Communication method and communication apparatus

Abstract

Provided in the embodiments of the present application are a communication method and a communication apparatus. The method comprises: a vertical federated learning server receiving first indication information; the vertical federated learning server determining second information on the basis of the first indication information; and the vertical federated learning server sending a first message to a vertical federated learning client, so as to instruct the vertical federated learning client to perform model training, wherein the first message comprises the second information. In this way, the inconsistency between a server and a client caused by a vertical federated learning server exclusively selecting and determining initial model information is avoided, thereby achieving the technical effect of negotiating initial model information between the server and the client, achieving the effect of optimized transmission of the initial model information, and reducing signaling overheads of re-negotiation.

Description

A communication method and communication device

[0001] This application claims priority to Chinese Patent Application No. 202410696326.6, filed with the State Intellectual Property Office of China on May 30, 2024, entitled “A Communication Method and Communication Device”, the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of communications, and more particularly to communication methods and communication devices. Background Technology

[0003] Functional network elements in communication networks can utilize artificial intelligence (AI) technology and big data in the communication network to output analytical data through model training and inference, thereby assisting in the formulation of communication network strategies and the adjustment of network resources.

[0004] For functional network elements in a carrier network to obtain datasets, data privacy protection may prevent these elements from accessing the data provided by the application provider, leading to infeasible or low-accuracy model training. To address this, distributed joint modeling can be used to break down data silos. For example, multiple functional network elements (or participants) can collaborate on modeling without sharing raw data through vertical federated learning (VFL), achieving AI collaboration.

[0005] For a specific federated learning task, all VFL participants need to negotiate and determine a machine learning (ML) model before model training begins. Each participant then performs machine learning based on the successfully negotiated ML model to obtain analytical data and assist in the network communication process.

[0006] However, the question of how to negotiate and determine a reasonable ML model between the Vertical Federated Learning Server and the Vertical Federated Learning Client, or among multiple participants in Vertical Federated Learning, remains to be studied. Summary of the Invention

[0007] The communication method and communication device provided in this application embodiment negotiate and determine a reasonable machine learning model between the vertical federated learning server and the vertical federated learning client, or among multiple vertical federated learning participants.

[0008] To achieve the above objectives, the embodiments of this application adopt the following technical solutions:

[0009] Firstly, a communication method is provided. This method can be executed by components of a network data analysis function network element, such as a processor, chip, or chip system of the network data analysis function network element, or by a logic module or software capable of implementing all or part of the network data analysis function network element. The method includes: a vertical federated learning server receiving first instruction information; the vertical federated learning server determining second information based on the first instruction information; the vertical federated learning server sending a first message to a vertical federated learning client to instruct the vertical federated learning client to perform model training; the first message including the second information.

[0010] In this embodiment of the invention, the preparation of a vertical federated learning server to start a training task includes being triggered by the vertical federated learning server based on internal conditions, or being triggered by the vertical federated learning server receiving an external instruction.

[0011] As an optional step in this embodiment, the vertical federated learning server initiates the process of discovering the vertical federated learning client.

[0012] In one possible implementation, the vertical federated learning server sends a discovery request message to the network repository function (NRF) element. This message carries information required by the vertical federated learning server regarding vertical federated learning, including information related to the application function type. This application function type information is used by the NRF element to discover and select the vertical federated learning client requested by the server. Accordingly, the NRF element obtains the application function (AF) identifier based on the discovery request message and sends it to the vertical federated learning server, carrying the AF identifier in a discovery request response message. The application function (AF) identifier serves as the identifier for the vertical federated learning client requested by the server.

[0013] The vertical federated learning server sends a vertical federated learning preparation request to the vertical federated learning client, which transmits one or more preparation information to the vertical federated learning client, including vertical federated learning task association identifier, vertical federated learning training preparation information, and vertical federated learning filtering information.

[0014] In possible implementations, the aforementioned preparation request message is sent by the vertical federated learning server to the vertical federated client through the network exposure function (NEF) network element, or it is sent directly by the vertical federated learning server to the vertical federated client.

[0015] The vertical federation client sends a vertical federation learning preparation response message to the vertical federation learning server.

[0016] In possible implementations, the aforementioned preparation response message is sent by the vertical federated learning client to the vertical federated server through the network exposure function (NEF) network element, or directly by the vertical federated learning client to the vertical federated server.

[0017] In one possible implementation, the longitudinal federated learning preparation response message includes first indication information. Optionally, the longitudinal federated learning preparation response message may further include application function (AF) identification information and / or model training feature space information.

[0018] The first indication information specifically includes:

[0019] The longitudinal federated learning client already has information about the initial model, or

[0020] The longitudinal federated learning client does not have information about the initial model, or

[0021] The longitudinal federated learning client requests information about the initial model, or

[0022] Vertical federated learning clients do not need to obtain information about the initial model, or

[0023] The longitudinal federated learning client refuses to receive information about the initial model, or

[0024] Vertical federated learning clients request information about the new model, or

[0025] The longitudinal federated learning client does not support information from the second initial model.

[0026] The second initial model is the initial model information that the vertical federated learning client has obtained before receiving the federated learning preparation request message mentioned above. It can be obtained by being generated locally by the federated learning client or by being sent to the federated learning client by the federated learning server.

[0027] As a possible implementation of the first instruction information, it can be explicit or implicit.

[0028] For example, if the vertical federated learning preparation response message does not contain explicit indication information, it is assumed by default that the vertical federated learning client needs to obtain the initial model information, while the message additionally carries indication information 0 or 1 to indicate that it does not need to obtain / rejects obtaining the initial model information.

[0029] The vertical federated learning server determines the second information based on the first instruction information. The second information includes specific vertical federated initial model information or information that does not require sending the initial model to the vertical federated learning client.

[0030] The vertical federated learning server sends a vertical federated learning training request message to the vertical federated learning client. The vertical federated learning training request message includes certain second information, which includes vertical federated initial model information, in order to instruct the vertical federated learning client to perform model training.

[0031] In a possible implementation, if the first indication information indicates that the vertical federated learning client already has the initial model information or the vertical federated learning client does not need to obtain the initial model information, then the vertical federated learning server will no longer send the initial model information to the vertical federated learning client in the vertical federated learning training request message.

[0032] In possible implementations, the aforementioned vertical federated training request message is sent from the vertical federated learning server to the vertical federated client via a network exposure function (NEF) network element, or it is sent directly from the vertical federated learning server to the vertical federated learning client.

[0033] The vertical federated learning client receives the initial vertical federated model information and begins local model training.

[0034] The vertical federated client sends a vertical federated learning training response message to the vertical federated learning server.

[0035] In possible implementations, the training response message is sent from the vertical federated learning client to the vertical federated learning server via a network exposure function (NEF) element, or directly from the vertical federated learning client to the vertical federated learning server.

[0036] Therefore, the above embodiments of the invention, by having the vertical federated learning client transmit first instruction information to the vertical federated learning server, enables the vertical federated learning server to determine whether it needs to send initial model information to the vertical federated learning client, and specifically which type and which initial model information to send. Based on the first instruction information, the vertical federated learning server accurately determines the initial learning model information to be sent to the vertical federated learning client (including whether it needs to be sent, and which model information to send). This avoids the inconsistency between the server and the client caused by the vertical federated learning server solely selecting and determining the initial model information, thereby achieving the technical effect of negotiating the initial model information between the server and the client, achieving the effect of optimized transmission of initial model information, and reducing the signaling overhead of renegotiation.

[0037] Secondly, a communication method is provided. This method can be executed by components of a network storage function element, such as a processor, chip, or chip system of the network storage function element, or by a logic module or software capable of implementing all or part of the network storage function element. The method includes:

[0038] The vertical federated learning server receives first instruction information; the vertical federated learning server determines second information based on the first instruction information; the vertical federated learning server sends a first message to the vertical federated learning client to instruct the vertical federated learning client to perform model training; the first message includes the second information.

[0039] In this embodiment of the invention, the preparation of a vertical federated learning server to start a training task includes being triggered by the vertical federated learning server based on internal conditions, or being triggered by the vertical federated learning server receiving an external instruction.

[0040] As an optional step in this embodiment, the vertical federated learning server initiates the process of discovering the vertical federated learning client.

[0041] In one possible implementation, the vertical federated learning server sends a discovery request message to the network repository function (NRF) element. This message carries information required by the vertical federated learning server regarding vertical federated learning, including information related to the application function type. This application function type information is used by the NRF element to discover and select the vertical federated learning client requested by the server. Accordingly, the NRF element obtains the application function (AF) identifier based on the discovery request message and sends it to the vertical federated learning server, carrying the AF identifier in a discovery request response message. The application function (AF) identifier serves as the identifier for the vertical federated learning client requested by the server.

[0042] The vertical federated learning server sends a vertical federated learning preparation request to the vertical federated learning client, which transmits one or more preparation information to the vertical federated learning client, including vertical federated learning task association identifier, vertical federated learning training preparation information, and vertical federated learning filtering information.

[0043] In possible implementations, the aforementioned preparation request message is sent by the vertical federated learning server to the vertical federated client through the network exposure function (NEF) network element, or it is sent directly by the vertical federated learning server to the vertical federated client.

[0044] The vertical federation client sends a vertical federation learning preparation response message to the vertical federation learning server.

[0045] In possible implementations, the aforementioned preparation response message is sent by the vertical federated learning client to the vertical federated server through the network exposure function (NEF) network element, or directly by the vertical federated learning client to the vertical federated server.

[0046] In one possible implementation, the longitudinal federated learning preparation response message includes first indication information. Optionally, the longitudinal federated learning preparation response message may further include application function (AF) identification information and / or model training feature space information.

[0047] The first indication information specifically includes:

[0048] The longitudinal federated learning client already has information about the initial model, or

[0049] The longitudinal federated learning client does not have information about the initial model, or

[0050] The longitudinal federated learning client requests information about the initial model, or

[0051] Vertical federated learning clients do not need to obtain information about the initial model, or

[0052] The longitudinal federated learning client refuses to receive information about the initial model, or

[0053] Vertical federated learning clients request information about the new model, or

[0054] The longitudinal federated learning client does not support information from the second initial model.

[0055] The second initial model is the initial model information that the vertical federated learning client has obtained before receiving the federated learning preparation request message mentioned above. It can be obtained by being generated locally by the federated learning client or by being sent to the federated learning client by the federated learning server.

[0056] As a possible implementation of the first instruction information, it can be explicit or implicit.

[0057] For example, if the vertical federated learning preparation response message does not contain explicit indication information, it is assumed by default that the vertical federated learning client needs to obtain the initial model information, while the message additionally carries indication information 0 or 1 to indicate that it does not need to obtain / rejects obtaining the initial model information.

[0058] The vertical federated learning server determines the second information based on the first instruction information. The second information includes specific vertical federated initial model information or information that does not require sending the initial model to the vertical federated learning client.

[0059] The vertical federated learning server sends a vertical federated learning training request message to the vertical federated learning client. The vertical federated learning training request message includes certain second information, which includes vertical federated initial model information, in order to instruct the vertical federated learning client to perform model training.

[0060] In a possible implementation, if the first indication information indicates that the vertical federated learning client already has the initial model information or the vertical federated learning client does not need to obtain the initial model information, then the vertical federated learning server will no longer send the initial model information to the vertical federated learning client in the vertical federated learning training request message.

[0061] In possible implementations, the aforementioned vertical federated training request message is sent from the vertical federated learning server to the vertical federated client via a network exposure function (NEF) network element, or it is sent directly from the vertical federated learning server to the vertical federated learning client.

[0062] The vertical federated learning client receives the initial vertical federated model information and begins local model training.

[0063] The vertical federated client sends a vertical federated learning training response message to the vertical federated learning server.

[0064] In possible implementations, the training response message is sent from the vertical federated learning client to the vertical federated learning server via a network exposure function (NEF) element, or directly from the vertical federated learning client to the vertical federated learning server.

[0065] Therefore, the above embodiments of the invention, by having the vertical federated learning client transmit first instruction information to the vertical federated learning server, enables the vertical federated learning server to determine whether it needs to send initial model information to the vertical federated learning client, and specifically which type and which initial model information to send. Based on the first instruction information, the vertical federated learning server accurately determines the initial learning model information to be sent to the vertical federated learning client (including whether it needs to be sent, and which model information to send). This avoids the inconsistency between the server and the client caused by the vertical federated learning server solely selecting and determining the initial model information, thereby achieving the technical effect of negotiating the initial model information between the server and the client, achieving the effect of optimized transmission of initial model information, and reducing the signaling overhead of renegotiation.

[0066] Thirdly, a communication method is provided. This method can be executed by components of a network data analysis function network element, such as a processor, chip, or chip system of the network data analysis function network element, or by a logic module or software capable of implementing all or part of the network data analysis function network element. The method includes: a vertical federated learning server receiving first instruction information; the vertical federated learning server determining second information based on the first instruction information; the vertical federated learning server sending a first message to a vertical federated learning client to instruct the vertical federated learning client to perform model training; the first message includes the second information.

[0067] In this embodiment of the invention, the preparation of a vertical federated learning server to start a training task includes being triggered by the vertical federated learning server based on internal conditions, or being triggered by the vertical federated learning server receiving an external instruction.

[0068] As an optional step in this embodiment, the vertical federated learning server initiates the process of discovering the vertical federated learning client.

[0069] In one possible implementation, the vertical federated learning server sends a discovery request message to the network repository function (NRF) element. This message carries information required by the vertical federated learning server regarding vertical federated learning, including application function type information. This application function type information is used by the NRF element to discover and select the vertical federated learning client requested by the server. Accordingly, the NRF element obtains the application function (AF) identifier based on the discovery request message and includes this identifier in a discovery request response message, which is then sent to the vertical federated learning server. The application function (AF) identifier serves as the identifier for the vertical federated learning client requested by the server.

[0070] The vertical federated learning server sends a vertical federated learning preparation request to the vertical federated learning client, which transmits one or more preparation information to the vertical federated learning client, including vertical federated learning task association identifier, vertical federated learning training preparation information, and vertical federated learning filtering information.

[0071] In possible implementations, the aforementioned preparation request message is sent by the vertical federated learning server to the vertical federated client through the network exposure function (NEF) network element, or it is sent directly by the vertical federated learning server to the vertical federated client.

[0072] The vertical federation client sends a vertical federation learning preparation response message to the vertical federation learning server.

[0073] In possible implementations, the aforementioned preparation response message is sent by the vertical federated learning client to the vertical federated server through the network exposure function (NEF) network element, or directly by the vertical federated learning client to the vertical federated server.

[0074] In one possible implementation, the longitudinal federated learning preparation response message includes first indication information. Optionally, the longitudinal federated learning preparation response message may further include application function (AF) identification information and / or model training feature space information.

[0075] The first indication information specifically includes:

[0076] The longitudinal federated learning client already has information about the initial model, or

[0077] The longitudinal federated learning client does not have information about the initial model, or

[0078] The longitudinal federated learning client requests information about the initial model, or

[0079] Vertical federated learning clients do not need to obtain information about the initial model, or

[0080] The longitudinal federated learning client refuses to receive information about the initial model, or

[0081] Vertical federated learning clients request information about the new model, or

[0082] The longitudinal federated learning client does not support information from the second initial model.

[0083] The second initial model is the initial model information that the vertical federated learning client has obtained before receiving the federated learning preparation request message mentioned above. It can be obtained by being generated locally by the federated learning client or by being sent to the federated learning client by the federated learning server.

[0084] As a possible implementation of the first instruction information, it can be explicit or implicit.

[0085] For example, if the vertical federated learning preparation response message contains or does not contain explicit indication information, it is assumed by default that the vertical federated learning client needs to obtain the initial model information, while the additional indication information 0 or 1 in the message indicates that it does not need to obtain / rejects obtaining the initial model information.

[0086] The vertical federated learning server determines the second information based on the first instruction information. The second information includes specific vertical federated initial model information or information that does not require sending the initial model to the vertical federated learning client.

[0087] The vertical federated learning server sends a vertical federated learning training request message to the vertical federated learning client. The vertical federated learning training request message includes certain second information, which includes vertical federated initial model information, in order to instruct the vertical federated learning client to perform model training.

[0088] In a possible implementation, if the first indication information indicates that the vertical federated learning client already has the initial model information or the vertical federated learning client does not need to obtain the initial model information, then the vertical federated learning server will no longer send the initial model information to the vertical federated learning client in the vertical federated learning training request message.

[0089] In possible implementations, the aforementioned vertical federated training request message is sent from the vertical federated learning server to the vertical federated client via a network exposure function (NEF) network element, or it is sent directly from the vertical federated learning server to the vertical federated learning client.

[0090] The vertical federated learning client receives the initial vertical federated model information and begins local model training.

[0091] The vertical federated client sends a vertical federated learning training response message to the vertical federated learning server.

[0092] In possible implementations, the training response message is sent from the vertical federated learning client to the vertical federated learning server via a network exposure function (NEF) element, or directly from the vertical federated learning client to the vertical federated learning server.

[0093] Therefore, the above embodiments of the invention, by having the vertical federated learning client transmit first instruction information to the vertical federated learning server, enables the vertical federated learning server to determine whether it needs to send initial model information to the vertical federated learning client, and specifically which type and which initial model information to send. Based on the first instruction information, the vertical federated learning server accurately determines the initial learning model information to be sent to the vertical federated learning client (including whether it needs to be sent, and which model information to send). This avoids the inconsistency between the server and the client caused by the vertical federated learning server solely selecting and determining the initial model information, thereby achieving the technical effect of negotiating the initial model information between the server and the client, achieving the effect of optimized transmission of initial model information, and reducing the signaling overhead of renegotiation.

[0094] Fourthly, a communication device is provided for implementing the various methods described above. This communication device can be a network data analysis function element in any of the above aspects or implementations, or a device containing the aforementioned network data analysis function element, or a device included in the aforementioned network data analysis function element, such as a chip. The communication device includes modules, units, or means corresponding to the above methods. These modules, units, or means can be implemented in hardware, software, or by hardware executing corresponding software. The hardware or software includes one or more modules or units corresponding to the above functions.

[0095] In some possible designs, the communication device may include a processing module and a transceiver module. The transceiver module, also referred to as a transceiver unit, is used to implement the transmission and / or reception functions in any of the above aspects and their possible implementations. The transceiver module may consist of transceiver circuits, transceivers, transceivers, or communication interfaces. The processing module can be used to implement the processing functions in any of the above aspects and their possible implementations.

[0096] In some possible designs, the transceiver module includes a sending module and a receiving module, which are used to implement the sending and receiving functions in any of the above aspects and any possible implementation methods.

[0097] Fifthly, a communication device is provided, comprising: at least one processor; said processor being configured to execute a computer program or instructions to cause the communication device to perform the method described in any of the preceding aspects.

[0098] In one possible implementation, the communication device further includes the memory. Optionally, the memory is coupled to the processor; the memory may be integrated with the processor, or the memory may be independent of the processor. Optionally, the processor is used to execute computer programs or instructions stored in the memory.

[0099] In one possible implementation, the memory is independent of the communication device.

[0100] In one possible implementation, the communication device further includes a communication interface for communicating with modules outside the communication device.

[0101] The communication device can be a network data analysis function network element in any of the above aspects or any implementation thereof, or a device containing the above network data analysis function network element, or a device contained in the above network data analysis function network element, such as a chip.

[0102] In a sixth aspect, a computer-readable storage medium is provided, which stores a computer program or instructions that, when executed on a communication device, enable the communication device to perform the methods described in any of the above aspects or any implementation thereof.

[0103] In a seventh aspect, a computer program product containing instructions is provided, which, when run on a communication device, enables the communication device to perform the methods described in any of the foregoing aspects or any implementation thereof.

[0104] Eighthly, a communication device (e.g., a chip or chip system) is provided, the communication device including a processor for implementing the functions involved in any of the above aspects or any implementation thereof.

[0105] In some possible designs, the communication device includes a memory for storing necessary program instructions and data.

[0106] In some possible designs, when the device is a chip system, it can be composed of chips or contain chips and other discrete components.

[0107] It is understood that when the communication device provided by any of the fourth to eighth aspects is a chip, the aforementioned sending action / function can be understood as an output, and the aforementioned receiving action / function can be understood as an input.

[0108] The technical effects of any of the design methods in aspects four through eight can be found in the technical effects of the different design methods in aspects one, two, and three above, and will not be repeated here.

[0109] Ninthly, a communication system is provided, comprising: a network data analysis function network element and a network storage function network element as described in any of the above aspects or any implementation thereof. Attached Figure Description

[0110] Figure 1 is a schematic diagram of the architecture of a communication system provided in an embodiment of this application;

[0111] Figure 2 is a schematic overview of an inventive solution provided in an embodiment of this application;

[0112] Figure 3 is an overview diagram of a four-stage vertical federated learning provided in an embodiment of this application;

[0113] Figure 4 is an embodiment of a method for obtaining a vertical federation model provided in this application.

[0114] Figure 5 is a second embodiment of a method for obtaining a vertical federation model provided in this application.

[0115] Figure 6 is a third embodiment of a method for obtaining a vertical federation model provided in this application.

[0116] Figure 7 is a fourth embodiment of a method for obtaining a vertical federation model provided in this application.

[0117] Figure 8 is a fifth embodiment of a method for obtaining a vertical federation model provided in this application.

[0118] Figure 9 is a sixth embodiment of a method for obtaining a vertical federation model provided in this application;

[0119] Figure 10 is a seventh embodiment of a method for obtaining a vertical federation model provided in this application.

[0120] Figure 11 is an eighth embodiment of a method for obtaining a vertical federated model provided in this application.

[0121] Figure 12 is an embodiment of the device provided in this application, embodiment 1.

[0122] Figure 13 is an embodiment of the device provided in this application, embodiment 2. Detailed Implementation

[0123] To facilitate understanding of the technical solutions provided in the embodiments of this application, a brief introduction to the relevant technical terms is given first. The brief introduction is as follows:

[0124] Functional network elements in communication networks can utilize artificial intelligence (AI) technology and big data in the communication network to output analytical data through model training and inference, thereby assisting in the formulation of communication network strategies and the adjustment of network resources.

[0125] For functional network elements in a carrier network to obtain datasets, data privacy protection may prevent these elements from accessing the data provided by the application provider, leading to infeasible or low-accuracy model training. To address this, distributed joint modeling can be used to break down data silos. For example, multiple functional network elements (or participants) can collaborate on modeling without sharing raw data through vertical federated learning (VFL), achieving AI collaboration.

[0126] The analysis ID can be used to characterize the type of analysis business or service, or simply service. This service is associated with a model; that is, the model can be used to execute the service. Alternatively, the analysis ID is associated with a model; that is, the model is used to execute the service corresponding to the analysis ID.

[0127] For a specific federated learning task, all participants in vertical federated learning (VFL) need to negotiate and determine a machine learning (ML) model before model training begins. Each participant then performs machine learning based on the successfully negotiated ML model to obtain analytical data and assist in the network communication process.

[0128] However, the question of how to negotiate and determine a reasonable machine learning (ML) model between the vertical federated learning server and the vertical federated learning client, or among multiple vertical federated learning participants, remains to be studied.

[0129] To address the aforementioned technical problems, the embodiments of this application propose the following technical solutions, which will be described below in conjunction with other accompanying drawings.

[0130] The technical solutions of this application embodiment can be applied to various communication systems, such as wireless network systems, vehicle-to-everything (V2X) communication systems, device-to-device (D2D) communication systems, vehicle-to-everything (V2X) communication systems, 4G mobile communication systems such as long term evolution (LTE) systems, worldwide interoperability for microwave access (WiMAX) communication systems, 5G mobile communication systems such as new radio (NR) systems, and future communication systems, etc.

[0131] In the embodiments of this application, "instruction" can include direct and indirect instructions, as well as explicit and implicit instructions. The information indicated by a certain piece of information (such as the first instruction information, second instruction information, or third instruction information below) is called the information to be instructed. In the specific implementation process, there are many ways to indicate the information to be instructed, such as, but not limited to, directly indicating the information to be instructed, such as the information to be instructed itself or its index. It can also indirectly indicate the information to be instructed by indicating other information, where there is a correlation between the other information and the information to be instructed. It can also indicate only a part of the information to be instructed, while the other parts are known or pre-agreed upon. For example, the instruction of specific information can be achieved by using a pre-agreed (e.g., protocol-defined) arrangement order of various pieces of information, thereby reducing instruction overhead to some extent. At the same time, common parts of various pieces of information can be identified and indicated uniformly to reduce the instruction overhead caused by individually indicating the same information.

[0132] Furthermore, the specific indication method can also be any existing indication method, such as, but not limited to, the above-mentioned indication methods and their various combinations. Specific details of various indication methods can be found in existing technologies, and will not be repeated here. As described above, for example, when multiple pieces of information of the same type need to be indicated, the indication methods for different pieces of information may differ. In the specific implementation process, the required indication method can be selected according to specific needs. This application embodiment does not limit the selected indication method; therefore, the indication methods involved in this application embodiment should be understood to cover various methods that enable the party to be indicated to obtain the information to be indicated.

[0133] "Predefined" or "pre-configured" can be achieved by pre-saving corresponding codes, tables, or other means that can be used to indicate relevant information in the device. This application does not limit the specific implementation method. "Saving" can refer to saving in one or more memories. These memories can be separate installations or integrated into the encoder, decoder, processor, or communication device. Alternatively, some memories can be separately installed, while others are integrated into the decoder, processor, or communication device. The type of memory can be any form of storage medium, and this application does not limit this.

[0134] The “protocol” mentioned in the embodiments of this application may refer to a protocol family in the field of communication, a standard protocol with a similar protocol family frame structure, or a related protocol applied to future communication systems. The embodiments of this application do not specifically limit this.

[0135] In the embodiments of this application, descriptions such as "when," "under the circumstances," "if," and "if" all refer to the device making corresponding processing under certain objective circumstances, and are not limited to a specific time. They do not require the device to make a judgment action during implementation, nor do they imply any other limitations.

[0136] In the description of the embodiments of this application, unless otherwise stated, " / " indicates that the objects before and after are in an "or" relationship. For example, A / B can represent A or B. "And / or" in the embodiments of this application is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone, where A and B can be singular or plural. Furthermore, in the description of the embodiments of this application, unless otherwise stated, "multiple" refers to two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple. Additionally, to facilitate a clear description of the technical solutions of the embodiments of this application, the terms "first" and "second" are used in the embodiments of this application to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first," "second," etc., do not limit the quantity or order of execution, and that "first," "second," etc., are not necessarily different. Furthermore, in the embodiments of this application, words such as "exemplary" or "for example" are used to indicate that something is being used as an example, illustration, or description. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner for ease of understanding.

[0137] The network architecture and business scenarios described in the embodiments of this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided in the embodiments of this application. As those skilled in the art will know, with the evolution of network architecture and the emergence of new business scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0138] To facilitate understanding of the embodiments of this application, Figure 1 is a diagram of a 5G network architecture based on a service-oriented interface provided in this application. This architecture includes user equipment (UE), radio access network (RAN), operation administration and maintenance (OAM), access and mobility management function (AMF), session management function (SMF), user plane function (UPF), policy control function (PCF), unified data management (UDM), NRF, NWDAF, NEF, AF, and other network elements.

[0139] in,

[0140] UE: Can be located within the beam / cell coverage area of ​​the access network device, which can provide communication services to the terminal device.

[0141] In Figure 1, the terminal device (UE) can be a device with wireless transceiver capabilities or a chip or chip system that can be installed on the device. It allows users to access the network and is used to provide voice and / or data connectivity to users. The terminal device can also be referred to as UE, subscriber unit, terminal, mobile station (MS), or mobile terminal (MT), etc.

[0142] For example, the terminal device UE in Figure 1 can be a mobile phone, a tablet computer, or a computer with wireless transceiver capabilities. Terminal devices can also be user stations, mobile stations, remote stations, remote terminal devices, mobile terminal devices, user terminal devices, wireless communication equipment, user agents, user devices, cellular phones, cordless phones, session initiation protocol (SIP) phones, wireless local loop (WLL) stations, personal digital assistants (PDAs), handheld devices with wireless communication capabilities, computing devices, processing devices connected to wireless modems, in-vehicle devices, wearable devices, terminal devices in the Internet of Things (IoT), home appliances, virtual reality (VR) terminals, augmented reality (AR) terminals, wireless terminals in industrial control, wireless terminals in autonomous driving, wireless terminals in telemedicine, wireless terminals in smart grids, wireless terminals in smart cities, wireless terminals in smart homes, vehicles with vehicle-to-vehicle (V2V) communication capabilities, intelligent connected vehicles, and unmanned aerial vehicle to unmanned aerial vehicle (UAV) communication. Unmanned aerial vehicles (U2U) with vehicle-to-U.S. communication capabilities, terminal devices in future networks, or terminal devices in future evolved public land mobile networks (PLMNs) are not restricted.

[0143] RAN: This can be any device deployed in the access network capable of wireless communication with terminal devices. It can also be a chip or chip system that can be configured in the aforementioned devices, a logical node or logical module, or a function implemented in software. It can be used to implement functions such as wireless physical control, resource scheduling and wireless resource management, wireless access control, and mobility management. Specifically, network devices can be devices that support wired access or devices that support wireless access.

[0144] For example, an access network device may consist of one or more access network (AN) / radio access network (RAN) nodes. AN / RAN nodes may be: evolved Node B (gNB), transmission reception point (TRP), evolved Node B (eNB), radio network controller (RNC), Node B (NB), base station controller (BSC), base transceiver station (BTS), home base station (e.g., home evolved Node B, or home Node B (HNB)), base band unit (BBU), or wireless fidelity (Wi-Fi) access point (AP), etc.

[0145] In another example, a radio access network device can also be a device that includes centralized unit (CU) nodes, distributed unit (DU) nodes, or both CU and DU nodes. For example, the access network device can be logically divided into CUs and DUs, with some protocol layer functions centrally controlled by the CU, and the remaining part or all of the protocol layer functions distributed in the DU, which is centrally controlled by the CU. Furthermore, the centralized unit (CU) can be further divided into a control plane (CU-CP) and a user plane (CU-UP). In different systems, the CU (including CU-CP or CU-UP) or DU can also have different names. For example, in an open radio access network (O-RAN) system, the CU can also be called an O-CU (open CU), the DU can also be called an O-DU, the CU-CP can also be called an O-CU-CP, and the CU-UP can also be called an O-CU-UP.

[0146] NWDAF network element: Possesses functions such as data collection, model training, data analysis, and model inference. It can collect relevant data from network function network elements, third-party service servers, terminal devices, or network management systems (e.g., OAM), perform data analysis or model training based on the relevant data, and provide data analysis results to network function network elements, third-party service servers, terminal devices, or network management systems, or provide trained models to other data analysis function network elements. Network data analysis function network elements can be divided into analysis logic functions and model training logic functions. The analysis logic function is the logical function within the network data analysis function network element, used to perform model inference, derive analysis results (i.e., derive statistical or predictive analysis results based on the analysis consumer's request), and provide analysis results. The model training logic function is the logical function within the network data analysis function network element, used to train models and provide training services (e.g., providing trained models). A network data analysis function network element may contain only analysis logic functions, only model training logic functions, or both. This application embodiment does not specifically limit this.

[0147] AMF network elements are primarily responsible for terminal device access authentication, mobility management, signaling interaction between various functional network elements, and termination of non-access stratum (NAS) layer signaling security. For example, they manage user registration status, reachability status, N1 / N2 interface signaling transmission, access authentication and authorization, user connection status, user registration and network entry, tracking area updates, cell handover user authentication, and key security.

[0148] SMF network elements primarily provide functions such as session management (e.g., session establishment, modification, and release), network protocol (IP) address allocation and management, and selection and control of user plane network elements.

[0149] UPF network elements are responsible for packet routing and forwarding, policy enforcement, traffic reporting, and Quality of Services (QoS) processing.

[0150] UDM network elements: manage user contracts, authorize access, and generate authentication information.

[0151] NRF network elements: provide the ability to register and discover network elements in the network.

[0152] PCF network element; mainly responsible for generating policies such as terminal device access policy and quality of service flow control policy, and can also provide the generated policies to access and mobility management function network elements or session management function network elements.

[0153] OAM (Operational Information Management) primarily performs daily network and service analysis, forecasting, planning, and configuration, as well as network and service testing and fault management. OAM can interact with RAN (Radio Array) to obtain UE location information measured by RAN or reported by UE.

[0154] Application Function (AF) network elements primarily serve as intermediary functional entities for interaction between application servers in the data network (DN) and network elements in the core network. They transmit application-side requests to the network side (e.g., quality of service requirements or user status event subscriptions). Application servers can use them to dynamically control network service quality and billing, and obtain operational information of a specific network element in the core network. In this embodiment, the application function network element can be a functional entity deployed by the operator (i.e., a trusted AF), or a functional entity deployed by a service provider. This service provider can be a third-party service provider (corresponding to an untrusted AF) or an internal service provider of the operator (corresponding to a trusted AF), without limitation.

[0155] NEF network element: mainly responsible for providing network capabilities and event access to external entities (such as untrusted AF network elements), as well as receiving relevant external information (such as receiving information provided by untrusted AF network elements).

[0156] To address the aforementioned issues, this application proposes a communication method comprising: a vertical federated learning server receiving first instruction information; the vertical federated learning server determining second information based on the first instruction information; and the vertical federated learning server sending a first message to a vertical federated learning client to instruct the client to perform model training; the first message including the second information. This avoids the inconsistency between the server and client caused by the vertical federated learning server solely selecting and determining the initial model information, thereby achieving the technical effect of negotiating the initial model information between the server and client, optimizing the transmission of initial model information, and reducing the signaling overhead of renegotiation.

[0157] The interaction process between various network elements / devices in the above-described communication system will be specifically described below with reference to Figures 2 to 13, through method embodiments. The information transmission method provided in this application embodiment can be applied to the communication system shown in Figure 1 above.

[0158] Figure 2 is a schematic overview of an inventive solution provided in an embodiment of this application;

[0159] S01, the vertical federated learning server receives message A from network functional entity A, which carries first instruction information;

[0160] In one possible implementation, network functional entity A is a network exposure function (NEF) network element, or network functional entity A is a network repository function (NRF) network element.

[0161] It should be noted that the vertical federation server in all embodiments of this application may also be called the vertical federation server or the vertical federation active participant.

[0162] The vertical federation server function and the vertical federation client function can be deployed in various forms in the network. Embodiments one to eight of this application are just examples.

[0163] The vertical federated learning server is implemented by the application function (AF) network element, while the vertical federated learning client is implemented by the network data function (NWDAF); or

[0164] The vertical federated learning client is implemented by the Application Function (AF) network element, while the vertical federated learning server is implemented by the Network Data Function (NWDAF); or

[0165] The vertical federated learning client is implemented using the Network Data Function (NWDAF), and the vertical federated learning server is also implemented using the Network Data Function (NWDAF).

[0166] It also includes any reasonable combination that can be extended or inferred by those skilled in the art through this application.

[0167] The first indication information specifically includes:

[0168] The longitudinal federated learning client already has information about the initial model, or

[0169] The longitudinal federated learning client does not have information about the initial model, or

[0170] The longitudinal federated learning client requests information about the initial model, or

[0171] Vertical federated learning clients do not need to obtain information about the initial model, or

[0172] The longitudinal federated learning client refuses to receive information about the initial model, or

[0173] The longitudinal federated learning client does not support information from the second initial model, or

[0174] The vertical federated learning server receives information from one or more vertical federated learning clients from the Network Storage Function (NRF).

[0175] S02, the vertical federated learning server determines second information based on the first instruction information received in step S01, the second information including initial model information and / or information of the first federated learning client.

[0176] S03, the vertical federated learning server sends a first message to the vertical federated learning client to instruct the vertical federated learning client to train the model;

[0177] The first message includes the second information.

[0178] S04, based on the second information received in S03, the longitudinal federated learning client trains the model.

[0179] When the second information is the information of the first federated learning client, specifically, when the vertical federated learning server sends the information of the first federated learning client to network function entity A, the information of the first federated learning client includes its identification information or IP address information or any other relevant information that can uniquely identify the federated learning client within a certain range. In this case, step S03 may only be used by network function entity A to find the federated learning client when sending the first message, and the federated learning client information will no longer be sent to the federated learning client itself.

[0180] Figure 3 is an overview diagram of a four-stage vertical federated learning provided in an embodiment of this application;

[0181] For a longitudinal federated learning task, the process from network function initialization to actual federated training involves the following four stages:

[0182] First: Network Function Registration Phase

[0183] 1) The vertical federated learning client sends a network function registration request message to the network storage function through the network open function. The registration request message contains the vertical federated learning capability information of the vertical federated learning client.

[0184] 2) The network storage function stores the vertical federated learning capability information of the obtained vertical federated learning client.

[0185] 3) The network storage function sends a network function registration response message to the vertical federated learning client through the network open function.

[0186] Second: Network Function Discovery Phase

[0187] 1) The process of the vertical federated learning server initiating the discovery of the vertical federated learning client;

[0188] In one possible implementation, the vertical federated learning server sends a discovery request message to the network repository function (NRF) element. This message carries information required by the vertical federated learning server regarding vertical federated learning, including information related to the application function type. This application function type information is used by the NRF element to discover and select the vertical federated learning client requested by the server. Accordingly, the NRF element obtains the application function (AF) identifier based on the discovery request message and sends it to the vertical federated learning server, carrying the AF identifier in a discovery request response message. The application function (AF) identifier serves as the identifier for the vertical federated learning client requested by the server.

[0189] Third: Preparation phase for vertical federated learning

[0190] 1) The vertical federated learning server sends a vertical federated learning preparation request to the vertical federated learning client, which transmits one or more preparation information to the vertical federated learning client, including vertical federated learning task association identifier, vertical federated learning training preparation information, vertical federated learning filtering information, etc.

[0191] In possible implementations, the aforementioned preparation request message is sent by the vertical federated learning server to the vertical federated client through the network exposure function (NEF) network element, or it is sent directly by the vertical federated learning server to the vertical federated client.

[0192] 2) The vertical federation client sends a vertical federation learning preparation response message to the vertical federation learning server.

[0193] In possible implementations, the aforementioned preparation response message is sent by the vertical federated learning client to the vertical federated server through the network exposure function (NEF) network element, or it is sent directly by the vertical federated learning client to the vertical federated server.

[0194] Fourth: Vertical Federated Training Phase

[0195] The vertical federated learning server sends a vertical federated learning training request message to the vertical federated learning client, thereby instructing the vertical federated learning client to train the model.

[0196] Figure 4 is an embodiment of a method for obtaining a vertical federated model provided in this application.

[0197] S101, The vertical federated learning server is ready to start a training task, including being triggered by the vertical federated learning server based on internal conditions, or being triggered by the vertical federated learning server receiving an external instruction.

[0198] S102, as an optional step in this embodiment, the vertical federated learning server initiates the process of discovering the vertical federated learning client.

[0199] In one possible implementation, the vertical federated learning server sends a discovery request message to the network repository function (NRF) element. This message carries information required by the vertical federated learning server regarding vertical federated learning, including information related to the application function type. This application function type information is used by the NRF element to discover and select the vertical federated learning client requested by the server. Accordingly, the NRF element obtains the application function (AF) identifier based on the discovery request message and sends it to the vertical federated learning server, carrying the AF identifier in a discovery request response message. The application function (AF) identifier serves as the identifier for the vertical federated learning client requested by the server.

[0200] In steps S103-S104, the vertical federated learning server sends a vertical federated learning preparation request to the vertical federated learning client, which transmits one or more preparation information to the vertical federated learning client, including vertical federated learning task association identifier, vertical federated learning training preparation information, and vertical federated learning filtering information.

[0201] In possible implementations, the aforementioned preparation request message is sent by the vertical federated learning server to the vertical federated client through the network exposure function (NEF) network element, or it is sent directly by the vertical federated learning server to the vertical federated client.

[0202] S105~S106, the vertical federation client sends a vertical federation learning preparation response message to the vertical federation learning server.

[0203] In possible implementations, the aforementioned preparation response message is sent by the vertical federated learning client to the vertical federated server through the network exposure function (NEF) network element, or directly by the vertical federated learning client to the vertical federated server.

[0204] In one possible implementation, the longitudinal federated learning preparation response message includes first indication information. Optionally, the longitudinal federated learning preparation response message may further include application function (AF) identification information and / or model training feature space information.

[0205] The first indication information specifically includes:

[0206] The longitudinal federated learning client already has information about the initial model, or

[0207] The longitudinal federated learning client does not have information about the initial model, or

[0208] The longitudinal federated learning client requests information about the initial model, or

[0209] Vertical federated learning clients do not need to obtain information about the initial model, or

[0210] The longitudinal federated learning client refuses to receive information about the initial model, or

[0211] Vertical federated learning clients request information about the new model, or

[0212] The longitudinal federated learning client does not support information from the second initial model.

[0213] The second initial model is the initial model information that the longitudinal federated learning client has obtained before receiving the federated learning preparation request message described in step S104. It can be obtained by being generated locally by the federated learning client or by being sent to the federated learning client by the federated learning server.

[0214] As a possible implementation of the first instruction information, it can be explicit or implicit.

[0215] For example, if the vertical federated learning response message contains or does not contain explicit indication information, it is assumed by default that the vertical federated learning client needs to obtain the initial model information, while the additional indication information 0 or 1 in the message indicates that it does not need to obtain / rejects obtaining the initial model information.

[0216] It is understandable that the specific implementation of the first instruction information is not limited to the examples above.

[0217] S107, the vertical federated learning server determines the second information based on the first instruction information, the second information including specific vertical federated initial model information or information that does not need to be sent to the vertical federated learning client.

[0218] S108-S109, the vertical federated learning server sends a vertical federated learning training request message to the vertical federated learning client. The vertical federated learning training request message includes the second information determined in step S107 above. The second information includes the vertical federated initial model information, thereby instructing the vertical federated learning client to perform model training.

[0219] In a possible implementation, if the first indication information in steps S105 and S106 indicates that the vertical federated learning client already has the initial model information or the vertical federated learning client does not need to obtain the initial model information, then in steps S108 to S109, the vertical federated learning server will no longer send the initial model information to the vertical federated learning client in the vertical federated learning training request message.

[0220] In possible implementations, the aforementioned vertical federated training request message is sent from the vertical federated learning server to the vertical federated client via a network exposure function (NEF) network element, or it is sent directly from the vertical federated learning server to the vertical federated learning client.

[0221] S110. The vertical federated learning client receives the initial vertical federated model information and begins local model training.

[0222] S111~S112. The vertical federated client sends a vertical federated learning training response message to the vertical federated learning server.

[0223] In possible implementations, the training response message is sent from the vertical federated learning client to the vertical federated learning server via a network exposure function (NEF) element, or directly from the vertical federated learning client to the vertical federated learning server.

[0224] Therefore, through steps S105-S106 and S107, S108, and S109 above, the vertical federated learning client transmits first instruction information to the vertical federated learning server, enabling the vertical federated learning server to determine whether initial model information needs to be sent to the vertical federated learning client, and which type and which initial model information to send. Based on the first instruction information, the vertical federated learning server accurately determines the initial learning model information to be sent to the vertical federated learning client (including whether it needs to be sent, and which model information to send). This avoids inconsistencies between the server and client caused by the vertical federated learning server solely selecting and determining the initial model information, thus achieving the technical effect of negotiating the initial model information between the server and the client, optimizing the transmission of initial model information, and reducing the signaling overhead of renegotiation.

[0225] It should be noted that the vertical federation server function and the vertical federation client function can be deployed in the network in various forms. Embodiments one to eight of this application are merely examples.

[0226] The vertical federated learning server is implemented by the application function (AF) network element, while the vertical federated learning client is implemented by the network data function (NWDAF); or

[0227] The vertical federated learning client is implemented by the Application Function (AF) network element, while the vertical federated learning server is implemented by the Network Data Function (NWDAF); or

[0228] The vertical federated learning client is implemented using the Network Data Function (NWDAF), and the vertical federated learning server is also implemented using the Network Data Function (NWDAF).

[0229] It also includes any reasonable combination that can be extended or inferred by those skilled in the art through this application.

[0230] It should be noted that the negotiation process for the initial model information in the above embodiments can be further extended to the negotiation of models in a general sense or the updating of machine learning models. This embodiment only uses the initial model as an example for illustration.

[0231] Figure 5 is a second embodiment of a method for obtaining a vertical federated model provided in this application.

[0232] S201, The vertical federated learning server is ready to start a training task, including being triggered by the vertical federated learning server based on internal conditions, or being triggered by the vertical federated learning server receiving an external instruction.

[0233] S202, as an optional step in this embodiment, the vertical federated learning server initiates the process of discovering the vertical federated learning client;

[0234] In one possible implementation, the vertical federated learning server sends a discovery request message to the network repository function (NRF) element. This message carries information required by the vertical federated learning server regarding vertical federated learning, including information related to the application function type. This application function type information is used by the NRF element to discover and select the vertical federated learning client requested by the server. Accordingly, the NRF element obtains the application function (AF) identifier based on the discovery request message and sends it to the vertical federated learning server, carrying the AF identifier in a discovery request response message. The application function (AF) identifier serves as the identifier for the vertical federated learning client requested by the server.

[0235] S203, the vertical federated learning server sends a vertical federated learning preparation request to the vertical federated learning client, which transmits one or more preparation information to the vertical federated learning client, including vertical federated learning task association identifier, vertical federated learning training preparation information, vertical federated learning filtering information, and the required model training feature space.

[0236] In a possible implementation, the aforementioned preparation request message is sent directly from the vertical federated learning server to the vertical federated client;

[0237] S204, the vertical federated learning client verifies one or more pieces of information in the preparation message in step S203, including but not limited to verifying the task association identifier, whether the model training feature space required by the vertical federated learning server matches the feature space supported by the client, etc.

[0238] S205, the vertical federation client sends a vertical federation learning preparation response message to the vertical federation learning server.

[0239] In a possible implementation, the aforementioned preparation response message is sent directly from the vertical federated learning client to the vertical federated server;

[0240] In one possible implementation, the longitudinal federated learning preparation response message includes first indication information. Optionally, the longitudinal federated learning preparation response message may further include application function (AF) identification information and / or model training feature space information.

[0241] The first indication information specifically includes:

[0242] The longitudinal federated learning client already has information about the initial model, or

[0243] The longitudinal federated learning client does not have information about the initial model, or

[0244] The longitudinal federated learning client requests information about the initial model, or

[0245] Vertical federated learning clients do not need to obtain information about the initial model, or

[0246] The longitudinal federated learning client refuses to receive information about the initial model, or

[0247] Vertical federated learning clients request information about the new model, or

[0248] The longitudinal federated learning client does not support information from the second initial model.

[0249] The second initial model is the initial model information that the longitudinal federated learning client has obtained before receiving the federated learning preparation request message described in step S203. It can be obtained by being generated locally by the federated learning client or by being sent to the federated learning client by the federated learning server.

[0250] As a possible implementation of the first instruction information, it can be explicit or implicit.

[0251] For example, if the vertical federated learning preparation response message contains or does not contain explicit indication information, it is assumed by default that the vertical federated learning client needs to obtain the initial model information, while the additional indication information 0 or 1 in the message indicates that it does not need to obtain / rejects obtaining the initial model information.

[0252] It is understandable that the specific implementation of the first instruction information is not limited to the examples above.

[0253] S206, the vertical federated learning server determines the second information based on the first instruction information, the second information including specific vertical federated initial model information or information that does not need to be sent to the vertical federated learning client.

[0254] S207, the vertical federated learning server sends a vertical federated learning training request message to the vertical federated learning client. The vertical federated learning training request message includes the second information determined in step S206 above. The second information includes the vertical federated initial model information, thereby instructing the vertical federated learning client to perform model training.

[0255] In a possible implementation, if the first indication information in step S205 indicates that the vertical federated learning client already has the initial model information or the vertical federated learning client does not need to obtain the initial model information, then in step 207, the vertical federated learning server no longer sends the initial model information to the vertical federated learning client in the vertical federated learning training request message.

[0256] In possible implementations, the aforementioned vertical federated training request message is sent from the vertical federated learning server to the vertical federated client via a network exposure function (NEF) network element, or it is sent directly from the vertical federated learning server to the vertical federated learning client.

[0257] S208. The vertical federated learning client receives the initial vertical federated model information and begins local model training.

[0258] S209. The vertical federated client sends a vertical federated learning training response message to the vertical federated learning server.

[0259] In possible implementations, the training response message is sent from the vertical federated learning client to the vertical federated learning server via a network exposure function (NEF) element, or directly from the vertical federated learning client to the vertical federated learning server.

[0260] Therefore, through steps S204, S205, and 206 / 207 above, the vertical federated learning client transmits first instruction information to the vertical federated learning server, enabling the vertical federated learning server to determine whether initial model information needs to be sent to the vertical federated learning client, and specifically which type and which initial model information to send. Based on the first instruction information, the vertical federated learning server accurately determines the initial learning model information to be sent to the vertical federated learning client (including whether it needs to be sent, and which model information to send). This avoids inconsistencies between the server and client caused by the vertical federated learning server solely selecting and determining the initial model information, thus achieving the technical effect of negotiating the initial model information between the server and the client, optimizing the transmission of initial model information, and reducing the signaling overhead of renegotiation.

[0261] It should be noted that the vertical federation server function and the vertical federation client function can be deployed in the network in various forms. Embodiments one to eight of this application are merely examples.

[0262] The vertical federated learning server is implemented by the Application Function (AF) network element, while the vertical federated learning client is implemented by the Network Data Function (NWDAF); or

[0263] The vertical federated learning client is implemented by the Application Function (AF) network element, while the vertical federated learning server is implemented by the Network Data Function (NWDAF); or

[0264] The vertical federated learning client is implemented using the Network Data Function (NWDAF), and the vertical federated learning server is also implemented using the Network Data Function (NWDAF).

[0265] It also includes any reasonable combination that can be extended or inferred by those skilled in the art through this application.

[0266] It should be noted that the negotiation process for the initial model information in the above embodiments can be further extended to the negotiation of models in a general sense or the updating of machine learning models. This embodiment only uses the initial model as an example for illustration.

[0267] Figure 6 is a third embodiment of a method for obtaining a vertical federated model provided in this application.

[0268] S301~S303, Network Function Registration Process.

[0269] S301, the vertical federated learning client sends a network function registration request message to the network storage function through the network open function. The registration request message contains the vertical federated learning capability information of the vertical federated learning client and the first instruction information.

[0270] The first indication information specifically includes:

[0271] The longitudinal federated learning client already has information about the initial model, or

[0272] The longitudinal federated learning client does not have information about the initial model, or

[0273] The longitudinal federated learning client requests information about the initial model, or

[0274] Vertical federated learning clients do not need to obtain information about the initial model, or

[0275] The longitudinal federated learning client refuses to receive information about the initial model, or

[0276] Vertical federated learning clients request information about the new model, or

[0277] The longitudinal federated learning client does not support information from the second initial model.

[0278] The second initial model is the initial model information that the longitudinal federated learning client has obtained before the above-mentioned network function registration process. It can be obtained by being generated locally by the federated learning client or by being sent to the federated learning client by the federated learning server.

[0279] As a possible implementation of the first instruction information, it can be explicit or implicit.

[0280] For example, if the vertical federated learning preparation response message contains or does not contain explicit indication information, it is assumed by default that the vertical federated learning client needs to obtain the initial model information, while the additional indication information 0 or 1 in the message indicates that it does not need to obtain / rejects obtaining the initial model information.

[0281] It is understandable that the specific implementation of the first instruction information is not limited to the examples above.

[0282] S302, the network storage function stores the vertical federated learning capability information and / or first instruction information of the vertical federated learning client obtained in step S301.

[0283] S303, the Network Storage Function (NRF) sends a Network Function Registration Response message to the Vertical Federated Learning Client via the Network Open Function.

[0284] S304~S305, the process of the vertical federated learning server initiating the discovery of the vertical federated learning client;

[0285] In one possible implementation, S304, the vertical federated learning server sends a discovery request message to the network repository function (NRF) network element. The discovery request message carries vertical federated learning-related information required by the vertical federated learning server, as well as application function type-related information. This application function type-related information is used by the network repository function network element to discover and select the vertical federated learning client requested by the vertical federated learning server.

[0286] In step S305, the network repository function (NRF) element obtains the application function (AF) identifier based on the discovery request message and sends it to the vertical federated learning server, carrying the application function (AF) identifier in the discovery request response message. The application function (AF) identifier is the identifier of the vertical federated learning client requested by the vertical federated learning server. Furthermore, the discovery request response message carries the first indication information transmitted in step S301 and stored in the network repository function (NRF) in step S302.

[0287] S306, Optional, Longitudinal Federated Learning Preparation Process.

[0288] In one possible implementation, the vertical federated learning server sends a vertical federated learning preparation request to the vertical federated learning client, which transmits one or more preparation information to the vertical federated learning client, including vertical federated learning task association identifier, vertical federated learning training preparation information, vertical federated learning filtering information, and the required model training feature space.

[0289] In a possible implementation, the aforementioned preparation request message is sent directly from the vertical federated learning server to the vertical federated client;

[0290] The vertical federation client sends a vertical federation learning preparation response message to the vertical federation learning server.

[0291] In a possible implementation, the aforementioned preparation response message is sent directly from the vertical federated learning client to the vertical federated server;

[0292] S307, the vertical federated learning server determines the second information based on the first instruction information obtained in step S305 above. The second information includes specific vertical federated initial model information or information that does not need to be sent to the vertical federated learning client.

[0293] S308-S309, the vertical federated learning server sends a vertical federated learning training request message to the vertical federated learning client. The vertical federated learning training request message includes the second information determined in step 307 above. The second information includes the vertical federated initial model information, thereby instructing the vertical federated learning client to perform model training.

[0294] In a possible implementation, if the first indication information in steps S301 and S305 indicates that the vertical federated learning client already has the initial model information or the vertical federated learning client does not need to obtain the initial model information, then in step S308, the vertical federated learning server will no longer send the initial model information to the vertical federated learning client in the vertical federated learning training request message.

[0295] In possible implementations, the aforementioned vertical federated training request message is sent from the vertical federated learning server to the vertical federated client via a network exposure function (NEF) network element, or it is sent directly from the vertical federated learning server to the vertical federated learning client.

[0296] S310. The vertical federated learning client receives the initial model information for vertical federation and begins local model training.

[0297] S311~S312. The vertical federated client sends a vertical federated learning training response message to the vertical federated learning server.

[0298] In possible implementations, the training response message is sent from the vertical federated learning client to the vertical federated learning server via a network exposure function (NEF) element, or directly from the vertical federated learning client to the vertical federated learning server.

[0299] Therefore, through steps S301, S302, S305 and S307, S308, and S309 above, the vertical federated learning client successfully transmits the first instruction information to the vertical federated learning server. This allows the vertical federated learning server to determine whether it needs to send initial model information to the vertical federated learning client, and specifically which type and which initial model information to send. Based on the first instruction information, the vertical federated learning server accurately determines the initial learning model information to be sent to the vertical federated learning client (including whether it needs to be sent and which model information to send). This avoids inconsistencies between the server and client caused by the vertical federated learning server solely selecting and determining the initial model information. This achieves the technical effect of negotiating the initial model information between the server and client, optimizing the transmission of initial model information and reducing the signaling overhead of renegotiation.

[0300] It should be noted that the vertical federation server function and the vertical federation client function can be deployed in the network in various forms. Embodiments one to eight of this application are merely examples.

[0301] The vertical federated learning server is implemented by the Application Function (AF) network element, while the vertical federated learning client is implemented by the Network Data Function (NWDAF); or

[0302] The vertical federated learning client is implemented by the Application Function (AF) network element, while the vertical federated learning server is implemented by the Network Data Function (NWDAF); or

[0303] The vertical federated learning client is implemented using the Network Data Function (NWDAF), and the vertical federated learning server is also implemented using the Network Data Function (NWDAF).

[0304] It also includes any reasonable combination that can be extended or inferred by those skilled in the art through this application.

[0305] It should be noted that the negotiation process for the initial model information in the above embodiments can be further extended to the negotiation of models in a general sense or the updating of machine learning models. This embodiment only uses the initial model as an example for illustration.

[0306] Figure 7 is a fourth embodiment of a method for obtaining a vertical federation model provided in this application.

[0307] Figure 7 provides an embodiment of a method for obtaining a vertical federation model, and its steps are the same as those in Figure 6.

[0308] The difference is that in Figure 7, the vertical federated learning server is implemented by the application function (AF) network element, while the vertical federated learning client is implemented by the network data function (NWDAF).

[0309] S401~S403, Network Function Registration Process.

[0310] S401, the vertical federated learning client sends a network function registration request message to the network storage function through the network open function. The registration request message contains the vertical federated learning capability information of the vertical federated learning client and the first instruction information.

[0311] The first indication information specifically includes:

[0312] The longitudinal federated learning client already has information about the initial model, or

[0313] The longitudinal federated learning client does not have information about the initial model, or

[0314] The longitudinal federated learning client requests information about the initial model, or

[0315] Vertical federated learning clients do not need to obtain information about the initial model, or

[0316] The longitudinal federated learning client refuses to receive information about the initial model, or

[0317] Vertical federated learning clients request information about the new model, or

[0318] The longitudinal federated learning client does not support information from the second initial model.

[0319] The second initial model is the initial model information that the longitudinal federated learning client has obtained before the above-mentioned network function registration process. It can be obtained by being generated locally by the federated learning client or by being sent to the federated learning client by the federated learning server.

[0320] As a possible implementation of the first instruction information, it can be explicit or implicit.

[0321] For example, if the vertical federated learning preparation response message contains or does not contain explicit indication information, it is assumed by default that the vertical federated learning client needs to obtain the initial model information, while the additional indication information 0 or 1 in the message indicates that it does not need to obtain / rejects obtaining the initial model information.

[0322] It is understandable that the specific implementation of the first instruction information is not limited to the examples above.

[0323] S402, the network storage function stores the vertical federated learning capability information and / or first instruction information of the vertical federated learning client obtained in step S401.

[0324] S403, the Network Storage Function (NRF) sends a Network Function Registration Response message to the Vertical Federated Learning client via the Network Open Function.

[0325] S404~S405, the process of the vertical federated learning server initiating the discovery of the vertical federated learning client;

[0326] In one possible implementation, S404, the vertical federated learning server sends a discovery request message to the network repository function (NRF) element. The discovery request message carries vertical federated learning-related information required by the vertical federated learning server, as well as application function type-related information. This application function type-related information is used by the network repository function element to discover and select the vertical federated learning client requested by the vertical federated learning server.

[0327] In step S405, the network repository function (NRF) element obtains the application function (AF) identifier based on the discovery request message and sends it to the vertical federated learning server, carrying the application function (AF) identifier in the discovery request response message. The application function (AF) identifier is the identifier of the vertical federated learning client requested by the vertical federated learning server. Furthermore, the discovery request response message carries the first indication information transmitted in step S401 and stored in the network repository function (NRF) in step S402.

[0328] S406, Optional, Vertical Federated Learning Preparation Process.

[0329] In one possible implementation, the vertical federated learning server sends a vertical federated learning preparation request to the vertical federated learning client, which transmits one or more preparation information to the vertical federated learning client, including vertical federated learning task association identifier, vertical federated learning training preparation information, vertical federated learning filtering information, and the required model training feature space.

[0330] In a possible implementation, the aforementioned preparation request message is sent directly from the vertical federated learning server to the vertical federated client;

[0331] The vertical federation client sends a vertical federation learning preparation response message to the vertical federation learning server.

[0332] In a possible implementation, the aforementioned preparation response message is sent directly from the vertical federated learning client to the vertical federated server;

[0333] S407, the vertical federated learning server determines the second information based on the first instruction information obtained in step S405 above. The second information includes specific vertical federated initial model information or information that does not need to be sent to the vertical federated learning client.

[0334] S408~S409, the vertical federated learning server sends a vertical federated learning training request message to the vertical federated learning client. The vertical federated learning training request message includes the second information determined in step 407 above. The second information includes the vertical federated initial model information, thereby instructing the vertical federated learning client to perform model training.

[0335] In a possible implementation, if the first indication information in steps S401 and S405 indicates that the vertical federated learning client already has the initial model information or the vertical federated learning client does not need to obtain the initial model information, then in step S308, the vertical federated learning server will no longer send the initial model information to the vertical federated learning client in the vertical federated learning training request message.

[0336] In possible implementations, the aforementioned vertical federated training request message is sent from the vertical federated learning server to the vertical federated client via a network exposure function (NEF) network element, or it is sent directly from the vertical federated learning server to the vertical federated learning client.

[0337] S410. The vertical federated learning client receives the initial vertical federated model information and begins local model training.

[0338] S411~S412. The vertical federated client sends a vertical federated learning training response message to the vertical federated learning server.

[0339] In possible implementations, the training response message is sent from the vertical federated learning client to the vertical federated learning server via a network exposure function (NEF) element, or directly from the vertical federated learning client to the vertical federated learning server.

[0340] Therefore, through steps S401, S402, S405 and S407, S408, and S409 above, the vertical federated learning client successfully transmits the first instruction information to the vertical federated learning server. This allows the vertical federated learning server to determine whether it needs to send initial model information to the vertical federated learning client, and specifically which type and which initial model information to send. Based on the first instruction information, the vertical federated learning server accurately determines the initial learning model information to be sent to the vertical federated learning client (including whether it needs to be sent and which model information to send). This avoids inconsistencies between the server and client caused by the vertical federated learning server solely selecting and determining the initial model information. This achieves the technical effect of negotiating the initial model information between the server and client, optimizing the transmission of initial model information and reducing the signaling overhead of renegotiation.

[0341] It should be noted that the vertical federation server function and the vertical federation client function can be deployed in the network in various forms. Embodiments one to eight of this application are merely examples.

[0342] The vertical federated learning server is implemented by the application function (AF) network element, while the vertical federated learning client is implemented by the network data function (NWDAF); or

[0343] The vertical federated learning client is implemented by the Application Function (AF) network element, while the vertical federated learning server is implemented by the Network Data Function (NWDAF); or

[0344] The vertical federated learning client is implemented using the Network Data Function (NWDAF), and the vertical federated learning server is also implemented using the Network Data Function (NWDAF).

[0345] It also includes any reasonable combination that can be extended or inferred by those skilled in the art through this application.

[0346] It should be noted that the negotiation process for the initial model information in the above embodiments can be further extended to the negotiation of models in a general sense or the updating of machine learning models. This embodiment only uses the initial model as an example for illustration.

[0347] Figure 8 is a fifth embodiment of a method for obtaining a vertical federated model provided in this application.

[0348] The fifth embodiment shown in Figure 8 shares similar key technical principles with the third embodiment shown in Figure 6. The difference is that the server and client for vertical federated learning are deployed in two Network Data Analysis Functions (NWDAFs), and there is no network element (NEF) capable of being opened in the network.

[0349] S501~S503, Network Function Registration Process.

[0350] S501, the vertical federated learning client sends a network function registration request message to the network storage function through the network open function. The registration request message contains the vertical federated learning capability information of the vertical federated learning client and the first instruction information.

[0351] The first indication information specifically includes:

[0352] The longitudinal federated learning client already has information about the initial model, or

[0353] The longitudinal federated learning client does not have information about the initial model, or

[0354] The longitudinal federated learning client requests information about the initial model, or

[0355] Vertical federated learning clients do not need to obtain information about the initial model, or

[0356] The longitudinal federated learning client refuses to receive information about the initial model, or

[0357] Vertical federated learning clients request information about the new model, or

[0358] The longitudinal federated learning client does not support information from the second initial model.

[0359] The second initial model is the initial model information that the longitudinal federated learning client has obtained before the above-mentioned network function registration process. It can be obtained by being generated locally by the federated learning client or by being sent to the federated learning client by the federated learning server.

[0360] As a possible implementation of the first instruction information, it can be explicit or implicit.

[0361] For example, if the vertical federated learning preparation response message contains or does not contain explicit indication information, it is assumed by default that the vertical federated learning client needs to obtain the initial model information, while the additional indication information 0 or 1 in the message indicates that it does not need to obtain / rejects obtaining the initial model information.

[0362] It is understandable that the specific implementation of the first instruction information is not limited to the examples above.

[0363] S502, the network storage function stores the vertical federated learning capability information and / or first instruction information of the vertical federated learning client obtained in step S501.

[0364] S503, the Network Storage Function (NRF) sends a Network Function Registration Response message to the Vertical Federated Learning client via the Network Open Function.

[0365] S504~S505, the process of the vertical federated learning server initiating the discovery of the vertical federated learning client;

[0366] In one possible implementation, S504, the vertical federated learning server sends a discovery request message to the network repository function (NRF) network element. The discovery request message carries vertical federated learning-related information required by the vertical federated learning server, as well as application function type-related information. This application function type-related information is used by the network repository function network element to discover and select the vertical federated learning client requested by the vertical federated learning server.

[0367] In step S505, the network repository function (NRF) element obtains the application function (AF) identifier based on the discovery request message and sends this identifier in the discovery request response message to the vertical federated learning server. The application function (AF) identifier is the identifier of the vertical federated learning client requested by the vertical federated learning server. Furthermore, the discovery request response message carries the first indication information transmitted in step S501 and stored in the network repository function (NRF) in step S502.

[0368] S506, Optional, Vertical Federated Learning Preparation Process.

[0369] In one possible implementation, the vertical federated learning server sends a vertical federated learning preparation request to the vertical federated learning client, which transmits one or more preparation information to the vertical federated learning client, including vertical federated learning task association identifier, vertical federated learning training preparation information, vertical federated learning filtering information, and the required model training feature space.

[0370] In a possible implementation, the aforementioned preparation request message is sent directly from the vertical federated learning server to the vertical federated client;

[0371] The vertical federation client sends a vertical federation learning preparation response message to the vertical federation learning server.

[0372] In a possible implementation, the aforementioned preparation response message is sent directly from the vertical federated learning client to the vertical federated server;

[0373] S507, the vertical federated learning server determines the second information based on the first instruction information obtained in step S505 above. The second information includes specific vertical federated initial model information or information that does not need to be sent to the vertical federated learning client.

[0374] S508, the vertical federated learning server sends a vertical federated learning training request message to the vertical federated learning client. The vertical federated learning training request message includes the second information determined in step 507 above. The second information includes the vertical federated initial model information, thereby instructing the vertical federated learning client to perform model training.

[0375] In a possible implementation, if the first indication information in steps S501 and S505 indicates that the vertical federated learning client already has the initial model information or the vertical federated learning client does not need to obtain the initial model information, then in step S308, the vertical federated learning server will no longer send the initial model information to the vertical federated learning client in the vertical federated learning training request message.

[0376] In possible implementations, the aforementioned vertical federated training request message is sent from the vertical federated learning server to the vertical federated client via a network exposure function (NEF) network element, or it is sent directly from the vertical federated learning server to the vertical federated learning client.

[0377] S509. The vertical federated learning client receives the initial vertical federated model information and begins local model training.

[0378] S510. The vertical federated client sends a vertical federated learning training response message to the vertical federated learning server.

[0379] In possible implementations, the training response message is sent from the vertical federated learning client to the vertical federated learning server via a network exposure function (NEF) element, or directly from the vertical federated learning client to the vertical federated learning server.

[0380] Therefore, through steps S501, S502, S505 and S507, S508 above, the vertical federated learning client successfully transmits the first instruction information to the vertical federated learning server. This allows the vertical federated learning server to determine whether it needs to send initial model information to the vertical federated learning client, and specifically which type and which initial model information to send. Based on the first instruction information, the vertical federated learning server accurately determines the initial learning model information to be sent to the vertical federated learning client (including whether it needs to be sent, and which model information to send). This avoids inconsistencies between the server and client caused by the vertical federated learning server solely selecting and determining the initial model information. It achieves the technical effect of negotiating the initial model information between the server and client, optimizing the transmission of initial model information and reducing the signaling overhead of further negotiation.

[0381] It should be noted that the vertical federation server function and the vertical federation client function can be deployed in the network in various forms. Embodiments one to eight of this application are merely examples.

[0382] The vertical federated learning server is implemented by the application function (AF) network element, while the vertical federated learning client is implemented by the network data function (NWDAF); or

[0383] The vertical federated learning client is implemented by the Application Function (AF) network element, while the vertical federated learning server is implemented by the Network Data Function (NWDAF); or

[0384] The vertical federated learning client is implemented using the Network Data Function (NWDAF), and the vertical federated learning server is also implemented using the Network Data Function (NWDAF).

[0385] It also includes any reasonable combination that can be extended or inferred by those skilled in the art through this application.

[0386] It should be noted that the negotiation process for the initial model information in the above embodiments can be further extended to the negotiation of models in a general sense or the updating of machine learning models. This embodiment only uses the initial model as an example for illustration.

[0387] Figure 9 is a sixth embodiment of a method for obtaining a vertical federation model provided in this application.

[0388] The sixth embodiment illustrated in Figure 9 aims to demonstrate that, in addition to the vertical federated learning client discovery process and the vertical federated learning preparation process, when the vertical federated learning client determines that it has a request to obtain an initial model or a model update requirement, it actively triggers a second request message to request model information from the vertical federated learning server.

[0389] S601, The vertical federated learning server is ready to start a training task, including being triggered by the vertical federated learning server based on internal conditions, or being triggered by the vertical federated learning server receiving an external instruction.

[0390] S602, as an optional step in this embodiment, the vertical federated learning server initiates the process of discovering the vertical federated learning client.

[0391] In one possible implementation, the vertical federated learning server sends a discovery request message to the network repository function (NRF) element. This message carries information required by the vertical federated learning server regarding vertical federated learning, including information related to the application function type. This application function type information is used by the NRF element to discover and select the vertical federated learning client requested by the server. Accordingly, the NRF element obtains the application function (AF) identifier based on the discovery request message and sends it to the vertical federated learning server, carrying the AF identifier in a discovery request response message. The application function (AF) identifier serves as the identifier for the vertical federated learning client requested by the server.

[0392] S603, the vertical federated learning server initiates the vertical federated learning preparation process to the vertical federated learning client.

[0393] The vertical federated learning server sends a vertical federated learning preparation request to the vertical federated learning client, which includes one or more preparation information such as the vertical federated learning task association identifier, the vertical federated learning training preparation information, the vertical federated learning filtering information, and the required model training feature space.

[0394] In a possible implementation, the aforementioned preparation request message is sent directly from the vertical federated learning server to the vertical federated client or via the Network Capability Open Element (NEF) to the vertical federated client.

[0395] The vertical federated learning client sends a vertical federated learning preparation response message to the vertical federated learning server.

[0396] In possible implementations, the aforementioned preparation response message is sent directly from the vertical federated learning client to the vertical federated server or through an intermediate network element, such as the Network Capability Open Element (NEF).

[0397] S604, The longitudinal federated learning client determines that there is no available initial machine learning model;

[0398] For example, a vertical federated learning client might determine that no initial machine learning model is available during the discovery or preparation process. Although the client obtains a third initial machine learning model from the server, or through local configuration or generation, it discovers that the feature space associated with this third model does not match or correspond to the feature space associated with its supported fourth model. For instance, the feature space of the third initial model might have a matrix dimension of 10*10, while the feature space of the fourth model supported by the client has a dimension of 8*8. Therefore, the client determines that no initial machine learning model is available.

[0399] In steps S605-S606, the vertical federated learning client actively requests to obtain an initial machine learning model by sending a second request message to the vertical federated learning server. The second request message contains first instruction information, and optionally, the second request also includes the application function (AF) identifier and / or model training feature space information.

[0400] The first indication information specifically includes:

[0401] The longitudinal federated learning client already has information about the initial model, or

[0402] The longitudinal federated learning client does not have information about the initial model, or

[0403] The longitudinal federated learning client requests information about the initial model, or

[0404] Vertical federated learning clients do not need to obtain information about the initial model, or

[0405] The longitudinal federated learning client refuses to receive information about the initial model, or

[0406] Vertical federated learning clients request information about the new model, or

[0407] The longitudinal federated learning client does not support information from the second initial model.

[0408] The second initial model is the initial model information that the longitudinal federated learning client has obtained before. It can be obtained by being generated locally by the federated learning client or by being sent to the federated learning client by the federated learning server.

[0409] As a possible implementation of the first instruction information, it can be explicit or implicit.

[0410] For example, if the vertical federated learning preparation response message contains or does not contain explicit indication information, it is assumed by default that the vertical federated learning client needs to obtain the initial model information, while the additional indication information 0 or 1 in the message indicates that it does not need to obtain / rejects obtaining the initial model information.

[0411] It is understandable that the specific implementation of the first instruction information is not limited to the examples above.

[0412] S607, the vertical federated learning server determines the second information based on the first instruction information, the second information including specific vertical federated initial model information or information that does not need to be sent to the vertical federated learning client.

[0413] In steps S608-S609, the vertical federated learning server sends a second request response message to the vertical federated learning client. The vertical federated learning training request message includes the second information determined in step S607 above. The second information includes the vertical federated initial model information, thereby instructing the vertical federated learning client to perform model training.

[0414] In a possible implementation, if the first indication information in steps S605 and S606 indicates that the vertical federated learning client already has the initial model information or the vertical federated learning client does not need to obtain the initial model information, then in step S608, the vertical federated learning server will no longer send the initial model information to the vertical federated learning client in the vertical federated learning training request message.

[0415] In possible implementations, the aforementioned second request message is sent by the vertical federated learning server to the vertical federated learning client through the network exposure function (NEF) network element, or it is sent directly by the vertical federated learning server to the vertical federated learning client.

[0416] Therefore, through steps S604, S605, S606 and S607 / S608 above, the vertical federated learning client transmits first instruction information to the vertical federated learning server, enabling the vertical federated learning server to determine whether initial model information needs to be sent to the vertical federated learning client, and which type and which initial model information to send. Based on the first instruction information, the vertical federated learning server accurately determines the initial learning model information to be sent to the vertical federated learning client (including whether it needs to be sent, and which model information to send). This avoids inconsistencies between the server and client caused by the vertical federated learning server solely selecting and determining the initial model information, thus achieving the technical effect of negotiating the initial model information between the server and the client, optimizing the transmission of initial model information, and reducing the signaling overhead of renegotiation.

[0417] It should be noted that the vertical federation server function and the vertical federation client function can be deployed in the network in various forms. Embodiments one to eight of this application are merely examples.

[0418] The vertical federated learning server is implemented by the application function (AF) network element, while the vertical federated learning client is implemented by the network data function (NWDAF); or

[0419] The vertical federated learning client is implemented by the Application Function (AF) network element, while the vertical federated learning server is implemented by the Network Data Function (NWDAF); or

[0420] The vertical federated learning client is implemented using the Network Data Function (NWDAF), and the vertical federated learning server is also implemented using the Network Data Function (NWDAF).

[0421] It also includes any reasonable combination that can be extended or inferred by those skilled in the art through this application.

[0422] It should be noted that the negotiation process for the initial model information in the above embodiments can be further extended to the negotiation of models in a general sense or the updating of machine learning models. This embodiment only uses the initial model as an example for illustration.

[0423] Figure 10 is a seventh embodiment of a method for obtaining a vertical federated model provided in this application.

[0424] The embodiment shown in Figure 10, 7, aims to proactively select federated learning clients that may support one or more machine learning models by selecting or reselecting vertical federated learning clients through the vertical federated learning server, thereby achieving the goal of negotiating initial model information between the server and the client.

[0425] S701~S703, Network Function Registration Process.

[0426] S701, the vertical federated learning client sends a network function registration request message to the network storage function through the network open function. The registration request message contains the vertical federated learning capability information of the vertical federated learning client and the fourth instruction information.

[0427] The fourth indication information specifically includes:

[0428] The longitudinal federated learning client already has information about the initial model, or

[0429] The longitudinal federated learning client does not have information about the initial model, or

[0430] The longitudinal federated learning client requests information about the initial model, or

[0431] Vertical federated learning clients do not need to obtain information about the initial model, or

[0432] The longitudinal federated learning client refuses to receive information about the initial model, or

[0433] Vertical federated learning clients request information about the new model, or

[0434] The longitudinal federated learning client does not support information from the second initial model.

[0435] The second initial model is the initial model information that the longitudinal federated learning client has obtained before the above-mentioned network function registration process. It can be obtained by being generated locally by the federated learning client or by being sent to the federated learning client by the federated learning server.

[0436] As a possible implementation of the fourth instruction information, it can be explicit or implicit.

[0437] For example, if the vertical federated learning preparation response message contains or does not contain explicit indication information, it is assumed by default that the vertical federated learning client needs to obtain the initial model information, while the additional indication information 0 or 1 in the message indicates that it does not need to obtain / rejects obtaining the initial model information.

[0438] It is understandable that the specific implementation of the fourth instruction information is not limited to the examples above.

[0439] S702, the network storage function stores the vertical federated learning capability information and / or first instruction information of the vertical federated learning client obtained in step S701.

[0440] S703, the Network Storage Function (NRF) sends a Network Function Registration Response message to the Vertical Federated Learning Client via the Network Open Function.

[0441] S704~S705, the process of the vertical federated learning server initiating the discovery of the vertical federated learning client;

[0442] In one possible implementation,

[0443] S704, the vertical federated learning server sends a second discovery request message to the network repository function (NRF) network element. The discovery request message carries vertical federated learning-related information required by the vertical federated learning server, as well as application function type-related information. This application function type-related information is used by the network repository function network element to discover and select the vertical federated learning client requested by the vertical federated learning server.

[0444] In step S705, the network repository function (NRF) element obtains the application function (AF) identifier based on the discovery request message and sends this identifier, along with a second discovery request response message, to the vertical federated learning server. The application function (AF) identifier is the identifier of the vertical federated learning client requested by the vertical federated learning server. Furthermore, the discovery request response message carries the fourth indication information transmitted in step S701 and stored in the network repository function (NRF) in step S702.

[0445] S706, based on the aforementioned fourth information, the vertical federated learning server determines whether to select the vertical federated learning client to participate in this vertical federated learning task. As one possible implementation, the vertical federated learning client can be identified using an Application Function Instance (AF instance ID). If a new vertical federated learning client needs to be selected, steps S707-S708 are triggered.

[0446] S707~S708, optional

[0447] S707, the vertical federated learning server sends a first discovery request message to the network repository function (NRF) element. This request message carries third information to indicate to the NRF information about federated learning clients that need to be discovered. The third information includes one or more of the following:

[0448] 1) The instruction request to obtain a vertical federated learning client requires support for obtaining an initial model from a vertical federated learning server, or

[0449] 2) The request indicates that the longitudinal federated learning client does not need to obtain information about the initial model, or

[0450] 3) Instructions to request information about an existing initial model in the longitudinal federated learning client, or

[0451] 4) Instructions to request information that the initial model does not exist in the longitudinal federated learning client, or

[0452] 5) Instructs requests to obtain information supporting specific federated learning client capabilities.

[0453] S708, the network repository function (NRF) element obtains the application function (AF) identifier based on the discovery request message, and sends the first indication information (e.g., the application function (AF) identifier) ​​in a first discovery request response message to the vertical federated learning server. The application function (AF) identifier is the identifier of one or more vertical federated learning clients requested by the vertical federated learning server.

[0454] S709, the vertical federated learning server determines the second information based on the first instruction information. The second information includes specific vertical federated initial model information, information that does not need to be sent to the vertical federated learning client, or the identifiers of one or more vertical federated learning clients.

[0455] In steps S710-S711, the vertical federated learning server sends a first message to one or more vertical federated learning clients determined in step S709. The first message includes the second information determined in step 709, which includes initial vertical federated learning model information or client identification information participating in vertical federated learning, thereby instructing the vertical federated learning clients to perform model training.

[0456] The vertical federated learning client identification information includes a client identifier ID, such as an application function ID, client address information, or FQDN information.

[0457] As one possible implementation, in step S710, the first message carries the Application Function (AF) identifier. This AF identifier is determined by the Vertical Federated Machine Learning Server in step S709 and falls under the category of the second information. After the S710 message is sent to the Network Open Function (NEF), the AF identifier, as a type of Vertical Federated Learning client identifier, is converted into an actually addressable FQDN (Fully Qualified Domain Name) identifier or IP address identifier, and may no longer be carried as information in the first message of step S711.

[0458] Optionally, the first message in steps S710 to S711 above may also include longitudinal federated learning initialization model information.

[0459] In a possible implementation, if the first indication information in step S708 indicates that the vertical federated learning client already has the initial model information or the vertical federated learning client does not need to obtain the initial model information, then in step S710, the vertical federated learning server will no longer send the initial model information to the vertical federated learning client in the vertical federated learning training request message.

[0460] In possible implementations, the aforementioned first request message is sent by the vertical federated learning server to the vertical federated learning client through a network exposure function (NEF) network element, or it is sent directly by the vertical federated learning server to the vertical federated learning client.

[0461] Therefore, through the above steps S701, S705, S706 and S707, S708, the federated learning server actively selects federated learning clients that may support one or more machine learning models by selecting or reselecting vertical federated learning clients, thereby achieving the goal of negotiating initial model information between the server and the client.

[0462] Based on the first instruction information, the vertical federated learning server accurately determines the initial learning model information that needs to be transmitted to the vertical federated learning client (including whether to send it and what kind of model information to send). This avoids the inconsistency between the server and the client caused by the vertical federated learning server selecting and determining the initial model information alone. In this way, the technical effect of negotiating the initial model information between the server and the client is achieved, the effect of optimizing the transmission of the initial model information is achieved, and the signaling overhead of renegotiation is reduced.

[0463] It should be noted that the vertical federation server function and the vertical federation client function can be deployed in the network in various forms. Embodiments one to eight of this application are merely examples.

[0464] The vertical federated learning server is implemented by the Application Function (AF) network element, while the vertical federated learning client is implemented by the Network Data Function (NWDAF); or

[0465] The vertical federated learning client is implemented by the Application Function (AF) network element, while the vertical federated learning server is implemented by the Network Data Function (NWDAF); or

[0466] The vertical federated learning client is implemented using the Network Data Function (NWDAF), and the vertical federated learning server is also implemented using the Network Data Function (NWDAF).

[0467] It also includes any reasonable combination that can be extended or inferred by those skilled in the art through this application.

[0468] It should be noted that the negotiation process for the initial model information in the above embodiments can be further extended to the negotiation of models in a general sense or the updating of machine learning models. This embodiment only uses the initial model as an example for illustration.

[0469] Figure 11 is an eighth embodiment of a method for obtaining a vertical federated model provided in this application.

[0470] The eighth embodiment shown in Figure 11 aims to address the problem of initial model negotiation by centering on a vertical federated learning client.

[0471] S801, the vertical federated learning server is ready to start a training task, including being triggered by the vertical federated learning server based on internal conditions, or being triggered by the vertical federated learning server receiving an external instruction.

[0472] S802, as an optional step in this embodiment, the vertical federated learning server initiates the process of discovering the vertical federated learning client;

[0473] In one possible implementation, the vertical federated learning server sends a discovery request message to the network repository function (NRF) element. This message carries information required by the vertical federated learning server regarding vertical federated learning, including information related to the application function type. This application function type information is used by the NRF element to discover and select the vertical federated learning client requested by the server. Accordingly, the NRF element obtains the application function (AF) identifier based on the discovery request message and sends it to the vertical federated learning server, carrying the AF identifier in a discovery request response message. The application function (AF) identifier serves as the identifier for the vertical federated learning client requested by the server.

[0474] S803~S804, the vertical federated learning server sends a vertical federated learning preparation request to the vertical federated learning client, which transmits one or more preparation information to the vertical federated learning client, including vertical federated learning task association identifier, vertical federated learning training preparation information, vertical federated learning filtering information, etc.

[0475] In possible implementations, the aforementioned preparation request message is sent by the vertical federated learning server to the vertical federated client through the network exposure function (NEF) network element, or it is sent directly by the vertical federated learning server to the vertical federated client.

[0476] S805~S806, the vertical federation client sends a vertical federation learning preparation response message to the vertical federation learning server.

[0477] In possible implementations, the aforementioned preparation response message is sent by the vertical federated learning client to the vertical federated server through the network exposure function (NEF) network element, or directly by the vertical federated learning client to the vertical federated server.

[0478] S807-S808, the vertical federated learning server sends a vertical federated learning training request message to the vertical federated learning client. The vertical federated learning training request message carries fifth information, which includes the analysis identifier (Analysis ID) and / or the vertical federated initial model information, thereby instructing the vertical federated learning client to perform model training.

[0479] The analysis identifier can be used to characterize the type of analysis business or analysis service. This service is associated with a model, meaning the model can be used to execute the service. Alternatively, the analysis identifier is associated with a model, meaning the model is used to execute the service corresponding to the analysis identifier.

[0480] In a possible implementation, the aforementioned vertical federated training request message is sent by the vertical federated learning server to the vertical federated client through the network exposure function (NEF) network element, or directly by the vertical federated learning server to the vertical federated learning client; S809. The vertical federated learning client receives the vertical federated initial model information and determines the first initial machine learning model information based on the second initial machine learning model in the fifth information and the third initial machine learning model in the locally stored sixth information.

[0481] Specifically, this includes, but is not limited to, the following processing:

[0482] 1. If the second initial machine learning model conflicts with the third initial machine learning model, the vertical federated learning client selects the second initial machine learning model as the first initial machine learning model; or

[0483] 2. If the vertical federated learning client does not have a local storage for or generate a third initial machine learning model, the vertical federated learning client selects to use the second initial machine learning model as the first initial machine learning model; or

[0484] 3. If, during steps S807 to S808, the vertical federated learning client does not receive any initial model information in the fifth message, then the vertical federated learning client selects to use the local third initial machine learning model as the first initial machine learning model; or

[0485] 4. If the vertical federated learning client determines that no model is available, it will actively request the vertical federated learning server to obtain the first initial machine learning model. The steps are shown in Figure 10, Example 6, steps S604 to S606 and steps S608 to S609. The details will not be repeated here.

[0486] S810, the vertical federated learning client begins local model training.

[0487] S811~S812. The vertical federated client sends a vertical federated learning training response message to the vertical federated learning server.

[0488] Optionally, the federated learning training response message carries the first initial machine learning model information determined by the longitudinal federated learning client in step S809.

[0489] In possible implementations, the training response message is sent from the vertical federated learning client to the vertical federated learning server via a network exposure function (NEF) element, or directly from the vertical federated learning client to the vertical federated learning server.

[0490] Therefore, by using steps S807-S808 and S809 above, the initial model is determined with the vertical federated learning client as the center. This avoids the inconsistency between the server and the client caused by the vertical federated learning server selecting and determining the initial model information alone. This achieves the technical effect of negotiating the initial model information between the server and the client, optimizes the transmission of the initial model information, and reduces the signaling overhead of renegotiation.

[0491] It should be noted that the vertical federation server function and the vertical federation client function can be deployed in the network in various forms. Embodiments one to eight of this application are merely examples.

[0492] The vertical federated learning server is implemented by the application function (AF) network element, while the vertical federated learning client is implemented by the network data function (NWDAF); or

[0493] The vertical federated learning client is implemented by the Application Function (AF) network element, while the vertical federated learning server is implemented by the Network Data Function (NWDAF); or

[0494] The vertical federated learning client is implemented using the Network Data Function (NWDAF), and the vertical federated learning server is also implemented using the Network Data Function (NWDAF).

[0495] It also includes any reasonable combination that can be extended or inferred by those skilled in the art through this application.

[0496] It should be noted that the negotiation process for the initial model information in the above embodiments can be further extended to the negotiation of models in a general sense or the updating of machine learning models. This embodiment only uses the initial model as an example for illustration.

[0497] The above mainly describes the solutions provided by the embodiments of this application from the perspective of interaction between various network elements. Correspondingly, the embodiments of this application also provide a communication device for implementing the various methods described above. This communication device can be a network data analysis function network element in the above method embodiments, or a device containing the above network data analysis function network element, or a component that can be used in a network data analysis function network element device; or, the communication device can be a network storage function network element in the above method embodiments, or a device containing the above network storage function network element, or a component that can be used in a network storage function network element. It is understood that, in order to implement the above functions, the communication device includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, in conjunction with the units and algorithm steps of the various examples described in the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0498] This application embodiment can divide the communication device into functional modules according to the above method embodiment. For example, each function can be divided into a separate functional module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module. It should be noted that the module division in this application embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.

[0499] Taking a communication device as an example of a network data analysis function network element or a network storage function network element in the above method embodiments, Figure 12 is a schematic diagram of the structure of a communication device provided in an embodiment of this application. As shown in Figure 12, the communication device 900 includes: a processing module 901 and a transceiver module 902. The processing module 901 is used to execute the processing functions of the network data analysis function network element or the network storage function network element in the above method embodiments. The transceiver module 902 is used to execute the transceiver functions of the network data analysis function network element, the network storage function network element, or the application function network element in the above method embodiments.

[0500] All relevant content of each step involved in the above method embodiments can be referenced from the functional description of the corresponding functional module, and will not be repeated here.

[0501] Since the communication device 900 provided in this embodiment can execute the above information transmission method, the technical effects it can achieve can be referred to the above method embodiment, and will not be repeated here.

[0502] In one possible design, the transceiver module 902 may include a receiving module and a transmitting module (not shown in Figure 12). The transceiver module is used to implement the transmitting and receiving functions of the communication device 900.

[0503] In one possible design, the communication device 900 may further include a storage module (not shown in FIG12) that stores programs or instructions. When the processing module 901 executes the program or instructions, the communication device 900 can perform the functions of the network data analysis function element or the network storage function element in any of the methods shown in FIG1 to FIG11.

[0504] It should be understood that the processing module 901 involved in the communication device 900 can be implemented by a processor or processor-related circuit components, and can be a processor or processing unit; the transceiver module 902 can be implemented by a transceiver or transceiver-related circuit components, and can be a transceiver or transceiver unit.

[0505] For example, FIG13 is a schematic diagram of another communication device provided in an embodiment of this application. This communication device may be a network data analysis function network element or a network storage function network element, or it may be a chip (system) or other component or assembly that can be disposed in the network data analysis function network element or the network storage function network element. As shown in FIG13, the communication device 1000 may include a processor 1001. In one possible design, the communication device 1000 may further include a memory 1002 and / or a transceiver 1003. The processor 1001 is coupled to the memory 1002 and the transceiver 1003, for example, they can be connected via a communication bus.

[0506] The following is a detailed description of each component of the communication device 1000, with reference to Figure 13:

[0507] The processor 1001 is the control center of the communication device 1000. It can be a single processor or a collective term for multiple processing elements. For example, the processor 1001 can be one or more central processing units (CPUs), application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement the embodiments of this application, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).

[0508] In one possible design, the processor 1001 can perform various functions of the communication device 1000 by running or executing software programs stored in the memory 902 and calling data stored in the memory 1002.

[0509] In a specific implementation, as one example, the processor 1001 may include one or more CPUs, such as CPU0 and CPU1 shown in FIG13.

[0510] In a specific implementation, as one embodiment, the communication device 900 may also include multiple processors, such as processors 1001 and 1004 shown in FIG. 9. Each of these processors may be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). Here, a processor may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0511] The memory 1002 is used to store the software program that executes the solution of this application, and is controlled by the processor 901 to execute it. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.

[0512] In one possible design, the memory 1002 can be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or it can be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto. The memory 1002 can be integrated with the processor 1001 or exist independently and coupled to the processor 1001; this application embodiment does not specifically limit this.

[0513] The transceiver 1003 is used for communication with other communication devices. For example, if the communication device 1000 is a network data analysis function network element, the transceiver 1003 can be used to communicate with network storage function network elements, etc. As another example, if the communication device 1000 is a network storage function network element, the transceiver 1003 can be used to communicate with network data analysis function network elements, etc.

[0514] In one possible design, transceiver 1003 may include a receiver and a transmitter (not shown separately in Figure 13). The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function.

[0515] In one possible design, the transceiver 1003 can be an input / output interface or interface circuit for inputting and / or outputting signals.

[0516] In one possible design, the transceiver 1003 can be integrated with the processor 1001, or it can exist independently and be coupled to the processor 1001. This application embodiment does not specifically limit this.

[0517] It should be noted that the structure of the communication device 1000 shown in Figure 13 does not constitute a limitation on the communication device. The actual communication device may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0518] Furthermore, the communication device 1000 can execute the above-described information transmission method, and therefore the technical effects it can achieve can be referred to the above-described method embodiments, which will not be repeated here.

[0519] In one possible implementation, this application also provides a computer-readable storage medium storing a computer program or instructions that, when executed by a computer, implement the functions of the above-described method embodiments.

[0520] In one possible implementation, this application also provides a computer program product that, when executed by a computer, implements the functions of the above-described method embodiments.

[0521] In one possible implementation, this application embodiment also provides a communication system, which includes the network data analysis function network element and the network storage function network element described in the above method embodiments.

[0522] In one possible implementation, this application also provides a communication method, which includes the method described in any of the above-described method embodiments or any implementation thereof.

[0523] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented using software programs, implementation can be, in whole or in part, in the form of a computer program product. This computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device including one or more servers, data centers, etc., that can be integrated with the medium. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium, or a semiconductor medium (e.g., a solid-state drive (SSD)).

[0524] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0525] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0526] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0527] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0528] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0529] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0530] Although this application has been described herein in conjunction with various embodiments, those skilled in the art, by reviewing the accompanying drawings, the disclosure, and the appended claims, will understand and implement other variations of the disclosed embodiments in carrying out the claimed application. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude multiple instances. A single processor or other unit can implement several functions listed in the claims. While different dependent claims may recite certain measures, this does not mean that these measures cannot be combined to produce good results.

[0531] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the scope of this application. Accordingly, this specification and drawings are merely exemplary illustrations of the application as defined by the appended claims, and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Thus, if such modifications and modifications fall within the scope of the claims and their equivalents, this application is also intended to include such modifications and modifications.

Claims

1. A communication method characterized by comprising: The method includes: The vertical federated learning server receives the first instruction information; The vertical federated learning server determines the second information based on the first instruction information; The vertical federated learning server sends a first message to the vertical federated learning client to instruct the vertical federated learning client to train the model. The first message includes the second information.

2. The method of claim 1, wherein, The first indication information includes: The vertical federated learning client already has information about the initial model, or The vertical federated learning client does not have information about the initial model, or The vertical federated learning client requests information about the initial model, or The vertical federated learning client does not need to obtain information about the initial model, or The vertical federated learning client refuses to obtain information about the initial model, or The vertical federated learning client does not support information from the second initial model, or The vertical federated learning server receives information from one or more vertical federated learning clients from the Network Storage Function (NRF).

3. The method according to claim 1, wherein the vertical federated learning server sends a first message to the vertical federated learning client, including the vertical federated learning server sending the first message to the federated learning client through a Network Function Open Network Element (NEF), or the vertical federated learning server sending the first message directly to the federated learning client.

4. The method of claim 3, wherein, The first message is a vertical federated learning training request message.

5. The method according to any one of claims 1 to 4, characterized in that, The second information includes initial model information and / or first longitudinal federated learning client information.

6. The method of claim 5, wherein, The first vertical federated learning client information includes client identification information, client address information, or FQDN information.

7. The method of claim 1, wherein, The vertical federated learning server receiving the first instruction information specifically includes: the vertical federated learning server receiving the first instruction information from a network storage function entity, or the vertical federated learning server receiving the first instruction information from a network open function entity, or the vertical federated learning server receiving the first instruction information from a vertical federated learning client.

8. The method according to claim 7, characterized in that, The vertical federated learning server receiving the first instruction information specifically includes: The vertical federated learning server obtains first indication information by receiving a vertical federated learning preparation response message from the Network Functions Open Functions entity, or The vertical federated learning server obtains the first indication information by receiving network element discovery response messages from the network storage function.

9. The method of claim 2, wherein, Before the longitudinal federated learning server receives information from one or more longitudinal federated learning clients from the Network Storage Function (NRF), it also includes: The vertical federated learning server sends a network element discovery request message to the network storage function (NRF).

10. The method of claim 9, wherein, The network element discovery request message further includes third information to indicate to the network storage function (NRF) information about federated learning clients that need to be discovered.

11. The method of claim 10, wherein, The third piece of information specifically includes: The instruction requests information about the federated learning client that requires the federated learning server to provide an initial model, or The instruction requests information about federated learning clients that will not reject the federated learning server's provision of an initial model, or The instruction requests information about federated learning clients that do not have a local model, or The instruction requests information to support specific federated learning client capabilities.

12. The method according to any one of claims 9-11, characterized in that, Before the vertical federated learning server sends the network element discovery request message to the network storage function (NRF), it also includes... The vertical federated learning server receives a second network function discovery request response message from the network storage function. The second network element discovery request response message contains fourth information. Based on the fourth information, the vertical federated learning server determines whether to select the current application function instance (AF instance) as the vertical federated learning client.

13. The method of claim 12, wherein, The fourth piece of information includes: The longitudinal federated learning client already has information about the initial model or The longitudinal federated learning client does not have information about the initial model or The longitudinal federated learning client requests information about the initial model, or Vertical federated learning clients do not need to obtain information about the initial model, or The longitudinal federated learning client does not support information from the second initial model, or The longitudinal federated learning client rejects information from the second initial model, or Information from one or more longitudinal federated learning clients of the Network Storage Function (NRF).

14. The method of claim 1, wherein, The vertical federated learning server receiving the first indication information is characterized in that: The vertical federated learning server receives the first indication information carried in the vertical federated learning preparation response message from the Network Open Functions Entity (NEF), or The vertical federated learning server receives the first indication information carried in the Network Function NF Discovery Request Response message from the Network Storage Function Entity (NRF), or The vertical federated learning server receives the first instruction information carried in the second request message from the network open function.

15. The method of any one of claims 1 to 14, wherein, The vertical federated learning server can be a network data analysis function element (NWDAF) or an application function entity (AF).

16. The method of any one of claims 1 to 14, wherein, The vertical federated learning client can be a network data analysis function element (NWDAF) or an application function entity (AF).

17. A method of communication, comprising: The method includes: The vertical federated learning client receives a vertical federated training request message, which carries fifth information; The longitudinal federated learning client determines the first initial model information based on the fifth piece of information and the sixth piece of information stored locally; The vertical federated learning client sends the initial model information to the network open function entity or the vertical federated learning client sends the initial model information to the vertical federated learning server.

18. The method of claim 17, wherein, The fifth piece of information includes analysis identifiers and / or second initial model information.

19. The method of claim 17 or 18, wherein, The sixth piece of information includes the third initial model information stored locally on the vertical federation client.

20. The method of any one of claims 17-19, wherein, The vertical federated learning client determines the first initial model information based on the fifth information and the locally stored sixth information, specifically including: If the fifth and sixth pieces of information are inconsistent, the vertical federated learning client selects to use the second initial model information from the fifth pieces of information as the first initial model information for model training, or If the sixth information does not include the third initial model information stored locally by the vertical federated client, the vertical federated learning client selects to use the second initial model information from the fifth information as the first initial model information for model training, or If the fifth piece of information does not include initial model information, the longitudinal federated learning client selects to use the third initial model information from the sixth piece of information as the first initial model information for model training, or If the fifth piece of information does not include initial model information and the sixth piece of information does not include initial model information, the vertical federated learning client sends a request message to the vertical federated learning server to obtain initial model information.

21. A communication device, characterized in that, The communication device includes a module or unit for performing the method according to any one of claims 1-20.

22. A communication device, characterized in that, The communication device includes a processor configured to cause the communication device to perform the method according to any one of claims 1-20 by means of logic circuits and / or executing instructions.

23. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes instructions that, when executed by a processor, cause the method according to any one of claims 1-20 to be implemented.

24. A computer program product, characterized in that, The computer program product includes instructions that, when executed on a computer, cause the computer to perform the method according to any one of claims 1-20.

Images