely hardware form, an entirely software form (including firmware, resident software, microcode, etc.), or a combination of software and hardware aspects.

[0032] For example, the disclosed methods and apparatus may be implemented as hardware circuits, including custom very large scale integration ("VLSI") circuits or gate arrays, off-the-shelf semiconductors such as logic chips, transistors, or other discrete components. The disclosed methods and apparatus may also be implemented in programmable hardware devices, such as field programmable gate arrays, programmable array logic, programmable logic devices, or the like. As another example, the disclosed methods and apparatus may comprise one or more physical or logical blocks of executable code, which may, for example, be organized as an object, procedure, or function.

[0033] Furthermore, the methods and apparatus may take the form of a program product embodied in one or more computer-readable storage devices storing machine-readable code, computer-readable code, and / or program code (hereinafter referred to as code). The storage device may be tangible, non-transitory, and / or non-transmitting. The storage device may not embody signals. In some arrangements, the storage device employs signals solely for accessing the code.

[0034] Any combination of one or more computer-readable media may be utilized. The computer-readable medium may be a computer-readable storage medium. The computer-readable storage medium may be a storage device that stores code. The storage device may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, holographic, micromechanical, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing.

[0035] More specific examples of storage devices (a non-exhaustive list) would include the following: an electrical connection having one or more conductors, a portable computer diskette, a hard disk, a random access memory ("RAM"), a read-only memory ("ROM"), an erasable programmable read-only memory ("EPROM" or flash memory), a portable compact disk read-only memory ("CD-ROM"), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer-readable storage medium may be any tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device.

[0036] References throughout this specification to instances of a particular method or apparatus, or similar language, mean that a particular feature, structure, or characteristic described in connection with the implementation of that instance is included in at least one embodiment of the method and apparatus described herein. Thus, references to instances of a particular method or apparatus, or similar language, may, but do not necessarily, refer to the same instance, but rather to "one or more but not all instances," unless expressly specified otherwise. The terms "comprise," "include," "have," and variations thereof mean "including, but not limited to," unless expressly specified otherwise. An enumerated listing of items does not imply that any or all items are mutually exclusive, unless expressly specified otherwise. The terms "a," "an," and "the" also mean "one or more," unless expressly specified otherwise.

[0037] As used herein, a list with the conjunction “and / or” includes any single item in the list or combination of items in the list. For example, a list of A, B, and / or C includes only A, only B, only C, a combination of A and B, a combination of B and C, a combination of A and C, or a combination of A, B, and C. As used herein, a list using the term “one or more of…” includes any single item in the list or combination of items in the list. For example, one or more of A, B, and C includes only A, only B, only C, a combination of A and B, a combination of B and C, a combination of A and C, or a combination of A, B, and C. As used herein, a list using the term “one of…” includes one and only one of any single item in the list. For example, “one of A, B, and C” includes only A, only B, or only C, and excludes the combination of A, B, and C. As used herein, “a member selected from the group consisting of A, B, and C” includes one and only one of A, B, or C, and excludes the combination of A, B, and C. As used herein, "a member selected from the group consisting of A, B, and C, and combinations thereof" includes only A, only B, only C, a combination of A and B, a combination of B and C, a combination of A and C, or a combination of A, B, and C.

[0038] In addition, the described features, structures, or characteristics described herein may be combined in any suitable manner. In the following description, many specific details (e.g., examples of programming, software modules, user selections, network transactions, database queries, database structures, hardware modules, hardware circuits, hardware chips, etc.) are provided to provide a thorough understanding of the present disclosure. However, those skilled in the relevant art will recognize that the disclosed methods and apparatus may be practiced without one or more of the specific details or using other methods, components, materials, etc. In other instances, well-known structures, materials, or operations are not shown or described in detail to avoid obscuring aspects of the present disclosure.

[0039] Aspects of the disclosed methods and devices are described below with reference to schematic flow charts and / or schematic block diagrams of methods, devices, systems, and program products. It should be understood that each block of the schematic flow charts and / or schematic block diagrams, and combinations of blocks in the schematic flow charts and / or schematic block diagrams, can be implemented by code. This code can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device to generate a machine, such that instructions executed by the processor of the computer or other programmable data processing device generate methods for implementing the functions / actions specified in the schematic flow charts and / or schematic block diagrams.

[0040] Code may also be stored in a storage device, which can direct a computer, other encodable data processing device or other device to operate in a specific manner, so that the instructions stored in the storage device produce an article of manufacture containing instructions for implementing the functions / actions specified in the schematic flowchart and / or schematic block diagram.

[0041] The code may also be loaded onto a computer, other programmable data processing device or other apparatus to cause a series of operating steps to be executed on the computer, other programmable device or other apparatus to produce a computer-implemented process, such that the code executed on the computer or other programmable device provides a process for implementing the functions / actions specified in the schematic flowchart and / or schematic block diagram.

[0042] The schematic flowcharts and / or schematic block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of devices, systems, methods, and program products. In this regard, each block in the schematic flowcharts and / or schematic block diagrams may represent a module, segment, or portion of code comprising one or more executable instructions of code for implementing the specified logical function.

[0043] It should also be noted that in some alternative implementations, the functions mentioned in the blocks may not occur in the order mentioned in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending on the functionality involved. Other steps and methods are contemplated that are equivalent in function, logic, or effect to one or more blocks, or portions thereof, of the illustrated figures.

[0044] In each figure, the description of an element may refer to an element of a previous figure. In all figures, the same numbers refer to the same elements.

[0045] Figure 1 An embodiment of a wireless communication system 100 for registering and discovering external federated learning clients is depicted. Such a system 100 may include an application entity within a trusted domain of a mobile network. Such a system 100 may include a core network or a core mobile network. In one embodiment, the wireless communication system 100 includes a remote unit 102 and a network unit 104. Even in Figure 1 Although a particular number of remote units 102 and network units 104 are depicted in FIG, one skilled in the art will recognize that any number of remote units 102 and network units 104 may be included in the wireless communication system 100.

[0046] In one embodiment, the remote unit 102 may comprise a computing device, such as a desktop computer, a laptop computer, a personal digital assistant ("PDA"), a tablet computer, a smartphone, a smart TV (e.g., a TV connected to the Internet), a set-top box, a game console, a security system (including a security camera), an in-vehicle computer, a network device (e.g., a router, a switch, a modem), an aerial vehicle, a drone, or the like. In some embodiments, the remote unit 102 comprises a wearable device, such as a smartwatch, a fitness band, an optical head-mounted display, or the like. Furthermore, the remote unit 102 may be referred to as a subscriber unit, a mobile device, a mobile station, a user, a terminal, a mobile terminal, a fixed terminal, a subscriber station, a UE, a user terminal, a device, or other terms used in the art. The remote unit 102 may communicate directly with one or more of the network units 104 via UL communication signals. In certain embodiments, the remote unit 102 may communicate directly with other remote units 102 via sidelink communication.

[0047] The network elements 104 may be distributed throughout a geographic area. In certain embodiments, the network elements 104 may also be referred to as access points, access terminals, base stations, base stations, node Bs, eNBs, gNBs, home node Bs, relay nodes, devices, core networks, air servers, radio access nodes, APs, NRs, network entities, access and mobility management functions (“AMFs”), unified data management functions (“UDMs”), unified data repositories (“UDRs”), UDM / UDRs, policy control functions (“PCFs”), radio access networks (“RANs”), network slice selection functions (“NSSFs”), operations, administration, and maintenance (“OAMs”), session

[0026] The terms "network element 104" and "network slice capability management function" may be used interchangeably. The network elements 104 may be network management functions ("SMF"), user plane functions ("UPF"), application functions, authentication server functions ("AUSF"), security anchor functions ("SEAF"), trusted non-3GPP gateway functions ("TNGF"), application functions, service enabler architecture layer ("SEAL") functions, vertical application enabler servers, edge enabler servers, edge configuration servers, mobile edge computing platform functions, mobile edge computing applications, application data analytics enabler servers, SEAL data delivery servers, middleware entities, network slice capability management servers, or any other term used in the art. The network elements 104 are typically part of a radio access network that includes one or more controllers communicatively coupled to one or more corresponding network elements 104. The radio access network is typically communicatively coupled to one or more core networks, which may be coupled to other networks, such as the Internet and public switched telephone networks, among other networks. These and other elements of the radio access and core networks are not illustrated but are generally well known to those of ordinary skill in the art.

[0048] In one embodiment, the wireless communication system 100 is compatible with the New Radio (NR) protocol standardized in 3GPP, wherein the network unit 104 transmits on the downlink (DL) using an orthogonal frequency division multiplexing ("OFDM") modulation scheme, and the remote unit 102 transmits on the uplink (UL) using a single carrier frequency division multiple access ("SC-FDMA") scheme or an OFDM scheme. More generally, however, the wireless communication system 100 may implement some other open or proprietary communication protocol, such as WiMAX, IEEE 802.11 variants, GSM, GPRS, UMTS, LTE variants, CDMA2000, ZigBee, Sigfoxx, and other protocols. The present disclosure is not intended to be limited to implementation of any particular wireless communication system architecture or protocol.

[0049] The network unit 104 can serve several remote units 102 within a service area, such as a cell or cell sector, via wireless communication links. The network unit 104 transmits DL communication signals in the time, frequency, and / or spatial domains to serve the remote units 102.

[0050] Figure 2 A user equipment device 200 is depicted that can be used to implement the methods described herein. The user equipment device 200 is used to implement one or more solutions described herein. The user equipment device 200 is one or more of the user equipment devices described in the embodiments herein. Specifically, for example, the user equipment device 200 may include a Figure 1 UE 102, from Figure 6 UE 680, by Figure 12 The user equipment device 200 includes a processor 205 , a memory 210 , an input device 215 , an output device 220 , and a transceiver 225 .

[0051] The input device 215 and the output device 220 are combined into a single device, such as a touch screen. In some implementations, the user equipment apparatus 200 does not include any input device 215 and / or output device 220. The user equipment apparatus 200 may include one or more of the processor 205, the memory 210, and the transceiver 225, and may not include the input device 215 and / or the output device 220.

[0052] As depicted, transceiver 225 includes at least one transmitter 230 and at least one receiver 235. Transceiver 225 can communicate with one or more cells (or wireless coverage areas) supported by one or more base units. Transceiver 225 can operate in unlicensed spectrum. Furthermore, transceiver 225 can include multiple UE panels supporting one or more beamforms. Furthermore, transceiver 225 can support at least one network interface 240 and / or application interface 245. Application interface 245 can support one or more APIs. Network interface 240 can support 3GPP reference points such as Uu, N1, PC5, etc. As will be appreciated by one of ordinary skill in the art, other network interfaces 240 can be supported.

[0053] The processor 205 may include any known controller capable of executing computer-readable instructions and / or capable of performing logical operations. For example, the processor 205 may be a microcontroller, a microprocessor, a central processing unit ("CPU"), a graphics processing unit ("GPU"), an auxiliary processing unit, a field programmable gate array ("FPGA"), or a similar programmable controller. The processor 205 may execute instructions stored in the memory 210 to perform the methods and routines described herein. The processor 205 is communicatively coupled to the memory 210, the input device 215, the output device 220, and the transceiver 225.

[0054] Processor 205 may control user equipment device 200 to implement the user equipment device behaviors described herein. Processor 205 may include an application processor (also referred to as a "main processor") that manages application domain and operating system ("OS") functions and a baseband processor (also referred to as a "baseband radio processor") that manages radio functions.

[0055] Memory 210 may be a computer-readable storage medium. Memory 210 may include volatile computer storage media. For example, memory 210 may include RAM, including dynamic RAM ("DRAM"), synchronous dynamic RAM ("SDRAM"), and / or static RAM ("SRAM"). Memory 210 may include non-volatile computer storage media. For example, memory 210 may include a hard drive, flash memory, or any other suitable non-volatile computer storage device. Memory 210 may include both volatile and non-volatile computer storage media.

[0056] Memory 210 may store data related to implementing the traffic class field, as described herein. Memory 210 may also store program code and related data, such as an operating system or other controller algorithms operating on device 200.

[0057] Input device 215 may include any known computer input device, including a touch panel, buttons, keyboard, stylus, microphone, or the like. Input device 215 may be integrated with output device 220, for example, as a touch screen or similar touch-sensitive display. Input device 215 may include a touch screen, allowing text to be entered using a virtual keyboard displayed on the touch screen and / or by handwriting on the touch screen. Input device 215 may include two or more different devices, such as a keyboard and a touch panel.

[0058] Output device 220 may be designed to output visual, auditory, and / or tactile signals. Output device 220 may include an electronically controllable display or display device capable of outputting visual data to a user. For example, output device 220 may include, but is not limited to, a liquid crystal display ("LCD"), a light emitting diode ("LED") display, an organic LED ("OLED") display, a projector, or a similar display device capable of outputting images, text, or the like to a user. As another non-limiting example, output device 220 may include a wearable display that is separate from but communicatively coupled to the rest of user equipment device 200, such as a smartwatch, smart glasses, a heads-up display, or the like. Furthermore, output device 220 may be a component of a smartphone, a personal digital assistant, a television, a desktop computer, a notebook (laptop) computer, a personal computer, a vehicle dashboard, or the like.

[0059] Output device 220 may include one or more speakers for producing sound. For example, output device 220 may generate an audible alarm or notification (e.g., a beep or chime). Output device 220 may include one or more haptic devices for generating vibration, motion, or other tactile feedback. All or part of output device 220 may be integrated with input device 215. For example, input device 215 and output device 220 may form a touch screen or similar touch-sensitive display. Output device 220 may be positioned near input device 215.

[0060] The transceiver 225 communicates with one or more network functions of the mobile communication network via one or more access networks. The transceiver 225 operates under the control of the processor 205 to transmit messages, data, and other signals, and also receives messages, data, and other signals. For example, the processor 205 may selectively activate the transceiver 225 (or portions thereof) at specific times to send and receive messages.

[0061] The transceiver 225 includes at least one transmitter 230 and at least one receiver 235. The one or more transmitters 230 may be used to provide uplink communication signals to a base unit of the wireless communication network. Similarly, the one or more receivers 235 may be used to receive downlink communication signals from the base unit. Although illustrated with only one transmitter 230 and one receiver 235, the user equipment device 200 may have any suitable number of transmitters 230 and receivers 235. Furthermore, the transmitter 230 and receiver 235 may be any suitable type of transmitter and receiver. The transceiver 225 may include a first transmitter / receiver pair for communicating with a mobile communication network over a licensed radio spectrum and a second transmitter / receiver pair for communicating with a mobile communication network over an unlicensed radio spectrum.

[0062] A first transmitter / receiver pair can be used to communicate with a mobile communication network over the licensed radio spectrum, and a second transmitter / receiver pair can be used to communicate with a mobile communication network over the unlicensed radio spectrum. The first transmitter / receiver pair and the second transmitter / receiver pair can be combined into a single transceiver unit, such as a single chip that performs functions for use with both the licensed and unlicensed radio spectrum. The first transmitter / receiver pair and the second transmitter / receiver pair can share one or more hardware components. For example, some of transceivers 225, transmitter 230, and receiver 235 can be implemented as physically separate components that access shared hardware resources and / or software resources, such as, for example, network interface 240.

[0063] One or more transmitters 230 and / or one or more receivers 235 may be implemented and / or integrated as a single hardware component, such as a multi-transceiver chip, a system on a chip, an application-specific integrated circuit ("ASIC"), or other type of hardware component. One or more transmitters 230 and / or one or more receivers 235 may be implemented and / or integrated as a multi-chip module. Other components, such as a network interface 240 or other hardware components / circuits, may be integrated into a single chip with any number of transmitters 230 and / or receivers 235. Transmitters 230 and receivers 235 may be logically configured as transceivers 225 using one or more common control signals, or as modular transmitters 230 and receivers 235 implemented in the same hardware chip or in a multi-chip module.

[0064] Figure 3 Detailed description of a network node 300 that may be used to implement the methods described herein. The network node 300 may be one embodiment of an entity in a wireless communication network (e.g., in one or more wireless communication networks described herein). For example, the network node 300 may include a Figure 6 NWDAF 610 (including MTLF 601 and FL server 602), from Figure 6 NEF 620, from Figure 11AnLF 1110 or NRF 1120 or NEF 1130 or MTLF 1140 from Figure 12 The network node 300 includes a processor 305 , a memory 310 , an input device 315 , an output device 320 , and a transceiver 325 .

[0065] Input device 315 and output device 320 may be combined into a single device, such as a touch screen. In some implementations, network node 300 does not include any input device 315 and / or output device 320. Network node 300 may include one or more of processor 305, memory 310, and transceiver 325, and may not include input device 315 and / or output device 320.

[0066] As depicted, transceiver 325 includes at least one transmitter 330 and at least one receiver 335. Here, transceiver 325 communicates with one or more remote units 200. In addition, transceiver 325 may support at least one network interface 340 and / or application interface 345. Application interface 345 may support one or more APIs. Network interface 340 may support 3GPP reference points such as Uu, N1, N2, and N3. As will be understood by one of ordinary skill in the art, other network interfaces 340 may be supported.

[0067] The processor 305 may include any known controller capable of executing computer-readable instructions and / or performing logical operations. For example, the processor 305 may be a microcontroller, a microprocessor, a CPU, a GPU, an auxiliary processing unit, an FPGA, or a similar programmable controller. The processor 305 may execute instructions stored in the memory 310 to perform the methods and routines described herein. The processor 305 is communicatively coupled to the memory 310, the input device 315, the output device 320, and the transceiver 325.

[0068] Memory 310 may be a computer-readable storage medium. Memory 310 may include volatile computer storage media. For example, memory 310 may include RAM, including dynamic RAM ("DRAM"), synchronous dynamic RAM ("SDRAM"), and / or static RAM ("SRAM"). Memory 310 may include non-volatile computer storage media. For example, memory 310 may include a hard drive, flash memory, or any other suitable non-volatile computer storage device. Memory 310 may include both volatile and non-volatile computer storage media.

[0069] The memory 310 may store data related to establishing a multipath unicast link and / or mobile operations. For example, the memory 310 may store parameters, configurations, resource assignments, policies, and the like, as described herein. The memory 310 may also store program code and related data, such as an operating system or other controller algorithms operating on the network node 300.

[0070] Input device 315 may include any known computer input device, including a touch panel, buttons, keyboard, stylus, microphone, or the like. Input device 315 may be integrated with output device 320, for example, as a touch screen or similar touch-sensitive display. Input device 315 may include a touch screen, allowing text to be entered using a virtual keyboard displayed on the touch screen and / or by handwriting on the touch screen. Input device 315 may include two or more different devices, such as a keyboard and a touch panel.

[0071] Output device 320 may be designed to output visual, auditory, and / or tactile signals. Output device 320 may include an electronically controllable display or display device capable of outputting visual data to a user. For example, output device 320 may include, but is not limited to, an LCD display, an LED display, an OLED display, a projector, or a similar display device capable of outputting images, text, or the like to a user. As another non-limiting example, output device 320 may include a wearable display that is separate from but communicatively coupled to the rest of network node 300, such as a smartwatch, smart glasses, a heads-up display, or the like. Furthermore, output device 320 may be a component of a smartphone, a personal digital assistant, a television, a desktop computer, a notebook (laptop) computer, a personal computer, a vehicle dashboard, or the like.

[0072] Output device 320 may include one or more speakers for producing sound. For example, output device 320 may generate an audible alarm or notification (e.g., a beep or chime). Output device 320 may include one or more haptic devices for generating vibration, motion, or other tactile feedback. All or part of output device 320 may be integrated with input device 315. For example, input device 315 and output device 320 may form a touch screen or similar touch-sensitive display. Output device 320 may be positioned near input device 315.

[0073] The transceiver 325 includes at least one transmitter 330 and at least one receiver 335. The one or more transmitters 330 can be used to communicate with the UE, as described herein. Similarly, the one or more receivers 335 can be used to communicate with network functions in the PLMN and / or RAN, as described herein. Although only one transmitter 330 and one receiver 335 are illustrated, the network node 300 can have any suitable number of transmitters 330 and receivers 335. Furthermore, the transmitter 330 and receiver 335 can be any suitable type of transmitter and receiver.

[0074] Figure 4 An embodiment 400 of an advanced SEAL (Advanced Automated Application Engineering) (ADAE) architecture as specified in 3GPP TS 23.434 is illustrated. Embodiment 400 includes a VAL UE 410, a 3GPP network system 420, a VAL server 430, and an application data analytics enablement server 440. VAL UE 410 is illustrated as including a VAL client 401 and an application data analytics enablement client 402. Separate VAL layers 450 and SEAL layers 460 are also illustrated in embodiment 400. VAL layer 450 itself includes VAL client 401 and VAL server 430. SEAL layer 460 includes application data analytics enablement client 402 and application data analytics enablement server 440. VAL client 401 is illustrated as communicating with VAL server 430 via 3GPP network system 420 using VAL-UU. VAL client 401 is also shown as communicating with application data analytics enablement client 401 using ADAE-C. Application data analytics enablement client 401 is further shown as communicating with application data analytics enablement server 440 via 3GPP network system 420 using ADAE-UU. Application data analytics enablement server 440 is further shown as communicating with 3GPP network system 420 using N33, N6, and ADAE-OAM. Application data analytics enablement server 440 is further shown as communicating with VAL server 430 using ADAE-S.

[0075] More specifically, embodiment 400 shows a high-level architecture for ADAE services. In this figure, a VAL server 430 communicates with an ADAES 440 via the ADAE-S reference point. ADAES 440, acting as an AF, can communicate with 5G core network functions (via the N33 reference point to the NEF and the N6 reference point to the UPF) and OAM (via the ADAE-OAM interface) in a 3GPP network system 420.

[0076] Figure 5An embodiment 500 of a coordinated ADAES deployment model as specified in 3GPP TS 23.436 is illustrated. Embodiment 500 includes EDN A1 510, EDN A2 520, a centralized DN (DNN-B) 530, and a PLMN 540. EDN A1 510 includes multiple EASs 511, EESs 512, and ADAE Server #1.1 513. EDN A2 520 includes multiple EASs 521, EESs 522, and ADAE Server #1.2 523. Centralized DN 530 includes multiple VAL servers 531 and a SEAL service 532 including ADAE Server #1 533. PLMN 540 includes an ADAE 1.1 service area 541 and an ADAE 1.2 service area 542. EDN A1 510 is shown interfacing with PLMN 540 through DNAIA1-m 550 and DNAIA1-n 560. EDN A2 520 is shown interfacing with PLMN 540 through NDAIA2-n 570. Centralized DN 530 is shown interfacing with PLMN 540 through DNAIB 580.

[0077] More specifically, in a coordinated deployment 500, as described in TS 23.436, multiple ADAES 513, 523, and 533 can be located at different external data networks (EDNs) / data networks (DNs) 510, 520, and 530 and deployed by the same ADAE provider. This coordinated deployment allows for local-global analytics derivation, which may be required to improve analytics confidence levels. The centrally deployed ADAES can also act as an ADAE analytics aggregator and control edge-deployed ADAES to derive analytics for different sub-regions. An example is load usage analytics for EDN #1 510 or EDN #2 520, which will help predict the performance of the VAL server 531 at the centrally located ADAES 533. Alternative deployments are suitable for ML-based analytics methods, such as supervised learning, where the centrally located ADAES 533 acts as the ML model training entity, and the edge-located ADAES 513 and 523 can act as the ML model inference entities (using edge data to improve prediction accuracy).

[0078] The statistics / forecasts of the edge-deployed ADAES 513 and 523 correspond to the ADAES service areas 541 and 541, which are equivalent to the EES / EAS service areas. The central ADAE server 533 covers all PLMN 540 areas and is used to coordinate with the distributed ADAES 513 and 523 (in the case of central ADAES performing aggregation) or jointly perform analysis (in the case of distributed analysis derivation, for example, for ML model training and reasoning in central and decentralized ADAES). Such analysis services can be provided to consumers at the central DN 530, such as the VAL server 531 or the SEAL service 532, or even on the PLMN 540 side (for example, NWDAF consumer service experience analysis).

[0079] The solution provided by this disclosure is to allow external FL clients to collaborate with the core network. An 'external' FL client is considered to be a FL client outside the core network. Such a FL client may, for example, reside at the AF / ADAES / application on the UE side (for the case of vertical domain FL).

[0080] The use case of an AF as a FL client is when the data producer resides on the DN side and the AF is expected to handle local ML model training rather than sending raw data. A specific case is when there is an analytics entity on the DN side, namely ADAES, which has the ability to perform analytics and can serve as a candidate FL client for the corresponding DN (or EDN).

[0081] Another use case is to allow UE-side applications to perform FL as clients. This specific case is for scenarios where local UE data is expected to be provided to the NWDAF for performing ML operations; however, such data may not be exposed to the 5GC from the device side, and using a FL client on the UE side may be preferred (for example, training data related to UE application performance (e.g., channel loss, rate, QoE), or for providing mobility patterns based on UE-side location reports). In such scenarios, the FL client can be an application on the UE or an enabler client on the UE side (e.g., ADAEC) that acts as a middleware application to perform FL for specific analysis events.

[0082] Currently, the NWDAF solution for federated learning explored in 3GPP SA2 assumes that the ML training entity resides in the same 5G core domain under the same management. However, if we assume that the application entity is performing ML training, then the data used locally can solve the problem related to the willingness of third parties to expose raw data to the MNO.

[0083] Figure 6An embodiment 600 of an ML federated learning deployment is illustrated. Embodiment 600 illustrates a NWDAF 610 including a MTLF 601, which includes a FL server 602. Also shown is a NEF 620, through which the NWDAF 610 can communicate with an AF / FL client 630, an ADAES FL client 650, and a device application / FL client 670. The AF / FL client 630 performs model training and data collection with data producers 640 in a vertical domain. The ADAES FL client 650 performs model training and data collection with a data producer VAL server 660. The device application / FL client 670 performs model training and data collection with a data producer UE 680. NWDAF 610 , NEF 620 , AF / FL client 630 , data producers at vertical domain 640 , ADAES FL client 650 , data producer VAL server 660 , device application / FL client 670 are all shown as part of federated learning 690 .

[0084] More specifically, NWDAF 610 includes MTLF 601, and more specifically, FL server 602. FL clients 630, 650, and 670 reside at a DN / third party, which can be trusted (e.g., trusted AF, middleware AF) or untrusted (e.g., UE application), or can also be a third-party analysis entity (e.g., ADAES). In this deployment, a key aspect is how to allow third-party FL clients 630, 650, and 670 to register their capabilities with the 5GC to enable discovery from the FL server 602. The high-level steps for registration / discovery will now be described.

[0085] An external function with AF capabilities (which may be a potential FL client) requests to register its FL client capabilities with the network repository. This can be performed during the AF registration phase or based on a trigger. Such a trigger can be a request for an analytics service from the NWDAF, the instantiation of an application, or a query from the NWDAF for data from an application. The trigger can be at least one of the following: the analytics service is initiated (FL is enabled); an application service that requires the analytics service is activated / starts running; analytics performance is lower than expected and the possibility of FL using a different FL type is indicated by the external function (or at the MTLF), which requires the analytics service to be running and the external function to be a consumer; the NWDAF identifies the need for external data to meet the analytics service performance target and generates a trigger event to query external candidate FL clients; if the target UE belongs to a different PLMN and is roaming to the HPLMN, or if the target UE belongs to the HPLMN and is roaming to the VPLMN, the NWDAF determines that the external FL client requires analytics data.

[0086] The network registry authorizes the request and storage of FL client profile information (in the case of a trusted AF), including the AFID / Application ID and address, as well as capabilities (which can serve as the FL client's analysis ID / Event ID, processing capabilities), exposure limits (whether this needs to be via the NEF or directly, depending on whether the AF is in the trusted domain), supported data types and availability, and the time and region supported by the FL client. If the candidate FL is untrusted, the FL client capabilities need to be mapped at the NEF; and the NEF needs to be registered with the corresponding FL capabilities associated with the client FL, assuming the external FL client is an AF. If the external FL client is a data producer belonging to a different PLMN, the request goes through another type of NEF.

[0087] NWDAF (AnLF) subscribes to the central MTLF for ML model provisioning, and the MTLF determines that FL is needed. It optionally determines the need for external data sources and the possibility of allowing external FL clients to use these sources.

[0088] MTLF performs a discovery request to the network registry, and also includes a preference / flag for discovering external FL clients. This preference / flag indicates whether the external FL client needs to be discovered, and may also indicate whether the FL client to be discovered is allowed to be a trusted source or an untrusted source.

[0089] The registry authorizes the request and lists the available FL clients, including external clients, and sends this as a discovery response to the MTLF.

[0090] For the purposes of this disclosure, references to 'exposed functions' include functions / entities such as NEF and CAPIF AEF; references to 'ML model training server functions' include functions / entities such as NWDAF MTLF; references to 'analysis functions' include functions / entities such as NWDAF AnLF; references to 'external ML model training client functions' include functions / entities such as FL client AF, UE, or ADAES; and references to 'service registry' include functions / entities such as NRF, UDR, CCF, external registry, edge platform registry, application registry, and DN registry. Furthermore, in this mechanism, MTLF is mentioned as the NF, however, this can be generalized so that the ML model training server function can also be within the DN / EDN side (e.g., a central ADAES performing this role).

[0091] The present invention discloses an application entity in a wireless communication system, comprising: a processor; and a memory coupled to the processor, the processor being configured to enable the application entity to: receive a query from at least one other network or application entity of the wireless communication system for acting as a federated learning participant for a specific machine learning-enabled analysis event, wherein the wireless communication system includes a core network; determine a federated learning client profile for the specific machine learning-enabled analysis event, wherein the federated learning client profile includes the ability of the application entity to act as a federated learning client for at least one data source outside the core network and / or act as a data source outside the core network for the specific machine learning-enabled analysis event; and transmit a request for registration information to a service registration entity, the registration information including the federated learning client profile and an identifier of the application entity.

[0092] In some embodiments, the processor is configured to cause the application entity to determine, when determining the federated learning client profile, whether a data source external to the core network is exposeable to at least one other network or application entity for purposes of a particular machine learning-enabled analytics event.

[0093] In some embodiments, the processor is further configured to cause the application entity to transmit an indication to at least one other network or application entity that machine learning using a data source external to the core network may be performed at an external data network external to the core network.

[0094] In some embodiments, a query to act as a federated learning participant includes a request for data on a specific machine learning-enabled analytics event.

[0095] In some embodiments, a query to act as a federated learning participant includes an indication that federated learning is to be used for a particular machine learning-enabled analytics event.

[0096] In some embodiments, the registration information includes an application function identifier; an application identifier; an application data analytics enabler service identifier; a user equipment identifier; identifiers of supported analytics types; a service area, including a geographic and / or topological area; an identifier of a local data source; and / or a preferred time period for federated learning. A 'topological area' corresponds to a network-defined area, such as a cell area or a tracking area or a group of cell areas. For example, a topological area can be identified by a cell ID.

[0097] In some embodiments, the service registration entity is a service registration entity selected from a list of service registries consisting of: a network repository function; a general application programming interface framework core function; a user data repository; an application registry; a data network registry; and an edge or cloud registry.

[0098] In some embodiments, the at least one other network or application entity includes at least one network or application entity selected from the list consisting of: a network exposure function; a data analysis function; and a machine learning model training logic function.

[0099] Figure 7 An embodiment 700 of a method in an application entity is illustrated. A first step 710 comprises receiving a query from at least one other network or application entity of a wireless communication system, comprising a core network, for a federated learning participant to act as for a particular machine learning enabled analytics event.

[0100] Another step 720 includes determining a federated learning client profile for the specific machine learning-enabled analytics event, wherein the federated learning client profile includes the ability of the application entity to act as a federated learning client for at least one data source external to the core network and / or to act as a data source external to the core network for the specific machine learning-enabled analytics event.

[0101] Another step 730 includes transmitting a request to register registration information to the service registration entity, the registration information including the federated learning client profile and an identifier of the application entity.

[0102] In some embodiments, determining the federated learning client profile includes determining that a data source external to the core network is exposeable to at least one other network or application entity for purposes of a particular machine learning-enabled analytics event.

[0103] In certain embodiments, method 700 may be performed by a processor that executes program code, such as a microcontroller, microprocessor, CPU, GPU, auxiliary processing unit, FPGA, or the like.

[0104] Some embodiments further include transmitting an indication to at least one other network or application entity that machine learning using a data source external to the core network may be performed at an external data network external to the core network.

[0105] In some embodiments, a query to act as a federated learning participant includes a request for data on a specific machine learning-enabled analytics event.

[0106] In some embodiments, a query to act as a federated learning participant includes an indication that federated learning is to be used for a particular machine learning-enabled analytics event.

[0107] In some embodiments, the registration information includes an application function identifier; an application identifier; an application data analytics enabler service identifier; a user equipment identifier; identifiers of supported analytics types; a service area, including a geographic and / or topological area; an identifier of a local data source; and / or a preferred time period for federated learning. A 'topological area' corresponds to a network-defined area, such as a cell area or a tracking area or a group of cell areas. For example, a topological area can be identified by a cell ID.

[0108] In some embodiments, the service registration entity is a service registration entity selected from a list of service registries consisting of: a network repository function; a general application programming interface framework core function; a user data repository; an application registry; a data network registry; and an edge or cloud registry.

[0109] In some embodiments, the at least one other network or application entity includes at least one network or application entity selected from the list consisting of: a network exposure function; a data analysis function; and a machine learning model training logic function.

[0110] The present invention also discloses a server entity in a wireless communication system, comprising: a processor; and a memory coupled to the processor, wherein the processor is configured to enable the server entity to: determine a federated learning requirement for a specific machine learning-enabled analysis event, wherein the federated learning requirement uses at least one application entity outside a core network of the wireless communication system; and transmit a request for discovering information of the at least one application entity to a service registration entity based on the federated learning requirement.

[0111] In some embodiments, the processor is configured to cause the server entity to query at least one application entity to act as a federated learning participant for a particular machine learning-enabled analytics event.

[0112] In some embodiments, a query to act as a federated learning participant includes a request for data on a specific machine learning-enabled analytics event.

[0113] In some embodiments, a query to act as a federated learning participant includes an indication that federated learning is to be used for a particular machine learning-enabled analytics event.

[0114] In some embodiments, the processor is configured to cause the server entity to receive a request for a trained machine learning model associated with a particular machine learning-enabled analytics event from the network entity when determining a federated learning requirement; and determine the federated learning requirement based at least in part on the request for the trained machine learning model.

[0115] In some embodiments, the request for a trained machine learning model includes an identifier of a specific machine learning-enabled analytics event; and the machine learning model region of interest.

[0116] In some embodiments, the request for discovery information includes an identifier of a specific machine learning-enabled analytics event; an indication of federated learning; a time period of interest for federated learning; and one or more federated learning clients for discovering at least one data source external to the core network for the specific machine learning-enabled analytics event and / or an indication of discovering at least one data source external to the core network. In some embodiments, the service registration entity is a service registration entity selected from a list of service registries consisting of: a network repository function; a general application programming interface framework core function; a user data repository; an application registry; a data network registry; and an edge or cloud registry.

[0117] In some embodiments, the server entity is selected from a list of server entities consisting of: network exposure functionality; data analysis functionality; and machine learning model training logic functionality.

[0118] In some embodiments, requests / responses / communications to / from untrusted sources may be sent via the NEF.

[0119] Figure 8 An embodiment 800 of a method in a server entity within a wireless communication system is described. A first step 810 includes determining a federated learning requirement for a specific machine learning-enabled analytics event, wherein the federated learning requirement utilizes at least one application entity external to a core network of the wireless communication system. Another step 820 includes transmitting a request to a service registration entity for discovery of information about the at least one application entity based on the federated learning requirement.

[0120] In certain embodiments, method 800 may be performed by a processor that executes program code, such as a microcontroller, microprocessor, CPU, GPU, auxiliary processing unit, FPGA, or the like.

[0121] Some embodiments further include querying at least one application entity to act as a federated learning participant for a particular machine learning-enabled analytics event.

[0122] In some embodiments, a query to act as a federated learning participant includes a request for data for a particular machine learning-enabled analytics event.

[0123] In some embodiments, a query to act as a federated learning participant includes an indication that federated learning is to be used for a particular machine learning-enabled analytics event.

[0124] In some embodiments, determining the federated learning requirement includes receiving a request from a network entity for a trained machine learning model associated with a particular machine learning-enabled analytics event; and determining the federated learning requirement based at least in part on the request for the trained machine learning model.

[0125] In some embodiments, the request for a trained machine learning model includes an identifier of a specific machine learning-enabled analytics event; and the machine learning model region of interest.

[0126] In some embodiments, the request for discovery information includes an identifier of a specific machine learning-enabled analytics event; an indication of federated learning; a time period of interest for federated learning; and one or more federated learning clients for discovering at least one data source external to the core network for the specific machine learning-enabled analytics event and / or an indication of discovering at least one data source external to the core network. In some embodiments, the service registration entity is a service registration entity selected from a list of service registries consisting of: a network repository function; a general application programming interface framework core function; a user data repository; an application registry; a data network registry; and an edge or cloud registry.

[0127] In some embodiments, the server entity is selected from a list of server entities consisting of: network exposure functionality; data analysis functionality; and machine learning model training logic functionality.

[0128] In some embodiments, requests / responses / communications to / from an untrusted source may include sending and / or receiving the requests / responses / communications via a NEF.

[0129] The present invention also discloses a service registration entity in a wireless communication system, which includes: a processor; and a memory coupled to the processor, wherein the processor is configured to enable the service registration entity to: receive a request for registration information from at least one application entity outside the core network of the wireless communication system, the registration information including the identity of the corresponding application entity and the corresponding federated learning client profile, wherein the federated learning client profile includes the ability of the corresponding application entity to act as a federated learning client of at least one data source outside the core network and / or act as a data source outside the core network for specific machine learning-enabled analysis events; and store the registration information in the memory.

[0130] In some embodiments, the processor is further configured to cause the service registration entity to: receive a request from the server entity for information on discovering a federated learning client that acts as at least one data source external to the core network and / or at least one application entity that acts as a data source external to the core network for a specific machine learning-enabled analytics event; retrieve discovery information from the memory, wherein the discovery information includes at least one of the stored registration information; and transmit the discovery information to the server entity.

[0131] In some embodiments, the processor is further configured to cause the service registration entity to authorize the request for discovery information.

[0132] In some embodiments, the registration information includes an application function identifier; an application identifier; an application data analytics enabler service identifier; a user equipment identifier; identifiers of supported analytics types; a service area, including a geographic and / or topological area; an identifier of a local data source; and / or a preferred time period for federated learning. A 'topological area' corresponds to a network-defined area, such as a cell area or a tracking area or a group of cell areas. For example, a topological area can be identified by a cell ID.

[0133] In some embodiments, the service registration entity is a service registration entity selected from a list of service registries consisting of: a network repository function; a general application programming interface framework core function; a user data repository; an application registry; a data network registry; and an edge or cloud registry.

[0134] In some embodiments, the service registry is a NEF. In such embodiments, the NEF must support the mapping of applications / entities to federated learning capabilities and their registration.

[0135] In some embodiments, the processor is further configured to cause the service registration entity to store, for each registration information, in the memory, an indication that the registration information is associated with an application entity external to the core network of the wireless communication system.

[0136] In some embodiments, the processor is configured to cause the service registration entity to transmit a response to the request to register the registration information to the at least one application entity.

[0137] In some embodiments, the NRF may point to an application registry, and the MTLF may send discovery requests directly to the application registry.

[0138] Figure 9 An embodiment 900 of a method in a service registry is described. A first step 910 includes receiving a request for registration information from at least one application entity external to a core network of a wireless communication system, the registration information including an identity of the corresponding application entity and a corresponding federated learning client profile, wherein the federated learning client profile includes the ability of the corresponding application entity to act as a federated learning client for at least one data source external to the core network and / or to act as a data source external to the core network for a specific machine learning-enabled analytics event. Another step 920 includes storing the registration information.

[0139] In certain embodiments, method 900 may be performed by a processor that executes program code, such as a microcontroller, microprocessor, CPU, GPU, auxiliary processing unit, FPGA, or the like.

[0140] Some embodiments include: receiving a request for discovery information of at least one application entity from a server entity, wherein the at least one application entity acts as a federated learning client of at least one data source external to a core network and / or acts as a data source external to the core network for a specific machine learning-enabled analytical event; extracting discovery information from a memory, wherein the discovery information includes at least one of the stored registration information; and transmitting the discovery information to the server entity.

[0141] Some embodiments include authorizing requests for discovery information.

[0142] In some embodiments, the registration information includes an application function identifier; an application identifier; an application data analytics enabler service identifier; a user equipment identifier; identifiers of supported analytics types; a service area, including a geographic and / or topological area; an identifier of a local data source; and / or a preferred time period for federated learning. A 'topological area' corresponds to a network-defined area, such as a cell area or a tracking area or a group of cell areas. For example, a topological area can be identified by a cell ID.

[0143] In some embodiments, the service registration entity is a service registration entity selected from a list of service registries consisting of: a network repository function; a general application programming interface framework core function; a user data repository; an application registry; a data network registry; and an edge or cloud registry.

[0144] In some embodiments, the service registry is a NEF. In such embodiments, the NEF must support the mapping of applications / entities to federated learning capabilities and their registration.

[0145] Some embodiments further comprise storing, for each registration message, an indication that the registration message relates to an application entity external to a core network of the wireless communication system.

[0146] Some embodiments include transmitting a response to the request to register registration information to at least one application entity.

[0147] In some embodiments, the NRF may point to an application registry, and the MTLF may send discovery requests directly to the application registry.

[0148] The present invention also discloses an exposure entity in a wireless communication system, comprising a memory and a processor coupled to the memory, wherein the processor is configured to cause the exposure entity to: map an identifier of an application entity to a federated learning client profile associated with the application entity, wherein the federated learning client profile includes the ability of the application entity to act as a federated learning client of at least one data source outside a core network of the wireless communication network and / or act as a data source outside the core network for specific machine learning-enabled analysis events.

[0149] In some embodiments, the processor is further configured to cause the exposure entity to receive a request from an application entity that acts as a federated learning client for at least one data source external to the core network and / or acts as a data source external to the core network for a specific machine learning-enabled analytics event; convert the request to the registration application entity based on the mapping; and send the converted registration request to the service registration entity.

[0150] Figure 10 An embodiment of a method in exposing functionality is described 1000. A first step 1010 includes mapping an identifier of an application entity to a federated learning client profile associated with the application entity, wherein the federated learning client profile includes the application entity's ability to act as a federated learning client for at least one data source external to a core network of a wireless communication network and / or to act as a data source external to the core network for a specific machine learning-enabled analytics event.

[0151] Some embodiments include: receiving a request to a federated learning client that acts as at least one data source external to a core network and / or an application entity that acts as a data source external to a core network for a specific machine learning-enabled analytical event; transforming the request to register the application entity based on a mapping; and sending the transformed registration request to a service registration entity.

[0152] Figure 11 An embodiment 1100 for registration and discovery of a trusted AF via an NRF is illustrated. This embodiment is for the case when the candidate FL client is a trusted AF. The figure illustrates AnLF 1110, NRF 1120, NEF 1130, MTLF FL server 1140, AF candidate FL client 1150, and data producer 1160.

[0153] In a first step 1101, the NWDAF MTLF 1140 attempts to subscribe candidate clients 1150 to a specific analysis event or ML model identity, or optionally queries a trusted AF (for a given application) to determine whether the data is available for ML-enabled analysis. The NWDAF MTLF 1140 also optionally indicates that ML is available. This can be accomplished in two ways: by calling anNaf-specific API or by calling an Nmtlf API to be defined in the eNA.

[0154] In the following step 1102, the trusted AF 1150 checks the data producer application 1160 and its policies / permissions to see if it is feasible to expose the data to the 5GC for ML-enabled analysis. Optionally, the AF 1150 can respond to the NWDAF MTLF 1140 that such exposure is not feasible, but the data can be trained locally on the DN side.

[0155] In the following step 1103, a trusted AF 1150 (e.g., AF, ADAES) with FL capabilities registers its FL client profile with the repository (NRF 1120). This may include the AF ID, App ID, or ADAES ID, supported analysis IDs, service area, data source ID list, and preferred time period for FL. This is referred to as an AF FL client registration request.

[0156] In the following step 1104 , the NRF 1120 stores the FL client profile and indicates that the FL client is an external entity.

[0157] In the following step 1105, the NRF 1120 also sends a registration response to the function acting as FL client 1150. This is illustrated as a registration response.

[0158] In another step 1106, the AnLF 1110 subscribes to the trained ML model associated with the analysis ID by calling the Nnwdaf_MLModelProvision(analysis ID, ML model filter (S-NSSAI, region of interest)) service operation. This is illustrated as Subscribe / MLModelProvision.

[0159] Since the MTLF 1140 with the FL server lacks data on the region of interest requested by the AnLF 1110 , it determines that the training should be based on the vertical FL client that has data on the region of interest and can be used for vertical federated learning via the NRF 1120 .

[0160] In another step 1107, the NWDAF / MTLF 1140 including the FL server detects that a FL client is needed, and in particular a FL client for unavailable data (eg, third-party data). This detection may be based on analyzing performance (eg, if this does not meet the requirements).

[0161] In another step 1108, the NWDAF / MTLF 1140, including the FL server, sends a discovery request to the NRF 1120 to discover available FL clients for the analysis ID, including external clients. This is done by calling the Nnrf_NFDiscovery_Request (analysis ID, vertical federated learning indication, service area, time period of interest, external FL client indication) service operation. This is illustrated as a discovery request (FL client type).

[0162] In other steps 1109a and 1109b, the NRF 1120 authorizes the discovery request and notifies the MTLF 1140 with the FL server of information about a plurality of external FL clients of different types (including AFs, UE applications) and information on how to reach these clients (e.g., via the NEF 1130, API information, etc.). This is described as a discovery response (external FL client information).

[0163] Figure 12 An embodiment 1200 for registration and discovery of an untrusted AF / UE is illustrated. This embodiment is for the case when the candidate FL client is an untrusted AF. The diagram illustrates AnLF 1220, NRF 1230, MTLF FL server 1240, NEF 1250, CCF / Application Registry / UDR 1260, untrusted AF candidate FL client 1270, and data producer application 1280.

[0164] In a first step 1201, the NWDAF MTLF 1240 attempts to subscribe to a candidate FL client 1270 to obtain data from an external application for a specific analysis event, for a data collection event, or for a specific ML model identity / profile. The request is provided to the NEF 1250 (illustrated as 1201a triggering / subscribing to data), and the NEF 1250 further queries the corresponding untrusted AF 1270 (illustrated as 1201b querying the AF for data or acting as a FL client for a given application) or the application at the UE (if using SNA) whether the data is available for ML-enabled analysis. The NWDAF MTLF 1240 also optionally indicates that FL is feasible.

[0165] In another step 1202, the AF / UE 1270 checks the data producer application 1280 and its policies / permissions to see if it is feasible to expose the data to the 5GC for ML-enabled analysis. Optionally, the AF / UE 1270 may respond to the NWDAF MTLF 1240 (via the NEF 1250) that such exposure is not feasible, but the data can be trained locally on the DN side.

[0166] In another step 1203, the FL-capable AF / UE 1270 registers its FL client profile with the repository 1260 (CAPIF CCF or edge / cloud registry). This may include the AF ID or AppID, supported analytics IDs, service area, list of data source IDs, and preferred time period for FL. In the case of a UE acting as a FL client, this includes the UE ID and capabilities. This is illustrated as an AF FL client registration request.

[0167] In another step 1204 , the application registry 1260 (CAPIF CCF or edge / cloud registry) stores the FL client profile and indicates that the FL client is an external entity.

[0168] In another step 1205, the repository 1260 also sends a registration response to the function acting as FL client 1270. This is illustrated as a registration response.

[0169] In another step 1206, the application repository 1260 may also provide some indication of possible FL clients other than 5GC to the NRF 1230 or UDR. This may be in the form of a registry ID / address and supported FL client types to allow the NWDAF MTLF 1240 to discover additional FL clients.

[0170] In another step 1207, the AnLF 1220 subscribes to the trained ML model associated with the analysis ID by calling the Nnwdaf_MLModelProvision(analysis ID, ML model filter (S-NSSAI, region of interest)) service operation. This is illustrated as Subscribe / MLModelProvision.

[0171] Alternatively, step 1206 can occur directly from the untrusted AF 1270 to the NRF 1230 via the NEF 1250. In that case, the NRF 1230 acts as a service registry. Therefore, step 1206 will replace steps 1203 to 1205 (step 1206 will be in the form of registration with the NRF 1230 via the NEF 1250). In this case, the NEF 1250 capability needs to be enhanced to have a mapping of the AF service ID or API caller ID to the FL client capability; therefore, the NEF 120 will register the capability with the NRF 1230 on behalf of the untrusted AF 1270.

[0172] In another step 1208, since the MTLF 1240 with the FL server lacks data on the area of interest requested by the AnLF 1220, it determines that training based on vertical federated learning is required, and is about to discover an FL client that has data on the area of interest and can be used for vertical federated learning via the NRF 1230 or via the application registry 1260.

[0173] The NWDAF / MTLF 1240 including the FL server detects that more FL clients are needed, and in particular FL clients for unavailable data (eg, third-party data). This detection may be based on analyzing performance (eg, if this does not meet the requirements).

[0174] In another step 1209, the NWDAF / MTLF 1240, including the FL server, sends a discovery request to the NRF 1230 to discover available FL clients for the analysis ID, including external clients. This is done by calling the Nnrf_NFDiscovery_Request (analysis ID, vertical federated learning indication, service area, time period of interest, external FL client indication) service operation. This is illustrated as a discovery request (FL client type).

[0175] In further steps 1210a and 1210b, the NRF 1230 authorizes the discovery request and notifies the MTLF 1240, which has a FL server, of information about a plurality of external FL clients of different types (including AFs and UE applications) and how to reach these clients (e.g., via the NEF 1250, API information, etc.). This may also include registration information to allow discovery of specific capabilities of the external FL clients. This is illustrated as a discovery response (external FL client information, registration information).

[0176] Different options are possible for steps 1207 to 1210. For example, NRF 1230, after receiving the FL client information from application registry 1260 (CCF, edge repository, edge UDR), provides all necessary information to MTLF 1240 via NEF 1250. Alternatively, NRF 1230 points to application registry 1260, and MTLF 1240 sends a discovery request to application registry 1260 via NEF 1250 (if application registry 1260 is not a UDR or is untrusted).

[0177] When a need for an external candidate FL client is detected, the MTLF 1240 may access the external registry 1260 (or this may be done via mapping by the NEF 1250).

[0178] The NRF 1230 may be allowed to act as the sole registry for untrusted AFs, and this requires the NEF 1250 to support mapping to FL capabilities and registration with the NRF 1230.

[0179] The problem addressed in this paper is how to enable application functions to participate in federated learning for ML-enabled NWDAF analysis. This participation allows application service providers and data producers to perform ML model training on the DN or UE side without exposing their data.

[0180] The present disclosure provides the required procedures for allowing an application to register its ability to act as a FL client and for allowing the NWDAF to discover external AF functions that act as FL clients for a given analytics event.

[0181] Other solutions do not assume that the FL client is outside the domain of the analysis function; therefore, registration and discovery do not consider the exposure aspects and the interactions between the network and DN / UE side.

[0182] Specifically, this disclosure describes how a trusted AF registers its capabilities with an NRF, and discovery of the NWDAF occurs via querying the NRF. Furthermore, this disclosure describes how an untrusted AF / application at a UE registers its capabilities, and discovery of the NWDAF occurs via querying the NEF / CAPIF / application registry.

[0183] More specifically, a method is provided for enabling an external application (which is itself associated with one or more data sources) to provide a federated learning client service (e.g., at an AF or external ML model training client function), the method comprising: receiving a trigger event for acting as a FL participant for an analysis event (the participant may be a data source or a FL client), wherein the trigger event is provided by at least one network or application entity; determining the capabilities of the application data source to act as a FL client for the analysis event; and registering with a service registry, wherein the registration includes providing its identification and capabilities as a FL client.

[0184] Further provided is a method for supporting a federated learning client service by an AF (e.g., at an ML model training server function), the method comprising: querying at least one AF to act as a FL participant (the participant may be a data source or a FL client) for an analysis event, wherein the query may include triggering the generation of an event; determining requirements for the FL using the at least one AF for analyzing the event; and discovering information about the at least one AF based on the determined requirements.

[0185] A method for supporting a federated learning client service by an AF (e.g., at a service registry) is further provided, the method comprising: receiving registration information from at least one AF, wherein the registration information includes application capabilities of acting as a FL client; storing the received registration information; receiving a request to discover registration information related to at least one AF acting as a FL client; extracting discovery information related to at least one AF acting as a FL client; and sending the extracted discovery information.

[0186] Further provided is a method for supporting registration and discovery of federated learning client services (e.g., at an exposed function), the method comprising: mapping an application identifier to at least one FL client capability; receiving a registration request from an external FL client; translating the request to register the external FL client based on the capability mapping; and sending the translated registration request to a service registry.

[0187] It should be noted that the above methods and apparatus illustrate rather than limit the present invention, and that those skilled in the art will be able to design many alternative arrangements without departing from the scope of the appended claims. The word "comprising" does not exclude the presence of elements or steps other than those listed in a claim, and "a" or "an" does not exclude a plurality, and a single processor or other unit may perform the functions of several units recited in a claim. Any reference signs in the claims should not be construed as limiting their scope.

[0188] Furthermore, although examples have been given in the context of specific communication standards, these examples are not intended to limit the communication standards to which the disclosed methods and apparatus may be applied. For example, although specific examples have been given in the context of 3GPP, the principles disclosed herein may also be applied to other wireless communication systems, and indeed to any communication system that uses routing rules.

[0189] The method may also be embodied in a set of instructions stored on a computer readable medium, which, when loaded into a computer processor, digital signal processor (DSP), or the like, causes the processor to carry out the aforementioned method.

[0190] The described methods and apparatus may be practiced in other specific forms. The described methods and apparatus should be considered in all respects as illustrative only and not restrictive. Therefore, the scope of the present invention is indicated by the appended claims rather than by the foregoing description. All changes that come within the meaning and range of equivalence of the claims are intended to be included within their scope.

[0191] The following abbreviations are relevant to the areas covered by this document: AF, application function; NF, network function; NWDAF, network data analysis function; OAM, operations and maintenance; UE, user equipment; MDAS, management domain analysis service; ADAES, application data analysis enabler service / server; ANLF, analysis logic function; MTLF, model training logic function; DNAI, data network access identifier; MOS, mean opinion score; MDT, minimized drive test; ADAEC, application data analysis enabler client; TRLF, trust rating logic function; and ML, machine learning.

Claims

1. An application entity in a wireless communication system, comprising: processor; and a memory coupled to the processor, the processor being configured to cause the application entity to: receiving a query from at least one other network or application entity of a wireless communication system to serve as a federated learning participant for a particular machine learning-enabled analytics event, wherein the wireless communication system includes a core network; determining a federated learning client profile for the specific machine learning-enabled analytics event, wherein the federated learning client profile includes the ability of the application entity to act as a federated learning client for at least one data source external to the core network and / or to act as a data source external to the core network for the specific machine learning-enabled analytics event; and A request for registering registration information is transmitted to a service registration entity, the registration information including the federated learning client configuration file and an identifier of the application entity.

2. The application entity according to claim 1, wherein the processor is configured to cause the application entity, when determining the federated learning client configuration file: Determining whether a data source external to the core network can be exposed to the at least one other network or application entity for purposes of the particular machine learning-enabled analytics event.

3. The application entity according to claim 2, wherein the processor is further configured to cause the application entity to: An indication is transmitted to the at least one other network or application entity that machine learning using the data source external to the core network can be performed at an external data network external to the core network.

4. The application entity according to any of the preceding claims, wherein the query to act as a federated learning participant comprises a request for data of the specific machine learning-enabled analytical event.

5. An application entity according to any of the preceding claims, wherein the query to act as a federated learning participant includes an indication that federated learning is to be used for the particular machine learning-enabled analytics event.

6. The application entity according to any one of the preceding claims, wherein the registration information comprises: Application function identifier; Application identifier; App Data Analytics Enabler service identifier; User equipment identifier; Identifiers of supported analysis types; Service area, including geographic and / or topological area; Identifier of the local data source; and / or The preferred time period for federated learning.

7. The application entity according to any one of the preceding claims, wherein the service registration entity is a service registration entity selected from a service registration table consisting of: Network repository functionality; Common application programming interface framework core functions; User data repository; Application registry; Data Network Registry; and Edge or cloud registry.

8. The application entity according to claim 1 , wherein the at least one other network or application entity comprises at least one network or application entity selected from the list consisting of: Network exposure capabilities; Data analysis capabilities; and Machine learning model training logic functions.

9. A server entity in a wireless communication system, comprising processor; and a memory coupled to the processor, the processor being configured to cause the server entity to: determining a federated learning requirement for a particular machine learning-enabled analytics event, wherein the federated learning requirement utilizes at least one application entity external to a core network of the wireless communication system; and A request for discovering information of the at least one application entity is transmitted to a service registration entity based on the federated learning requirement.

10. The server entity according to claim 9, wherein the processor is configured to cause the server entity to: The at least one application entity is queried to serve as a federated learning participant for the specific machine learning-enabled analytics event.

11. The server entity of claim 10, wherein the query to act as a federated learning participant comprises a request for data for the specific machine learning-enabled analytics event.

12. The server entity of any of claims 10-11, wherein the query to act as a federated learning participant includes an indication that federated learning is to be used for the particular machine learning-enabled analytics event.

13. The server entity of any one of claims 9 to 12, wherein the processor is configured to cause the server entity, when determining a federated learning requirement: receiving, from a network entity, a request for a trained machine learning model associated with the particular machine learning-enabled analytics event; and The federated learning requirement is determined based at least in part on the request for the trained machine learning model.

14. The server entity of claim 13, wherein the request for a trained machine learning model comprises: an identifier for the specific machine learning-enabled analytics event; and Areas of interest for machine learning models.

15. The server entity according to claim 14, wherein the request for discovery information comprises: said identifier of said particular machine learning-enabled analytics event; Instructions for federated learning; The time period of interest for federated learning; One or more federated learning clients for discovering at least one data source external to the core network and / or instructions for discovering at least one data source external to the core network for the specific machine learning-enabled analytics event.

16. The server entity according to any one of claims 9 to 15, wherein the service registration entity is a service registration entity selected from a service registration table consisting of: Network repository functionality; Common application programming interface framework core functions; User data repository; Application registry; Data Network Registry; and Edge or cloud registry.

17. The server entity according to any one of claims 9 to 16, wherein the server entity is selected from the list of server entities consisting of: Network exposure capabilities; Data analysis capabilities; and Machine learning model training logic functions.

18. A service registration entity in a wireless communication system, comprising: processor; and a memory coupled to the processor, the processor being configured to cause the service registration entity to: receiving, from at least one application entity external to a core network of the wireless communication system, a request for registration information, the registration information including an identity of the corresponding application entity and a corresponding federated learning client profile, wherein the federated learning client profile includes a capability of the corresponding application entity to act as a federated learning client for at least one data source external to the core network and / or to act as a data source external to the core network for a specific machine learning-enabled analytics event; and The registration information is stored in the memory.

19. The service registration entity according to claim 18, wherein the processor is further configured to cause the service registration entity to: receiving, from a server entity, a request for information to discover a federated learning client acting as at least one data source external to the core network and / or at least one application entity acting as a data source external to the core network for a specific machine learning-enabled analytics event; extracting discovery information from the memory, wherein the discovery information includes at least one of the stored registration information; and The discovery information is transmitted to the server entity.

20. The service registration entity according to claim 19, wherein the processor is further configured to cause the service registration entity to: The request for discovery information is authorized.

21. The service registration entity according to any one of claims 18 to 20, wherein the registration information comprises: Application function identifier; Application identifier; App Data Analytics Enabler service identifier; User equipment identifier; Identifiers of supported analysis types; Service area, including geographic and / or topological area; Identifier of the local data source; and / or The preferred time period for federated learning.

22. The service registration entity according to any one of claims 18 to 21, wherein the service registration entity is a service registration entity selected from a service registration table consisting of: Network repository functionality; Common application programming interface framework core functions; User data repository; Application registry; Data Network Registry; and Edge or cloud registry.

23. The service registration entity according to any one of claims 18 to 22, wherein the processor is further configured to cause the service registration entity to: An indication that each registration information is associated with an application entity external to the core network of the wireless communication system is stored in the memory for each registration information.

24. The service registration entity according to any one of claims 18 to 23, wherein the processor is configured to cause the service registration entity to: A response to the request to register the registration information is transmitted to the at least one application entity.

25. A method in an application entity, wherein the application entity is in a wireless communication system, the method comprising: receiving a query from at least one other network or application entity of the wireless communication system to serve as a federated learning participant for a particular machine learning-enabled analytics event, wherein the wireless communication system includes a core mobile network; determining a federated learning client profile for the specific machine learning-enabled analytics event, wherein the federated learning client profile includes the ability of the application entity to act as a federated learning client for at least one data source external to the core network and / or to act as a data source external to the core network for the specific machine learning-enabled analytics event; and A request for registering registration information is transmitted to a service registration entity, the registration information including the federated learning client configuration file and an identifier of the application entity.

26. The method of claim 25, wherein determining a federated learning client profile for the specific machine learning-enabled analytics event comprises: Determining whether a data source external to the core network can be exposed to the at least one other network or application entity for purposes of the particular machine learning-enabled analytics event.

27. The method of claim 26, further comprising: An indication is transmitted to the at least one other network or application entity that machine learning using the data source external to the core network can be performed at an external data network external to the core network.

28. The method of any one of claims 25-27, wherein the query to act as a federated learning participant comprises a request for data for the specific machine learning-enabled analytics event.

29. The method of any one of claims 25 to 28, wherein the query to serve as a federated learning participant includes an indication that federated learning is to be used for the particular machine learning-enabled analytics event.

30. The method according to any one of claims 25 to 29, wherein the registration information comprises: Application function identifier; Application identifier; App Data Analytics Enabler service identifier; User equipment identifier; Identifiers of supported analysis types; Service area, including geographic and / or topological area; Identifier of the local data source; and / or The preferred time period for federated learning.

31. The method according to any one of claims 25 to 30, wherein the service registration entity is a service registration entity selected from a service registration table consisting of: Network repository functionality; Common application programming interface framework core functions; User data repository; Application registry; Data Network Registry; and Edge or cloud registry.

32. The method according to any one of claims 25 to 31, wherein the at least one other network or application entity comprises at least one network or application entity selected from the list consisting of: Network exposure capabilities; Data analysis capabilities; and Machine learning model training logic functions.

Images