Communication method and device

Abstract

The embodiment of the invention provides a communication method and device, and the method comprises the steps: a first network element receives a first request from a second network element, and the first request is used for requesting the first network element to participate in a first federated learning task; the first network element obtains first information, wherein the first information is used for indicating that the second network element can provide a global model corresponding to the first federated learning task for the first network element; the first network element sends a first response of the first request to the second network element according to the first information, and the first response is used for indicating the first network element to determine to participate in the first federated learning task. Therefore, the first network element (namely the client) can clearly determine whether the first network element has the authority of acquiring the global model, and the first network element can decide whether to participate in the first federated learning task based on whether the first network element can acquire the global model, thereby ensuring that the first network element can effectively participate in the federated learning task, and improving the user satisfaction of the client.

Description

Technical Field

[0001] This application relates to the field of communications, and in particular, to a communication method and apparatus. Background Art

[0002] In a core network, a network data analytics function (NWDAF) network element (hereinafter referred to as an NWDAF network element or NWDAF) may be a decentralized entity device that holds local data, and multiple NWDAFs may perform federated learning (FL).

[0003] Before starting each federated learning task, a server NWDAF may send a request to participate in the federated learning task to multiple client NWDAFs; after the server NWDAF receives responses from at least one client NWDAF agreeing to participate in the federated learning task, the server NWDAF may select at least one client NWDAF from the client NWDAFs that have agreed to participate in the federated learning task to form a federated learning group corresponding to the federated learning task. Among them, the federated learning group includes one server NWDAF and at least one client NWDAF. Further, the server NWDAF and the client NWDAFs in the federated learning group may perform multiple rounds of iterative training (without exchanging local data) based on an initial model, so as to obtain a global model corresponding to the federated learning task.

[0004] After multiple rounds of iterative training are completed, the server NWDAF may obtain the global model, and the server NWDAF may send the global model to one or more client NWDAFs. Correspondingly, during the process of obtaining the global model, the client NWDAF is always in a passive acquisition state. Therefore, after participating in the federated learning task, the client NWDAF may not be able to obtain the global model corresponding to the federated learning task. When the client NWDAF participates in the federated learning task but cannot obtain the global model, the user experience of the client NWDAF is poor and the commercial interests are damaged. Summary of the Invention

[0005] Embodiments of this application provide a communication method and apparatus, which are used to clarify the permission of a client to obtain a global model, ensure the effective participation of the client in the federated learning task, and improve the user satisfaction of the client.

[0006] In a first aspect, the present application provides a communication method. This method is applied to a first network element, or a component in the first network element (such as a processor, a chip, a chip system, a circuit, or others), or a software module. Taking the case where this method is applied to the first network element as an example, the method may include: The first network element receives a first request from a second network element, and the first request is used to request the first network element to participate in a first federated learning task; the first network element obtains first information, and the first information is used to indicate that the second network element can provide the global model corresponding to the first federated learning task to the first network element; the first network element sends a first response to the first request to the second network element according to the first information, and the first response is used to indicate that the first network element determines to participate in the first federated learning task.

[0007] By using this method, the first network element (i.e., the client) can clarify whether it has the permission to obtain the global model. The first network element can make a decision on whether to participate in the first federated learning task based on whether it can obtain the global model, so as to ensure that the first network element can effectively participate in the federated learning task and improve the user satisfaction of the client.

[0008] In a possible design, the first information includes at least one network element identifier, and the at least one network element identifier includes the identifier of the first network element; or, the first information includes at least one analysis identifier, and the at least one analysis identifier includes the analysis identifier corresponding to the first federated learning task; or, the first information includes at least one interoperability identifier, and the at least one interoperability identifier includes the interoperability identifier corresponding to the first network element. In this way, the first information can indicate the permissions of devices with different granularities to obtain the global model, and can improve the flexibility of the first information.

[0009] In a possible design, the process of obtaining the first information described above includes: The first network element sends a second request to the NRF network element, and the second request is used to query whether the second network element can provide the global model corresponding to the first federated learning task to the first network element; the first network element receives the first information from the NRF network element. In this way, the first network element can obtain the first information through the NRF network element.

[0010] In a possible design, the first request may include the first information. In this way, the first network element can directly obtain the first information through the first request, saving signaling overhead.

[0011] In a possible design, the first network element may also receive model information from the second network element, and the model information is used to obtain the global model corresponding to the first federated learning task.

[0012] In this way, the second network element can actively send the model information to the first network element. Correspondingly, the first network element can receive the model information, improving the user satisfaction of the client corresponding to the first network element.

[0013] In a possible design, before the first network element receives model information from the second network element, the first network element may also send a first address to the second network element, and the first address is used to receive the model information.

[0014] In this way, the second network element may send the model information to the first network element through the first address, and the first network element may receive the model information through the first address.

[0015] In a possible design, the first network element may also send a third request to the second network element. The third request is used to request to obtain model information, and the model information is used to obtain the global model corresponding to the first federated learning task. The first network element receives a second response to the third request from the second network element. The second response includes the model information, or the second response includes indication information indicating that the model acquisition fails.

[0016] In this way, after receiving the third request for obtaining model information sent by the first network element, the second network element may send a second response corresponding to the third request to the first network element, and the second response includes the model information, which improves the flexibility of the first network element to obtain the model information.

[0017] In a possible design, the third request includes identification information corresponding to the first federated learning task, and the second response is determined according to the identification information. In this way, the identification information corresponding to the first federated learning task is carried in the third request, and the identification information may be used as an index for the second network element to query the global model, so that the second network element can accurately send the model information of the global model of the first federated learning task to the first network element. In addition, the identification information corresponding to the first federated learning task is carried in the third request, and the identification information may be used as a verification value for the second network element to verify the first network element, so as to prevent network elements not related to the first federated learning task from obtaining the model information of the global model corresponding to the first federated learning task, and improve communication security.

[0018] In a possible design, the first request may include identification information; and / or, the first network element may also receive identification information from the second network element. In this way, the first network element may obtain the foregoing identification information through multiple channels, improving flexibility.

[0019] In a possible design, the identification information includes at least one of the following: the model identification of the initial model corresponding to the first federated learning task; the model identification of the model in any round of iterative training process corresponding to the first federated learning task; the model identification of the global model corresponding to the first federated learning task; or, the task identification of the first federated learning task.

[0020] In this way, the identification information may be the model identification of the model or the task identification of the federated learning task, improving flexibility.

[0021] In a possible design, the first network element may also send second information to the NRF network element, where the second information is used to indicate that the first network element needs to obtain the global model corresponding to the federated learning task in which the first network element participates.

[0022] In this way, the NRF network element can obtain the second information from the first network element, so that the second network element can obtain the second information through the NRF network element. Furthermore, the second network element can select the network element (such as the first network element) participating in the federated learning task, improving the effectiveness of the client participating in the federated learning task.

[0023] In a second aspect, the present application provides a communication method, which is applied to the second network element, or a component in the second network element (such as a processor, a chip, a chip system, a circuit, or others), or a software module. Taking the method applied to the second network element as an example, the method includes: the second network element sends a first request to the first network element, where the first request is used to request the first network element to participate in a first federated learning task; the second network element sends first information to the first network element, where the first information is used to indicate that the second network element can provide the global model corresponding to the first federated learning task to the first network element; the second network element receives a first response from the first network element, where the first response is the response information of the first request, and the first response is used to indicate that the first network element determines to participate in the first federated learning task; the first response is determined according to the first information.

[0024] In a possible design, the first information includes at least one network element identifier, and the at least one network element identifier includes the identifier of the first network element; or, the first information includes at least one analysis identifier, and the at least one analysis identifier includes the analysis identifier corresponding to the first federated learning task; or, the first information includes at least one interoperability identifier, and the at least one interoperability identifier includes the interoperability identifier corresponding to the first network element.

[0025] In a possible design, the second network element may also send the first information to the NRF network element, so that the first network element can obtain the first information from the NRF network element. It can also be understood that the second network element sends the first information to the first network element through the NRF network element.

[0026] In a possible design, the first request includes the first information. In this way, the second network element can send the first information to the first network element through the first request.

[0027] In a possible design, the second network element may also send model information to the first network element, where the model information is used to obtain the global model corresponding to the first federated learning task.

[0028] In a possible design, before the second network element sends the model information to the first network element, the second network element may also receive a first address from the first network element, where the first address is used for the first network element to receive the model information.

[0029] In a possible design, the second network element may also receive a third request from the first network element. The third request is used to request to obtain model information, and the model information is used to obtain the global model corresponding to the first federated learning task. The second network element sends a second response to the third request to the first network element. The second response includes the model information, or the second response includes indication information indicating that the model acquisition fails.

[0030] In a possible design, the process of the second network element sending the second response to the third request to the first network element includes: when the third request includes the identification information corresponding to the first federated learning task, the second network element sends the second response to the first network element. In this way, the second network element can determine whether the first network element has the permission to obtain the global model corresponding to the first federated learning task based on whether the third request includes the identification information corresponding to the first federated learning task, so as to determine whether to send the second response to the first network element, improving the accuracy of communication. In other words, only when the second network element determines that the first network element has the permission to obtain the global model according to the third request, the second network element will send the second response to the first network element, improving the accuracy of communication.

[0031] In a possible design, the first request includes identification information; and / or, the second network element sends the identification information to the first network element.

[0032] In a possible design, the identification information includes at least one of the following: the model identification of the initial model corresponding to the first federated learning task; the model identification of the model in any round of iterative training process corresponding to the first federated learning task; the model identification of the global model corresponding to the first federated learning task; or, the task identification of the first federated learning task.

[0033] In a possible design, before the second network element sends the first request to the first network element, the second network element may send a fourth request to the NRF network element. The fourth request is used to query whether the first network element needs to obtain the global model corresponding to the federated learning task participated by the first network element. The second network element receives a third response to the fourth request from the NRF network element. The third response is used to indicate that the first network element needs to obtain the global model corresponding to the federated learning task participated by the first network element. In this way, the second network element can obtain the second information of the first network element through the NRF network element, and then the second network element can determine whether to select the first network element to participate in the federated learning task according to the second information, improving the effectiveness of the client participating in the federated learning task.

[0034] In a third aspect, the present application provides a communication method. This method is applied to an NRF network element, or a component in the NRF network element (such as a processor, a chip, a chip system, a circuit, or others), or a software module. Taking the case where this method is applied to an NRF network element as an example, the method includes: The NRF network element receives a second request from a first network element. The second request is used to query whether a second network element can provide the global model corresponding to the first federated learning task to the first network element; The NRF network element sends first information to the first network element. The first information is used to indicate that the second network element can provide the global model corresponding to the first federated learning task to the first network element.

[0035] In a possible design, the NRF network element may also receive first information from the second network element.

[0036] In a possible design, the NRF network element may also receive a fourth request from the second network element. The fourth request is used to query whether the first network element needs to obtain the global model corresponding to the federated learning task participated by the first network element; The NRF network element sends a third response to the fourth request to the second network element. The third response is used to indicate that the first network element needs to obtain the global model corresponding to the federated learning task participated by the first network element.

[0037] In a possible design, the NRF network element may also receive second information from the first network element. The second information is used to indicate that the first network element needs to obtain the global model corresponding to the federated learning task participated by the first network element.

[0038] In a fourth aspect, the present application provides a communication method. This method is applied to a first network element, or a component in the first network element (such as a processor, a chip, a chip system, a circuit, or others), or a software module. Taking the case where this method is applied to the first network element as an example, the method may include: The first network element receives a first request from the second network element. The first request is used to request the first network element to participate in the first federated learning task; The first network element sends a first response to the first request to the second network element. The first response is used to indicate that the first network element determines to participate in the first federated learning task or refuses to participate in the first federated learning task. The first response is also used to indicate that the first network element needs to obtain the global model corresponding to the federated learning task participated by the first network element.

[0039] Optionally, the first response may include one or more of the following: a first indication information or a model acquisition address. Wherein, the first indication information is used to indicate that the first network element needs to obtain the global model corresponding to the first federated learning task; The model acquisition address is used to indicate the address where the first network element receives the model information.

[0040] Using this method, the first network element (i.e., the client) can send a response to the second network element to indicate that the first network element needs to obtain the global model corresponding to the federated learning task in which the first network element participates, so that the second network element can determine whether the first network element needs to obtain the global model.

[0041] In a fifth aspect, the present application further provides a communication device. The communication device can execute the methods or the solutions in each possible design shown in the first aspect, the second aspect, the third aspect, or the fourth aspect above. The communication device can be a chip or a circuit capable of executing the functions corresponding to the above methods, or a device including the chip or the circuit.

[0042] In a possible design, the communication device includes a communication unit for receiving and / or sending data; the communication device further includes a processing unit for implementing the methods in any one of the possible designs shown in the first aspect, the second aspect, the third aspect, or the fourth aspect above. The foregoing functions can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more units corresponding to the above functions.

[0043] In a sixth aspect, the present application further provides a communication device. The communication device can execute the methods or the solutions in each possible design shown in the first aspect, the second aspect, the third aspect, or the fourth aspect above. The communication device includes: a processor. When the processor executes instructions, the communication device or the device equipped with the communication device executes the methods in any one of the possible designs shown in the first aspect, the second aspect, the third aspect, or the fourth aspect above.

[0044] Optionally, the communication device may further include a memory for storing computer-executable program code, and the program agent may include the foregoing instructions. The memory can be placed inside the communication device or outside the communication device, which is not limited in the present application. The memory can be coupled to the processor.

[0045] Among them, the communication device may further include a communication interface. Optionally, if the communication device is a chip or a circuit, the communication interface can be the input / output interface of the chip, such as input / output pins, etc.

[0046] In a seventh aspect, the present application provides a communication system, which includes at least one of the following: the first network element executing the method in the first aspect or the first network element executing the method in the fourth aspect, the second network element executing the method in the second aspect, and the NRF network element executing the method in the third aspect.

[0047] In an eighth aspect, the present application provides a computer-readable storage medium storing a computer program, which, when run on a computer, causes the computer to execute the method in any one of the possible designs shown in the first, second, third, or fourth aspect above.

[0048] In a ninth aspect, the present application provides a computer program product, where a computer-readable storage medium stores computer-executable instructions, which, when called by a computer, cause the computer to execute the method in any one of the possible designs shown in the first, second, third, or fourth aspect above.

[0049] In a tenth aspect, the present application provides a chip, which includes a processor for executing the method in any one of the possible designs shown in the first, second, third, or fourth aspect above. Optionally, the chip may further include a communication interface for inputting and / or outputting signaling or data. Optionally, the chip may further include a memory for storing the aforementioned computer program; the processor is coupled to the memory, and the processor can read the computer program stored in the memory to execute the method in any one of the possible designs shown in the first, second, third, or fourth aspect above.

[0050] In addition, for the technical effects brought by the second to tenth aspects, reference may be made to the descriptions of the various possible solutions in the first aspect above, which will not be elaborated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a schematic diagram of the architecture of a communication system provided by an embodiment of the present application;

[0052] Figure 2 It is a schematic flowchart of a communication method provided by an embodiment of the present application;

[0053] Figure 3 It is a schematic flowchart of another communication method provided by an embodiment of the present application;

[0054] Figure 4 It is an example diagram of a communication method provided by an embodiment of the present application;

[0055] Figure 5 It is an example diagram of another communication method provided by an embodiment of the present application;

[0056] Figure 6 It is an example diagram of another communication method provided by an embodiment of the present application;

[0057] Figure 7 It is a schematic diagram of the structure of a communication device provided by an embodiment of the present application;

[0058] Figure 8 This is a schematic structural diagram of another communication device provided by an embodiment of the present application. Detailed implementation manners

[0059] In order to make the objectives, technical solutions and beneficial effects of the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0060] In the description of the present application, unless otherwise specified, " / " means "or". For example, A / B may represent A or B. The "and / or" in the present application is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the present application, "at least one item" means one item or more, and "a plurality of items" means two or more. In the description of the present application, words such as "first" and "second" are only used for the purpose of distinguishing descriptions, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating or implying order.

[0061] Embodiments of the present application can be applied to various communication systems, such as: global system for mobile communications (GSM) system, code division multiple access (CDMA) system, wideband code division multiple access (WCDMA) system, general packet radio service (GPRS), long term evolution (LTE) system, LTE frequency division duplex (FDD) system, LTE time division duplex (TDD), universal mobile telecommunications system (UMTS), worldwide interoperability for microwave access (WIMAX) communication system, fifth generation (5G) system or new radio (NR), or applied to future communication systems or other similar communication systems, etc.

[0062] Figure 1 It is a schematic diagram of the architecture of a communication system. Figure 1 In the architecture of the shown communication system, it may include terminal devices, access network devices, and core network devices. The terminal devices access the data network (DN) through the access network devices and the core network devices. Among them, the core network devices include multiple network functions (NFs) or network elements. The core network devices can also be referred to as core network elements. For example, the core network devices may include some or all of the following network elements (or core network elements): unified data management (UDM) network element, unified data repository (UDR) network element, application function (AF) network element, policy control function (PCF) network element, access and mobility management function (AMF) network element, session management function (SMF) network element, user plane function (UPF) network element, network data analytics function (NWDAF) network element, network repository function (NRF) network element (not shown in the figure), etc.

[0063] It should be understood that the present application does not limit the number of network elements included in the communication system. For example, Figure 1 the communication system in

[0064] The terminal device can be a user equipment (UE), a mobile station, a mobile terminal, etc. The terminal device can be widely applied to various scenarios, such as device-to-device (D2D), vehicle to everything (V2X) communication, machine-type communication (MTC), internet of things (IOT), virtual reality, augmented reality, industrial control, autonomous driving, telemedicine, smart grid, smart furniture, smart office, smart wearables, smart transportation, smart city, etc. The terminal device can be a mobile phone, a tablet computer, a computer with wireless transceiver function, a wearable device, a vehicle, an urban air vehicle (such as a drone, a helicopter, etc.), a ship, a robot, a robotic arm, a smart home device, etc.

[0065] The access network device can be a radio access network (RAN) device. For example: a base station, an evolved NodeB (eNodeB), a transmission reception point (TRP), a next generation NodeB (gNB) in a 5G mobile communication system, a next generation base station in a 6th generation (6G) mobile communication system, a base station in a future mobile communication system, or an access node in a wireless fidelity (WiFi) system, etc.; it can also be a module or unit that completes part of the base station functions. For example, it can be a central unit (CU) or a distributed unit (DU). The radio access network device can be a macro base station, a micro base station or an indoor station, and can also be a relay node or a donor node, etc. The embodiments of the present application do not limit the specific technologies and specific device forms adopted by the radio access network device.

[0066] The access network device and the terminal device can be fixed in position or movable. The access network device and the terminal device can be deployed on land, including indoor or outdoor, handheld or vehicle-mounted; they can also be deployed on water; and can also be deployed on airplanes, balloons and artificial satellites in the air. The embodiments of the present application do not limit the application scenarios of the access network device and the terminal device.

[0067] The following briefly introduces some core network devices:

[0068] The AMF network element, abbreviated as AMF, includes functions such as performing mobility management, access authentication / authorization, etc. In addition, the AMF is also responsible for transmitting user policies between the terminal device and the PCF.

[0069] The SMF network element, abbreviated as SMF, includes functions such as performing session management, executing the control policies issued by the PCF, selecting the UPF, and allocating the Internet Protocol (IP) address of the terminal device.

[0070] The UPF network element, abbreviated as UPF, as the interface to the data network, includes functions such as completing user plane data forwarding, session / flow-level billing statistics, and bandwidth limitation.

[0071] The UDM network element, abbreviated as UDM, includes functions such as performing management of subscription data and user access authorization.

[0072] The UDR network element, abbreviated as UDR, includes functions such as accessing and storing various types of data, such as subscription data, policy data, and application data.

[0073] The NEF network element, abbreviated as NEF, is used to support the opening of capabilities and events.

[0074] The AF network element, abbreviated as AF, transmits the requirements from the application side to the network side. For example, quality of service (QoS) requirements or user status event subscriptions, etc. The AF can be a third-party functional entity or an application server deployed by the operator.

[0075] The PCF network element, abbreviated as PCF, includes policy control functions such as being responsible for charging at the session and service flow levels, QoS bandwidth guarantee, mobility management, and terminal device policy decision-making.

[0076] The NRF network element, abbreviated as NRF, can be used to provide network element discovery functions. Based on the requests of other network elements, it provides network element information corresponding to the network element type. The NRF network element also provides network element management services, such as network element registration, update, deregistration, and network element status subscription and push.

[0077] The NWDAF network element, abbreviated as NWDAF, is mainly used to collect data (including one or more of terminal device data, access network device data, core network element data, and third-party application device data). Among them, these data can be the data of the terminal device, access network device, core network element, or third-party application device itself, or the data of the terminal device on the access network device, the core network element, or the third-party application device. Then, it performs data analysis based on the collected data and outputs the data analysis results for use in network, network management device, and application execution policy decisions. NWDAF can use machine learning models for data analysis. In the embodiments of this application, an NWDAF can be a separate network element or co-located with other network elements. For example, the NWDAF can be set in the PCF network element or the AMF network element.

[0078] In the 3rd generation partnership project (3GPP) Release 17, the training function and inference function of the NWDAF are split. An NWDAF can support only the model training function, or only the data inference function, or both the model training function and the data inference function.

[0079] The following exemplarily introduces two NWDAFs that support different functions.

[0080] In this application, the network element with the model training function can be an NWDAF that supports the model training function, also referred to as the training NWDAF, or an NWDAF that supports the model training logical function (MTLF), abbreviated as MTLF. Exemplarily, the MTLF can perform model training based on the acquired data to obtain the trained model.

[0081] The NWDAF can be an NWDAF that supports the data inference function, also referred to as the inference NWDAF, or an NWDAF that supports the analytics logical function (AnLF), abbreviated as AnLF.

[0082] Exemplarily, AnLF can request a model from MTLF through the model subscription (MLModelProvision_Subscribe) service or message. The model can be obtained by MTLF through training based on relevant data of the model. Furthermore, AnLF can input the input data into the trained model to obtain analysis results or inference data. It can be understood that MTLF can be regarded as NWDAF that at least supports the model training function. As a possible implementation method, MTLF can also support the data inference function. AnLF can be regarded as NWDAF that at least supports the data inference function. As a possible implementation method, AnLF can also support the model training function.

[0083] In the embodiments of the present application, AnLF and MTLF will be used as examples for illustration, but it does not constitute a limitation on NAWDF.

[0084] It can be understood that the above network elements are examples of one implementation method. The present application does not exclude that in a 6G or newer wireless communication system, there may be network elements or devices with the functions of the above network elements having other names or other forms.

[0085] It can be understood that the above network elements or functions can be either network elements in hardware devices, software functions running on dedicated hardware, or virtualized functions instantiated on a platform (such as a cloud platform). As a possible implementation method, the above network elements or functions can be implemented by one device, jointly implemented by multiple devices, or can also be a functional module within one device. The embodiments of the present application do not make specific limitations on this.

[0086] Figure 1 Among them, Nudr, Npcf, Namf, Nudm, Nsmf, Naf, and Nnwdaf are service - oriented interfaces provided by the above - mentioned UDR, PCF, AMF, UDM, SMF, AF, and NWDAF respectively, for calling corresponding service - oriented operations. N1, N2, N3, N4, and N6 are interface serial numbers, and the meanings of these interface serial numbers are as follows:

[0087] 1), N1: The interface between the AMF network element and the terminal device, which can be used to transmit non - access stratum (NAS) signaling (such as including QoS rules from the AMF network element) to the terminal device, etc.

[0088] 2), N2: The interface between the AMF network element and the access network device, which can be used to transmit radio bearer control information from the core network side to the access network device, etc.

[0089] 3), N3: The interface between the access network device and the UPF network element, mainly used to transmit uplink and downlink user plane data between the access network device and the UPF network element.

[0090] 4) N4: The interface between the SMF network element and the UPF network element can be used to transfer information between the control plane and the user plane, including the distribution of forwarding rules, QoS rules, traffic statistics rules, etc. from the control plane to the user plane and the information reporting of the user plane.

[0091] 5) N6: The interface between the UPF network element and the DN is used to transfer the uplink and downlink user data streams between the UPF network element and the DN.

[0092] The following explains the basic technical concepts involved in this application:

[0093] 1. Analysis ID: It is used to identify an analysis service (or analysis service, abbreviated as service). In the embodiments of this application, the aforementioned service is related to a data model, that is, the data model can be used to execute the service. In other words, the analysis ID is related to the data model, that is, the data model is used to execute the service corresponding to the analysis ID. The data model can also be called a machine learning model or a training model, that is, a model generated by training using data.

[0094] The following uses an example to illustrate the analysis ID. MTLF is related to the analysis ID, that is, the model provided by MTLF supports the execution of the service corresponding to the analysis ID. Exemplarily, MTLF can be related to one or more analysis IDs. It can be understood that this MTLF can provide models for the services corresponding to each of the one or more analysis IDs. For example, MTLF1 is related to analysis ID 1 and analysis ID 2, that is, MTLF1 corresponds to analysis ID 1 and analysis ID 2, then MTLF1 can provide models for the services corresponding to analysis ID 1 and analysis ID 2.

[0095] 2. Vendor ID: It is used to identify a device manufacturer. One vendor ID can correspond to one or more network elements (network functions (NF)), such as NWDAF, NWDAF containing MTLF. For example, MTLF1 and MTLF2 correspond to vendor ID 1, that is, MTLF1 and MTLF2 belong to the same manufacturer, and the vendor ID of this manufacturer is vendor ID 1.

[0096] For the NF, the vendor ID corresponds to the manufacturer of the network element. It can also be understood as network element information (NFinformation), and the network element information identifies the vendor information of the network element.

[0097] 3. Interoperability Indicator: It is used to identify the list of NWDAF (such as MTLF or AnLF) providers (or suppliers, manufacturers) that can obtain models from each other. Among them, the interoperability indicator can also be called an interoperability indication, or a machine learning (ML) model interoperability indicator, or an MLModel interoperability Indicator.

[0098] For example, the interoperability indicator includes one or more manufacturer identifiers (such as a list of manufacturers or a list of manufacturer identifiers), or the interoperability indicator is associated with the one or more manufacturer identifiers. Suppose MTLF has an interoperability indicator, and this interoperability indicator corresponds to the first list of manufacturer identifiers. Then, the manufacturers in the first list of manufacturer identifiers are allowed to obtain models from MTLF; or, the manufacturers in the first list of manufacturer identifiers are allowed to retrieve or use the models provided by MTLF; or, the manufacturers in the first list of manufacturer identifiers can obtain models from MTLF; or, the manufacturers in the first list of manufacturer identifiers can retrieve or use the models provided by MTLF; or, the interoperability indicator also indicates that MTLF supports the NWDAF of the manufacturers in the first list of manufacturer identifiers to request the models provided by MTLF. Another example is that the interoperability indicator can also correspond to MTLF, or to an analysis identifier, or to the analysis identifier of MLTF; in other words, the interoperability indicator is related to MTLF, or the interoperability indicator is related to the analysis identifier.

[0099] Optionally, an MTLF may have one or more interoperability identifiers. If an MTLF has multiple interoperability identifiers, the multiple interoperability identifiers respectively correspond to different analysis identifiers. For example, assume that the NWDAF network element (taking MTLF NFID 1 as an example) corresponds to analysis identifier 1 and analysis identifier 2. Then, the MTLF to which MTLF NF ID 1 belongs has interoperability identifier 1 and interoperability identifier 2. Interoperability identifier 1 corresponds to analysis identifier 1, and interoperability identifier 2 corresponds to analysis identifier 2. Further taking an example, if the MTLF to which MTLF NF ID 1 belongs and the MTLF to which MTLF NF ID 2 belongs both belong to the list of vendors that support interoperability, then the MTLF to which MTLF NF ID 2 belongs may have interoperability identifier 1 and interoperability identifier 2, where interoperability identifier 1 corresponds to analysis identifier 1, and / or, interoperability identifier 2 corresponds to analysis identifier 2, that is, MTLFs of the same vendor may have the same interoperability identifier for the same analysis identifier. Additionally, if the MTLF to which MTLF NF ID 1 belongs and the MTLF to which MTLF NF ID 2 belongs do not belong to the same list of vendors that support interoperability, then the MTLF to which MTLF NF ID 2 belongs may have interoperability identifier 3 and interoperability identifier 4, where interoperability identifier 3 corresponds to analysis identifier 1, and / or, interoperability identifier 4 corresponds to analysis identifier 2, that is, MTLFs of the same vendor may have different interoperability identifiers for the same analysis identifier.

[0100] Exemplarily, analysis identifier 1 is related to model 1, that is, model 1 is used to execute the service corresponding to analysis identifier 1. Analysis identifier 1 is related to interoperability identifier 1, that is, model 1 is related to interoperability identifier 1. Assume that interoperability identifier 1 includes the identifiers of vendor 1 and vendor 2, that is, model 1 can be provided for use by vendor 1 and vendor 2, or it can be understood that if the vendor of NWDAF is vendor 1 or vendor 2, then this NWDAF can use model 1.

[0101] In order to clarify the client's permission to obtain the global model, ensure the client's effective participation in the federated learning task, and improve the user satisfaction of the client, the embodiments of the present application provide a communication method. This communication method can be implemented in the communication system shown above. This communication method involves a first network element, a second network element, and an NRF network element; the first network element and the second network element may be the NWDAF network element shown above, or may be other network elements, which are not limited in the present application. For example, the first network element may be a client network element, and the second network element may be a server network element. Figure 1 Shown above, or may be other network elements, which are not limited in the present application. For example, the first network element may be a client network element, and the second network element may be a server network element. Figure 1 Shown above, or may be other network elements, which are not limited in the present application. For example, the first network element may be a client network element, and the second network element may be a server network element.

[0102] The communication method provided by the embodiments of the present application will be introduced below with reference to the accompanying drawings. Figure 2A communication method provided by an embodiment of this application may include the following steps:

[0103] S201: A second network element sends a first request to a first network element; correspondingly, the first network element receives the first request from the second network element. The first request is used to request the first network element to participate in a first federated learning task.

[0104] Optionally, the first request may include identification information corresponding to the first federated learning task.

[0105] Optionally, the first federated learning task may be replaced with: other names such as first federated learning, or first federated learning function, or first federated learning process, or first federated learning activity, etc.

[0106] Optionally, "participate" may be replaced with "join". In the embodiment of this application, requesting to participate in the first federated learning task may be understood as: requesting an entity (or network element or device or node, etc.) (such as the first network element) that has the ability to participate in the first federated learning task to participate in the first federated learning task. Or rather, in the embodiment of this application, requesting the first network element to participate in the first federated learning task may be understood as: requesting the first network element to determine that the first network element has the ability to participate in the first federated learning task. Or rather, in the embodiment of this application, requesting the first network element to participate in the first federated learning task may be understood as: requesting the first network element to agree to participate (or join) the first federated learning task.

[0107] S202: The first network element obtains first information.

[0108] Optionally, the first information is used to indicate that the second network element can provide the global model corresponding to the first federated learning task to the first network element. It should be understood that "the second network element can provide the global model corresponding to the first federated learning task to the first network element" includes but is not limited to indicating any of the following situations: the second network element can actively send the global model corresponding to the first federated learning task to the first network element; the second network element can send the global model corresponding to the first federated learning task to the first network element after receiving a request to obtain the global model; the second network element can authorize the first network element to obtain the global model corresponding to the first federated learning task; or, the first network element has the permission to obtain the global model corresponding to the first federated learning task.

[0109] In the embodiment of this application, the global model corresponding to the federated learning task refers to the model generated after the client network element and the server network element jointly execute the federated learning task.

[0110] In some examples, the first information may also indicate the permissions of devices with different granularities (such as multiple network elements including the first network element) to obtain the global model.

[0111] Among them, the first information may include at least one network element identifier, and the at least one network element identifier includes the identifier of the first network element; alternatively, the first information may include at least one analysis identifier, and the at least one analysis identifier includes the analysis identifier corresponding to the first federated learning task; alternatively, the first information may include at least one interoperability identifier, and the at least one interoperability identifier includes the interoperability identifier corresponding to the first network element. It should be understood that the first information includes but is not limited to the identifiers of the foregoing several granularities.

[0112] In this way, the first information can indicate the permissions of devices with different granularities to obtain the global model, which can improve the flexibility of the first information. For example, when the first information includes at least one network element identifier, the first information is the indication information at the network element granularity, that is, the first information indicates that the second network element can provide the global model corresponding to the first federated learning task to the network elements corresponding to the at least one network element identifier. Another example is that when the first information includes at least one analysis identifier, the first information is the indication information at the analysis identifier granularity, that is, the first information indicates that the second network element can provide the global model corresponding to the first federated learning task to the network elements corresponding to the at least one analysis identifier. Another example is that when the first information includes at least one interoperability identifier, the first information is the indication information at the interoperability identifier granularity, and the interoperability identifier includes at least one manufacturer identifier, then the first information indicates that the second network element can provide the global model corresponding to the first federated learning task to the network elements belonging to the manufacturer corresponding to the manufacturer identifier.

[0113] In a possible design, as Figure 3 shown, the implementation manners for the first network element to obtain the first information include but are not limited to manner A1 and manner A2:

[0114] Manner A1 (including S202-a1):

[0115] S202-a1: The second network element sends the first information to the first network element. Correspondingly, the first network element receives the first information from the second network element.

[0116] Optionally, after receiving the first information from the second network element, the first network element may further send a confirmation message of the first information to the second network element.

[0117] Optionally, the first information may be carried in the first request in S201. It should be understood that at this time, S201 and S202-a1 may be the same step; that is to say, when adopting manner A1, S201 and S202 may be combined into one step. Based on this, when the first request includes the first information, the first network element may parse the first information from the first request after executing S201.

[0118] Optionally, the first information may also be carried in other messages sent by the second network element to the first network element.

[0119] Mode A2 (including at least one of the following steps: S202-b1, S202-b2, S202-b3): The second network element may send the first information to the first network element via the NRF network element.

[0120] S202-b2: The first network element sends a second request to the NRF network element; correspondingly, the NRF network element receives the second request from the first network element; the second request is used to query whether the second network element can provide the global model corresponding to the first federated learning task to the first network element. In some examples, the second request in S202-b2 may be a network element discovery request in the network element discovery process of the first network element, and this network element discovery request is used to request to obtain the network element configuration file (NF profile) (including one or more configuration policies) of at least one network element (including the second network element), and one or more configuration policies in this NF profile may include the aforementioned "querying whether the second network element can provide the global model corresponding to the first federated learning task to the first network element". The "network element discovery process" in the embodiments of the present application may refer to the traditional technical solutions in the art and will not be elaborated here.

[0121] S202-b3: The NRF network element sends the first information to the first network element; correspondingly, the first network element receives the first information from the NRF network element. In some examples, the first information in S202-b3 may be carried in the network element discovery response of the network element discovery process, or the first information in S202-b3 is the network element discovery response of the network element discovery process; wherein, the network element discovery response is used to send the NF profile of the second network element.

[0122] In this way, by using the aforementioned S202-b2 and S202-b3, the first network element can obtain the first information through the NRF network element. For example, the first network element can obtain the first information through the network element discovery process.

[0123] Optionally, before executing S202-b2, the second network element may also execute S202-b1.

[0124] S202-b1: The second network element sends the first information to the NRF network element so that the first network element can obtain the first information from the NRF network element. Correspondingly, the NRF network element receives the first information from the second network element. The NRF network element may also store this first information. In some examples, S202-b1 may be implemented in the network element registration process of the second network element. In other words, the first information may be carried in the network element registration request of the second network element. The "network element registration process" in the embodiments of the present application may refer to the traditional technical solutions in the art and will not be elaborated here.

[0125] In this way, the NRF network element can obtain the first information from the second network element for subsequent provision to the first network element.

[0126] S203: The first network element sends a first response to the first request to the second network element according to the first information; correspondingly, the second network element receives the first response from the first network element. The first response is used to instruct the first network element to determine to participate in the first federated learning task.

[0127] By using the method shown in the foregoing S201 to S203, the first network element (i.e., the client) can clarify whether it has the permission to obtain the global model, or clarify whether it will obtain the global model. The first network element can decide whether to participate in the first federated learning task based on whether it can obtain the global model, so as to ensure that the first network element can effectively participate in the federated learning task and improve the user satisfaction of the client.

[0128] In a possible example, after the first network element receives the first request from the second network element, when the first network element needs to obtain the global model corresponding to the first federated learning task and the first network element has not obtained the first information, the first network element sends a response to the second network element indicating that it refuses to participate in the first federated learning task.

[0129] In a possible example, when the first network element does not need to obtain the global model corresponding to the first federated learning task, regardless of whether the first network element has obtained the first information, the first network element can send a first response to the second network element indicating that the first network element determines to participate in the first federated learning task. That is to say, after executing S201, without executing S202, the first network element can send a first response to the second network element indicating that the first network element determines to participate in the first federated learning task.

[0130] In another possible example, when the first network element needs to obtain the global model corresponding to the first federated learning task, the first network element can also not obtain the first information, that is, does not execute S202. In the scenario of this example, the first network element can send a first response to the first request to the second network element without relying on the first information. In other words, after executing S201, the first network element can send a first response to the second network element. In the scenario of the foregoing example, the first response can also be used to instruct the first network element to obtain the global model corresponding to the federated learning task in which the first network element participates. In this way, the first network element (i.e., the client) can send a first response to the second network element indicating that the first network element needs to obtain the global model corresponding to the federated learning task in which the first network element participates, so that the second network element can clarify whether the first network element needs to obtain the global model. Further, the first network element (i.e., the client) can obtain the global model corresponding to the first federated learning task (for example: execute the steps of S205-c1 to S205-c2, or execute the steps of S205-d1 to S205-d2).

[0131] Optionally, the first response includes one or more of the following: first indication information or a model acquisition address. The first indication information is used to indicate that the first network element needs to acquire model information corresponding to the first federated learning task. Alternatively, the first indication information is used to indicate that the first network element requests to acquire model information corresponding to the first federated learning task. The model acquisition address is used to indicate the address for the first network element to acquire model information corresponding to the first federated learning task.

[0132] In some examples, the first network element and the second network element may pre - agree on the following information in advance through standard pre - configuration or negotiation: The first network element's agreement to participate in the first federated learning task indicates that the first network element needs to acquire the global model corresponding to the first federated learning task.

[0133] In the scenario of the foregoing example, the first response can also be used to indicate that the first network element needs to acquire the global model corresponding to the federated learning task in which the first network element participates.

[0134] In the case of no logical conflict, the above - mentioned multiple examples can be cross - referenced and combined into new embodiments, which are not limited in this application.

[0135] In a possible design, as Figure 3 shown, after executing S203 or any of the foregoing examples (not shown in the figure), the first network element and the second network element may further execute S204: The first network element and the second network element execute the first federated learning task (the first federated learning task includes one or more rounds of training). Optionally, the second network element may further send the task identifier of the first federated learning task and / or the model identifier of the global model corresponding to the first federated learning task to the first network element.

[0136] Exemplarily, when performing the first round of training, the second network element may send an initial model to the first network element; the second network element may also send the model identifier of the initial model to the first network element.

[0137] Further, when performing any subsequent round of training, the second network element may send a process model to the first network element; the second network element may also send the model identifier of the process model to the first network element.

[0138] In a possible design, as Figure 3As shown, the first network element and the second network element may also perform the steps from S205-c1 to S205-c2, or the first network element and the second network element may also perform the steps from S205-d1 to S205-d2. That is to say, after the first network element and the second network element complete the first federated learning task (such as performing one or more rounds of training), the first network element may obtain model information; the two ways for the first network element to obtain model information include way B1 and way B2. Among them, the model information may be a global model, or the model information may be parameter information of the global model, or the model information may be information required to obtain the global model.

[0139] Way B1 (including at least one of the following steps: S205-c1, S205-c2):

[0140] S205-c2: The second network element may send the model information to the first network element; correspondingly, the first network element receives the model information from the second network element, and the model information is used to obtain the global model corresponding to the first federated learning task.

[0141] In some embodiments, the second network element may send the model information to all network elements participating in the first federated learning task; or, the second network element may send the model information to all network elements participating in the last round of training in the first federated learning task.

[0142] In this way, the first network element may obtain the model information by way B1, that is, the second network element may actively send the model information to the first network element.

[0143] Optionally, before the second network element sends the model information to the first network element, the first network element may also perform S205-c1.

[0144] S205-c1: The first network element may also send a first address to the second network element; correspondingly, the second network element receives the first address from the first network element. The first address is used to receive the model information. In some examples, the first address in S205-c1 may be carried in the confirmation message sent by the first network element to the second network element in the foregoing way A1.

[0145] In this way, the second network element may send the model information to the first network element through the first address. For example, the second network element may store the model information in the storage space corresponding to the first address, and the second network element may obtain the model information through the storage space corresponding to the first location.

[0146] Way B2 (including S205-d1 and S205-d2):

[0147] S205-d1: The first network element may send a third request to the second network element; correspondingly, the second network element receives the third request from the first network element. The third request is used to request to obtain the model information.

[0148] S205-d2: The second network element may send a second response to the third request to the first network element; correspondingly, the first network element receives the second response to the third request from the second network element. Wherein, the second response includes model information, or the second response includes indication information for indicating that the model acquisition fails.

[0149] In this way, the first network element may obtain model information by way B2, that is, the first network element may send a third request for obtaining model information to the second network element, so that the second network element sends a second response corresponding to the third request to the first network element, and the second response includes model information, improving the flexibility of the first network element to obtain model information.

[0150] In some examples, when executing S204, the first network element may exit the first federated learning task halfway due to internal reasons, that is, the first network element does not participate in the whole process of executing the first federated learning task in S204. At this time, the second network element may not actively send model information to the first network element, and the first network element may obtain model information by sending a third request to the second network element. Assume that when the second network element receives the third request, the first federated learning task is in an unfinished state, then the second network element may send a second response to the first network element, and the second response includes indication information that the model acquisition fails. The second response may further include the error occurrence time, the waiting time (that is, the time interval between receiving the first request and sending the second response), and the error reason (for example, "the first federated learning task is in an unfinished state at the current moment" or "the global model is not generated at the current moment").

[0151] Optionally, the third request may include identification information corresponding to the first federated learning task. At this time, the second response may be determined according to the foregoing identification information.

[0152] In some examples, when the third request includes identification information corresponding to the first federated learning task, the second network element sends a second response to the first network element. On the contrary, when the third request received by the second network element does not include identification information corresponding to the first federated learning task, the second network element may not send the foregoing second response to the first network element, or the second response includes indication information for indicating that the model acquisition fails.

[0153] Wherein, the identification information may include at least one of the following: the model identification of the initial model corresponding to the first federated learning task; the model identification of the model in any round of iterative training process corresponding to the first federated learning task; the model identification of the global model corresponding to the first federated learning task; or, the task identification of the first federated learning task.

[0154] It should be understood that the second network element may execute multiple federated learning tasks, and thus may store global models corresponding to multiple federated learning tasks respectively. In the foregoing method B2, identification information corresponding to the first federated learning task is carried in the third request, and this identification information can be used as an index for the second network element to query the global model. Thus, the second network element can accurately send the model information of the global model of the first federated learning task to the first network element, improving communication accuracy. In addition, carrying the identification information corresponding to the first federated learning task in the third request, this identification information can be used by the second network element to verify the verification value of the first network element, thereby preventing network elements not related to the first federated learning task from obtaining the model information of the global model corresponding to the first federated learning task, improving communication security.

[0155] Optionally, the method for the first network element to obtain the identification information includes but is not limited to method C1 and method C2:

[0156] Method C1: When the first request includes the identification information, the first network element can parse and obtain the identification information from the first request after executing S201.

[0157] Method C2: The second network element sends the identification information to the first network element; correspondingly, the first network element can also receive the identification information from the second network element. For example, during the execution of the first federated learning task by the first network element and the second network element, the first network element and the second network element perform the actions in the foregoing method C2.

[0158] In a possible design, as Figure 3 shown, the first network element can also perform the following steps (including S200-e1):

[0159] S200-e1: The first network element can also send second information to the NRF network element; correspondingly, the NRF network element receives the second information from the first network element. The second information is used to indicate that the first network element needs to obtain the global models corresponding to some or all of the federated learning tasks participated by the first network element. The NRF network element can also store the second information.

[0160] In some embodiments, the second information can also indicate the permissions of the global models corresponding to the federated learning tasks initiated by devices with different granularities (such as multiple network elements including the second network element) for the first network element. For example, the second information can be used to indicate that the first network element needs to obtain the global models corresponding to the federated learning tasks initiated by the network element(s) corresponding to a certain network element identifier; for another example, the second information can be used to indicate that the first network element needs to obtain the global models corresponding to a certain analysis identifier; for another example, the second information can be used to indicate that the first network element needs to obtain the global models corresponding to the federated learning tasks initiated by the network element(s) belonging to a certain interoperability identifier.

[0161] In this way, the second information can indicate the permissions of the global models corresponding to the federated learning tasks initiated by devices with different granularities, which can improve the flexibility of the second information.

[0162] It should be noted that the execution order of the foregoing S200-e1 can be executed before the foregoing S201. In some examples, S200-e1 can be implemented in the network element registration process of the first network element. In other words, the second information can be carried in the network element registration request of the first network element; or, the foregoing S200-e1 can be executed in any one of the links from S201 to S205-d2; or, the foregoing S200-e1 can be executed after the foregoing S205-d2. In this way, the NRF network element can obtain the second information from the first network element.

[0163] Optionally, before executing S201, the second network element can also obtain the second information through the following steps (S200-e2 and S200-e3).

[0164] S200-e2: The second network element can also send a fourth request to the NRF network element; correspondingly, the NRF network element receives the fourth request from the second network element. The fourth request is used to query whether the first network element needs to obtain the global model corresponding to the federated learning task participated by the first network element. In some examples, the fourth request in S200-e2 can be the network element discovery request in the network element discovery process of the second network element. This network element discovery request can be used to request to obtain the NF profile (including one or more configuration policies) of at least one network element (including the first network element), and the one or more configuration policies can include the foregoing "query whether the first network element needs to obtain the global model corresponding to the federated learning task participated by the first network element".

[0165] S200-e3: The NRF network element sends a third response to the fourth request to the second network element; correspondingly, the second network element can also receive the third response to the fourth request from the NRF network element. The third response is used to indicate that the first network element needs to obtain the global model corresponding to the federated learning task participated by the first network element. For example, the third response can include the second information, or the second network element can determine the second information according to the third response. In some examples, the third response in S200-e3 can be carried in the network element discovery response of the network element discovery process, or the third response in S200-e3 is the network element discovery response of the network element discovery process; wherein, the network element discovery response is used to send the NF profile of the first network element.

[0166] In this way, by using the foregoing S200-e2 and S200-e3, the second network element can obtain the second information through the NRF network element. Using the same method, the second network element can obtain the second information corresponding to multiple client network elements through the NRF network element.

[0167] In some examples, the second network element may select some of the multiple client network elements according to the obtained permission requirements of the multiple client network elements (whether it is necessary to obtain the permission to the global model corresponding to the federated learning task participated by the network element), and send a request to participate in the federated learning task to the selected client network elements, so as to improve the effectiveness of the client network elements participating in the federated learning task. Among them, the way for the second network element to select client network elements may be: when the second network element can provide the global model corresponding to the federated learning task to the client network element, the second network element may select some client network elements with the permission requirement to obtain the global model; when the second network element cannot provide the global model corresponding to the federated learning task to the client network element, the second network element may select some client network elements without the permission requirement to obtain the global model.

[0168] Based on the communication method shown in S201 to S205-d2 above, as Figure 4 、 Figure 5 and Figure 6 shown, the present application provides the following three possible examples of communication methods. In the example of this communication method, the network elements executing the federated learning task include: a server network element, a client network element 1, a client network element 2, and an NRF network element.

[0169] Example 1

[0170] Through S401-a1 to S401-a3, the network element registration process of the server network element is executed. This process may refer to S202-b1 to S202-b3 above.

[0171] S401-a1: The server network element sends a network element registration request to the NRF network element; the network element registration request includes the configuration policy of the server network element. This configuration policy can be used to indicate the network element authorized to obtain the global model, that is, this configuration policy can be used to indicate that at least one client network element has the permission to obtain the global model. The configuration policy of the server network element may refer to the description of the first information above. In some examples, the network element registration request initiated by the server network element may include the NF profile of the server network element, and this NF profile may include the foregoing configuration policy.

[0172] Assume that the server network element can provide the global model corresponding to the federated learning task to at least one client network element, then this configuration policy may include the identifiers corresponding to the at least one client network element. For example, this configuration policy includes the identifiers of at least one client network element, the analysis identifiers corresponding to at least one federated learning task, or the interoperability identifiers corresponding to at least one client network element.

[0173] S401-a2: The NRF network element stores the configuration policy of the server network element.

[0174] S401-a3: The NRF network element sends a network element registration response of the server network element to the server network element.

[0175] Through S402-b1 to S402-b3, the network element registration process of client network element 1 is executed. This process can refer to the aforementioned S200-e1 to S200-e3.

[0176] S402-b1: Client network element 1 sends a network element registration request to the NRF network element; the network element registration request includes the configuration policy of this client network element. This configuration policy is used to indicate that client network element 1 needs to obtain the global model corresponding to the federated learning task participated by client network element 1. The configuration policy of this client network element 1 can refer to the aforementioned description of the second information. In some examples, the network element registration request initiated by client network element 1 may include the NF profile of this client network element 1, and this NF profile may include the aforementioned configuration policy.

[0177] S402-b2: The NRF network element stores the configuration policy of client network element 1.

[0178] S402-b3: The NRF network element sends a network element registration response of client network element 1 to client network element 1.

[0179] Optionally, the network element registration process of client network element 2 (or other client network elements) can be executed through actions similar to S402-b1 to S402-b3 (not shown in the figure).

[0180] S403: The server network element can select at least one client network (such as including client network element 1) according to the configuration policies of multiple client network elements, so that the at least one client network element participates in federated learning task A.

[0181] For example, assuming that the configuration policy of the server network element indicates that the client network elements with the permission to obtain the global model corresponding to the federated learning task include client network element 1 (or client network element 2), then the client network elements selected by the server network element may include client network element 1 (or client network element 2).

[0182] For another example, assume that the configuration policy of the server network element indicates that the client network elements with the permission to obtain the global model corresponding to the federated learning task do not include client network element 1 and client network element 2. The configuration policy of client network element 1 indicates that client network element 1 needs to obtain the global model corresponding to the federated learning task in which client network element 1 participates. The configuration policy of client network element 2 indicates that client network element 2 does not need to obtain the global model corresponding to the federated learning task in which client network element 2 participates (or the configuration policy of client network element 2 does not indicate that client network element 2 needs to obtain the global model corresponding to the federated learning task in which client network element 2 participates). Then, the client network elements selected by the server network element include client network element 1 and do not include client network element 2.

[0183] Optionally, before executing S403, the server network element can also obtain the configuration policies of the foregoing multiple client network elements through a network element discovery process (not shown in the figure). The method for the server network element to obtain the configuration policies of multiple client network elements through the network element discovery process can refer to the foregoing S200-e2 and S200-e3, which will not be elaborated here. Alternatively, before executing S403, the server network element can also pre-store the configuration policies of multiple client network elements.

[0184] Through S404-c1 to S404-c3, determine whether client network element 1 participates in federated learning task A. This process can refer to the foregoing S201 and S203.

[0185] S404-c1: The server network element sends a request to client network element 1 to participate in federated learning task A.

[0186] S404-c2: Client network element 1 determines a response (agree to participate or refuse to participate) to participate in federated learning task A according to the configuration policy of the server network element. It should be understood that S404-c2 can refer to the process of determining the first response in the foregoing S203. Among them, the configuration policy can include the first information in the foregoing communication method, or the configuration policy can be the first information in the foregoing communication method.

[0187] For example, assume that the configuration policy of client network element 1 indicates that client network element 1 does not need to obtain the global model corresponding to the federated learning task in which client network element 1 participates (or the configuration policy of client network element 1 does not indicate that client network element 1 needs to obtain the global model corresponding to the federated learning task in which client network element 1 participates). Then, client network element 1 determines a response for indicating agreement to participate in federated learning task A.

[0188] For another example, assume that the configuration policy of client network element 1 indicates that client network element 1 needs to obtain the global model corresponding to the federated learning task in which client network element 1 participates, and the configuration policy of the server indicates that client network element 1 has the permission to obtain the global model. Then, client network element 1 determines a response for indicating consent to participate in federated learning task A. Conversely, assume that the configuration policy of client network element 1 indicates that client network element 1 needs to obtain the global model corresponding to the federated learning task in which client network element 1 participates, and the configuration policy of the server indicates that client network element 1 does not have the permission to obtain the global model. Then, client network element 1 determines a response for indicating refusal to participate in federated learning task A.

[0189] Before executing S404-c2, client network element 1 can also obtain the configuration policy of the aforementioned server network element through a network element discovery process (not shown in the figure). The method by which client network element 1 can also obtain the configuration policy of the aforementioned server network element through the network element discovery process can refer to S202-b2 and S202-b3 in the aforementioned method A2. Alternatively, before executing S404-c2, client network element 1 can also pre-store the configuration policy of the aforementioned server network element.

[0190] S404-c3: Client network element 1 sends a response to participate in federated learning A to the server network element.

[0191] Optionally, through actions similar to S404-c1 to S404-c3, the server network element can determine whether other client network elements among the at least one client network element selected in S403 participate in federated learning task A (not shown in the figure), so as to determine at least one client network element participating in federated learning task A.

[0192] Through S405-d1 and S405-d2, model training is iteratively executed, that is, federated learning task A is executed. This process can refer to the aforementioned S204.

[0193] S405-d1: The server network element sends a model training message to client network element 1. In the first iteration, the model training message includes an initial model and a model identifier corresponding to the initial model; in subsequent iterations, the model training message includes a process model and a model identifier corresponding to the process model.

[0194] S405-d2: Client network element 1 sends a training feedback message to the server network element.

[0195] It should be understood that the process model in S405-d1 is generated based on the training feedback information of at least one client network element in the previous round of training.

[0196] Optionally, through actions similar to S405-d1 and S405-d2, the server network element and at least one client network element can iteratively perform multiple rounds of model training to determine the global model corresponding to the federated learning task A.

[0197] During the foregoing iterative model training process, some client network elements (such as client network element 1) can participate in the federated learning task A throughout the process, and some client network elements (such as client network element 2) may drop out of the federated learning task A midway due to internal reasons.

[0198] S406: The server network element saves the model information of the global model obtained after multiple rounds of iterative training.

[0199] Through S407-e1 (or S407-f1 and S407-f2), client network element 1 obtains the model information. This process can refer to the foregoing S205-c1 to S205-c2, or refer to the foregoing S205-d1 to S205-d2.

[0200] S407-e1: The server network element sends the model information to client network element 1. This process can refer to the foregoing method B1.

[0201] Optionally, the server network element can also send the model information to other client network elements participating in the federated learning task A (not shown in the figure).

[0202] S407-f1: Client network element 1 sends a global model request to the server network element. The global model request can include the model identifier of the initial model and / or the model identifier of the process model in S405-d1. This process can refer to S205-d1 in the foregoing method B2.

[0203] S407-f2: The server network element sends the model information to client network element 1. This process can refer to S205-d2 in the foregoing method B2.

[0204] For example, assume that the server network element determines that the foregoing global model request does not include the model identifier of the initial model and does not include the model identifier of the process model. Then the server network element does not execute S407-f2, or the server network element sends indication information for indicating that the global model acquisition fails to client network element 1.

[0205] Optionally, through actions similar to S407-f1 and S407-f2, other client network elements (such as client network element 2) can obtain the foregoing model information (not shown in the figure).

[0206] Example 2

[0207] Through S501-a1 to S501-a3, perform the network element registration process of the server network element. This process can refer to the aforementioned S401-a1 to S401-a3.

[0208] Through S502-b1 to S502-b3, perform the network element registration process of client network element 1. This process can refer to the aforementioned S402-b1 to S402-b3. Optionally, through actions similar to S502-b1 to S502-b3, the server network element and client network element 2 (or other client network elements) can also perform the network element registration process of client network element 2 (or other client network elements) (not shown in the figure).

[0209] S503: The server network element can select at least one client network (such as including client network element 1) according to the configuration policies of multiple client network elements. This process can refer to the aforementioned S403.

[0210] Through S504-c1 to S504-c3, determine whether client network element 1 participates in the federated learning task A. This process can refer to the aforementioned S404-c1 to S404-c3. Optionally, through actions similar to S504-c1 to S504-c3, the server network element can determine whether other client network elements among the at least one client network element selected in S503 participate in the federated learning task A (not shown in the figure), so as to determine at least one client network element participating in the federated learning task A.

[0211] Through S505-d1 and S505-d2, iteratively perform model training, that is, perform the federated learning task A. This process can refer to the aforementioned S405-d1 and S405-d2. Optionally, through actions similar to S505-d1 and S505-d2, the server network element and at least one client network element can iteratively perform multiple rounds of model training, so as to determine the global model corresponding to the federated learning task A.

[0212] Among them, S505-d1 includes: the server network element sends a model training message to client network element 1; the difference between S505-d1 and S405-d1 in Example 1 is that in each iteration, the model training message can include the model identifier of the global model corresponding to the federated learning task A. Optionally, in the first iteration, the model training message can also include the initial model and the model identifier corresponding to the initial model; in subsequent iterations, the model training message can also include the process model and the model identifier corresponding to the process model.

[0213] S506: The server network element saves the model information of the global model obtained after multiple rounds of iterative training. This process can refer to the aforementioned S406.

[0214] Through S507-e1 (or S507-f1 and S507-f2), the client network element 1 obtains model information. This process can refer to the aforementioned S407-e1 (or S407-f1 and S407-f2). Optionally, through an action similar to S507-e1 (or S507-f1 and S507-f2), other client network elements (such as client network element 2) can obtain the aforementioned model information (not shown in the figure).

[0215] Among them, S507-f2 includes: the server network element sends model information to the client network element 1. For example, the difference between S507-f2 and S407-f2 in Example 1 is that, assuming that the server network element determines that the model identifier of the global model is not included in the aforementioned global model request, the server network element does not execute S507-f2, or the server network element sends indication information for indicating the failure of global model acquisition to the client network element 1.

[0216] Example 3

[0217] Through S601-a1 to S601-a3, the network element registration process of the server network element is executed. This process can refer to the aforementioned S401-a1 to S401-a3.

[0218] Through S602-b1 to S602-b3, the network element registration process of the client network element 1 is executed. This process can refer to the aforementioned S402-b1 to S402-b3. Optionally, through an action similar to S602-b1 to S602-b3, the server network element and the client network element 2 (or other client network elements) can also execute the network element registration process of the client network element 2 (or other client network elements) (not shown in the figure).

[0219] S603: The server network element can select at least one client network (such as including the client network element 1) according to the configuration policies of multiple client network elements. This process can refer to the aforementioned S403.

[0220] Through S604-c1 to S604-c3, it is determined whether the client network element 1 participates in the federated learning task A. This process can refer to the aforementioned S404-c1 to S404-c3. Optionally, through an action similar to S604-c1 to S604-c3, the server network element can also determine whether other client network elements among the at least one client network element selected in S603 participate in the federated learning task A (not shown in the figure), so as to determine at least one client network element participating in the federated learning task A.

[0221] Among them, S604-c3 includes: The client network element 1 sends a response to participate in Federated Learning A to the server network element. The difference between S604-c3 and S404-c3 in the first example and S504-c3 in the second example is that the response to participate in Federated Learning A may include the global model receiving address.

[0222] Through S605-d1 and S605-d2, the model training is iteratively executed, that is, the Federated Learning task A is executed. This process can refer to S405-d1 and S405-d2 in the foregoing first example, or refer to S505-d1 and S505-d2 in the foregoing second example. Optionally, through actions similar to S605-d1 and S605-d2, the server network element and at least one client network element can iteratively execute multiple rounds of model training to determine the global model corresponding to the Federated Learning task A.

[0223] Among them, S605-d2 includes: The client network element 1 sends a training feedback message to the server network element. The difference between S605-d2 and S405-d2 in the first example and S505-d2 in the second example is that the training feedback message may include the global model receiving address.

[0224] S606: The server network element saves the model information of the global model obtained after multiple rounds of iterative training. This process can refer to the foregoing S406.

[0225] Through S607-e1, the client network element 1 obtains the model information. This process can refer to the foregoing S407-e1. Optionally, the server network element can also send the model information to other client network elements participating in the Federated Learning task A (not shown in the figure).

[0226] Among them, S607-e1 includes: The server network element sends the model information to the client network element 1. The difference between S607-e1 and S407-e1 in the first example and S507-e1 in the second example is that when the "response to participate in Federated Learning A" in S604-c3 or the "training feedback message" in S605-d2 carries the global model receiving address, the server network element can send the model information to the global model receiving address, so as to achieve the purpose of the server network element sending the model information to the client network element 1.

[0227] In each embodiment of the present application, if there is no special description and logical conflict, the terms and / or descriptions between different embodiments are consistent and can be referenced to each other. The technical features in different embodiments can be combined to form new embodiments according to their internal logical relationships.

[0228] The method provided by the embodiments of the present application is introduced above in combination with the accompanying drawings. The communication device provided by the embodiments of the present application is introduced below in combination with the accompanying drawings.

[0229] Based on the same inventive concept, the present application also provides a communication device for implementing the communication method provided in the above embodiments. Referring to Figure 7 as shown, the communication device 700 includes a communication unit 701 and a processing unit 702. The communication unit 701 is configured to receive and / or transmit data; the processing unit 702 is configured to implement Figure 2 or Figure 3 the steps in the communication method as shown.

[0230] In a possible example, when the communication device 700 is used to implement the function of the above-mentioned Figure 2 or Figure 3 shown first network element, the communication unit 701 is configured to: receive a first request from a second network element, where the first request is used to request the communication device 700 to participate in a first federated learning task; the communication unit 701 is further configured to: obtain first information, where the first information is used to indicate that the second network element can provide the global model corresponding to the first federated learning task to the communication device 700; the processing unit 702 is configured to control the communication unit 701 to perform the following steps through the communication unit 701: according to the first information, control the communication unit 701 to send a first response to the first request to the second network element, where the first response is used to indicate that the communication device 700 determines to participate in the first federated learning task. The processing unit 702 may be configured to control the transceiver operations of the communication unit 701.

[0231] In a possible design, the first information includes at least one network element identifier, and the at least one network element identifier includes the identifier of the communication device 700; alternatively, the first information includes at least one analysis identifier, and the at least one analysis identifier includes the analysis identifier corresponding to the first federated learning task; alternatively, the first information includes at least one interoperability identifier, and the at least one interoperability identifier includes the interoperability identifier corresponding to the communication device 700.

[0232] In a possible design, the communication unit 701 may further be configured to: send a second request to the NRF network element, where the second request is used to query whether the second network element can provide the global model corresponding to the first federated learning task to the communication device 700; receive the first information from the NRF network element.

[0233] In a possible design, the first request includes the first information.

[0234] In a possible design, the communication unit 701 is further configured to: receive model information from the second network element, where the model information is used to obtain the global model corresponding to the first federated learning task.

[0235] In a possible design, the communication unit 701 is further configured to: before receiving model information from a second network element, send a first address to the second network element, where the first address is used to receive the model information.

[0236] In a possible design, the communication unit 701 is further configured to: send a third request to the second network element, where the third request is used to request to obtain model information, and the model information is used to obtain the global model corresponding to the first federated learning task; the communication unit 701 is further configured to: receive a second response to the third request from the second network element, where the second response includes the model information, or the second response includes indication information indicating that the model acquisition fails.

[0237] In a possible design, the third request includes identification information corresponding to the first federated learning task, and the second response is determined according to the identification information.

[0238] In a possible design, the first request includes identification information; and / or, the communication unit 701 is further configured to: receive identification information from the second network element.

[0239] In a possible design, the identification information includes at least one of the following: the model identification of the initial model corresponding to the first federated learning task; the model identification of the model in any round of iterative training process corresponding to the first federated learning task; the model identification of the global model corresponding to the first federated learning task; or, the task identification of the first federated learning task.

[0240] In a possible design, the communication unit 701 is further configured to: send second information to the NRF network element, where the second information is used to indicate that the communication device 700 needs to obtain the global model corresponding to the federated learning task participated by the communication device 700.

[0241] In a possible example, when the communication device 700 is used to implement the functions of the second network element shown above Figure 2 or Figure 3 the communication unit 701 is configured to: send a first request to the first network element, where the first request is used to request the first network element to participate in the first federated learning task; the communication unit 701 is further configured to: send first information to the first network element, where the first information is used to indicate that the communication device 700 can provide the global model corresponding to the first federated learning task to the first network element; the communication unit 701 is further configured to: receive a first response from the first network element, where the first response is the response information of the first request, and the first response is used to indicate that the first network element determines to participate in the first federated learning task; the first response is determined according to the first information. The processing unit 702 may be configured to control the transceiver operations of the communication unit 701.

[0242] In a possible design, the first information includes at least one network element identifier, and the at least one network element identifier includes the identifier of the first network element; alternatively, the first information includes at least one analysis identifier, and the at least one analysis identifier includes the analysis identifier corresponding to the first federated learning task; alternatively, the first information includes at least one interoperability identifier, and the at least one interoperability identifier includes the interoperability identifier corresponding to the first network element.

[0243] In a possible design, the communication unit 701 is further configured to: send the first information to the NRF network element, so that the first network element obtains the first information from the NRF network element.

[0244] In a possible design, the first request includes the first information.

[0245] In a possible design, the communication unit 701 is further configured to: send model information to the first network element, where the model information is used to obtain the global model corresponding to the first federated learning task.

[0246] In a possible design, the communication unit 701 is further configured to: before sending the model information to the first network element, receive a first address from the first network element, where the first address is used for the first network element to receive the model information.

[0247] In a possible design, the communication unit 701 is further configured to: receive a third request from the first network element, where the third request is used to request to obtain model information, and the model information is used to obtain the global model corresponding to the first federated learning task; the communication unit 701 is further configured to: send a second response to the third request to the first network element; the second response includes the model information, or the second response includes indication information for indicating that the model acquisition fails.

[0248] In a possible design, the communication unit 701 may further be configured to: when the third request includes the identification information corresponding to the first federated learning task, send the second response to the first network element.

[0249] In a possible design, the first request includes identification information; and / or, the communication unit 701 is further configured to: send the identification information to the first network element.

[0250] In a possible design, the identification information includes at least one of the following: the model identifier of the initial model corresponding to the first federated learning task; the model identifier of the model in any round of iterative training process corresponding to the first federated learning task; the model identifier of the global model corresponding to the first federated learning task; or, the task identifier of the first federated learning task.

[0251] In a possible design, the communication unit 701 is further configured to: before sending a first request to a first network element, send a fourth request to the NRF network element, where the fourth request is used to query whether the first network element needs to obtain a global model corresponding to a federated learning task participated by the first network element; and receive a third response to the fourth request from the NRF network element, where the third response is used to indicate that the first network element needs to obtain a global model corresponding to a federated learning task participated by the first network element.

[0252] In a possible example, when the communication device 700 is used to implement the functions of the NRF network element shown above Figure 2 or Figure 3 the communication unit 701 is configured to: receive a second request from a first network element, where the second request is used to query whether a second network element can provide a global model corresponding to a first federated learning task to the first network element; the communication unit 701 is further configured to: send a first piece of information to the first network element, where the first piece of information is used to indicate that the second network element can provide a global model corresponding to a first federated learning task to the first network element. The processing unit 702 may be configured to control the transceiver operations of the communication unit 701.

[0253] In a possible design, the communication unit 701 is further configured to: receive a first piece of information from a second network element.

[0254] In a possible design, the communication unit 701 is further configured to: receive a fourth request from a second network element, where the fourth request is used to query whether a first network element needs to obtain a global model corresponding to a federated learning task participated by the first network element; the communication unit 701 is further configured to: send a third response to the fourth request to the second network element, where the third response is used to indicate that the first network element needs to obtain a global model corresponding to a federated learning task participated by the first network element.

[0255] In a possible design, the communication unit 701 is further configured to: receive a second piece of information from a first network element, where the second piece of information is used to indicate that the first network element needs to obtain a global model corresponding to a federated learning task participated by the first network element.

[0256] Based on the same technical concept, an embodiment of the present application further provides another communication device 800, and the communication device 800 may implement the communication method provided in the above embodiments. Refer to Figure 8 As shown, the communication device 800 includes a processor 801. Optionally, the communication device 800 further includes a memory 802 and / or a communication interface 803. The memory may be placed inside the communication device or outside the communication device, which is not limited in this application. Among them, the communication interface 803, the processor 801, and the memory 802 are interconnected with each other. Exemplarily, the communication device 800 may be the first network element, the second network element, or the NRF network element shown in the embodiments of the present application.

[0257] Optionally, the communication interface 803, the processor 801, and the memory 802 are interconnected with each other through a bus 804. The bus 804 may be a peripheral component interconnect (PCI) bus, an extended industry standard architecture (EISA) bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For the sake of simplicity of representation, Figure 8 it is only represented by a thick line in the figure, but it does not mean that there is only one bus or one type of bus.

[0258] The communication interface 803 is configured to receive and / or transmit signals to implement communication with other devices outside the communication device.

[0259] The processor 801 may be configured to execute any of the foregoing Figures 2 to 6 communication methods. The communication method may refer to the descriptions in the foregoing embodiments and will not be elaborated herein. Among them, the processor 801 may be a central processing unit (CPU), a network processor (NP), or a combination of a CPU and an NP, etc. The processor 801 may further include a hardware chip. The foregoing hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The foregoing PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof. When implementing the foregoing functions, the processor 801 may be implemented by hardware, and of course, it may also implement the corresponding software through hardware.

[0260] The memory 802 is used to store program instructions and the like. Specifically, the program instructions may include program code, and the program code includes computer operation instructions. The memory 802 may include a random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The processor 801 executes the program instructions stored in the memory 802 to implement the above functions, thereby implementing the method provided in the above embodiments.

[0261] Based on the same technical concept, an embodiment of the present application further provides a computer program, which, when running on a computer, causes the computer to execute the method provided in the above embodiments.

[0262] Based on the same technical concept, an embodiment of the present application further provides a computer-readable storage medium, in which a computer program is stored, and when the computer program runs on a computer, it causes the computer to execute the method provided in the above embodiments.

[0263] Among them, the storage medium may be any available medium that can be accessed by a computer. Taking this as an example but not limited to: the computer-readable medium may include RAM, ROM, EEPROM, CD-ROM or other optical disc storage, magnetic disk storage medium or other magnetic storage devices, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer.

[0264] Based on the same technical concept, an embodiment of the present application further provides a chip, which is used to read the computer program stored in the memory and implement the method provided in the above embodiments.

[0265] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program code.

[0266] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to the application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices produce means for implementing the functions specified in a process Figure 1 a process or multiple processes and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0267] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means that implement the functions specified in a process Figure 1 a process or multiple processes and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0268] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operating steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in a process Figure 1 a process or multiple processes and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0269] Obviously, those skilled in the art can make various changes and modifications to this application without departing from the scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalent technologies, this application is also intended to include these changes and modifications.

Claims

1. A communication method, applied to a first network element, characterized in that The method includes: Receiving a first request from a second network element, where the first request is used to request the first network element to participate in a first federated learning task; Obtaining first information, where the first information is used to indicate that the second network element can provide the global model corresponding to the first federated learning task to the first network element; According to the first information, sending a first response to the first request to the second network element, where the first response is used to indicate that the first network element determines to participate in the first federated learning task.

2. The method according to claim 1, wherein: The first information includes at least one network element identifier, and the at least one network element identifier includes the identifier of the first network element; or The first information includes at least one analysis identifier, and the at least one analysis identifier includes the analysis identifier corresponding to the first federated learning task; or The first information includes at least one interoperability identifier, and the at least one interoperability identifier includes the interoperability identifier corresponding to the first network element.

3. The method according to claim 1 or 2, characterized in that, The obtaining of the first information includes: Sending a second request to a Network Repository Function (NRF) network element, where the second request is used to query whether the second network element can provide the global model corresponding to the first federated learning task to the first network element; Receiving the first information from the NRF network element.

4. The method according to claim 1 or 2, wherein: The first request includes the first information.

5. The method according to any one of claims 1-4, characterized in that The method further includes: Receiving model information from the second network element, where the model information is used to obtain the global model corresponding to the first federated learning task.

6. The method according to claim 5, wherein Before receiving the model information from the second network element, it further includes: Sending a first address to the second network element, where the first address is used to receive the model information.

7. The method according to any one of claims 1-4, characterized in that, The method further includes: Sending a third request to the second network element, where the third request is used to request to obtain model information, and the model information is used to obtain the global model corresponding to the first federated learning task; Receiving a second response to the third request from the second network element, where the second response includes the model information, or the second response includes indication information for indicating that the model acquisition fails.

8. The method according to claim 7, wherein: The third request includes the identification information corresponding to the first federated learning task, and the second response is determined according to the identification information.

9. The method according to claim 8, wherein: The first request includes the identification information; and / or The method further includes: receiving the identification information from the second network element.

10. The method according to claim 8 or 9, characterized in that, The identification information includes at least one of the following: The model identifier of the initial model corresponding to the first federated learning task; The model identifier of the model in any round of iterative training process corresponding to the first federated learning task; The model identifier of the global model corresponding to the first federated learning task; or The task identifier of the first federated learning task.

11. The method according to any one of claims 1 to 10, characterized in that, The method further includes: Sending second information to the NRF network element, where the second information is used to indicate that the first network element needs to obtain the global model corresponding to the federated learning task participated by the first network element.

12. A communication method, applied to a second network element, characterized in that The method includes: Sending a first request to a first network element, where the first request is used to request the first network element to participate in a first federated learning task; Sending first information to the first network element, where the first information is used to indicate that a second network element can provide a global model corresponding to the first federated learning task to the first network element; Receiving a first response from the first network element, where the first response is a response message of the first request, and the first response is used to indicate that the first network element determines to participate in the first federated learning task; the first response is determined according to the first information.

13. The method according to claim 12, wherein The first information includes at least one network element identifier, and the at least one network element identifier includes an identifier of the first network element; or The first information includes at least one analysis identifier, and the at least one analysis identifier includes an analysis identifier corresponding to the first federated learning task; or The first information includes at least one interoperability identifier, and the at least one interoperability identifier includes an interoperability identifier corresponding to the first network element.

14. The method according to claim 12 or 13, characterized in that, The method further includes: Sending the first information to a Network Repository Function (NRF) network element.

15. The method according to claim 12 or 13, wherein The first request includes the first information.

16. The method according to any one of claims 12 - 15, characterized in that, The method further includes: Sending model information to the first network element, where the model information is used to obtain a global model corresponding to the first federated learning task.

17. The method according to claim 16, wherein Before sending the model information to the first network element, it further includes: Receiving a first address from the first network element, where the first address is used for the first network element to receive the model information.

18. The method according to any one of claims 12-15, characterized in that, The method further includes: Receiving a third request from the first network element, where the third request is used to request to obtain model information, and the model information is used to obtain a global model corresponding to the first federated learning task; Sending a second response to the third request of the first network element; the second response includes the model information, or the second response includes indication information for indicating that model acquisition fails.

19. The method according to claim 18, wherein Sending the second response to the third request of the first network element includes: When the third request includes identification information corresponding to the first federated learning task, sending the second response to the first network element.

20. The method according to claim 19, wherein The first request includes the identification information; and / or The method further includes: sending the identification information to the first network element.

21. The method according to claim 19 or 20, characterized in that, The identification information includes at least one of the following: A model identifier of an initial model corresponding to the first federated learning task; A model identifier of a model in any round of iterative training process corresponding to the first federated learning task; A model identifier of a global model corresponding to the first federated learning task; or A task identifier of the first federated learning task.

22. The method according to any one of claims 12-21, characterized in that, Before sending the first request to the first network element, it further includes: Sending a fourth request to a Network Repository Function (NRF) network element, where the fourth request is used to query whether the first network element needs to obtain a global model corresponding to a federated learning task participated by the first network element. Receive a third response to the fourth request from the NRF network element, where the third response is used to indicate that the first network element needs to obtain the global model corresponding to the federated learning task participated in by the first network element.

23. A communication method, applied to a Network Repository Function (NRF) network element, characterized in that, The method includes: Receive a second request from a first network element, where the second request is used to query whether a second network element can provide the global model corresponding to a first federated learning task to the first network element; Send first information to the first network element, where the first information is used to indicate that the second network element can provide the global model corresponding to the first federated learning task to the first network element.

24. The method according to claim 23, wherein The method further includes: Receive the first information from the second network element.

25. The method according to claim 23 or 24, characterized in that The method further includes: Receive a fourth request from the second network element, where the fourth request is used to query whether the first network element needs to obtain the global model corresponding to the federated learning task participated in by the first network element; Send a third response to the fourth request to the second network element, where the third response is used to indicate that the first network element needs to obtain the global model corresponding to the federated learning task participated in by the first network element.

26. The method according to any one of claims 23-25, characterized in that, The method further includes: Receive second information from the first network element, where the second information is used to indicate that the first network element needs to obtain the global model corresponding to the federated learning task participated in by the first network element.

27. A communication device, characterized in that, Includes: A communication unit and a processing unit; The communication unit is used to receive and / or send data; The processing unit is used to execute the method according to any one of claims 1-11, or is used to execute the method according to any one of claims 12-22; or is used to execute the method according to any one of claims 23-26.

28. A communication device, characterized in that, Includes: At least one processor; The at least one processor is used to execute the method according to any one of claims 1-11, or is used to execute the method according to any one of claims 12-22; or is used to execute the method according to any one of claims 23-26.

29. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are called by a computer, the computer is used to execute the method according to any one of claims 1-11, or is used to execute the method according to any one of claims 12-22, or is used to execute the method according to any one of claims 23-26.

30. A chip system, characterized in that, Includes a processor; The processor is used to execute a computer-executable program, so that the device installed with the chip system is used to execute the method according to any one of claims 1-11, or is used to execute the method according to any one of claims 12-22, or is used to execute the method according to any one of claims 23-26.

Images