Vertical federated learning methods, vertical federated learning system, communication entity and storage medium

Abstract

Disclosed in the present application are vertical federated learning methods, a vertical federated learning system, a communication entity and a storage medium. A method comprises: sending a vertical federated learning joining request to a vertical federated learning client (701); receiving vertical federated learning response information fed back by the vertical federated learning client on the basis of the vertical federated learning joining request (702); and on the basis of the vertical federated learning response information, determining the vertical federated learning client, which participates in vertical federated learning (703).

Description

Method, system, communication entity and storage medium for vertical federated learning Technical Field

[0001] The present application relates to the field of communication technology, for example, to a method, system, communication entity and storage medium for vertical federated learning. Background Art

[0002] TS23.288 defines the Network Data Analytics Function (NWDAF). NWDAF can perform statistical data and machine learning tasks. NWDAF can interact with different entities for different purposes.

[0003] Currently, although the TS23.288 protocol defines NWDAF for model training based on federated learning, it does not support vertical federated learning processes.

[0004] Summary of the Invention

[0005] This embodiment of the present application provides a method for vertical federated learning, which is applied to a vertical federated learning server. The method includes:

[0006] Send a vertical federated learning joining request to the vertical federated learning client;

[0007] Receiving vertical federated learning response information fed back by the vertical federated learning client in response to the vertical federated learning joining request;

[0008] According to the vertical federated learning response information, a vertical federated learning client participating in the vertical federated learning is determined.

[0009] This embodiment of the present application provides a method for vertical federated learning, which is applied to a vertical federated learning client. The method includes:

[0010] Receive a vertical federated learning joining request sent by a vertical federated learning server;

[0011] According to the vertical federated learning joining request, vertical federated learning response information is fed back to the vertical federated learning server.

[0012] This embodiment of the present application provides a method for vertical federated learning, which is applied to NEF. The method includes:

[0013] Receive a federated learning capability type sent by a communication entity; wherein the federated learning capability type includes at least one of the following: whether it supports serving as a federated learning server, whether it supports serving as a federated learning client, or whether it supports serving as both a federated learning server and a federated learning client; or the federated learning capability type includes at least one of the following: whether it supports serving as a vertical federated learning server, whether it supports serving as a vertical federated learning client, or whether it supports serving as both a vertical federated learning server and a vertical federated learning client;

[0014] The federated learning capability type is stored.

[0015] The embodiment of the present application provides a vertical federated learning system, comprising: a vertical federated learning server and a vertical federated learning client;

[0016] The vertical federated learning server is used to execute the vertical federated learning method executed by the vertical federated learning server in any of the above vertical federated learning methods;

[0017] The vertical federated learning client is used to execute the vertical federated learning method executed by the vertical federated learning client in any of the above-mentioned vertical federated learning methods.

[0018] An embodiment of the present application provides a communication entity, including: a processor; the processor is used to implement the method of vertical federated learning of any of the above embodiments when executing a computer program.

[0019] An embodiment of the present application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method of vertical federated learning in any of the above embodiments.

[0020] With respect to the above embodiments and other aspects of the present application and their implementation, further description is provided in the accompanying drawings, detailed description and claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] FIG1 is a schematic diagram of the architecture of a core network provided by an embodiment;

[0022] FIG2 is a schematic diagram of an architecture for collecting data by a vertical federated learning communication entity provided by an embodiment;

[0023] FIG3 is a schematic diagram of an architecture for collecting data by another vertical federated learning communication entity provided by an embodiment;

[0024] FIG4 is a schematic diagram of an architecture for providing data by a vertical federated learning communication entity according to an embodiment;

[0025] FIG5 is a schematic diagram of an architecture of another vertical federated learning communication entity providing data according to an embodiment;

[0026] FIG6 is a schematic diagram of the structure of a vertical federated learning system provided by an embodiment;

[0027] FIG7 is a schematic diagram of a flow chart of a method for vertical federated learning provided by an embodiment;

[0028] FIG8 is a flow chart of another method for vertical federated learning provided by an embodiment;

[0029] FIG9 is a schematic flow chart of a method for vertical federated learning provided by an embodiment;

[0030] FIG10 is a schematic flow chart of another method for vertical federated learning provided by an embodiment;

[0031] FIG11a is a signaling interaction diagram of a method for vertical federated learning provided by an embodiment;

[0032] FIG11b is a signaling interaction diagram of another method for vertical federated learning provided by an embodiment;

[0033] FIG11c is a signaling interaction diagram of another method for vertical federated learning provided by an embodiment;

[0034] FIG12a is a signaling interaction diagram of another method for vertical federated learning provided by an embodiment;

[0035] FIG12 b is a signaling interaction diagram of another method for vertical federated learning provided by an embodiment;

[0036] FIG12c is a signaling interaction diagram of another method for vertical federated learning provided by an embodiment;

[0037] FIG13 is a signaling interaction diagram of another method for vertical federated learning provided by an embodiment;

[0038] FIG14 is a signaling interaction diagram of another method for vertical federated learning provided by an embodiment;

[0039] FIG15 is a signaling interaction diagram of another method for vertical federated learning provided by an embodiment;

[0040] FIG16 is a schematic diagram of a flow chart of a method for vertical federated learning provided by an embodiment;

[0041] FIG17a is a signaling interaction diagram of a method for vertical federated learning provided by an embodiment;

[0042] FIG17 b is a signaling interaction diagram of another method for vertical federated learning provided by an embodiment;

[0043] FIG18 is a schematic diagram of the structure of a device for vertical federated learning provided by an embodiment;

[0044] FIG19 is a schematic diagram of the structure of another apparatus for vertical federated learning provided by an embodiment;

[0045] FIG20 is a schematic diagram of the structure of another apparatus for vertical federated learning provided by an embodiment;

[0046] FIG21 is a schematic structural diagram of a communication entity provided by an embodiment. DETAILED DESCRIPTION

[0047] It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. The embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0048] The vertical federated learning (VFL) method provided in this application can be applied to fifth-generation mobile communication technology (5th-generation, 5G) systems, LTE and 5G hybrid architecture systems, 5G New Radio (New Radio, NR) systems, and new communication systems emerging in future communication developments, such as sixth-generation mobile communication technology (6th-generation, 6G) systems. Optionally, the vertical federated learning method can be applied in the core network of the communication system, outside the core network, or between the core network and outside the core network.

[0049] Figure 1 is a schematic diagram of a core network architecture provided by an embodiment. As shown in Figure 1, user equipment (UE) is connected to the core network's network functions via the radio access network (RAN). The RAN manages radio resources, transmits user data received over the N3 interface to the UE, and transmits user data from the UE over the N3 interface. The RAN maps between dedicated radio bearers (DRBs) and Quality of Service (QoS) traffic in protocol data unit (PDU) sessions.

[0050] The Access and Mobility Management Function (AMF) includes the following functions: registration management, connection management, reachability management, and mobility management. This function also performs access authentication and access authorization. The AMF is a non-access stratum (NAS) security terminal and forwards SM NAS between the UE and the Session Management Function (SMF).

[0051] The SMF includes the following functions: session establishment, modification, and release; UE Internet Protocol (IP) address allocation and management (including optional authorization functions); user plane function (UPF) selection and control; and downlink data notification. The SMF controls the UPF through the N4 interface. The SMF provides the UPF with packet detection rules (PDRs) to instruct it on how to detect user data traffic. It also provides forwarding action rules (FARs), QoS enforcement rules (QERs), and usage reporting rules (URRs) to instruct the UPF on how to forward user data traffic, perform QoS processing, and perform usage reporting on user data traffic detected using the PDRs.

[0052] The UPF includes the following functions: serving as an anchor point for intra- and inter-Radio Access Technology (RAT) mobility, packet routing and forwarding, traffic usage reporting, user plane QoS processing, downlink packet buffering, and downlink data notification triggering. The GTP-U tunnel is used for the N3 interface between the RAN and the UPF. The GTP-U tunnel is per PDU session. For downlink traffic, the UPF binds the downlink traffic to the QoS traffic within the PDU session GTP-U tunnel using the FAR received from the SMF. For uplink traffic, the RAN transfers the user plane traffic to the QoS flow identified by the UE.

[0053] The Policy Control Function (PCF) provides QoS policy rules to the control plane functions for enforcement. The PCF converts application function (AF) requests into policy control and charging control (PCC) rules applicable to the PDU session.

[0054] Unified Data Management (UDM) performs the generation of 3rd Generation Partnership Project (3GPP) authentication and key agreement (AKA) credentials, authorization of access to subscription-based data, UE registration with the serving Network Function (NF) (e.g., provisioning the UE's storage service, AMF, and PDU session storage service, SMF), and subscription management. UDM accesses the Unified Data Repository (UDR) to retrieve UE subscription data and stores the UE context in the UDR. UDM and UDR can be deployed together.

[0055] The method for vertical federated learning provided in the embodiment of the present application can be applied to a federated learning communication entity. The federated learning communication entity can be an NWDAF or an AF. The NWDAF is a 5G core network (5G Core, 5GC) NF located in the control plane, which performs statistical data and machine learning related tasks in the 5GC. The AF in the embodiment of the present application can be an AF within the core network or an AF outside the core network.

[0056] NWDAF can interact with different communication entities for different purposes: 1. Collect data based on event subscriptions provided by AMF, SMF, UPF, PCF, UDM, Network Slice Admission Control Function (NSACF), AF (directly or through Network Exposure Function (NEF)) and Operation Administration and Maintenance (OAM); 2. (Optional) Perform analysis and data collection using the Data Collection Coordination Function (DCCF); 3. Retrieve information from the data repository, for example, retrieve user-related information UDR through UDM or retrieve PFD information through NEF (Packet Flow Descriptions Function (PFDF)); 4. Collect location information data from the Location Services (LCS) system; 5. (Optional) Store and retrieve information from the Analytic Data Repository Function (ADRF); 6. (Optional) Store and retrieve information from the Messaging Framework Adaptor Function (MFFA) Function (MFAF) analyzes and collects data; 7. Retrieve information about NF, for example, retrieve NF-related information from Network Repository Function (NRF); 8. Provide analysis to consumers on demand; 9. Provide batch data related to analysis identity (ID); 10. Provide accuracy information of analysis ID; 11. Provide ML model accuracy information or ML model accuracy degradation indication about machine learning (ML) models.

[0057] A single instance or multiple instances of NWDAF can be deployed in a Public Land Mobile Network (PLMN). NWDAF can include the following logical functions: Analytics Logical Function (AnLF), which is a logical function in NWDAF used to perform reasoning, derive analytical information (i.e., derive statistics and / or predictions based on analytical consumer requests) and expose analytical services (i.e., Nnwdaf_AnalyticsSubscription or Nnwdaf_AnalyticsInfo); Model Training Logical Function (MTLF), which is another logical function in NWDAF used to train ML models and expose new training services (e.g., providing trained ML models). An NWDAF can include one MTLF, one AnLF, or both logical functions.

[0058] The DCCF is also a NF on the 5GC control plane. It coordinates the collection and distribution of data requested by NF consumers. It prevents data sources from processing multiple subscriptions to the same data and prevents multiple notifications containing the same information from being sent due to uncoordinated requests from data consumers. The DCCF applies to: NWDAFs requesting data from data sources (e.g., for computational analysis); NF consumers requesting analysis from NWDAF data sources; NF consumers requesting data from ADRF data sources; and ADRFs receiving data from NF data sources.

[0059] Figure 2 is a schematic diagram of the architecture for collecting data by a vertical federated learning communication entity, provided by one embodiment. Figure 3 is a schematic diagram of the architecture for collecting data by another vertical federated learning communication entity, provided by one embodiment. Figure 2 shows that NWDAF can collect data from any NF (also called a data source). Figure 3 shows that NWDAF can collect data from any NF through DCCF. In other words, Figures 2 and 3 show two data collection structures to facilitate NWDAF data analysis and model training.

[0060] FIG4 is a schematic diagram of an architecture for providing data by a vertical federated learning communication entity, provided by one embodiment. FIG5 is a schematic diagram of an architecture for providing data by another vertical federated learning communication entity, provided by one embodiment. FIG4 shows that NWDAF can provide data to any NF. FIG5 shows that NWDAF can provide data to any NF through DCCF. The data here includes at least one of the following: analysis results, statistical results, and prediction information. That is, FIG4 and FIG5 show two possible network data analysis exposure architectures for use by any NF consumer that subscribes to or requests analysis.

[0061] Federated learning includes vertical federated learning and horizontal federated learning. Vertical federated learning is suitable for situations where sample data features overlap slightly but sample identifiers overlap significantly. In other words, in vertical federated learning, the training data for different client models share the same sample space but different feature spaces. Horizontal federated learning is suitable for situations where sample data features overlap significantly but sample identifiers overlap less. In other words, in horizontal federated learning, the training data for different client models share the same feature space but different sample spaces.

[0062] In an embodiment of the present application, a method, system, communication entity and storage medium for vertical federated learning are provided. Through the interaction between a vertical federated learning server and a vertical federated learning client, the vertical federated learning clients participating in the vertical federated learning can be determined, thereby realizing vertical federated learning.

[0063] The following describes the method, system, communication entity and technical effects of vertical federated learning.

[0064] Figure 6 is a structural diagram of a vertical federated learning system provided by an embodiment. As shown in Figure 6, the vertical federated learning system includes: a vertical federated learning server and multiple vertical federated learning clients. The vertical federated learning server in this embodiment can be: NWDAF, AF inside the core network, or AF outside the core network. The vertical federated learning client in this embodiment can include at least one of the following: NWDAF, AF inside the core network, or AF outside the core network. The content of the interaction between the vertical federated learning server and the vertical federated learning client can include at least one of the following: information of the machine learning model, machine learning parameters. The machine learning parameters in this embodiment include: loss information, gradient information, etc. The information of the machine learning model in this embodiment includes: intermediate training results trained by the vertical federated learning client.

[0065] It should be noted that the vertical federated learning server in this embodiment can also be called an active participant of vertical federated learning, and the vertical federated learning client in this embodiment can also be called a passive participant of vertical federated learning.

[0066] Vertical federated learning is applicable to various scenarios, for example, 5GC NF and AF work together to achieve user device experience prediction, public network integration mode (PNI-NPN), etc.

[0067] The method of vertical federated learning provided in this embodiment is used to describe how to construct the vertical federated learning system shown in FIG6 .

[0068] Figure 7 is a schematic flow chart of a method for vertical federated learning provided by one embodiment. The method for vertical federated learning provided by this embodiment is applied to a vertical federated learning server. For example, the vertical federated learning server shown in Figure 6 . As shown in Figure 7 , the method includes the following steps.

[0069] Step 701: Send a vertical federated learning joining request to the vertical federated learning client.

[0070] In one implementation, before step 701, the vertical federated learning server may send a vertical federated learning client discovery request to a third-party communication entity and receive information about vertical federated learning clients determined by the third-party communication entity based on the vertical federated learning client discovery request. Accordingly, based on this implementation, step 701 may be implemented by sending a vertical federated learning join request to at least some of the vertical federated learning clients determined by the third-party communication entity, based on a selection rule.

[0071] Optionally, in this implementation, the information of the vertical federated learning client determined by the third-party communication entity is the information of the client determined by the third-party communication entity according to the selection rule and the vertical federated learning client discovery request.

[0072] The selection rule may be a preset rule, the contents of which will be described in detail in the subsequent embodiments.

[0073] In another implementation, the vertical federated learning client in step 701 may be at least one vertical federated learning client within the service area of ​​the vertical federated learning server. In step 701, the vertical federated learning server sends a vertical federated learning joining request to at least one vertical federated learning client within its service area.

[0074] The vertical federated learning joining request in this embodiment is used to request the vertical federated learning client to join the vertical federated learning where the vertical federated learning server is located.

[0075] When the vertical federated learning server determines that vertical federated learning is required, step 701 is executed. The vertical federated learning server determines that vertical federated learning is required in at least one of the following ways: the vertical federated learning server receives a model training request from a consumer-side communication entity (also referred to as a service client); the vertical federated learning server receives a model training request from a federated learning client; or the vertical federated learning server determines that it needs to perform model training.

[0076] Step 702: Receive vertical federated learning response information fed back by the vertical federated learning client in response to the vertical federated learning joining request.

[0077] After the vertical federated learning client receives the vertical federated learning joining request sent by the vertical federated learning server, it can send vertical federated learning response information to the vertical federated learning server according to the vertical federated learning joining request.

[0078] In step 702, the vertical federated learning server receives vertical federated learning response information fed back by the vertical federated learning client in response to the vertical federated learning joining request.

[0079] Optionally, the vertical federated learning response information in this embodiment includes: response information indicating joining the vertical federated learning and response information indicating not joining the vertical federated learning.

[0080] Step 703: Determine the vertical federated learning client that participates in the vertical federated learning based on the vertical federated learning response information.

[0081] In one implementation, the vertical federated learning server determines the vertical federated learning client that is instructed to join the vertical federated learning by the corresponding vertical federated learning response information as the vertical federated learning client that participates in the vertical federated learning.

[0082] In another implementation, the vertical federated learning server determines the vertical federated learning clients that participate in the vertical federated learning from the vertical federated learning clients that have been instructed to join the vertical federated learning by the vertical federated learning response information according to the selection rule.

[0083] It should be noted that the number of vertical federated learning clients participating in vertical federated learning in this embodiment may be one or more. For example, the number of vertical federated learning clients participating in vertical federated learning in this embodiment may be two or more.

[0084] After step 703, it is equivalent to establishing a vertical federated learning system where the vertical federated learning server is located. Subsequently, the specific process of vertical federated learning can be carried out based on the system to perform model training. A possible training process is: the vertical federated learning server sends a model training request to the vertical federated learning client participating in the vertical federated learning. The vertical federated learning client participating in the vertical federated learning collects training data and performs model training based on the model training request. During the training process, it is possible to interact with the vertical federated learning server to transmit information about the machine learning model and / or machine learning parameters. The training data of each vertical federated learning client participating in the vertical federated learning in this embodiment has the same sample space and different feature spaces.

[0085] The method for vertical federated learning provided in this embodiment is applied to a vertical federated learning server and includes: sending a vertical federated learning participation request to a vertical federated learning client; receiving vertical federated learning response information from the vertical federated learning client in response to the vertical federated learning participation request; and determining the vertical federated learning client participating in the vertical federated learning based on the vertical federated learning response information. This method for vertical federated learning, through interaction between the vertical federated learning server and the vertical federated learning client, can determine the vertical federated learning client participating in the vertical federated learning, thereby implementing vertical federated learning.

[0086] Figure 8 is a flowchart illustrating another method for vertical federated learning provided by one embodiment. Building on the embodiment shown in Figure 7 and various optional implementations, this embodiment details how to determine a vertical federated learning server and how to specifically determine a vertical federated learning client. As shown in Figure 8, the method for vertical federated learning provided by this embodiment includes the following steps.

[0087] Step 801: Send federated learning capability type or federated learning capability update information to a third-party communication entity.

[0088] Among them, the federated learning capability type and the federated learning capability update information both include at least one of the following: whether it supports being a vertical federated learning server, whether it supports being a vertical federated learning client, and whether it supports being both a vertical federated learning server and a vertical federated learning client.

[0089] Optionally, the vertical federated learning server in this embodiment includes at least one of the following: NWDAF, AF within the core network, or AF outside the core network.

[0090] In one implementation, when the vertical federated learning server is an application function outside the core network, the third-party communication entity is NEF or NRF. When the third-party communication entity is NRF, the vertical federated learning server can send the federated learning capability type or federated learning capability update information to the NRF through NEF. Furthermore, the vertical federated learning server sends the federated learning capability type or federated learning capability update information to the NEF. The NEF receives and stores the federated learning capability type or federated learning capability update information sent by the vertical federated learning server. The NEF can also send the federated learning capability type or the federated learning capability update information to the NRF. The NRF receives and stores the federated learning capability type or the federated learning capability update information.

[0091] In another implementation, when the vertical federated learning server is NWDAF or AF within the core network, the third-party communication entity is NRF.

[0092] After receiving the federated learning capability type or federated learning capability update information, the third-party communication entity stores it.

[0093] Optionally, the third-party communication entity may also send a response message of the federated learning capability type to the vertical federated learning server, or send a response message of the federated learning capability update information to the vertical federated learning server. Correspondingly, the vertical federated learning server receives the response message of the federated learning capability type or the response message of the federated learning capability update information sent by the third-party communication entity.

[0094] Optionally, the vertical federated learning server in this embodiment may be a server determined by the source communication entity. Further, the vertical federated learning server in this embodiment is a server determined by the source communication entity based on the selection rule and the vertical federated learning server determined by the third-party communication entity based on the selection rule.

[0095] Step 802: Receive a request from a source communication entity to serve as a vertical federated learning server.

[0096] In the first implementation, the source communication entity is the communication entity that receives the model subscription request sent by the consumer communication entity. The model subscription request includes an analysis identifier (Identity, ID). The consumer communication entity in this embodiment can be a 5GC NF, AF, or OAM.

[0097] The analysis ID in this embodiment is used to indicate the model to be trained.

[0098] Optionally, the model subscription request in this embodiment further includes at least one of the following: other information indicating the purpose of the model, a model metric, an expected value corresponding to the model metric, an identifier of the consumer-side communication entity, and address information of the consumer-side communication entity. For example, the expected value corresponding to the model metric may be an accuracy rate of 90%.

[0099] In this implementation, if the source communication entity determines, based on the model subscription request, that the model corresponding to the analysis ID needs to be trained through vertical federated learning, and that the source communication entity itself cannot serve as a vertical federated learning server, the source communication entity selects a vertical federated learning server through a third-party communication entity. A possible implementation process may be: the source communication entity sends information requesting a federated learning server to the third-party communication entity; the third-party communication entity determines the information of the vertical federated learning server based on the selection rules and the information requesting the federated learning server, and sends the determined information to the source communication entity; the source communication entity determines the vertical federated learning server based on the selection rules and the vertical federated learning server determined by the third-party communication entity based on the selection rules; the source communication entity sends a request to the determined vertical federated learning server to serve as a vertical federated learning server; the vertical federated learning server receives the request sent by the source communication entity to serve as a vertical federated learning server.

[0100] Optionally, the request as a vertical federated learning server includes at least one of the following: an identifier of a consumer-end communication entity, and address information of the consumer-end communication entity.

[0101] Optionally, the model metric in this embodiment refers to a standard used to measure the model, such as accuracy, precision, and recall.

[0102] Accuracy is a percentage that is the number of correct predictions divided by the total number of predictions. Model metrics may also include at least one of the following: mean squared error (MSE), root mean squared error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and R-squared.

[0103] The mean square error is the expected value of the square of the difference between the parameter estimate and the true parameter value. The mean square error of the model (denoted as mlModelMSE) is used to indicate the average square error of the model and can be calculated as:

[0104] The root mean square error of the model (denoted as mlModelRMSE) is used to indicate the root mean square error of the model and can be calculated as:

[0105] The mean absolute error of the model (denoted as mlModelMAE) is used to indicate the mean absolute error of the model and can be calculated as:

[0106] The mean absolute proportional error of the model (denoted as mlModelMAPE) is used to indicate the mean absolute proportional error of the model, which can be calculated as:

[0107] The R square of the model (denoted as R 2 ) is used to indicate the quality of fit of the model and can be calculated as:

[0108] In the above formula, n represents the number of samples, Y is the calibration data (label / ground truth data), γ represents the model's predicted output, and ym is the mean of the samples. All data types involved in the above formula are floating-point numbers.

[0109] It can be understood that the number of vertical federated learning servers determined by the third-party communication entity based on the selection rules and the information requested from the federated learning server is greater than or equal to the number of vertical federated learning servers determined by the communication entity based on the selection rules and the vertical federated learning servers determined by the third-party communication entity based on the selection rules.

[0110] In the second implementation, the source communication entity is a vertical federated learning client that needs to perform model training. Based on this implementation, the vertical federated learning method provided in this embodiment further includes the following steps: determining the source communication entity as a vertical federated learning client that participates in the vertical federated learning.

[0111] In this implementation, if the source communication entity possesses some data but lacks incomplete features or calibration data (called ground truth data or labels), it selects a vertical federated learning server through a third-party communication entity. The specific process is similar to that in the first implementation, except that the request to the vertical federated learning server in this implementation includes at least one of the following: the source communication entity's identifier or the source communication entity's address information.

[0112] Step 803: According to the selection rule, send response information as a vertical federated learning server to the source communication entity.

[0113] The vertical federated learning server determines that it can participate in the vertical federated learning as a server according to the selection rules, and sends a response message as a vertical federated learning server to the source communication entity.

[0114] Based on the second implementation in step 802, before step 804, the method for vertical federated learning provided in this embodiment further includes the following step: receiving a vertical federated learning request sent by a source communication entity. The vertical federated learning request includes at least one of the following: an analysis ID, information about the model corresponding to the analysis ID, the type of the model corresponding to the analysis ID, training environment configuration requirements required for the model corresponding to the analysis ID, an identifier of the source communication entity, and a set identifier (set ID) of the source communication entity.

[0115] Optionally, the identifier of the source communication entity in this embodiment may be an NF ID or a vendor ID.

[0116] Of course, based on the first implementation in step 802, before step 804, a vertical federated learning request sent by the source communication entity may also be received. The vertical federated learning request includes at least one of the following: an analysis ID, information about the model corresponding to the analysis ID, the type of the model corresponding to the analysis ID, training environment configuration requirements required for the model corresponding to the analysis ID, an identifier of the consumer-end communication entity, and a set ID of the consumer-end communication entity.

[0117] Step 804: Send a vertical federated learning client discovery request to a third-party communication entity.

[0118] Optionally, after step 803 , the vertical federated learning server may spontaneously send a vertical federated learning client discovery request to the third-party communication entity.

[0119] Optionally, the vertical federated learning server may send a vertical federated learning client discovery request to the third-party communication entity after receiving the vertical federated learning request sent by the source communication entity.

[0120] Step 805: Receive information of the vertical federated learning client determined by the third-party communication entity according to the vertical federated learning client discovery request.

[0121] Optionally, the information of the vertical federated learning client determined by the third-party communication entity is information of the client determined by the third-party communication entity according to a selection rule and a vertical federated learning client discovery request.

[0122] After receiving the vertical federated learning client discovery request, the third-party communication entity determines the information of the vertical federated learning client according to the selection rule and the vertical federated learning client discovery request.

[0123] Step 806: According to the selection rule, a vertical federated learning joining request is sent to at least some of the vertical federated learning clients determined by the third-party communication entity.

[0124] After receiving the information of the vertical federated learning clients determined by the third-party communication entity, the vertical federated learning server determines at least some of the clients from these vertical federated learning clients according to the selection rules, and sends vertical federated learning joining requests to these at least some of the vertical federated learning clients.

[0125] Based on the first implementation method of step 802, the vertical federated learning joining request in step 806 includes at least one of the following: the identifier of the consumer-end communication entity, the address information of the consumer-end communication entity, the model metric, the expected value corresponding to the model metric, information related to the loss type to be transmitted for vertical federated learning, and the maximum response time for the vertical federated learning client to return the intermediate training results after receiving the loss information in each round of training.

[0126] Based on the second implementation method of step 802, the vertical federated learning joining request in step 806 includes at least one of the following: the identification of the source communication entity, the address information of the source communication entity, the model metric, the expected value corresponding to the model metric, information related to the loss type to be transmitted for vertical federated learning, and the maximum response time for the vertical federated learning client to return the intermediate training results after receiving the loss information in each round of training.

[0127] Step 807: Receive vertical federated learning response information fed back by the vertical federated learning client according to the vertical federated learning joining request.

[0128] The implementation process and technical principle of step 807 are similar to those of step 702 and will not be repeated here.

[0129] Step 808: According to the selection rule, determine the vertical federated learning clients that participate in the vertical federated learning from the vertical federated learning clients whose vertical federated learning response information indicates joining the vertical federated learning.

[0130] In one implementation, in step 808, the vertical federated learning server may first determine that the corresponding vertical federated learning response information is a vertical federated learning client indicating joining the vertical federated learning, and then determine the vertical federated learning client participating in the vertical federated learning from these vertical federated learning clients indicating joining the vertical federated learning based on the selection rules.

[0131] In another implementation, in step 808, the vertical federated learning server may first determine the vertical federated learning clients that meet the selection rules, and then determine the vertical federated learning clients that participate in the vertical federated learning as the vertical federated learning clients that indicate joining the vertical federated learning using the corresponding vertical federated learning response information among these vertical federated learning clients that meet the selection rules.

[0132] Optionally, the selection rules involved in this embodiment include at least one of the following: the type of communication entity (also referred to as NF type), the federated learning capability type of the communication entity, the service area corresponding to the communication entity (also referred to as the service area where the NF is located), whether the communication entity is currently performing other federated learning or model training tasks, analysis ID, the type of communication entity that can collect data sources for local model training, the time period that can be used for vertical federated learning, the interoperability index of the model corresponding to the analysis ID, the information of the model corresponding to the analysis ID, the type of model corresponding to the analysis ID, the training environment configuration requirements required for the model corresponding to the analysis ID, the data availability of the communication entity, the time availability of the communication entity, the computing power information of the communication entity and the communication capability of the communication entity, the information of the network function served by the communication entity, the slicing information of the communication entity, the model sharing information of the communication entity, the gradient calculation method, the gradient optimization algorithm, the loss function type, the gradient descent algorithm, and the maximum response time for the vertical federated learning client to return the intermediate training results after receiving the loss information in each round of training.

[0133] It should be noted that the selection rules involved in the above process may be the same, different, or partially the same. In practical applications, the selection rules can be flexibly determined according to needs.

[0134] The method for vertical federated learning provided in this embodiment provides a method for determining participating vertical federated learning clients. The participating vertical federated learning clients determined based on this method can meet the needs of subsequent vertical federated learning, thereby improving the efficiency and success rate of vertical federated learning. Furthermore, the signaling interaction method for determining participating vertical federated learning clients in this embodiment is concise and efficient, further improving the efficiency of vertical federated learning.

[0135] Figure 9 is a flowchart illustrating a method for vertical federated learning provided by one embodiment. The method for vertical federated learning provided by this embodiment is applied to a vertical federated learning client. For example, the vertical federated learning client shown in Figure 6 . As shown in Figure 9 , the method includes the following steps.

[0136] Step 901: Receive a vertical federated learning joining request sent by a vertical federated learning server.

[0137] Step 901 and step 701 are corresponding features.

[0138] In one implementation, the vertical federated learning client in this embodiment is a client determined by the vertical federated learning server based on the information of the vertical federated learning client determined by the third-party communication entity and a selection rule.

[0139] In another implementation, the vertical federated learning client in this embodiment is a client in the service area of ​​the vertical federated learning server.

[0140] The vertical federated learning joining request in this embodiment is used to request the vertical federated learning client to join the vertical federated learning where the vertical federated learning server is located.

[0141] Step 902: Feedback vertical federated learning response information to the vertical federated learning server based on the vertical federated learning joining request.

[0142] Step 902 and step 702 are corresponding features.

[0143] After receiving the vertical federated learning joining request, the vertical federated learning client can feedback vertical federated learning response information to the vertical federated learning server according to the vertical federated learning joining request.

[0144] Optionally, the vertical federated learning response information in this embodiment includes: response information indicating joining the vertical federated learning and response information indicating not joining the vertical federated learning.

[0145] The method for vertical federated learning provided in this embodiment is applied to a vertical federated learning client and includes receiving a vertical federated learning participation request from a vertical federated learning server and, based on the vertical federated learning participation request, feeding back vertical federated learning response information to the vertical federated learning server. This method for vertical federated learning, through interaction between the vertical federated learning client and the vertical federated learning server, can determine which vertical federated learning clients are participating in the vertical federated learning process, thereby implementing vertical federated learning.

[0146] Figure 10 is a flowchart illustrating another method for vertical federated learning provided by one embodiment. This embodiment, based on the embodiment shown in Figure 9 and various optional implementations, provides a detailed description of the other steps included in this method. As shown in Figure 10, the method for vertical federated learning provided by this embodiment includes the following steps.

[0147] Step 1001: Send federated learning capability type or federated learning capability update information to a third-party communication entity.

[0148] Among them, the federated learning capability type and the federated learning capability update information both include at least one of the following: whether it supports being a vertical federated learning server, whether it supports being a vertical federated learning client, and whether it supports being both a vertical federated learning server and a vertical federated learning client.

[0149] Optionally, the vertical federated learning client in this embodiment includes at least one of the following: NWDAF, AF within the core network, or AF outside the core network.

[0150] In one implementation, when the vertical federated learning client is an application function outside the core network, the third-party communication entity is NEF or NRF. When the third-party communication entity is NRF, the vertical federated learning client can send the federated learning capability type or federated learning capability update information to the NRF through NEF. Furthermore, the vertical federated learning client NEF sends the federated learning capability type or federated learning capability update information. NEF receives and stores the federated learning capability type or federated learning capability update information sent by the vertical federated learning client. NEF can also send the federated learning capability type or the federated learning capability update information to NRF. NRF receives and stores the federated learning capability type or the federated learning capability update information.

[0151] In another implementation, when the vertical federated learning client is an NWDAF or an AF within the core network, the third-party communication entity is an NRF.

[0152] After receiving the federated learning capability type or federated learning capability update information, the third-party communication entity stores it.

[0153] Optionally, the third-party communication entity may also send a response message of the federated learning capability type to the vertical federated learning client, or send a response message of the federated learning capability update information to the vertical federated learning client. Correspondingly, the vertical federated learning client receives the response message of the federated learning capability type or the response message of the federated learning capability update information sent by the third-party communication entity.

[0154] Optionally, based on the second implementation method in step 802, when the vertical federated learning client is a vertical federated learning client that needs to perform model training, the method also includes the following steps: sending information requesting a federated learning server to a third-party communication entity; receiving information of a vertical federated learning server determined by the third-party communication entity according to a selection rule; and sending a request as a vertical federated learning server to at least some of the vertical federated learning servers determined by the third-party communication entity according to the selection rule.

[0155] Furthermore, based on the above implementation, the method for vertical federated learning provided in this embodiment further includes: receiving response information sent by at least some of the vertical federated learning servers as vertical federated learning servers according to a selection rule. Optionally, when there are multiple vertical federated learning servers, the vertical federated learning client may determine one vertical federated learning server from among them according to the selection rule. When there is only one vertical federated learning server, the vertical federated learning client determines that vertical federated learning server as the server participating in the vertical federated learning, i.e., the vertical federated learning server in the embodiments shown in Figures 7 and 8.

[0156] Optionally, based on the above implementation, the method for vertical federated learning provided in this embodiment further includes: sending a vertical federated learning request to the vertical federated learning server. The vertical federated learning request includes at least one of the following: an analysis ID, information about the model corresponding to the analysis ID, the type of the model corresponding to the analysis ID, training environment configuration requirements required for the model corresponding to the analysis ID, an identifier of the vertical federated learning client, and a collective identifier of the vertical federated learning clients.

[0157] Step 1002: Receive a vertical federated learning joining request sent by a vertical federated learning server.

[0158] Optionally, the vertical federated learning client is a client determined by the vertical federated learning server based on information of the vertical federated learning client determined by the third-party communication entity and a selection rule.

[0159] Furthermore, the information of the vertical federated learning client determined by the third-party communication entity is the information of the client determined by the third-party communication entity according to the selection rule.

[0160] The implementation process and technical principle of step 1002 are similar to those of step 901 and will not be repeated here.

[0161] Step 1003: Determine whether to join the vertical federated learning where the vertical federated learning server is located based on the selection rule and the vertical federated learning joining request.

[0162] After receiving the vertical federated learning joining request, the vertical federated learning client in this embodiment determines whether to join the vertical federated learning where the vertical federated learning server is located according to the selection rule and the information included in the vertical federated learning joining request.

[0163] Step 1004: When it is determined to join the vertical federated learning where the vertical federated learning server is located, response information for instructing to join the vertical federated learning is sent to the vertical federated learning server.

[0164] If the vertical federated learning client determines to join the vertical federated learning where the vertical federated learning server is located according to the selection rule and the vertical federated learning joining request, it sends response information to the vertical federated learning server to instruct it to join the vertical federated learning.

[0165] If the vertical federated learning client determines not to join the vertical federated learning where the vertical federated learning server is located based on the selection rule and the vertical federated learning joining request, it sends response information to the vertical federated learning server to indicate not to join the vertical federated learning.

[0166] It should be noted that in the scenario where the vertical federated learning client is a vertical federated learning client that needs to perform model training, in step 1004, the vertical federated learning client sends a response message to the vertical federated learning server indicating joining the vertical federated learning in order to participate in the vertical federated learning.

[0167] Optionally, the selection rules include at least one of the following: the type of communication entity, the federated learning capability type of the communication entity, the service area corresponding to the communication entity, whether the communication entity is currently performing other federated learning or model training tasks, the analysis identification ID, the type of communication entity that can collect data sources for local model training, the time period that can be used for vertical federated learning, the interoperability indicators of the model corresponding to the analysis ID, the information of the model corresponding to the analysis ID, the type of model corresponding to the analysis ID, the training environment configuration requirements required for the model corresponding to the analysis ID, the data availability of the communication entity, the time availability of the communication entity, the computing power information of the communication entity and the communication capability of the communication entity, the information of the network function served by the communication entity, the slicing information of the communication entity, the model sharing information of the communication entity, the gradient calculation method, the gradient optimization algorithm, the loss function type, the gradient descent algorithm, and the maximum response time for the vertical federated learning client to return the intermediate training results after receiving the loss information in each round of training.

[0168] It should be noted that the selection rules involved in the above process may be the same, different, or partially the same. In practical applications, the selection rules can be flexibly determined according to needs.

[0169] The method for vertical federated learning provided in this embodiment provides a method for determining participating vertical federated learning clients. The participating vertical federated learning clients determined based on this method can meet the needs of subsequent vertical federated learning, thereby improving the efficiency and success rate of vertical federated learning. Furthermore, the signaling interaction method for determining participating vertical federated learning clients in this embodiment is concise and efficient, further improving the efficiency of vertical federated learning.

[0170] The following describes the method of vertical federated learning provided by this embodiment from the perspective of signaling interaction. Figures 11a to 11c are used to describe the process of a vertical federated learning communication entity reporting the type of federated learning capability to a third-party communication entity. Figures 12a to 12c are used to describe the process of a vertical federated learning communication entity updating the federated learning capability to a third-party communication entity. Figure 13 is a process for a consumer-side communication entity to initiate the selection of a vertical federated learning server. Figure 14 is a process for a vertical federated learning client that needs to perform model training to initiate the selection of a vertical federated learning server. Figure 15 is a process for a vertical federated learning server to select a vertical federated learning client.

[0171] Figure 11a is a signaling interaction diagram for a method for vertical federated learning provided by one embodiment. As shown in Figure 11a, the federated learning communication entity in this embodiment is the NWDAF or an AF within the core network. The third-party communication entity is the NRF. The process by which the vertical federated learning communication entity reports its federated learning capability type to the third-party communication entity includes the following steps.

[0172] Step 1101a: The NWDAF or the AF in the core network sends the federated learning capability type to the NRF.

[0173] Optionally, the NWDAF or the AF in the core network may send the federated learning capability type through the Nnrf_NFManagement_NFRegister_request service operation. The specific content of the federated learning capability type is as shown in the above embodiment and will not be repeated here.

[0174] Step 1102a: The NRF stores the received federated learning capability type.

[0175] Step 1103a: The NRF sends a registration response message to the NWDAF or the AF in the core network.

[0176] Optionally, the NRF may send a registration response message via the Nnrf_NFManagement_NFRegister_response service operation.

[0177] After the above steps 1101a to 1103a, the NWDAF or the AF in the core network realizes the vertical federated learning capability type reporting to the NRF.

[0178] Figure 11b is a signaling interaction diagram for another method of vertical federated learning provided by one embodiment. As shown in Figure 11b, the federated learning communication entity in this embodiment is an AF outside the core network. The third-party communication entity is a NEF. The process of the vertical federated learning communication entity reporting the federated learning capability type to the third-party communication entity includes the following steps.

[0179] Step 1101b: The AF outside the core network sends the federated learning capability type to the NEF.

[0180] Step 1102b: The NEF stores the received federated learning capability type.

[0181] Step 1103b: The NEF sends a registration response message to the AF outside the core network.

[0182] After the above steps 1101b to 1103b, the AF outside the core network realizes the vertical federated learning capability type reporting to the NEF.

[0183] Figure 11c is a signaling interaction diagram for another method of vertical federated learning provided by one embodiment. As shown in Figure 11c, the federated learning communication entity in this embodiment is an AF outside the core network. The third-party communication entity is an NRF. The process of the vertical federated learning communication entity reporting the federated learning capability type to the third-party communication entity includes the following steps.

[0184] Step 1101c: The AF outside the core network sends the federated learning capability type to the NEF.

[0185] Step 1102c: The NEF stores the received federated learning capability type.

[0186] Step 1103c: The NEF sends a registration response message to the AF outside the core network.

[0187] Step 1104c: The NEF sends the received federated learning capability type to the NRF.

[0188] Step 1105c: The NRF stores the received federated learning capability type.

[0189] Step 1106c: The NRF sends a registration response message to the NEF.

[0190] It should be noted that step 1103c may be performed after step 1106c, that is, after the NEF receives the registration response information sent by the NRF, it sends the registration response information to the AF outside the core network.

[0191] After the above steps 1101c to 1106c, the AF outside the core network realizes the vertical federated learning capability type reporting to the NRF through the NEF.

[0192] The registration response information in FIG. 11 a to FIG. 11 c is used to indicate whether the registration is successful.

[0193] Figure 12a is a signaling interaction diagram for another method of vertical federated learning provided by one embodiment. As shown in Figure 12a, the federated learning communication entity in this embodiment is the NWDAF or an AF within the core network. The third-party communication entity is the NRF. The process of updating the federated learning capabilities of the vertical federated learning communication entity to the third-party communication entity includes the following steps.

[0194] Step 1201a: The NWDAF or the AF in the core network sends federated learning capability update information to the NRF.

[0195] Optionally, the NWDAF or the AF in the core network may send the federated learning capability update information through the Nnrf_NFManagement_NFUpdate_request service operation. The specific content of the federated learning capability update information is as shown in the above embodiment and will not be repeated here.

[0196] Step 1202a: The NRF stores the received federated learning capability update information.

[0197] Step 1203a: The NRF sends an update response message to the NWDAF or the AF in the core network.

[0198] Optionally, the NRF may send an update response message via the Nnrf_NFManagement_NFUpdate_response service operation.

[0199] After the above steps 1201a to 1203a, the NWDAF or the AF in the core network realizes the update of the federated learning capability to the NRF.

[0200] Figure 12b is a signaling interaction diagram for another method of vertical federated learning provided by one embodiment. As shown in Figure 12b , the federated learning communication entity in this embodiment is an AF outside the core network. The third-party communication entity is a NEF. The process of updating the federated learning capability of the vertical federated learning communication entity to the third-party communication entity includes the following steps.

[0201] Step 1201b: The AF outside the core network sends federated learning capability update information to the NEF.

[0202] Step 1202b: The NEF stores the received federated learning capability update information.

[0203] Step 1203b: The NEF sends an update response message to the AF outside the core network.

[0204] After the above steps 1201b to 1203b, the AF outside the core network realizes the federated learning capability update information to the NEF.

[0205] Figure 12c is a signaling interaction diagram for another method of vertical federated learning provided by one embodiment. As shown in Figure 12c, the federated learning communication entity in this embodiment is an AF outside the core network. The third-party communication entity is an NRF. The process of updating the federated learning capability of the vertical federated learning communication entity to the third-party communication entity includes the following steps.

[0206] Step 1201c: The AF outside the core network sends federated learning capability update information to the NEF.

[0207] Step 1202c: The NEF stores the received federated learning capability update information.

[0208] Step 1203c: The NEF sends an update response message to the AF outside the core network.

[0209] Step 1204c: The NEF sends the received federated learning capability update information to the NRF.

[0210] Step 1205c: The NRF stores the received federated learning capability update information.

[0211] Step 1206c: The NRF sends an update response message to the NEF.

[0212] It should be noted that step 1203c may be performed after step 1206c, that is, after the NEF receives the update response information sent by the NRF, it sends the update response information to the AF outside the core network.

[0213] After the above steps 1201c to 1206c, the AF outside the core network realizes the vertical federated learning capability type reporting to the NRF through the NEF.

[0214] The update response information in FIG. 12 a to FIG. 12 c is used to indicate whether the update is successful.

[0215] Figure 13 is a signaling interaction diagram of another method of vertical federated learning provided by an embodiment. It shows the process of the consumer-side communication entity initiating the selection of a vertical federated learning server. In the scenario where the vertical federated learning server and the vertical federated learning client in this embodiment are NWDAF or AF within the core network, the third-party communication entity in this embodiment is NRF. In the scenario where the vertical federated learning server and the vertical federated learning client in this embodiment are AF outside the core network, the third-party communication entity in this embodiment is NRF or NEF. As shown in Figure 13, the process includes the following steps.

[0216] Step 1300: The consumer communication entity sends a model subscription request to the source communication entity.

[0217] The model subscription request includes at least one of the following: analysis ID, other information indicating the purpose of the model, model metrics, expected values ​​corresponding to the model metrics, identifier of the consumer-end communication entity, and address information of the consumer-end communication entity.

[0218] If the source communication entity has the capability of a vertical federated learning server, for example, it has calibration data (ground truth data / label), then the subsequent steps are skipped and the vertical federated learning client selection process is directly performed, as shown in Figure 15. If the model corresponding to the decision analysis ID of the source communication entity requires vertical federated learning for training, and the source communication entity itself cannot serve as a VFL server (for example, it does not have ground truth data / label), then another NWDAF or AF is selected from a third-party communication entity as the VFL server. Optionally, the source communication entity in this embodiment can be an NWDAF or AF.

[0219] Step 1301: The source communication entity selects a vertical federated learning server through a third-party communication entity.

[0220] A possible process for selecting a vertical federated learning server may include the following steps: a source communication entity sends information requesting a federated learning server to a third-party communication entity; the third-party communication entity determines the information of the vertical federated learning server based on the selection rules and the information requesting the federated learning server, and sends the determined information to the source communication entity; the source communication entity determines the vertical federated learning server based on the selection rules and the vertical federated learning server determined by the third-party communication entity based on the selection rules.

[0221] Optionally, when the third-party communication entity determines the information of the vertical federated learning server based on the selection rules and the information of the requested federated learning server, the selection rules used may include at least one of the following information: NF type (for example, NWDAF or AF), whether the NF supports serving as a server for vertical federated learning, the service area where the NF is located, the NF information served by the VFL server (for example, NWDAF or AF), the NF slice information, the NF model sharing information, whether the NF is performing other federated learning or other model training tasks, analysis ID (model usage), etc.

[0222] After determining the information of the vertical federated learning server according to the selection rule and the information of the requested federated learning server, the third-party communication entity may return a list including the information of the determined vertical federated learning server to the source communication entity.

[0223] When the source communication entity determines the vertical federated learning server based on the selection rules and the list of information about the vertical federated learning server returned by the third-party communication entity, the selection rules used may include at least one of the following information: NF type (for example, NWDAF or AF), the service area corresponding to the NF, whether the NF is performing other federated learning or model training tasks, data and time availability, computing and communication capabilities, and model interoperability information, etc.

[0224] Step 1302: The source communication entity sends a request to the vertical federated learning server selected by it to serve as a vertical federated learning server.

[0225] Optionally, the request as a vertical federated learning server includes at least one of the following: identification or address information of a consumer-end communication entity.

[0226] Step 1303: The vertical federated learning server sends response information as the vertical federated learning server to the source communication entity.

[0227] The vertical federated learning server sends a response message to the source communication entity indicating that it is a vertical federated learning server or not a vertical federated learning server based on a selection rule. The selection rule includes at least one of the following: data and time availability, computing and communication capabilities, and model interoperability information.

[0228] The source communication entity may determine a vertical federated learning server to participate in vertical federated learning from the vertical federated learning servers that send response information as vertical federated learning servers according to a selection rule. The source communication entity may also send information about the determined vertical federated learning server to the consumer communication entity.

[0229] Through the above steps 1300 to 1303, the process of the consumer-side communication entity initiating the selection of the vertical federated learning server can be implemented.

[0230] Figure 14 is a signaling interaction diagram of another method of vertical federated learning provided by an embodiment. It shows the process of a vertical federated learning client that needs to perform model training initiating the selection of a vertical federated learning server. In this embodiment, the source communication entity can be referred to as a passive participant, the vertical federated learning server can be referred to as an active participant, and the vertical federated learning client can be referred to as a passive participant. In the scenario where the vertical federated learning server and the vertical federated learning client in this embodiment are NWDAF or AF within the core network, the third-party communication entity in this embodiment is NRF. In the scenario where the vertical federated learning server and the vertical federated learning client in this embodiment are AF outside the core network, the third-party communication entity in this embodiment is NRF or NEF. As shown in Figure 14, the process includes the following steps.

[0231] Step 1401: The source communication entity performs a process of selecting a vertical federated learning server through a third-party communication entity.

[0232] The specific process is detailed in steps 1301 to 1303. No further details will be given here.

[0233] The source communication entity in this embodiment is a communication entity that needs to perform model training, such as NWDAF or AF.

[0234] It should be noted that if the communication entity that needs to perform model training initiates the process of selecting a vertical federated learning server, the process of step 1401 does not include step 1300 and the step of returning the determined vertical federated learning server to the consumer-end communication entity.

[0235] Step 1402: The source communication entity sends a vertical federated learning request to the vertical federated learning server.

[0236] Optionally, the vertical federated learning request includes at least one of the following: an analysis ID, information about the model corresponding to the analysis ID, the type of the model corresponding to the analysis ID, training environment configuration requirements for the model corresponding to the analysis ID, an identifier of the source communication entity, and a set identifier (set ID) of the source communication entity. For example, the model types may include random forests, support vector machines (SVMs), and neural networks.

[0237] Step 1403: After receiving the vertical federated learning request, the vertical federated learning server selects a vertical federated learning client.

[0238] The specific process of the vertical federated learning server selecting the vertical federated learning client will be described in Figure 15.

[0239] Step 1404: The vertical federated learning server sends a vertical federated learning joining request to the vertical federated learning client.

[0240] Step 1405: The vertical federated learning client sends vertical federated learning response information to the vertical federated learning server.

[0241] Step 1406: The vertical federated learning server sends a federated learning establishment success / failure indication to the source communication entity.

[0242] Through the above steps 1401 to 1406, the process of initiating the selection of a vertical federated learning server by a vertical federated learning client that needs to perform model training can be implemented.

[0243] Figure 15 is a signaling interaction diagram of another method of vertical federated learning provided by an embodiment. It shows the process of a vertical federated learning server selecting a client to participate in vertical federated learning. In the scenario where the vertical federated learning server and the vertical federated learning client in this embodiment are NWDAF or AF within the core network, the third-party communication entity in this embodiment is NRF. In the scenario where the vertical federated learning server and the vertical federated learning client in this embodiment are AF outside the core network, the third-party communication entity in this embodiment is NRF or NEF. Figure 15 is an example of the number of vertical federated learning clients being n. n is an integer greater than 1. As shown in Figure 15, the process includes the following steps.

[0244] Step 1501a: The vertical federated learning server sends the federated learning capability type to the third-party communication entity.

[0245] Step 1501b: The vertical federated learning client sends the federated learning capability type to the third-party communication entity.

[0246] In step 1501a and step 1501b, the federated learning capability type may be sent via the Nnrf_NFManagement_NFRegister_request service operation. The specific content of the federated learning capability type is as described in the above embodiment and will not be repeated here.

[0247] Step 1502: The third-party communication entity stores the received federated learning capability type.

[0248] Step 1503a: The third-party communication entity sends a registration response message to the vertical federated learning client.

[0249] Step 1503b: The third-party communication entity sends a registration response message to the vertical federated learning server.

[0250] Optionally, in step 1503a and step 1503b, the third-party communication entity may send registration response information through the Nnrf_NFManagement_NFRegister_response service operation.

[0251] The process from step 1501a to step 1503b may also be referred to as a registration process of the federated learning communication entity.

[0252] After the vertical federated learning server is determined, it can perform the following operations to select the vertical federated learning client.

[0253] Step 1504: The vertical federated learning server sends a vertical federated learning client discovery request to the third-party communication entity.

[0254] Optionally, the vertical federated learning server can send a vertical federated learning client discovery request by calling the Nnrf_NFDiscovery_Request service operation.

[0255] Step 1505: The third-party communication entity authorizes the vertical federated learning client to discover the service.

[0256] After receiving the vertical federated learning client discovery request, the third-party communication entity determines the information of the vertical federated learning client according to the selection rules and the vertical federated learning client discovery request. Optionally, the selection rules here may include at least one of the following: NF type (e.g., NWDAF or AF), whether the NF supports being a client for vertical federated learning, the service area corresponding to the NF, whether the NF is performing other federated learning, the purpose of the model trained by the model training task VFL (analysis ID), the NF type that can collect data sources for local model training, the time period during which the client can be used for VFL training, model interoperability, model information or model type information to be trained with VFL (e.g., random neural network, support vector machine (SVM), neural network), and the training environment configuration requirements required for the model.

[0257] In step 1505, the third-party communication entity determines a group of vertical federated learning clients that can perform vertical federated learning.

[0258] Step 1506: The third-party communication entity sends the information of the vertical federated learning client determined by it to the vertical federated learning server.

[0259] Optionally, the third-party communication entity feeds back the information of the vertical federated learning client determined by it to the vertical federated learning server through the Nnrf_NFDiscovery_RequestResponse service operation.

[0260] Steps 1504 to 1506 may be referred to as a discovery process of a federated learning communication entity.

[0261] Step 1507: The vertical federated learning server sends a vertical federated learning joining request to at least some of the vertical federated learning clients determined by the third-party communication entity according to the selection rule.

[0262] Optionally, the selection rules involved in step 1507 include at least one of the following: NF type (e.g., NWDAF or AF), the service area corresponding to the NF, whether the NF is performing other federated learning or model training tasks, data and time availability, computing and communication capabilities, model interoperability information, and model information or model type information to be trained on VFL (e.g., random neural network, SVM, neural network), training environment configuration requirements required for the model, etc.

[0263] Optionally, the process of selecting a vertical federated learning server based on different triggers, that is, the two implementation methods in step 802, in one implementation method, the vertical federated learning joining request includes at least one of the following: the identifier of the consumer-end communication entity, the address information of the consumer-end communication entity, the model metric, the expected value corresponding to the model metric, information related to the loss type to be transmitted for vertical federated learning, and the maximum response time for the vertical federated learning client to return the intermediate training results after receiving the loss information in each round of training. In another implementation method, the vertical federated learning joining request includes at least one of the following: the identifier of the source communication entity, the address information of the source communication entity, the model metric, the expected value corresponding to the model metric, information related to the loss type to be transmitted for vertical federated learning, and the maximum response time for the vertical federated learning client to return the intermediate training results after receiving the loss information in each round of training.

[0264] Optionally, the vertical federated learning join request can also be called federatedlearningPreparationrequest.

[0265] Step 1508: The vertical federated learning client decides whether to join the vertical federated learning.

[0266] The vertical federated learning client determines whether to join the vertical federated learning server based on the selection rules and the vertical federated learning joining request. The selection rules include at least one of the following: data and time availability, computing and communication capabilities, model interoperability information, model information or model type information (e.g., random neural network, support vector machine, neural network) to be trained on VFL, and the training environment configuration requirements required by the model.

[0267] Step 1509: The vertical federated learning client sends vertical federated learning response information to the vertical federated learning server.

[0268] Step 15010: The vertical federated learning server determines the vertical federated learning clients participating in the vertical federated learning based on the vertical federated learning response information.

[0269] Optionally, the above steps 1504 to 15010 may be repeatedly executed until the vertical federated learning server determines that the number of clients participating in the vertical federated learning meets the requirement.

[0270] The above steps 1501a to 15010 implement the process of the vertical federated learning server selecting the clients participating in the vertical federated learning.

[0271] This embodiment also provides a vertical federated learning system, comprising: a vertical federated learning server and a vertical federated learning client. The vertical federated learning server is configured to execute the steps performed by the vertical federated learning server in the vertical federated learning method provided in any of the above embodiments. The vertical federated learning client is configured to execute the steps performed by the vertical federated learning client in the vertical federated learning method provided in any of the above embodiments. The implementation principles and technical effects are similar to those of the above embodiments and will not be further elaborated here.

[0272] FIG16 is a flow chart of a method for vertical federated learning provided by an embodiment. The method for vertical federated learning provided in this embodiment is applied to NEF. As shown in FIG16 , the method for vertical federated learning includes the following steps.

[0273] Step 1601: Receive the federated learning capability type sent by the communication entity.

[0274] The federated learning capability type includes at least one of the following: whether it supports serving as a federated learning server, whether it supports serving as a federated learning client, or whether it supports serving as both a federated learning server and a federated learning client. Alternatively, the federated learning capability type includes at least one of the following: whether it supports serving as a vertical federated learning server, whether it supports serving as a vertical federated learning client, or whether it supports serving as both a vertical federated learning server and a vertical federated learning client.

[0275] Step 1602: Store the federated learning capability type.

[0276] Optionally, after step 1602, the following step is also included: sending response information of the federated learning capability type to the communication entity.

[0277] Optionally, after step 1601, the following steps are further included: sending the federated learning capability type to the NRF; and receiving response information of the federated learning capability type sent by the NRF.

[0278] In one implementation, the method for vertical federated learning provided in this embodiment also includes the following steps: receiving federated learning capability update information sent by a communication entity, wherein the federated learning capability update information includes at least one of the following: whether it supports being a federated learning server, whether it supports being a federated learning client, or whether it supports being a federated learning server and a federated learning client at the same time, or the federated learning capability type includes at least one of the following: whether it supports being a vertical federated learning server, whether it supports being a vertical federated learning client, or whether it supports being a vertical federated learning server and a vertical federated learning client at the same time; and storing the federated learning capability update information.

[0279] Optionally, based on the above implementation, the method may further include the following step: sending response information of the federated learning capability update information to the communication entity.

[0280] Optionally, based on the above implementation, the method may further include the following steps: sending federated learning capability update information to the NRF; and receiving response information of the federated learning capability update information sent by the NRF.

[0281] Optionally, based on the method of vertical federated learning of Figures 7 to 15, the method of vertical federated learning provided by this embodiment may also include at least one of the following steps: receiving a federated learning client selection request sent by a communication entity, determining the information of the federated learning client based on the federated learning client selection request, and sending the information of the federated learning client to the communication entity; receiving information requesting a federated learning server sent by the communication entity, determining the information of the federated learning server based on the information requesting the federated learning server, and sending the information of the federated learning server to the communication entity.

[0282] Optionally, the communication entity in this embodiment is an application function outside the core network.

[0283] Figure 17a is a signaling interaction diagram of a method for vertical federated learning provided by an embodiment. As shown in Figure 17a, it includes the following steps.

[0284] Step 1701a: The communication entity sends the federated learning capability type or federated learning capability update information to the NEF.

[0285] Step 1702a: The NEF stores the federated learning capability type or federated learning capability update information.

[0286] Step 1703a: The NEF sends a response message to the communication entity.

[0287] The response message in step 1703a is a registration response message or an update response message.

[0288] Optionally, FIG17a may further include the following process: the NRF receives a federated learning client selection request sent by the communication entity, or receives information requesting a federated learning server sent by the communication entity.

[0289] Figure 17b is a signaling interaction diagram of another method for vertical federated learning provided by an embodiment. As shown in Figure 17b, it includes the following steps.

[0290] Step 1701b: The communication entity sends the federated learning capability type or federated learning capability update information to the NEF.

[0291] Step 1702b: The NEF stores the federated learning capability type or federated learning capability update information.

[0292] Step 1703b: The NEF sends a response message to the communication entity.

[0293] The response message in step 1703b is a registration response message or an update response message.

[0294] Step 1704b: The NEF sends the received federated learning capability type or federated learning capability update information to the NRF.

[0295] Step 1705b: The NRF stores the received federated learning capability type or federated learning capability update information.

[0296] Step 1706b: The NRF sends a response message to the NEF.

[0297] The response message in step 1706b is a registration response message or an update response message.

[0298] It should be noted that step 1703b may be performed after step 1706b, that is, after the NEF receives the response information sent by the NRF, it sends the response information to the communication entity.

[0299] The method for vertical federated learning provided in this embodiment enables the NEF to receive and store the federated learning capability type or federated learning capability update information sent by the communication entity, thereby facilitating subsequent vertical federated learning through the NEF.

[0300] Figure 18 is a schematic diagram of the structure of a device for vertical federated learning, provided in one embodiment. This device is provided in a vertical federated learning server. As shown in Figure 18 , the device for vertical federated learning provided in this embodiment includes the following modules: a first sending module 1801, a first receiving module 1802, and a determining module 1803.

[0301] The first sending module 1801 is configured to send a vertical federated learning joining request to a vertical federated learning client.

[0302] The first receiving module 1802 is configured to receive vertical federated learning response information fed back by the vertical federated learning client according to the vertical federated learning joining request.

[0303] The determination module 1803 is configured to determine the vertical federated learning client participating in the vertical federated learning according to the vertical federated learning response information.

[0304] In one embodiment, the first sending module 1801 is further configured to send a vertical federated learning client discovery request to a third-party communication entity. The first receiving module 1802 is configured to receive information about a vertical federated learning client determined by the third-party communication entity based on the vertical federated learning client discovery request.

[0305] In one embodiment, the first sending module 1801 is configured to: send the vertical federated learning joining request to at least some of the vertical federated learning clients determined by the third-party communication entity according to a selection rule.

[0306] In one embodiment, the information of the vertical federated learning client determined by the third-party communication entity is the information of the client determined by the third-party communication entity according to a selection rule and the vertical federated learning client discovery request.

[0307] In one embodiment, the determination module 1803 is configured to: determine the vertical federated learning client participating in the vertical federated learning from the vertical federated learning clients whose vertical federated learning response information indicates joining the vertical federated learning according to a selection rule.

[0308] In one embodiment, the vertical federated learning server is a server determined by the source communication entity based on a selection rule and a vertical federated learning server determined by the third-party communication entity based on the selection rule.

[0309] In one embodiment, the first receiving module 1802 is configured to receive a request from the source communication entity to serve as a vertical federated learning server. The first sending module 1801 is configured to send response information to the source communication entity to serve as a vertical federated learning server according to a selection rule.

[0310] In one embodiment, the selection rules include at least one of the following: the type of communication entity, the type of federated learning capability of the communication entity, the service area corresponding to the communication entity, whether the communication entity is currently performing other federated learning or model training tasks, the analysis ID, the type of communication entity that can collect data sources for local model training, the time period that can be used for vertical federated learning, the interoperability index of the model corresponding to the analysis ID, the information of the model corresponding to the analysis ID, the type of model corresponding to the analysis ID, the training environment configuration requirements required for the model corresponding to the analysis ID, the data availability of the communication entity, the time availability of the communication entity, the computing power information of the communication entity and the communication capability of the communication entity, the information of the network function served by the communication entity, the slicing information of the communication entity, the model sharing information of the communication entity, the gradient calculation method, the gradient optimization algorithm, the loss function type, the gradient descent algorithm, and the maximum response time for the vertical federated learning client to return the intermediate training results after receiving the loss information in each round of training.

[0311] In one embodiment, the source communication entity is a communication entity that receives a model subscription request sent by a consumer communication entity, wherein the model subscription request includes an analysis ID.

[0312] In one embodiment, the vertical federated learning joining request includes at least one of the following: the identifier of the consumer-end communication entity, the address information of the consumer-end communication entity, the model metric, the expected value corresponding to the model metric, information related to the loss type to be transmitted for vertical federated learning, and the maximum response time for the vertical federated learning client to return the intermediate training results after receiving the loss information in each round of training.

[0313] In one embodiment, the determination module 1803 is further configured to: determine the source communication entity as a vertical federated learning client participating in vertical federated learning.

[0314] In one embodiment, the first receiving module 1802 is further configured to receive a vertical federated learning request sent by the source communication entity. The vertical federated learning request includes at least one of the following: an analysis ID, information about a model corresponding to the analysis ID, a type of the model corresponding to the analysis ID, training environment configuration requirements required for the model corresponding to the analysis ID, an identifier of the source communication entity, and a collective identifier of the source communication entity.

[0315] In one embodiment, the first sending module 1801 is further configured to send a federated learning capability type or federated learning capability update information to the third-party communication entity. The federated learning capability type and federated learning capability update information each include at least one of the following: whether the function as a vertical federated learning server is supported, whether the function as a vertical federated learning client is supported, or whether the function as both a vertical federated learning server and a vertical federated learning client is supported.

[0316] In one embodiment, the vertical federated learning server includes at least one of the following: an NWDAF, an application function within the core network, or an application function outside the core network. If the vertical federated learning server is an application function outside the core network, the third-party communication entity is an NEF or an NRF. If the vertical federated learning server is an NWDAF or an application function within the core network, the third-party communication entity is an NRF.

[0317] The device for vertical federated learning provided in this embodiment can implement the method for vertical federated learning implemented by the vertical federated learning server in the above embodiment. The implementation principle and technical effects are similar to those of the above embodiment and will not be repeated here.

[0318] Figure 19 is a schematic diagram of the structure of another apparatus for vertical federated learning provided by one embodiment. This apparatus is provided in a vertical federated learning client. As shown in Figure 19, the apparatus for vertical federated learning provided by this embodiment includes the following modules: a second receiving module 1901 and a second sending module 1902.

[0319] The second receiving module 1901 is configured to receive a vertical federated learning joining request sent by a vertical federated learning server.

[0320] The second sending module 1902 is configured to feed back vertical federated learning response information to the vertical federated learning server according to the vertical federated learning joining request.

[0321] In one embodiment, the vertical federated learning client is a client determined by the vertical federated learning server based on information of the vertical federated learning client determined by a third-party communication entity and a selection rule.

[0322] In one embodiment, the information of the vertical federated learning client determined by the third-party communication entity is the information of the client determined by the third-party communication entity according to a selection rule.

[0323] The second sending module 1902 is configured to: determine whether to join the vertical federated learning where the vertical federated learning server is located based on the selection rule and the vertical federated learning joining request; if it is determined to join the vertical federated learning where the vertical federated learning server is located, send response information to the vertical federated learning server to indicate joining the vertical federated learning.

[0324] In one embodiment, the vertical federated learning client is a vertical federated learning client that requires model training. The second sending module 1902 is further configured to send information requesting a federated learning server to a third-party communication entity. The second receiving module 1901 is further configured to receive information about vertical federated learning servers determined by the third-party communication entity based on a selection rule. The second sending module 1902 is further configured to send a request to become a vertical federated learning server to at least some of the vertical federated learning servers determined by the third-party communication entity based on the selection rule.

[0325] In one embodiment, the second receiving module 1901 is further configured to receive response information sent by at least some of the vertical federated learning servers as vertical federated learning servers according to the selection rule.

[0326] In one embodiment, the selection rules include at least one of the following: the type of communication entity, the type of federated learning capability of the communication entity, the service area corresponding to the communication entity, whether the communication entity is currently performing other federated learning or model training tasks, the analysis ID, the type of communication entity that can collect data sources for local model training, the time period that can be used for vertical federated learning, the interoperability index of the model corresponding to the analysis ID, the information of the model corresponding to the analysis ID, the type of model corresponding to the analysis ID, the training environment configuration requirements required for the model corresponding to the analysis ID, the data availability of the communication entity, the time availability of the communication entity, the computing power information of the communication entity and the communication capability of the communication entity, the information of the network function served by the communication entity, the slicing information of the communication entity, the model sharing information of the communication entity, the gradient calculation method, the gradient optimization algorithm, the loss function type, the gradient descent algorithm, and the maximum response time for the vertical federated learning client to return the intermediate training results after receiving the loss information in each round of training.

[0327] In one embodiment, the second sending module 1902 is further configured to send a vertical federated learning request to the vertical federated learning server. The vertical federated learning request includes at least one of the following: an analysis ID, information about the model corresponding to the analysis ID, the type of the model corresponding to the analysis ID, training environment configuration requirements required for the model corresponding to the analysis ID, an identifier of the vertical federated learning client, and a collective identifier of the vertical federated learning clients.

[0328] In one embodiment, the second sending module 1902 is further configured to send a federated learning capability type or federated learning capability update information to the third-party communication entity. The federated learning capability type and federated learning capability update information each include at least one of the following: whether the function as a vertical federated learning server is supported, whether the function as a vertical federated learning client is supported, or whether the function as both a vertical federated learning server and a vertical federated learning client is supported.

[0329] In one embodiment, the vertical federated learning client includes at least one of the following: an NWDAF, an application function within the core network, or an application function outside the core network. If the vertical federated learning client is an application function outside the core network, the third-party communication entity is an NEF or an NRF. If the vertical federated learning client is an NWDAF or an application function within the core network, the third-party communication entity is an NRF.

[0330] The device for vertical federated learning provided in this embodiment can implement the method for vertical federated learning implemented by the vertical federated learning client in the above embodiment. The implementation principle and technical effects are similar to those of the above embodiment and will not be repeated here.

[0331] FIG20 is a schematic diagram of the structure of another apparatus for vertical federated learning provided by one embodiment. This apparatus is applied to NEF. As shown in FIG20 , the apparatus for vertical federated learning provided by this embodiment includes the following modules: a third receiving module 2001 and a storage module 2002.

[0332] The third receiving module 2001 is configured to receive a federated learning capability type sent by a communication entity.

[0333] Among them, the federated learning capability type includes at least one of the following: whether it supports being a federated learning server, whether it supports being a federated learning client, whether it supports being a federated learning server and a federated learning client at the same time, or the federated learning capability type includes at least one of the following: whether it supports being a vertical federated learning server, whether it supports being a vertical federated learning client, whether it supports being a vertical federated learning server and a vertical federated learning client at the same time.

[0334] The storage module 2002 is configured to store the federated learning capability type.

[0335] In one embodiment, the apparatus further includes: a third sending module configured to send response information of the federated learning capability type to the communication entity.

[0336] In one embodiment, the third sending module is further configured to send the federated learning capability type to the NRF. The third receiving module 2001 is further configured to receive response information of the federated learning capability type sent by the NRF.

[0337] In one embodiment, the third receiving module 2001 is further configured to receive federated learning capability update information sent by the communication entity. The federated learning capability update information includes at least one of the following: whether the entity supports serving as a federated learning server, whether the entity supports serving as a federated learning client, or whether the entity supports serving as both a federated learning server and a federated learning client; or the federated learning capability type includes at least one of the following: whether the entity supports serving as a vertical federated learning server, whether the entity supports serving as a vertical federated learning client, or whether the entity supports serving as both a vertical federated learning server and a vertical federated learning client. The storage module 2002 is further configured to store the federated learning capability update information.

[0338] In one embodiment, the third sending module is further configured to send response information of the federated learning capability update information to the communication entity.

[0339] In one embodiment, the third sending module is further configured to send the federated learning capability update information to the NRF. The third receiving module 2001 is further configured to receive response information of the federated learning capability update information sent by the NRF.

[0340] In one embodiment, the third receiving module 2001 is further configured to receive a federated learning client selection request sent by the communication entity, determine the information of the federated learning client based on the federated learning client selection request, and send the information of the federated learning client to the communication entity.

[0341] The third receiving module 2001 is further configured to receive information requesting a federated learning server sent by the communication entity, determine information of the federated learning server based on the information requesting the federated learning server, and send the information of the federated learning server to the communication entity.

[0342] In one embodiment, the communication entity is an application function outside the core network.

[0343] The apparatus for vertical federated learning provided in this embodiment can implement the method for vertical federated learning implemented by NEF in the above embodiment. The implementation principle and technical effects are similar to those in the above embodiment and will not be repeated here.

[0344] The embodiment of the present application further provides a communication entity, including: a processor, the processor being configured to implement the method provided in any embodiment of the present application when executing a computer program. The communication entity in this embodiment can be a vertical federated learning server, a vertical federated learning client, or a NEF.

[0345] Figure 21 is a schematic diagram of the structure of a communication entity provided by one embodiment. As shown in Figure 21, the communication entity includes a processor 60, a memory 61, and a communication interface 62. The number of processors 60 in the communication entity can be one or more, and Figure 21 uses one processor 60 as an example. The processor 60, memory 61, and communication interface 62 in the communication entity can be connected via a bus or other means, and Figure 21 uses a bus connection as an example. The bus represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or a local bus using any of a variety of bus structures.

[0346] The memory 61, as a computer-readable storage medium, can be configured to store software programs, computer-executable programs, and modules, such as program instructions / modules corresponding to the methods in the embodiments of the present application. The processor 60 executes the software programs, instructions, and modules stored in the memory 61 to execute at least one functional application and data processing of the communication entity, thereby implementing the above-mentioned method.

[0347] Memory 61 may include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function; the data storage area may store data generated based on the use of the terminal. Furthermore, memory 61 may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state memory device. In some instances, memory 61 may include memory remotely located relative to processor 60, and such remote memory may be connected to a communication entity via a network. Examples of such networks include, but are not limited to, the Internet, an intranet, a network, a mobile communication network, and combinations thereof.

[0348] The communication interface 62 can be configured to receive and send data.

[0349] An embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method provided in any embodiment of the present application is implemented.

[0350] The computer storage medium of the embodiment of the present application can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to: an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. The computer-readable storage medium includes (a non-exhaustive list): an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it.

[0351] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, the data signal carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transfer a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0352] The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wire, optical cable, radio frequency (RF), etc., or any suitable combination of the foregoing.

[0353] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination of multiple programming languages, including object-oriented programming languages ​​(such as Java, Smalltalk, C++, Ruby, Go), and conventional procedural programming languages ​​(such as "C" or similar programming languages). The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0354] It will be appreciated by those skilled in the art that the term user terminal covers any suitable type of wireless user equipment, such as a mobile phone, a portable data processing device, a portable web browser or a vehicle-mounted mobile station.

[0355] In general, various embodiments of the present application may be implemented in hardware or dedicated circuits, software, logic, or any combination thereof. For example, some aspects may be implemented in hardware, while other aspects may be implemented in firmware or software that can be executed by a controller, microprocessor, or other computing device, although the present application is not limited thereto.

[0356] Embodiments of the present application may be implemented by executing computer program instructions by a data processor of a mobile device, for example, in a processor entity, or by hardware, or by a combination of software and hardware. The computer program instructions may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages.

[0357] Any block diagram of a logical flow in the drawings of this application may represent program steps, or may represent interconnected logical circuits, modules and functions, or may represent a combination of program steps and logical circuits, modules and functions. A computer program may be stored on a memory. The memory may be of any type suitable for the local technical environment and may be implemented using any suitable data storage technology, such as, but not limited to, read-only memory (ROM), random access memory (RAM), optical memory devices and systems (digital versatile discs DVD or CD), etc. Computer-readable media may include non-transitory storage media. A data processor may be of any type suitable for the local technical environment, such as, but not limited to, a general-purpose computer, a special-purpose computer, a microprocessor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a programmable logic device (FPGA), and a processor based on a multi-core processor architecture.

Claims

1. A method for vertical federated learning, applied to a vertical federated learning server, comprising: Send a vertical federated learning joining request to the vertical federated learning client; Receiving vertical federated learning response information fed back by the vertical federated learning client in response to the vertical federated learning joining request; According to the vertical federated learning response information, a vertical federated learning client participating in the vertical federated learning is determined.

2. The method according to claim 1, further comprising: Send a vertical federated learning client discovery request to the third-party communication entity; Receive information of a vertical federated learning client determined by the third-party communication entity according to the vertical federated learning client discovery request.

3. The method according to claim 2, wherein: The sending of a vertical federated learning joining request to the vertical federated learning client includes: According to the selection rule, the vertical federated learning joining request is sent to at least some of the vertical federated learning clients among the vertical federated learning clients determined by the third-party communication entity.

4. The method according to claim 2, wherein: The information of the vertical federated learning client determined by the third-party communication entity is the information of the client determined by the third-party communication entity according to the selection rule and the vertical federated learning client discovery request.

5. The method according to claim 1, wherein The determining, according to the vertical federated learning response information, a vertical federated learning client participating in the vertical federated learning includes: According to the selection rule, the vertical federated learning clients participating in the vertical federated learning are determined from the vertical federated learning clients whose vertical federated learning response information indicates joining the vertical federated learning.

6. The method according to claim 1, wherein The vertical federated learning server is a server determined by the source communication entity according to the selection rule and the third-party communication entity based on the selection rule.

7. The method according to claim 6, further comprising: receiving a request sent by the source communication entity to serve as a vertical federated learning server; According to the selection rule, response information as a vertical federated learning server is sent to the source communication entity.

8. The method according to any one of claims 3 to 7, wherein: The selection rules include at least one of the following: the type of communication entity, the type of federated learning capability of the communication entity, the service area corresponding to the communication entity, whether the communication entity is currently performing other federated learning or model training tasks, the analysis identification ID, the type of communication entity that can collect data sources for local model training, the time period that can be used for vertical federated learning, the interoperability index of the model corresponding to the analysis ID, the information of the model corresponding to the analysis ID, the type of model corresponding to the analysis ID, the training environment configuration requirements required for the model corresponding to the analysis ID, the data availability of the communication entity, the time availability of the communication entity, the computing power information of the communication entity and the communication capability of the communication entity, the information of the network function served by the communication entity, the slicing information of the communication entity, the model sharing information of the communication entity, the gradient calculation method, the gradient optimization algorithm, the loss function type, the gradient descent algorithm, and the maximum response time for the vertical federated learning client to return the intermediate training results after receiving the loss information in each round of training.

9. The method according to claim 6, wherein: The source communication entity is a communication entity that receives a model subscription request sent by a consumer communication entity, wherein the model subscription request includes an analysis ID.

10. The method according to claim 9, wherein: The vertical federated learning joining request includes at least one of the following: the identifier of the consumer-end communication entity, the address information of the consumer-end communication entity, the model metric, the expected value corresponding to the model metric, information related to the loss type to be transmitted for vertical federated learning, and the maximum response time for the vertical federated learning client to return the intermediate training results after receiving the loss information in each round of training.

11. The method according to claim 6, wherein: The source communication entity is a vertical federated learning client that needs to perform model training, and the method further includes: The source communication entity is determined as the vertical federated learning client participating in the vertical federated learning.

12. The method according to claim 11, further comprising: Receive a vertical federated learning request sent by the source communication entity; wherein the vertical federated learning request includes at least one of the following: analysis ID, information of the model corresponding to the analysis ID, the type of the model corresponding to the analysis ID, the training environment configuration requirements required for the model corresponding to the analysis ID, the identifier of the source communication entity, and the collective identifier of the source communication entity.

13. The method according to claim 1, further comprising: Send a federated learning capability type or federated learning capability update information to a third-party communication entity; wherein the federated learning capability type and the federated learning capability update information respectively include at least one of the following: whether it supports serving as a vertical federated learning server, whether it supports serving as a vertical federated learning client, and whether it supports serving as both a vertical federated learning server and a vertical federated learning client.

14. The method according to claim 1, wherein The vertical federated learning server includes at least one of the following: a network data analysis function NWDAF, an application function within the core network, or an application function outside the core network; In the case where the vertical federated learning server is an application function outside the core network, the third-party communication entity is a network exposure function NEF or a network storage function NRF; In the case where the vertical federated learning server is an NWDAF or an application function within the core network, the third-party communication entity is NRF.

15. A method for vertical federated learning, applied to a vertical federated learning client, the method comprising: Receive a vertical federated learning joining request sent by a vertical federated learning server; According to the vertical federated learning joining request, vertical federated learning response information is fed back to the vertical federated learning server.

16. The method according to claim 15, wherein The vertical federated learning client is a client determined by the vertical federated learning server according to the information of the vertical federated learning client determined by the third-party communication entity and a selection rule.

17. The method according to claim 16, wherein The information of the vertical federated learning client determined by the third-party communication entity is the information of the client determined by the third-party communication entity according to the selection rule.

18. The method according to claim 15, wherein Feedback of vertical federated learning response information to the vertical federated learning server according to the vertical federated learning joining request includes: Determine whether to join the vertical federated learning where the vertical federated learning server is located according to the selection rule and the vertical federated learning joining request; In the case of determining to join the vertical federated learning where the vertical federated learning server is located, response information for instructing to join the vertical federated learning is sent to the vertical federated learning server.

19. The method according to claim 15, wherein The vertical federated learning client is a vertical federated learning client that needs to perform model training, and the method further includes: Sending a request to the federated learning server to the third-party communication entity; Receiving information of a vertical federated learning server determined by the third-party communication entity according to a selection rule; According to the selection rule, a request to serve as a vertical federated learning server is sent to at least some of the vertical federated learning servers determined by the third-party communication entity.

20. The method according to claim 19, further comprising: Receive response information sent by at least some of the vertical federated learning servers as vertical federated learning servers according to the selection rule.

21. The method according to any one of claims 16 to 20, wherein: The selection rules include at least one of the following: the type of communication entity, the type of federated learning capability of the communication entity, the service area corresponding to the communication entity, whether the communication entity is currently performing other federated learning or model training tasks, the analysis identification ID, the type of communication entity that can collect data sources for local model training, the time period that can be used for vertical federated learning, the interoperability index of the model corresponding to the analysis ID, the information of the model corresponding to the analysis ID, the type of model corresponding to the analysis ID, the training environment configuration requirements required for the model corresponding to the analysis ID, the data availability of the communication entity, the time availability of the communication entity, the computing power information of the communication entity and the communication capability of the communication entity, the information of the network function served by the communication entity, the slicing information of the communication entity, the model sharing information of the communication entity, the gradient calculation method, the gradient optimization algorithm, the loss function type, the gradient descent algorithm, and the maximum response time for the vertical federated learning client to return the intermediate training results after receiving the loss information in each round of training.

22. The method of claim 19, further comprising: Send a vertical federated learning request to the vertical federated learning server; wherein, the vertical federated learning request includes at least one of the following: analysis ID, information of the model corresponding to the analysis ID, type of the model corresponding to the analysis ID, training environment configuration requirements required for the model corresponding to the analysis ID, the identifier of the vertical federated learning client, and the collective identifier of the vertical federated learning client.

23. The method of claim 15, further comprising: Send a federated learning capability type or federated learning capability update information to a third-party communication entity; wherein the federated learning capability type and the federated learning capability update information respectively include at least one of the following: whether it supports serving as a vertical federated learning server, whether it supports serving as a vertical federated learning client, and whether it supports serving as both a vertical federated learning server and a vertical federated learning client.

24. The method according to claim 15, wherein The vertical federated learning client includes at least one of the following: a network data analysis function NWDAF, an application function within the core network, or an application function outside the core network; In the case where the vertical federated learning client is an application function outside the core network, the third-party communication entity is a network exposure function NEF or a network storage function NRF; In the case where the vertical federated learning client is NWDAF or an application function within the core network, the third-party communication entity is NRF.

25. A method for vertical federated learning, applied to a network open function (NEF), comprising: Receive a federated learning capability type sent by a communication entity; wherein the federated learning capability type includes at least one of the following: whether it supports serving as a federated learning server, whether it supports serving as a federated learning client, or whether it supports serving as both a federated learning server and a federated learning client; or the federated learning capability type includes at least one of the following: whether it supports serving as a vertical federated learning server, whether it supports serving as a vertical federated learning client, or whether it supports serving as both a vertical federated learning server and a vertical federated learning client; The federated learning capability type is stored.

26. The method according to claim 25, further comprising: Sending federated learning capability type response information to the communication entity.

27. The method of claim 25, further comprising: Sending the federated learning capability type to the network storage function NRF; Receive the response information of the federated learning capability type sent by the NRF.

28. The method of claim 25, further comprising: Receiving federated learning capability update information sent by a communication entity; wherein the federated learning capability update information includes at least one of the following: whether it supports serving as a federated learning server, whether it supports serving as a federated learning client, or whether it supports serving as both a federated learning server and a federated learning client; or the federated learning capability type includes at least one of the following: whether it supports serving as a vertical federated learning server, whether it supports serving as a vertical federated learning client, or whether it supports serving as both a vertical federated learning server and a vertical federated learning client; The federated learning capability update information is stored.

29. The method according to claim 28, further comprising: Sending response information of federated learning capability update information to the communication entity.

30. The method of claim 28, further comprising: Sending the federated learning capability update information to NRF; Receive response information of the federated learning capability update information sent by the NRF.

31. The method of claim 25, further comprising at least one of the following: receiving a federated learning client selection request sent by the communication entity, determining information of a federated learning client according to the federated learning client selection request, and sending the information of the federated learning client to the communication entity; Receive information requesting a federated learning server sent by the communication entity, determine information of the federated learning server according to the information requesting the federated learning server, and send the information of the federated learning server to the communication entity.

32. The method of claim 25, wherein: The communication entity is an application function outside the core network.

33. A vertical federated learning system, comprising a vertical federated learning server and a vertical federated learning client; The vertical federated learning server is used to execute the vertical federated learning method according to any one of claims 1 to 14; The vertical federated learning client is used to execute the vertical federated learning method as described in any one of claims 15 to 24.

34. A communication entity comprising: processor; The processor is used to implement the vertical federated learning method as described in any one of claims 1 to 14 when executing a computer program, or to implement the vertical federated learning method as described in any one of claims 15 to 24, or to implement the vertical federated learning method as described in any one of claims 25 to 32.

35. A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the computer program implements the vertical federated learning method as described in any one of claims 1 to 14, or the vertical federated learning method as described in any one of claims 15 to 24, or the vertical federated learning method as described in any one of claims 25 to 32.

Images