Methods, apparatus, and systems for selecting federated learning members

Abstract

This disclosure relates to methods, apparatus, systems, and computer-storable media for selecting federated learning members, and pertains to the field of communication technology. The method for selecting federated learning members includes: receiving an analysis subscription or request from a Network Data Analysis Function (NWDAF) consumer, the analysis subscription or request including a current analysis identifier; determining current federated learning requirement information corresponding to the current analysis identifier, the current federated learning requirement information including at least one of a list of algorithms or models for the current federated learning, an encryption method for the current federated learning, and a list of expected times for the current federated learning, as well as a list of Regions of Interest (AOIs); and selecting a target edge NWDAF network element from multiple edge NWDAF network elements as a federated learning member corresponding to the current analysis identifier, based on the current federated learning requirement information. According to this disclosure, the accuracy of selecting federated learning members can be improved, thereby improving the accuracy of related analysis.

Description

Technical Field

[0001] This disclosure relates to the field of communication technology, and in particular to methods, apparatus and systems for selecting federated learning members, and computer-storable media. Background Technology

[0002] In related technologies, based on existing 3GPP specifications, the central NWDAF (Network Data Analytics Function) network element selects federated learning members corresponding to the current analysis identifier through the registration information of existing edge NWDAF network elements. Summary of the Invention

[0003] In related technologies, the registration information in the existing 3GPP specifications does not involve information related to federated learning, which makes it impossible to select federated learning members more accurately, thus reducing the accuracy of related analysis.

[0004] To address the aforementioned technical problems, this disclosure proposes a solution that can improve the accuracy of selecting federated learning members, thereby improving the accuracy of related analysis.

[0005] According to a first aspect of this disclosure, a method for selecting federated learning members is provided, comprising: receiving an analysis subscription or request from a Network Data Analysis Function (NWDAF) consumer, the analysis subscription or request including a current analysis identifier; determining current federated learning requirement information corresponding to the current analysis identifier, the current federated learning requirement information including at least one of a list of algorithms or models for the current federated learning, an encryption method for the current federated learning, and a list of expected times for the current federated learning, as well as a list of Areas of Interest (AOIs); and selecting a target edge NWDAF network element from multiple edge NWDAF network elements as a federated learning member corresponding to the current analysis identifier, based on the current federated learning requirement information.

[0006] In some embodiments, selecting a target edge NWDAF network element from multiple edge NWDAF network elements as a federated learning member corresponding to the current analysis identifier, based on the current federated learning requirement information, includes: sending the current federated learning requirement information corresponding to the current analysis identifier to a network repository function (NRF) network element, wherein the current federated learning requirement information is used to assist the NRF network element in selecting candidate edge NWDAF network elements that match the current federated learning requirement information from multiple edge NWDAF network elements based on the correspondence between the network element identifiers, analysis identifiers, and federated learning requirement information of the multiple edge NWDAF network elements; receiving the network element identifiers of the candidate edge NWDAF network elements from the NRF network element; and selecting the target edge NWDAF network element from the candidate edge NWDAF network elements based on the network element identifiers of the candidate edge NWDAF network elements from the NRF network element.

[0007] In some embodiments, selecting the target edge NWDAF network element from the candidate edge NWDAF network elements includes: selecting the target edge NWDAF network element from the candidate edge NWDAF network elements based on the idle time of the candidate edge NWDAF network elements from the NRF network element.

[0008] In some embodiments, selecting the target edge NWDAF network element from the candidate edge NWDAF network elements based on the idle time of the candidate edge NWDAF network elements from the NRF network element includes: selecting candidate edge NWDAF network elements whose idle time meets a preset time condition as the target edge NWDAF network element.

[0009] In some embodiments, selecting candidate edge NWDAF network elements whose idle time meets a preset time condition includes: selecting a preset number of candidate edge NWDAF network elements with the longest idle time as target edge NWDAF network elements; or selecting a preset number of candidate edge NWDAF network elements whose idle time is concentrated within a preset time range as target edge NWDAF network elements; or selecting a preset number of candidate edge NWDAF network elements whose idle time is both the longest and concentrated within a preset time range as target edge NWDAF network elements.

[0010] In some embodiments, the correspondence between the network element identifier, analysis identifier, and federated learning requirement information of the edge NWDAF network element is stored in the NRF network element during the registration process of the multiple edge NWDAF network elements to the NRF network element.

[0011] In some embodiments, the method for selecting federated learning members further includes: sending a registration request to the NRF network element, the registration request including a correspondence between the network element identifier, the analysis identifier, and the federated learning requirement information.

[0012] In some embodiments, sending the current federated learning requirement information corresponding to the current analysis identifier to the NRF network element includes: sending a request for discovering federated learning members to the NRF network element, wherein the request for discovering federated learning members includes the current federated learning requirement information corresponding to the current analysis identifier.

[0013] In some embodiments, receiving the network element identifier of the candidate edge NWDAF network element from the NRF network element includes: receiving a response from the NRF network element for discovering federated learning members, wherein the response for discovering federated learning members includes the network element identifier of the candidate edge NWDAF network element, and the response for discovering federated learning members is a response to a request for discovering federated learning members.

[0014] In some embodiments, the method for selecting federated learning members is performed by the central NWDAF network element.

[0015] According to a second aspect of this disclosure, an apparatus for selecting federated learning members is provided, comprising: a receiving module configured to receive an analysis subscription or request from a Network Data Analysis Function (NWDAF) consumer, the analysis subscription or request including a current analysis identifier; a determining module configured to determine current federated learning requirement information corresponding to the current analysis identifier, the current federated learning requirement information including at least one of a list of algorithms or models for the current federated learning, an encryption method for the current federated learning, and a list of expected times for the current federated learning, as well as a list of regions of interest (AOIs); and a selecting module configured to select a target edge NWDAF network element from a plurality of edge NWDAF network elements as a federated learning member corresponding to the current analysis identifier, based on the current federated learning requirement information.

[0016] According to a third aspect of this disclosure, an apparatus for selecting federated learning members is provided, comprising: a memory; and a processor coupled to the memory, the processor being configured to perform the method for selecting federated learning members as described in any of the above embodiments based on instructions stored in the memory.

[0017] According to a fourth aspect of this disclosure, a system for selecting federated learning members is provided, comprising: the means for selecting federated learning members as described in any of the foregoing embodiments.

[0018] In some embodiments, the system for selecting federated learning members further includes: a Network Repository Function (NRF) network element configured to: receive current federated learning requirement information corresponding to the current analysis identifier; select candidate edge NWDAF network elements from a plurality of edge NWDAF network elements that match the current federated learning requirement information based on the correspondence between the network element identifier, analysis identifier, and federated learning requirement information of the edge NWDAF network elements; and send the network element identifier of the candidate edge NWDAF network element to the means for selecting federated learning members.

[0019] According to a fifth aspect of this disclosure, a computer-storeable medium is provided having computer program instructions stored thereon that, when executed by a processor, implement the method for selecting federated learning members as described in any of the above embodiments.

[0020] In the above embodiments, the accuracy of selecting federated learning members can be improved, thereby improving the accuracy of related analysis. Attached Figure Description

[0021] The accompanying drawings, which form part of this specification, illustrate embodiments of this disclosure and, together with the specification, serve to explain the principles of this disclosure.

[0022] This disclosure will become clearer with reference to the accompanying drawings and the following detailed description, wherein:

[0023] Figure 1 This is a flowchart illustrating a method for selecting federated learning members according to some embodiments of the present disclosure;

[0024] Figure 2 A block diagram illustrating an apparatus for selecting federated learning members according to some embodiments of the present disclosure;

[0025] Figure 3 This is a block diagram illustrating an apparatus for selecting federated learning members according to other embodiments of the present disclosure;

[0026] Figure 4 This is a block diagram illustrating a system for selecting federated learning members according to some embodiments of the present disclosure;

[0027] Figure 5 This is a signaling diagram illustrating a method for selecting federated learning members according to some embodiments of this disclosure;

[0028] Figure 6 This is a block diagram illustrating a computer system for implementing some embodiments of the present disclosure. Detailed Implementation

[0029] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present disclosure.

[0030] At the same time, it should be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn according to actual scale.

[0031] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit this disclosure or its application or use.

[0032] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0033] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0034] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0035] Figure 1 This is a flowchart illustrating a method for selecting federated learning members according to some embodiments of this disclosure.

[0036] like Figure 1 As shown, the method for selecting federated learning members includes steps S110-S130. In some embodiments, the method for selecting federated learning members is performed by a central NWDAF network element. The central NWDAF network element is an NWDAF network element that receives analytics subscriptions or analytics requests from NWDAF consumers.

[0037] In step S110, an analytics subscription or request from an NWDAF consumer is received. The analytics subscription or request includes the current analytics identifier. The analytics identifier is, for example, an analytics ID (Identity document).

[0038] In step S120, the current federated learning requirement information corresponding to the current analysis identifier is determined. The current federated learning requirement information includes at least one of the following: a list of algorithms or models used for the current federated learning, an encryption method used for the current federated learning, and a list of expected times for the current federated learning, as well as a list of AOIs (Area of ​​Interest).

[0039] In step S130, based on the current federated learning requirements, a target edge NWDAF network element is selected from multiple edge NWDAF network elements as a federated learning member corresponding to the current analysis identifier.

[0040] In the above embodiments, the target edge NWDAF network elements are determined as federated learning members based on the current federated learning requirements. This makes the federated learning members more compatible and matched with the current federated learning requirements corresponding to the current analysis identifier, thereby improving the accuracy of federated learning member selection, and consequently improving the accuracy of related analysis and enhancing the service experience. Furthermore, federated learning can also protect privacy and improve privacy security.

[0041] In some embodiments, step S130 can be implemented as follows.

[0042] First, the current federated learning requirement information corresponding to the current analysis identifier is sent to the NRF (Network Repository Function) network element. This requirement information assists the NRF network element in selecting candidate edge NWDAF network elements that match the current federated learning requirement information from multiple edge NWDAF network elements, based on the correspondence between the network element identifiers, analysis identifiers, and federated learning requirement information of multiple edge NWDAF network elements.

[0043] In some embodiments, the correspondence between the network element identifier, analysis identifier, and federated learning requirement information of the edge NWDAF network element is stored in the NRF network element during the registration process of multiple edge NWDAF network elements to the NRF network element.

[0044] In some embodiments, a request to discover federated learning members can be sent to the NRF network element. The request includes current federated learning requirement information corresponding to the current analysis identifier. For example, the request to discover federated learning members can be sent to the NRF network element by sending an Nnrf_NFDiscovery_Request message. In some embodiments, the Nnrf_NFDiscovery_Request message may also carry S-NSSAI (Single Network Slice Selection Assistance Information).

[0045] In some embodiments, before sending the current federated learning requirement information corresponding to the current analysis identifier to the NRF network element, the central NWDAF network element determines whether to perform federated learning to complete model training based on the current analysis identifier. For example, if the data corresponding to the current analysis identifier involves user privacy data, the determination result is yes. If the determination result is yes, the current federated learning requirement information corresponding to the current analysis identifier is sent to the NRF network element.

[0046] Then, receive the network element identifier of the candidate edge NWDAF network element from the NRF network element.

[0047] In some embodiments, a response for discovering federated learning members is received from an NRF network element. The response for discovering federated learning members includes the network element identifier of the candidate edge NWDAF network element. The response for discovering federated learning members is a response to a request for discovering federated learning members. For example, the response for discovering federated learning members is received by receiving an Nnrf_NFDiscovery_RequestResponse message. In some embodiments, the response for discovering federated learning members also includes the idle time of the candidate edge NWDAF network element.

[0048] Finally, based on the network element identifiers of the candidate edge NWDAF network elements from the NRF network elements, the target edge NWDAF network element is selected from the candidate edge NWDAF network elements.

[0049] In some embodiments, a target edge NWDAF network element can be selected from candidate edge NWDAF network elements based on their idle time from the NRF network element. For example, candidate edge NWDAF network elements whose idle time meets a preset time condition can be selected as the target edge NWDAF network element.

[0050] In some embodiments, a preset number of candidate edge NWDAF network elements with the longest idle time can be selected as target edge NWDAF network elements. This method can save communication bandwidth.

[0051] In some embodiments, a predetermined number of candidate edge NWDAF network elements whose idle time is concentrated within a predetermined time range can be selected as target edge NWDAF network elements. This approach can improve time utilization efficiency, thereby improving the efficiency of subsequent related analysis.

[0052] In some embodiments, a preset number of candidate edge NWDAF network elements with the longest idle time and concentrated within a preset time range can be selected as target edge NWDAF network elements.

[0053] In some embodiments, the central NWDAF element can also send a registration request to the NRF element. The registration request includes the mapping between the central NWDAF element's element identifier, analysis identifier, and federated learning requirement information. For example, the registration request is sent using the Nnrf_NFManagement_NFRegister message.

[0054] In some embodiments, the central NWDAF network element can also receive registration responses from NRF network elements. For example, the registration response is sent using the Nnrf_NFManagement_NFRegister Response message. The registration response is consistent with existing 3GPP specifications and will not be described further here.

[0055] In some embodiments, when multiple NWDAF instances are deployed in a PLMN, the multiple NWDAFs may need to undergo federated learning. The central NWDAF needs to select appropriate edge NWDAFs as federated learning members in each training round, and the methods for selecting federated learning members in any of the embodiments of this disclosure can be used.

[0056] In some embodiments of a 6G network, the method for selecting federated learning members in any of the embodiments of this disclosure can also be applied to the scenario of member selection when the central AI network element and the edge AI network element in the core network are performing federated learning. The central AI network element and the edge AI network element are the central NWDAF network element and the edge NWDAF network element, respectively.

[0057] In some embodiments, both the central NWDAF element and the edge NWDAF element have a Model Training Logical Function (MTLF). Any NWDAF element with a Model Training Logical Function can function as either a central NWDAF element or an edge NWDAF element.

[0058] Figure 2 A block diagram of an apparatus for selecting federated learning members according to some embodiments of the present disclosure is shown.

[0059] like Figure 2 As shown, the device 21 for selecting federated learning members includes a receiving module 211, a determining module 212, and a selecting module 213. In some embodiments, the device 21 for selecting federated learning members is deployed in a central NWDAF network element.

[0060] The receiving module 211 is configured to receive analysis subscriptions or requests from consumers of the Network Data Analysis Function (NWDAF), for example, to perform... Figure 1 The step S110 shown. The analytics subscription or request includes the current analytics identifier.

[0061] Module 212 is configured to determine the current federated learning requirements corresponding to the current analysis identifier, for example, by performing actions such as... Figure 1 The step S120 shown. The current federated learning requirements information includes at least one of the following: a list of algorithms or models used for the current federated learning, an encryption method used for the current federated learning, and a list of expected times for the current federated learning, as well as a list of regions of interest (AOIs).

[0062] Selection module 213 is configured to select a target edge NWDAF element from multiple edge NWDAF elements based on the current federated learning requirements, as a federated learning member corresponding to the current analysis identifier, for example, performing the following... Figure 1 The step S130 shown.

[0063] Figure 3 This is a block diagram illustrating an apparatus for selecting federated learning members according to other embodiments of the present disclosure.

[0064] like Figure 3 As shown, the apparatus 31 for selecting federated learning members includes a memory 311 and a processor 312 coupled to the memory 311. The memory 311 is used to store instructions for performing methods corresponding to embodiments of the method for selecting federated learning members. The processor 312 is configured to perform methods for selecting federated learning members in any of the embodiments of this disclosure based on the instructions stored in the memory 311.

[0065] Figure 4 This is a block diagram illustrating a system for selecting federated learning members according to some embodiments of the present disclosure.

[0066] like Figure 4 As shown, the system 4 for selecting federated learning members includes means 41 for selecting federated learning members. For example, means 41 for selecting federated learning members is means 21 or 31 for selecting federated learning members in any of the embodiments of this disclosure.

[0067] In some embodiments, the apparatus 4 for selecting federated learning members further includes an NRF network element 42. The NRF network element 42 is configured to receive current federated learning requirement information corresponding to the current analysis identifier; select candidate edge NWDAF network elements that match the current federated learning requirement information from multiple edge NWDAF network elements according to the correspondence between the network element identifier, analysis identifier and federated learning requirement information of the edge NWDAF network elements; and send the network element identifier of the candidate edge NWDAF network elements to the apparatus 41 for selecting federated learning members.

[0068] Figure 5 This is a signaling diagram illustrating a method for selecting federated learning members according to some embodiments of this disclosure.

[0069] like Figure 5 As shown, the method for selecting federated learning members includes steps S501-S507.

[0070] In step S501, the central NWDAF network element and the edge NWDAF network element send registration requests to the NRF network element. The registration request includes the correspondence between the network element identifier, analysis identifier, and federated learning requirement information of the central or edge NWDAF network element. For example, the registration request is sent using the Nnrf_NFManagement_NFRegister message.

[0071] In step S502, the NRF network element sends registration responses to both the central NWDAF network element and the edge NWDAF network element. For example, the registration response can be sent using the Nnrf_NFManagement_NFRegister Response message. The registration response is consistent with existing 3GPP specifications and will not be elaborated further here.

[0072] In step S503, the NWDAF consumer sends an analytics subscription or analytics request to the central NWDAF consumer. The analytics subscription or request includes the current analytics identifier.

[0073] In step S504, the central NWDAF network element determines whether to perform federated learning. In some embodiments, the central NWDAF network element determines whether to perform federated learning to complete model training based on the current analysis identifier. For example, if the data corresponding to the current analysis identifier involves user privacy data, the determination result is yes. If the determination result is yes, the central NWDAF network element executes step S505.

[0074] In step S505, the central NWDAF network element sends a request to the NRF network element to discover federated learning members. The request includes current federated learning requirement information corresponding to the current analysis identifier. For example, the request to discover federated learning members is sent to the NRF network element by sending an Nnrf_NFDiscovery_Request message. In some embodiments, the Nnrf_NFDiscovery_Request message may also carry S-NSSAI (Single Network Slice Selection Assistance Information).

[0075] In step S506, the NRF network element sends a response for discovering federated learning members to the central NWDAF network element. The response for discovering federated learning members includes the network element identifier of the candidate edge NWDAF network element. The response for discovering federated learning members is a response to a request for discovering federated learning members. For example, the response for discovering federated learning members is sent by sending an Nnrf_NFDiscovery_Request Response message. In some embodiments, the response for discovering federated learning members also includes the idle time of the candidate edge NWDAF network element.

[0076] In step S507, the central NWDAF network element selects the target edge NWDAF network element from the candidate edge NWDAF network elements based on the network element identifiers of the candidate edge NWDAF network elements from the NRF network element.

[0077] Figure 5 Some detailed steps in the illustrated embodiments can be referred to Figure 1 The specific implementation examples are not described here.

[0078] Figure 6 This is a block diagram illustrating a computer system for implementing some embodiments of the present disclosure.

[0079] like Figure 6 As shown, the computer system 60 can be represented in the form of a general computing device. The computer system 60 includes a memory 610, a processor 620, and a bus 600 connecting different system components.

[0080] The memory 610 may include, for example, system memory, non-volatile storage media, etc. The system memory may store, for example, an operating system, applications, a boot loader, and other programs. The system memory may include volatile storage media, such as random access memory (RAM) and / or cache memory. The non-volatile storage media may store, for example, instructions for performing at least one of the methods for selecting federated learning members in a corresponding embodiment. Non-volatile storage media include, but are not limited to, disk storage, optical storage, flash memory, etc.

[0081] The processor 620 can be implemented using a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete hardware components such as discrete gates or transistors. Accordingly, each module, such as the decision module and the determination module, can be implemented by executing instructions in the central processing unit (CPU) memory to perform the corresponding steps, or by implementing dedicated circuitry to perform the corresponding steps.

[0082] Bus 600 can use any of the various bus architectures. For example, bus architectures include, but are not limited to, Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MCA) bus, and Peripheral Component Interconnect (PCI) bus.

[0083] The computer system 60 may also include an input / output interface 630, a network interface 640, and a storage interface 650. These interfaces 630, 640, and 650, as well as the memory 610 and processor 620, can be connected via a bus 600. The input / output interface 630 provides a connection interface for input / output devices such as a monitor, mouse, and keyboard. The network interface 640 provides a connection interface for various networked devices. The storage interface 650 provides a connection interface for external storage devices such as floppy disks, USB flash drives, and SD cards.

[0084] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus, and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations thereof, can be implemented by computer-readable program instructions.

[0085] These computer-readable program instructions are provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable device to produce a machine, such that execution of the instructions by the processor produces means for implementing the functions specified in one or more boxes of the flowchart and / or block diagram.

[0086] These computer-readable program instructions may also be stored in a computer-readable storage medium. These instructions cause a computer to work in a particular manner to produce an article of manufacture, including instructions that implement the functions specified in one or more boxes in a flowchart and / or block diagram.

[0087] This disclosure may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects.

[0088] The methods, apparatus, systems, and computer-storable media for selecting federated learning members in the above embodiments can improve the accuracy of selecting federated learning members, thereby improving the accuracy of related analysis.

[0089] The methods, apparatus, systems, and computer-storable media for selecting federated learning members according to this disclosure have been described in detail. To avoid obscuring the concept of this disclosure, some details known in the art have not been described. Those skilled in the art will fully understand how to implement the technical solutions disclosed herein based on the above description.

Claims

1. A method for selecting members of a federated learning system, comprising: Receive analytics subscriptions or requests from NWDAF consumers, where the analytics subscriptions or requests include the current analytics identifier; Determine the current federated learning requirement information corresponding to the current analysis identifier. The current federated learning requirement information includes at least one of the following: a list of algorithms or models for the current federated learning, an encryption method for the current federated learning, and a list of expected times for the current federated learning, as well as a list of regions of interest (AOIs). Based on the current federated learning requirement information, a target edge NWDAF network element is selected from multiple edge NWDAF network elements as a federated learning member corresponding to the current analysis identifier. This includes: sending the current federated learning requirement information corresponding to the current analysis identifier to the network repository function (NRF) network element. The current federated learning requirement information is used to assist the NRF network element in selecting candidate edge NWDAF network elements that match the current federated learning requirement information from multiple edge NWDAF network elements based on the correspondence between the network element identifiers, analysis identifiers, and federated learning requirement information of the multiple edge NWDAF network elements; receiving the network element identifiers of the candidate edge NWDAF network elements from the NRF network element; and selecting the target edge NWDAF network element from the candidate edge NWDAF network elements based on the network element identifiers of the candidate edge NWDAF network elements from the NRF network element.

2. The method for selecting federated learning members according to claim 1, wherein, Selecting the target edge NWDAF network element from the candidate edge NWDAF network elements includes: The target edge NWDAF network element is selected from the candidate edge NWDAF network elements based on the idle time of the candidate edge NWDAF network elements from the NRF network element.

3. The method for selecting federated learning members according to claim 2, wherein, Selecting the target edge NWDAF network element from the candidate edge NWDAF network elements based on the idle time of the candidate edge NWDAF network elements includes: Candidate edge NWDAF network elements whose idle time meets the preset time conditions are selected as the target edge NWDAF network elements.

4. The method for selecting federated learning members according to claim 3, wherein, Candidate edge NWDAF network elements whose idle time meets the preset time conditions include: Select a preset number of candidate edge NWDAF network elements with the longest idle time as the target edge NWDAF network element; or Select a predetermined number of candidate edge NWDAF network elements whose idle time is concentrated within a predetermined time range, as the target edge NWDAF network element; or Select a preset number of candidate edge NWDAF network elements that have the longest idle time and are concentrated within a preset time range, as the target edge NWDAF network elements.

5. The method for selecting federated learning members according to claim 1, wherein, The correspondence between the network element identifier, analysis identifier, and federated learning requirement information of the edge NWDAF network element is stored in the NRF network element during the registration process of the multiple edge NWDAF network elements to the NRF network element.

6. The method for selecting federated learning members according to claim 1, further comprising: A registration request is sent to the NRF network element, and the registration request includes the correspondence between the network element identifier, the analysis identifier, and the federated learning requirement information.

7. The method for selecting federated learning members according to claim 1, wherein, Sending the current federated learning requirement information corresponding to the current analysis identifier to the NRF network element includes: A request for discovering federated learning members is sent to the NRF network element, the request for discovering federated learning members including the current federated learning requirement information corresponding to the current analysis identifier.

8. The method for selecting federated learning members according to claim 7, wherein, The network element identifier received from the NRF network element for the candidate edge NWDAF network element includes: Receive a response from the NRF network element for discovering federated learning members. The response for discovering federated learning members includes the network element identifier of the candidate edge NWDAF network element. The response for discovering federated learning members is a response to a request for discovering federated learning members.

9. The method for selecting federated learning members according to claim 1, wherein, The method for selecting federated learning members is performed by the central NWDAF network element.

10. An apparatus for selecting federated learning members, comprising: The receiving module is configured to receive analysis subscriptions or requests from consumers of the Network Data Analysis Function (NWDAF), wherein the analysis subscriptions or requests include a current analysis identifier. The determination module is configured to determine the current federated learning requirement information corresponding to the current analysis identifier. The current federated learning requirement information includes at least one of the following: a list of algorithms or models for the current federated learning, an encryption method for the current federated learning, and a list of expected times for the current federated learning, as well as a list of regions of interest (AOIs). The selection module is configured to select a target edge NWDAF network element from multiple edge NWDAF network elements as a federated learning member corresponding to the current analysis identifier, based on the current federated learning requirement information. This includes: sending the current federated learning requirement information corresponding to the current analysis identifier to the network repository function (NRF) network element; the current federated learning requirement information assists the NRF network element in selecting candidate edge NWDAF network elements that match the current federated learning requirement information from multiple edge NWDAF network elements based on the correspondence between the network element identifiers, analysis identifiers, and federated learning requirement information of the multiple edge NWDAF network elements; receiving the network element identifiers of the candidate edge NWDAF network elements from the NRF network element; and selecting the target edge NWDAF network element from the candidate edge NWDAF network elements based on the network element identifiers of the candidate edge NWDAF network elements from the NRF network element.

11. An apparatus for selecting federated learning members, comprising: Memory; as well as A processor coupled to the memory, the processor being configured to perform the method for selecting federated learning members as described in any one of claims 1 to 9, based on instructions stored in the memory.

12. A system for selecting members of a federated learning system, comprising: The apparatus for selecting federated learning members as described in claim 10 or 11.

13. The system for selecting federated learning members according to claim 12, further comprising: The network warehousing function (NRF) element is configured as follows: Receive the current federated learning requirement information corresponding to the current analysis identifier; Based on the correspondence between the network element identifier, analysis identifier and federated learning requirement information of the edge NWDAF network element, candidate edge NWDAF network elements that match the current federated learning requirement information are selected from multiple edge NWDAF network elements. The network element identifier of the candidate edge NWDAF network element is sent to the device for selecting federated learning members.

14. A computer-storeable medium having stored thereon computer program instructions that, when executed by a processor, implement the method for selecting federated learning members as described in any one of claims 1 to 9.

Images