Information configuration method, configuration device, network equipment, medium and program product

Abstract

The present disclosure provides an information configuration method, a configuration device, a network device, a medium and a program product, and relates to the technical field of networks. The configuration method of vertical federated learning information comprises the following steps: querying a NWDAF Function configuration file to obtain a vertical federated learning VFL information list, the VFL information list comprising a machine learning analysis task ID and corresponding VFL capability information, a VFL interoperation indicator and a feature identifier, the VFL capability information representing that the NWDAF is a VFL server and / or a client, the VFL interoperation indicator being used for matching network functions NFs with the same cooperative capability, and the feature identifier representing the data features of the machine learning analysis task; sending a first VFL registration request to a network storage function NRF, the first VFL registration request carrying the VFL information list; and receiving a first registration success response fed back by the NRF. Through the technical solution of the present disclosure, the network intelligent analysis task can be reliably carried out, thereby helping to improve the service experience of users.

Description

Technical Field

[0001] This disclosure relates to the field of network technology, and in particular to a method for configuring vertical federated learning information, a device for configuring vertical federated learning information, a network device, a computer-readable storage medium, and a computer program product. Background Technology

[0002] 3GPP Rel-19 (Mobile Communication International Standard Version 19) proposes the NWDAF (Network Data Analysis Function) management and vertical federated learning process, clarifying that NWDAF can assume the role of federated learning server or client. NWDAF needs to rely on the configuration data of the network management system to determine the types of vertical federated learning information it supports. However, the configuration scheme and management mechanism for vertical federated learning information have not yet been defined, making it impossible to associate and bind vertical federated learning information with NWDAF network element configuration information.

[0003] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0004] The purpose of this disclosure is to provide a method, apparatus, network device, storage medium, and computer program product for configuring vertical federated learning information, which at least to some extent overcomes the problem in related technologies that it is impossible to associate and bind vertical federated learning information with NWDAF network element configuration information.

[0005] Other features and advantages of this disclosure will become apparent from the following detailed description, or may be learned in part by practice of this disclosure.

[0006] According to one aspect of this disclosure, a method for configuring vertical federated learning information is provided, applied to a network data analysis function (NWDAF), comprising: calling the recognizable name (DN) of machine learning analysis information (MlAnalyticsInfo), querying the NWDAF configuration file to obtain a list of vertical federated learning (VFL) information, wherein the VFL information list includes a machine learning analysis task ID and corresponding VFL capability information, a VFL interoperability indicator, and a feature identifier, wherein the machine learning analysis task ID represents a network data analysis and / or machine learning task to be vertically federated, the VFL capability information represents the NWDAF as a VFL server and / or VFL client, the VFL interoperability indicator is used to match network functions (NFs) with the same collaborative capabilities, and the feature identifier represents the data characteristics of the machine learning analysis task; sending a first VFL registration request to a network storage function (NRF), the first VFL registration request carrying the VFL information list; and receiving a first registration success response from the NRF.

[0007] In one embodiment of this disclosure, the recognizable name DN of the machine learning analysis information MlAnalyticsInfo is invoked to query the NWDAFFunction configuration file to obtain a list of vertical federated learning (VFL) information. This includes: locating the corresponding configuration node in the NWDAFFunction configuration file based on the recognizable name DN of the MlAnalyticsInfo; reading the attribute parameters of all VFL-related information object class (IOC) instances mounted under the configuration node, the attribute parameters including the VFL capability information, the VFL interoperability indicator, and the feature identifier; and using the machine learning analysis task ID as a grouping index to associate and read the attribute parameters belonging to the same group to obtain the VFL information list.

[0008] In one embodiment of this disclosure, when the NWDAF acts as the VFL server, the feature identifier includes at least one of the following: network traffic to be analyzed, user behavior, network slice, service latency, and terminal access characteristics; when the NWDAF acts as the VFL client, the feature identifier includes at least one of the following: locally stored network traffic, user behavior, network slice, service latency, and terminal access characteristics.

[0009] In one embodiment of this disclosure, when the NWDAF acts as the VFL client, the attribute parameters read further include a client aggregation capability identifier, which indicates whether the VFL client has the ability to aggregate intermediate computation results from other VFL clients.

[0010] In one embodiment of this disclosure, when the NWDAF acts as the VFL server, it further includes: sending a Network Function (NF) Discovery Request to the NRF, the NF Discovery Request being used to request the NRF to filter VFL clients that meet the conditions for vertical federated learning collaboration; and receiving a client list fed back by the NRF, the client list including at least one of NWDAF clients, Trusted Application Functions (AFs), and Untrusted AFs.

[0011] In one embodiment of this disclosure, sending a Network Function (NF) discovery request to the NRF includes: triggering or receiving a VFL client discovery instruction sent by a consumer via the NWDAF; performing local access verification based on operator policy, target analytics task ID, the VFL interoperability indicator, and service area to determine whether VFL is enabled for the target analytics task; if VFL is enabled, using network element type, the target analytics task ID, VFL client capabilities, and the VFL interoperability indicator as filtering conditions, and generating the NF discovery request based on the filtering conditions; and sending the NF discovery request to the NRF.

[0012] According to another aspect of this disclosure, a method for configuring vertical federated learning information, applied to an NRF, is provided, comprising: receiving a first VFL registration request sent by an NWDAF, the first VFL registration request carrying a VFL information list, the VFL information list including a machine learning analysis task ID and corresponding VFL capability information, a VFL interoperability indicator, and a feature identifier, wherein the machine learning analysis task ID represents a network data analysis and / or machine learning task to be performed for vertical federated learning, the VFL capability information represents the NWDAF as a VFL server and / or VFL client, the VFL interoperability indicator is used to match network functions (NFs) with the same collaborative capabilities, and the feature identifier represents the data characteristics of the machine learning analysis task; registering the NWDAF with VFLs based on the VFL information list; and sending a first registration success response to the NWDAF.

[0013] In one embodiment of this disclosure, the method further includes: receiving a second VFL registration request sent by a trusted AF and / or a third VFL registration request sent by a Network Open Function (NEF), wherein the third VFL registration request is sent by an untrusted AF to the NEF; performing VFL registration and filing for the trusted AF based on the second VFL registration request, and / or performing VFL registration and filing for the untrusted AF based on the third VFL registration request; sending a second registration success response to the trusted AF and / or sending a third registration success response to the NEF.

[0014] In one embodiment of this disclosure, the method further includes: receiving a Network Function (NF) discovery request sent by a VFL server, the NF discovery request being used to request the NRF to filter VFL clients that meet the vertical federated learning collaboration conditions, the VFL server being any one of an NWDAF server, a trusted AF, or a NEF; performing an authentication operation on the NF discovery request; if the authentication is successful, filtering the registered NF files to obtain a list of clients that meet the vertical federated learning collaboration conditions; and feeding back the client list to the VFL server, the client list including at least one of NWDAF clients, trusted AFs, and untrusted AFs.

[0015] In one embodiment of this disclosure, the authentication operation for the NF discovery request includes verifying the network element identity and network access qualifications of the VFL server; and performing discovery service authorization verification on the VFL server.

[0016] In one embodiment of this disclosure, the registered VFL clients are screened to obtain a list of clients that meet the vertical federated learning collaboration conditions. This includes: matching all the registered NF files based on the network element type, target analysis task ID, VFL client capabilities, and VFL interoperability indicators carried in the NF discovery request, and filtering out clients that meet the vertical federated learning collaboration conditions to obtain the client list.

[0017] According to another aspect of this disclosure, a method for configuring vertical federated learning information is provided, applied to a Network Function for Exploration and Exploration (NEF), comprising: receiving a third Virtual Function (VFL) registration request sent by an untrusted AF; sending the third VFL registration request to a Network Function for Exploration and Exploration (NRF) to enable the NRF to register the untrusted AF, wherein the NRF also stores a list of VFL information sent by the NWDAF, the VFL information list including a machine learning analysis task ID and corresponding VFL capability information, a VFL interoperability indicator, and a feature identifier, wherein the machine learning analysis task ID represents a network data analysis and / or machine learning task to be performed for vertical federated learning, the VFL capability information represents the NWDAF as a VFL server and / or VFL client, the VFL interoperability indicator is used to match network functions (NFs) with the same collaborative capabilities, and the feature identifier represents the data characteristics of the machine learning analysis task; and receiving a third registration success response from the NRF, wherein one of the untrusted AF and the NWDAF acts as a VFL server and the other as a VFL client.

[0018] In one embodiment of this disclosure, the method further includes: receiving an NF discovery request sent by the untrusted AF, the NF discovery request being used to request the NRF to filter VFL clients that meet the vertical federated learning collaboration conditions; authenticating the NF discovery request based on operator policies to verify whether the untrusted AF has permission to access the VFL client corresponding to the machine learning analysis task ID; sending the NF discovery request to the NRF after successful authentication; and receiving a client list fed back by the NRF, the client list including the NWDAF that meets the vertical federated learning collaboration conditions.

[0019] In one embodiment of this disclosure, the NF discovery request includes a target analysis task ID, a request for VFL capability information, the VFL interoperability indicator, and optional supported feature identifiers.

[0020] In one embodiment of this disclosure, the method further includes: extracting VFL information and service parameter information for each NWDAF from the client list; determining, based on the VFL information, NWDAFs having the same target analysis task ID, the same VFL interoperability indicator, and the feature identifier matching the requirements, as candidate NWDAFs; filtering the candidate NWDAFs based on at least one of load status, service capacity, and network reachability in the service parameter information to obtain a target NWDAF, and assigning a temporary external identifier to the target NWDAF; and sending a client discovery response carrying the temporary external identifier to the untrusted AF.

[0021] In one embodiment of this disclosure, the candidate NWDAFs are filtered based on at least one of the load status, service capacity, and network reachability in the service parameter information to obtain target NWDAFs. The method further includes: if the VFL information of multiple target NWDAFs includes a VFL client aggregation capability identifier, then one of the multiple target NWDAFs is designated as an aggregation node and fed back to the untrusted AF.

[0022] According to another aspect of this disclosure, a configuration device for vertical federated learning information is provided, applied to a network data analysis function (NWDAF), comprising: a calling module, configured to call the identifiable name (DN) of machine learning analysis information (MlAnalyticsInfo), query the NWDAF configuration file, and obtain a list of vertical federated learning (VFL) information, wherein the VFL information list includes a machine learning analysis task ID and corresponding VFL capability information, a VFL interoperability indicator, and a feature identifier, wherein the machine learning analysis task ID represents a network data analysis and / or machine learning task to be vertically federated, the VFL capability information represents the NWDAF as a VFL server and / or VFL client, the VFL interoperability indicator is used to match network functions (NFs) with the same collaborative capabilities, and the feature identifier represents the data characteristics of the machine learning analysis task; a first sending module, configured to send a first VFL registration request to a network storage function (NRF), the first VFL registration request carrying the VFL information list; and a first receiving module, configured to receive a first registration success response from the NRF.

[0023] According to another aspect of this disclosure, a configuration apparatus for vertical federated learning information is provided, applied to an NRF, comprising: a second receiving module, configured to receive a first VFL registration request sent by an NWDAF, the first VFL registration request carrying a VFL information list, the VFL information list including a machine learning analysis task ID and corresponding VFL capability information, a VFL interoperability indicator, and a feature identifier, the machine learning analysis task ID representing a network data analysis and / or machine learning task to be performed for vertical federated learning, the VFL capability information representing the NWDAF as a VFL server and / or VFL client, the VFL interoperability indicator being used to match network functions (NFs) with the same collaborative capabilities, and the feature identifier representing the data characteristics of the machine learning analysis task; a registration module, configured to register the NWDAF with VFLs based on the VFL information list; and a second sending module, configured to send a first registration success response to the NWDAF.

[0024] According to another aspect of this disclosure, a configuration apparatus for vertical federated learning information is provided, applied to a Network Function for Optimization (NEF), comprising: a third receiving module for receiving a third VFL registration request sent by an untrusted AF; a third sending module for sending the third VFL registration request to a Network Function for Optimization (NRF) to enable the NRF to register the untrusted AF, wherein the NRF also stores a list of VFL information sent by the NWDAF, the VFL information list including a machine learning analysis task ID and corresponding VFL capability information, a VFL interoperability indicator, and a feature identifier, wherein the machine learning analysis task ID represents a network data analysis and / or machine learning task to be performed for vertical federated learning, the VFL capability information represents the NWDAF as a VFL server and / or VFL client, the VFL interoperability indicator is used to match network functions (NFs) with the same cooperative capabilities, and the feature identifier represents the data characteristics of the machine learning analysis task; and a fourth receiving module for receiving a third registration success response from the NRF, wherein one of the untrusted AF and the NWDAF acts as a VFL server and the other as a VFL client.

[0025] According to another aspect of this disclosure, a network device is provided, comprising: a processor; and a memory for storing executable instructions of the processor; the processor being configured to perform a configuration method for vertical federated learning information of the first aspect by executing the executable instructions.

[0026] According to another aspect of this disclosure, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the above-described method for configuring vertical federated learning information.

[0027] According to another aspect of this disclosure, a computer program product is provided, on which a computer program is stored, which, when executed by a processor, implements the above-described method for configuring vertical federated learning information.

[0028] The vertical federated learning information configuration scheme provided in the embodiments of this disclosure uses NWDAF to call the identifiable name DN corresponding to MlAnalyticsInfo, retrieves the local NWDAFunction configuration file, and extracts the VFL capability information and VFL of the corresponding machine learning analysis task. Interoperability indicators and feature identifiers are used to generate a VFL information list. This method only retrieves the necessary vertical federated learning configuration items, without loading the full data, which reduces the resource consumption of NWDAF. The NWDAF reports the VFL information list to NRF through the first VFL registration request to complete the registration and filing. It can establish an association between scattered VFL service parameters and NWDAF configuration files and complete network-side registration, realizing the binding of VFL information and network element configuration, and building a standardized configuration management model. Furthermore, based on unified configuration and registration data, VFL capabilities, collaboration rules, and data characteristics can be uniformly queried and maintained. NRF can also identify NWDAFs with VFL functions. When the VFL server initiates a node discovery request, such NWDAFs can be listed as alternative clients, providing data support for network element retrieval, capability matching, and cross-node collaborative operations. This guides the VFL server and client to execute corresponding learning strategies, ensuring the stable operation of network intelligent analysis tasks and optimizing the overall business experience.

[0029] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0030] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure. It is obvious that the drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0031] Figure 1 This diagram illustrates a flowchart of a method for configuring vertical federated learning information according to an embodiment of the present disclosure. Figure 2 This diagram illustrates a data structure related to vertical federated learning in an NWDAF according to an embodiment of this disclosure; Figure 3 This diagram illustrates a flowchart of another method for configuring vertical federated learning information in an embodiment of this disclosure. Figure 4 This illustration shows a flowchart of another method for configuring vertical federated learning information in an embodiment of the present disclosure; Figure 5 A flowchart illustrating another method for configuring vertical federated learning information in an embodiment of this disclosure is shown. Figure 6A flowchart illustrating another method for configuring vertical federated learning information in an embodiment of this disclosure is shown. Figure 7 This diagram illustrates a configuration apparatus for vertical federated learning information in an embodiment of the present disclosure. Figure 8 This diagram illustrates another configuration device for vertical federated learning information in an embodiment of the present disclosure. Figure 9 A schematic diagram of a configuration device for another type of vertical federated learning information is shown in an embodiment of this disclosure. Figure 10 A structural block diagram of a computer device according to an embodiment of the present disclosure is shown. Detailed Implementation

[0032] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that this disclosure will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0033] Furthermore, the accompanying drawings are merely illustrative of this disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0034] Vertical Federated Learning (VFL) is a distributed machine learning paradigm suitable for scenarios where participating data samples overlap significantly and data features differ considerably. Multiple participating network elements possess network data features of different dimensions. Each party does not need to transfer original data across network elements; instead, they conduct joint training and inference solely through the intermediate results of the interactive model. This approach accomplishes intelligent analysis tasks while effectively protecting data privacy.

[0035] The current 3GPP Rel-19 defines the NWDAF management specifications and Vertical Federated Learning (VFL) service processes, covering two application scenarios: NWDAF acting as a VFL server and VFL client. According to the requirements of the 3GPPSA2 working group, NWDAF must rely on the configuration data of the network management system to determine the types of vertical federated learning information it supports.

[0036] However, the existing standard system has obvious shortcomings: under the SA5 Network Management Orchestration and Billing Working Group framework, the configuration scheme and management mechanism for vertical federated learning information have not yet been defined, and it is impossible to realize the association and binding of vertical federated learning information with NWDAF network element configuration information.

[0037] Existing network management systems are based on the Network Resource Model (NRM) and manage areas such as the radio access network, core network, and network slicing. They define various Information Object Classes (IOCs) to constrain attribute semantics and interaction behaviors. Taking the core network NRM as an example, the NWDAFunction IOC is defined in protocol TS 28.541. By configuring attributes such as TAI and analysis task ID, it supports NWDAF in completing basic functions such as network element registration, network function discovery, model training, and data analysis. However, this IOC does not extend VFL-related configuration capabilities.

[0038] Based on the scheme disclosed herein, vertical federated learning information can be configured into the NWDAF network element, enabling the NWDAF to determine the supported vertical federated learning information when required by the NRF (Network Repository Function), thereby instructing the vertical federated learning server / client to execute the corresponding learning strategy and ensuring the user's business experience.

[0039] The network elements involved in this disclosure are explained below.

[0040] NWDAF (Network Data Analytics Function): It can act as a VFL server or client to complete VFL information registration and participate in federated learning operations.

[0041] NRF (Network Repository Function): Responsible for the registration of VFL information of network elements across the entire network, request authentication, and screening of collaborative clients.

[0042] AF (Application Function): Represents a third-party business application, which is divided into trusted and untrusted categories and can participate in VFL-related processes.

[0043] NEF (Network Exposure Function): As an external gateway, it proxies untrusted AFs to forward requests, verify permissions, and manage temporary identifiers.

[0044] NF (Network Function): A general term for various service network elements in a network.

[0045] The following will describe in more detail the steps of the method for configuring vertical federated learning information in this example embodiment, with reference to the accompanying drawings and embodiments.

[0046] Figure 1 A flowchart illustrating a method for configuring vertical federated learning information in an embodiment of this disclosure is shown.

[0047] like Figure 1 As shown, a method for configuring vertical federated learning information according to an embodiment of this disclosure, applied to a network data analysis function (NWDAF), includes: Step S102: Call the recognizable name DN of the machine learning analysis information MlAnalyticsInfo, query the NWDAFFunction configuration file, and obtain the vertical federated learning VFL information list. The VFL information list includes the machine learning analysis task ID and the corresponding VFL capability information, VFL interoperability indicator, and feature identifier. The machine learning analysis task ID represents the network data analysis and / or machine learning task to be performed in vertical federated learning. The VFL capability information represents NWDAF as a VFL server and / or VFL client. The VFL interoperability indicator is used to match network functions NF with the same collaborative capabilities. The feature identifier represents the data characteristics of the machine learning analysis task.

[0048] In some embodiments, machine learning analysis information MlAnalyticsInfo is a collection of information used to carry the configuration and description of network data analysis and machine learning related tasks, and is used to define the basic attributes and operating rules of various network analysis tasks.

[0049] In some embodiments, the Distinguished Name (DN) is an identifier used to locate the corresponding configuration object or data node within the network function, enabling quick lookup and access to the configuration file.

[0050] In some embodiments, NWDAFFunction is a function configuration file corresponding to the network data analysis function, which is used to store all configuration parameters such as various analysis tasks, service capabilities, and collaborative rules supported by the network element.

[0051] In some embodiments, the machine learning analysis task ID is used to uniquely distinguish different vertical federated learning tasks; VFL capability information is used to indicate that the current network element can assume one or both roles of VFL server and VFL client; VFL interoperability indicator is used to filter network functions with matching and collaborative capabilities to ensure normal cross-network element linkage. When multiple NFs (i.e., NWDAF and / or AF) collaboratively execute VFL, only NFs that share the same VFL interoperability indicator (vflInterInfo) can identify each other's feature ID (featureId) and other VFL configuration parameters; feature identifier is used to indicate the data feature type used by the task.

[0052] Step S104: Send a first VFL registration request to the Network Storage Function (NRF). The first VFL registration request carries a list of VFL information.

[0053] In some embodiments, the first VFL registration request is a signaling request initiated by the NWDAF to the NRF. To support VFL training and inference, when an NWDAF that supports VFL registers with the NRF, if the NWDAF is a VFL server or a VFL client, the registration must include a VFL interoperability indicator (vflInterInfo) corresponding to each Analytics ID and an optional feature identifier (featureId). The registration can be initiated as a server or as a client. When a VFL server or NF consumer discovers an NWDAF that supports VFL through the NRF, it will consider the above registration information.

[0054] After receiving the request, the NRF extracts the VFL information list from the request, completes the registration and filing of the corresponding network element information, and stores the relevant data for subsequent querying and matching.

[0055] Step S106: Receive the first registration success response from the NRF.

[0056] In some embodiments, the first registration success response is a feedback signaling returned by the NRF to the NWDAF, used to inform the NWDAF that the VFL information registration process is complete.

[0057] In this embodiment, the NWDAF calls the identifiable name DN corresponding to MlAnalyticsInfo to retrieve the local NWDAFunction configuration file and extract the VFL capability information and VFL of the corresponding machine learning analysis task. Interoperability indicators and feature identifiers are used to generate a VFL information list. This method only retrieves the necessary vertical federated learning configuration items, without loading the full data, which reduces the resource consumption of NWDAF. The NWDAF reports the VFL information list to NRF through the first VFL registration request to complete the registration and filing. It can establish an association between scattered VFL service parameters and NWDAF configuration files and complete network-side registration, realizing the binding of VFL information and network element configuration, and building a standardized configuration management model. Furthermore, based on unified configuration and registration data, VFL capabilities, collaboration rules, and data characteristics can be uniformly queried and maintained. NRF can also identify NWDAFs with VFL functions. When the VFL server initiates a node discovery request, such NWDAFs can be listed as alternative clients, providing data support for network element retrieval, capability matching, and cross-node collaborative operations. This guides the VFL server and client to execute corresponding learning strategies, ensuring the stable operation of network intelligent analysis tasks and optimizing the overall business experience.

[0058] In one embodiment of this disclosure, the recognizable name (DN) of the machine learning analysis information MlAnalyticsInfo is invoked to query the NWDAFFunction configuration file, obtaining a list of vertical federated learning (VFL) information, including: Based on the identifiable name DN of MlAnalyticsInfo, locate the corresponding configuration node in the NWDAFFunction configuration file; read the attribute parameters of all information object class IOC instances related to VFL mounted under the configuration node, including VFL capability information, VFL interoperability indicators and feature identifiers; use the machine learning analysis task ID as the group index to read the attribute parameters belonging to the same group to obtain the VFL information list.

[0059] In some embodiments, each vertical federated learning information is managed as an Information Object Class (IOC). NWDAF locates the corresponding configuration node in the NWDAFFunction configuration file by referencing the identifiable name (DN) of the machine learning analysis information MlAnalyticsInfo, and then obtains the attribute parameters and values ​​corresponding to the IOC instance, such as the attribute value "TRUE". In addition, the network management system integrates the generated vertical federated learning VFL information list and sets it as the newly added MlAnalyticsInfo data type in the NWDAFFunction configuration file to achieve unified storage, retrieval and standardized management of VFL-related data.

[0060] In some embodiments, Figure 2 A class diagram of an NWDAFFunction configuration object is shown. NWDAFFunction is equivalent to the configuration root node of NWDAF and is directly associated with the MlAnalyticsInfo class. It represents the set of machine learning analysis capabilities supported by this network element. MlAnalyticsInfo describes a specific machine learning analysis task. vflInterInfo represents the VFL interoperability indicator, storing the collaborative interaction specifications of the task, including supported VFL protocol versions, communication modes, collaborative algorithm types, etc., used to match NFs with the same collaborative capabilities. featureId is a string type that can store the data feature types used by the analysis task (such as network traffic, user behavior, network slicing, service latency, terminal access characteristics, etc.), used to limit the data dimensions of the task, and also used to match data compatibility with other nodes. vflClientAggrCap is vflClientAggrCap, i.e., the client aggregation capability identifier, representing whether the current NWDAF, when acting as a VFL client, has the ability to aggregate intermediate computation results from other VFL clients. A value of True indicates that it has aggregation capability and can act as a secondary aggregation node; a value of False indicates that it does not have aggregation capability and only participates in computation as a normal client.

[0061] In some embodiments, vertical federated learning information is encapsulated as an Information Object Class (IOC) instance. NWDAF locates the target configuration node in the NWDAFFunction configuration file based on the identifiable name (DN) corresponding to MlAnalyticsInfo, reads the attribute parameters of all VFL-related IOC instances under that node, and then classifies and integrates the parameters using the machine learning analysis task ID as a grouping index, finally generating a structured VFL information list. Based on the architecture design of DN and IOC, NWDAF can call one or more DNs to obtain vertical federated learning information corresponding to different dimensions and tasks.

[0062] In this embodiment, by managing and configuring the vertical federated learning information associated with each NWDAF as an independent IOC, and locating the configuration node through the identifiable name (DN), the system accurately reads various attribute parameters of the required VFL-related IOC instances. Instead of loading all configuration information, it only calls the vertical federated learning information of specified items, effectively saving NWDAF network element resources. Parameters are grouped and integrated using the machine learning analysis task ID as an index, and scattered VFL configuration parameters are organized into a standardized VFL information list according to the task dimension. All types of VFL-related data are centrally stored in the NWDAFFunction configuration architecture. This approach provides a reliable and organized data source for NWDAF to report VFL registration information to NRF, and also achieves refined and modular configuration management of VFL information, significantly improving the standardization and data integrity of VFL information configuration.

[0063] In one embodiment of this disclosure, when NWDAF acts as a VFL server, the feature identifier includes at least one of the following: network traffic to be analyzed, user behavior, network slicing, service latency, and terminal access characteristics.

[0064] In this embodiment, the feature identifiers on the VFL server side can clearly define the data dimensions on which the VFL server conducts joint analysis, delineate the data scope of the analysis task, facilitate the connection and matching of collaborative nodes, and ensure that the vertical federated learning task proceeds normally around the predetermined network analysis content.

[0065] When NWDAF acts as a VFL client, its signature includes at least one of the following: locally stored network traffic, user behavior, network slicing, service latency, and terminal access characteristics.

[0066] In this embodiment, the feature identifier on the VFL client side is used to indicate the local data types that the VFL client can participate in the computation, clarify the data content it can provide, facilitate capability matching with the VFL server and other collaborating nodes, and support the joint analysis work.

[0067] In one embodiment of this disclosure, when NWDAF acts as a VFL client, the attribute parameters read also include a client aggregation capability identifier, which indicates whether the VFL client has the ability to aggregate intermediate computation results from other VFL clients.

[0068] In some embodiments, the vertical federated learning information under the machine learning analysis list (MlAnalyticsList) configured by the network management system for the NWDAF indicates the configuration parameter information that each VFL-enabled NWDAF needs to include when registering with the NRF and needs to be considered when being discovered by the NRF, including the VFL interoperability indicator (vflInterInfo), feature identifier (featureId), and intermediate result aggregation capability (vflClientAggrCap).

[0069] The VFL interoperability indicator (vflInterInfo) is an important piece of information that assists multiple NFs in coordinating the execution of VFLs. Its constraint is that if an NWDAF supports VFL interoperability with a certain Analytics ID, it must include this attribute; otherwise, the NWDAF is not allowed to execute VFL operations.

[0070] The feature identifier (featureId) represents the feature information supported by NWDAF as a VFL client for a specific analysis task ID (i.e., VFL training and / or inference). Its constraint is that if a VFL CapabilityType (vflCapabilityType) exists, this attribute must be included.

[0071] The intermediate result aggregation capability (vflClientAggrCap) is constrained by the fact that the NWDAF supports VFL client capabilities that aggregate intermediate results from other VFL clients, and the VFL capability type (vflCapabilityType) is "VFL_CLIENT" or "VFL_SERVER_AND_CLIENT".

[0072] In this embodiment, the client aggregation capability identifier is used to indicate whether the current VFL client can aggregate the intermediate computation results output by other VFL clients. The network side can assign corresponding computation responsibilities based on the identifier and reasonably plan the multi-client collaborative working mode.

[0073] In one embodiment of this disclosure, when NWDAF acts as a VFL server, it further includes: sending a Network Function (NF) Discovery Request to the NRF, the NF Discovery Request being used to request the NRF to filter VFL clients that meet the conditions for vertical federated learning collaboration; and receiving a client list fed back by the NRF, the client list including at least one of NWDAF clients, Trusted Application Functions (AFs), and Untrusted AFs.

[0074] In some embodiments, a trusted application function (AF) can be understood as an application function entity that has been authenticated by the network side and whose permissions are controlled. An untrusted application function entity is an application function entity that has not passed the network's complete security authentication or whose access permissions and running status are uncertain. The network will not fully open the interaction permissions. When participating in vertical federated learning tasks, it is necessary to add protection mechanisms such as data verification and access restrictions.

[0075] In some embodiments, if a vertical federated learning task needs to be performed, the NWDAF, acting as the VFL server, actively initiates a network function discovery request to the NRF. The discovery request is used to request the NRF to retrieve registered VFL client nodes that meet the collaboration requirements. After receiving the client list returned by the NRF, the participants in this joint task can be confirmed. The list covers various network elements such as NWDAFs of the same type, trusted application function AFs, and untrusted AFs, in order to complete the screening of collaborative nodes.

[0076] In this embodiment, NWDAF, as a VFL server, can proactively initiate a node search process, use NRF to filter collaborative clients and obtain the corresponding list. NWDAF can connect to various types of nodes such as NWDAF, trusted AF, and untrusted AF, expanding the available data for analysis tasks, extending the business coverage, and clarifying the network functions of all participating tasks, thereby helping to ensure the continuous and stable operation of vertical federated learning analysis services.

[0077] In one embodiment of this disclosure, sending a Network Function (NF) discovery request to the NRF includes: The NWDAF triggers or receives VFL client discovery commands sent by consumers, performs local access verification based on operator policies, target analysis task ID, VFL interoperability indicator, and service area to determine whether VFL is enabled for the target analysis task; if VFL is enabled, it uses network element type, target analysis task ID, VFL client capabilities, and VFL interoperability indicator as filtering conditions, and generates an NF discovery request based on the filtering conditions; and sends the NF discovery request to the NRF.

[0078] In some embodiments, the VFL client discovery process may include two triggering methods: initiated autonomously by the NWDAF, which acts as the VFL server, or initiated by an instruction issued by an external consumer.

[0079] In some embodiments, after the process is started, the operator's policies can be checked to see if the current business scenario and data usage rules allow the vertical federated learning to be enabled. Then, the target analysis task ID is matched to confirm whether the task type matches the applicable scenario of VFL technology. Furthermore, the VFL interoperability indicator is compared to determine whether the current network element's interaction capability can support cross-node collaborative computing. Finally, the service area information is combined to check whether the task execution scope meets the regional control requirements. If the detection results determine that the requirements are met, the vertical federated learning task is enabled.

[0080] After determining the task to be initiated, NWDAF defines the candidate range based on network element type, including network functions such as NWDAF, trusted AF, and untrusted AF that can participate in VFL tasks into the screening objects; for the target analysis task ID, it binds rules to the corresponding analysis task, limiting node search to only for this task; for VFL client capabilities, it marks the data processing, computation execution and other capabilities required by the task; for VFL interoperability indicators, it defines the unified interaction specifications and communication standards that need to be unified between nodes, and the screening conditions are obtained based on the above constraints.

[0081] In this embodiment, two working modes, server self-triggered and external command triggered, are used to adapt to the startup requirements under different business scenarios. The pre-entry verification process can comprehensively judge the applicability of VFL technology by combining operational specifications, task attributes, and service scope, preventing the blind startup of joint tasks. Multi-dimensional screening conditions can match the client nodes required by the task, improve the compatibility between nodes, reduce unnecessary signaling interactions, improve the operational efficiency of the node discovery process, and ensure the reliability of the node selection process of vertical federated learning.

[0082] like Figure 3 As shown, a method for configuring vertical federated learning information according to another embodiment of this disclosure, applied to NRF, includes: Step S302: Receive the first VFL registration request sent by NWDAF. The first VFL registration request carries a VFL information list. The VFL information list includes a machine learning analysis task ID and the corresponding VFL capability information, a VFL interoperability indicator, and a feature identifier. The machine learning analysis task ID represents the network data analysis and / or machine learning task to be performed for vertical federated learning. The VFL capability information represents NWDAF as a VFL server and / or VFL client. The VFL interoperability indicator is used to match network functions (NFs) with the same collaborative capabilities. The feature identifier represents the data characteristics of the machine learning analysis task.

[0083] Step S304: Register and file the NWDAF VFL based on the VFL information list.

[0084] In some embodiments, after receiving the first VFL registration request sent by the NWDAF, the NRF extracts the VFL information list in the request, and enters the task ID, VFL capability information, VFL interoperability indicator, feature identifier, and other contents included in the list into the network function configuration file of the corresponding NWDAF for retention, thereby completing the VFL registration and filing operation of the network element.

[0085] Step S306: Send the first registration success response to NWDAF.

[0086] In this embodiment, by receiving the registration data of NWDAF from the NRF and completing the filing, the VFL-related information is centrally stored in the network element configuration file, realizing unified management of VFL-related information. When carrying out node search and collaborative matching work in the future, the filing data can be directly retrieved, providing data support for cross-network element vertical federated learning collaborative interaction.

[0087] In one embodiment of this disclosure, the method further includes: receiving a second VFL registration request sent by a trusted AF and / or a third VFL registration request sent by a Network Open Function (NEF), wherein the third VFL registration request is sent by an untrusted AF to the NEF; performing VFL registration and filing for the trusted AF based on the second VFL registration request, and / or performing VFL registration and filing for the untrusted AF based on the third VFL registration request; sending a second registration success response to the trusted AF and / or sending a third registration success response to the NEF.

[0088] In some embodiments, the second VFL registration request is initiated by the trusted AF, and the message carries information such as the application function's own VFL capabilities, interaction rules, and data characteristics, which is used to register its own VFL-related qualifications and parameters with the NRF.

[0089] The third VFL registration request corresponds to the registration requirement of an untrusted AF. Since the untrusted AF has not completed the full network security authentication, it does not have the authority to interact directly with the NRF. This type of network element will first send the registration request to the NEF, and then the NEF will forward it to the NRF, using the NEF to complete the secure relay and request preprocessing.

[0090] In this embodiment, NRF supports multiple network functions such as NWDAF, trusted AF, and untrusted AF to complete VFL registration and filing. It sets differentiated registration paths for network elements with different security levels, which not only completes the information collection of network elements that can participate in vertical federated learning, but also adds a relay control link for network elements with lower security levels, standardizes the registration process of different types of network elements, and improves the security and order of the overall network operation.

[0091] In one embodiment of this disclosure, the method further includes: receiving a Network Function (NF) discovery request sent by a VFL server, wherein the NF discovery request is used to request the NRF to filter VFL clients that meet the conditions for vertical federated learning collaboration, and the VFL server is any one of an NWDAF server, a trusted AF, or a NEF; performing an authentication operation on the NF discovery request; if the authentication is successful, filtering the registered NF files to obtain a list of clients that meet the conditions for vertical federated learning collaboration; and feeding back the client list to the VFL server, wherein the client list includes at least one of NWDAF clients, trusted AFs, and untrusted AFs.

[0092] In this embodiment, after receiving the NF discovery request from the VFL server, the NRF first executes an authentication process to verify the legitimate identity and usage rights of the initiator. When the authentication process is completed and the result is successful, the NRF retrieves the local network function files that have completed VFL registration and filing, compares and filters them one by one according to the collaboration conditions in the request, and generates a list of clients that meet the requirements after the filtering is completed. The list is then sent back to the VFL server or NEF that initiated the request for subsequent collaboration tasks. The addition of an authentication step before node retrieval can detect query requests initiated by unauthorized entities. Combined with the filtering work based on network element files, it can quickly locate client nodes that meet the collaboration requirements, ensuring the reliability of nodes participating in vertical federated learning.

[0093] In one embodiment of this disclosure, the authentication operation for an NF discovery request includes: Verify the network element identity and network access qualifications of the VFL server.

[0094] In some embodiments, the NRF retrieves the network element basic files stored locally, compares the network element identifier, device code and other information carried in the NF discovery request, verifies whether the VFL server that initiated the request is a legitimate network access device, and can also check the network access license record and operating status of the network element to confirm that the device is in a normal network access state, thereby completing the verification of the network element identity and network access qualification.

[0095] Perform discovery service authorization verification on the VFL server.

[0096] In some embodiments, NRF retrieves the network permission configuration policy, queries the permission list corresponding to the current VFL server, verifies whether the network element has been assigned network function discovery related operation permissions, and confirms that the permission configuration allows it to initiate client lookup requests, thus determining that the authorization verification has passed; if the corresponding permission is not configured, the verification has failed.

[0097] In one embodiment of this disclosure, the registered VFL clients are screened to obtain a list of clients that meet the conditions for vertical federated learning collaboration. This includes: matching all registered NF files based on the network element type, target analysis task ID, VFL client capabilities, and VFL interoperability indicators carried in the NF discovery request, and filtering out clients that meet the conditions for vertical federated learning collaboration to obtain the client list.

[0098] In some embodiments, NRF extracts four items from the NF discovery request: network element type, target analysis task ID, VFL client capability, and VFL interoperability indicator. These are used as matching criteria. NRF then iterates through all locally registered network function files and compares the corresponding fields of each file. Only network functions whose contents match the matching criteria are identified as VFL clients that meet the cooperation conditions, and these are then compiled into a client list.

[0099] In this practical embodiment, a multi-field, line-by-line matching method is adopted to filter clients based on task requirements, network element attributes, and interaction rules, so that the selected nodes can be adapted to the current vertical federated learning task.

[0100] like Figure 4 As shown, a method for configuring vertical federated learning information according to another embodiment of this disclosure, applied to NEF, includes: Step S402: Receive a third VFL registration request sent by an untrusted AF.

[0101] Step S404: Send a third VFL registration request to the NRF so that the NRF can register the untrusted AF. The NRF also stores a list of VFL information sent by the NWDAF. The VFL information list includes the machine learning analysis task ID and the corresponding VFL capability information, VFL interoperability indicator and feature identifier. The machine learning analysis task ID represents the network data analysis and / or machine learning task to be performed for vertical federated learning. The VFL capability information represents the NWDAF as a VFL server and / or VFL client. The VFL interoperability indicator is used to match network functions (NFs) with the same collaborative capabilities. The feature identifier represents the data characteristics of the machine learning analysis task.

[0102] Step S406: Receive the third registration success response from the NRF, wherein one of the untrusted AF and NWDAF acts as the VFL server and the other acts as the VFL client.

[0103] In this embodiment, the untrusted AF lacks the access qualification to directly interact with the NRF via signaling and cannot independently complete VFL capability registration. This prevents such network elements from directly participating in vertical federated learning collaborative tasks. By forwarding the third VFL registration request of the untrusted AF through the NEF, the NRF can normally complete the VFL registration and filing of the untrusted AF. Since the untrusted AF can only select the NWDAF as the VFL client when acting as a VFL server, combined with the VFL information list sent by the NWDAF, the NEF can realize the construction of the VFL network when the untrusted AF acts as a VFL server. This opens up the collaborative channel between low-trust level application functions and internal network NWDAF nodes, supports restricted network elements in successfully building a dedicated vertical federated learning task network, and constructs a federated collaborative architecture for network elements with different security levels.

[0104] In one embodiment of this disclosure, the method further includes: receiving an NF discovery request sent by an untrusted AF, the NF discovery request being used to request the NRF to filter VFL clients that meet the vertical federated learning collaboration conditions; authenticating the NF discovery request based on operator policies to verify whether the untrusted AF has permission to access the VFL client corresponding to the machine learning analysis task ID; sending the NF discovery request to the NRF after successful authentication; and receiving a client list from the NRF, the client list including NWDAFs that meet the vertical federated learning collaboration conditions.

[0105] In some embodiments, the NEF local storage operator's preset permission control policy records the task access permissions, data interaction scope, and collaborative call constraints corresponding to different untrusted AFs. When an untrusted AF initiates an NF discovery request with a specified machine learning analysis task ID, the NEF extracts the target analysis task ID and the network element identity information of the initiator from the request, compares it with the local storage operator's policy, queries whether the current untrusted AF has the client node query permission for the corresponding task ID, and determines whether the untrusted AF can call the corresponding VFL collaborative resources. Only when the policy matches and the permission entry is legally present is the authentication result determined to be passed.

[0106] In this embodiment, NEF adds a policy-level authentication step to NF discovery requests from untrusted AFs. This allows for permission constraints on node query behavior of low-trust-level network elements, preventing untrusted AFs from indiscriminately and unrestrictedly querying VFL client resources across the entire network. This mechanism can precisely constrain the task resource access scope of untrusted AFs, eliminating unauthorized queries and illegal resource calls. While allowing untrusted AFs to participate in vertical federated learning tasks, it ensures the access security and orderly management of VFL node resources across the entire network.

[0107] In one embodiment of this disclosure, the NF discovery request includes a target analysis task ID, a request for VFL capability information, a VFL interoperability indicator, and optional supported feature identifiers.

[0108] In some embodiments, the multiple fields carried in the NF discovery request are used to determine the constraints for matching vertical federated learning nodes. Among them, the target analysis task ID is used to lock the machine learning analysis task to be executed and limit the task scenario for node matching; the VFL capability information requirements are used to clarify the server or client working capabilities required by the candidate NWDAF; the VFL interoperability indicator is used to unify the collaborative interaction specifications between nodes and ensure that the node operation logic and communication rules are compatible with each other; the optional supported feature identifiers are used to supplement and limit the data analysis feature dimensions that the node can provide, improve the adaptability of the node's business layer, and the combination of multiple fields can convey the collaborative requirements of untrusted AFs and provide a matching basis for NRF to match suitable VFL clients.

[0109] In one embodiment of this disclosure, the method further includes: extracting VFL information and service parameter information for each NWDAF from the client list; determining, based on the VFL information, NWDAFs with the same target analysis task ID, the same VFL interoperability indicator, and feature identifiers matching the requirements as candidate NWDAFs; filtering the candidate NWDAFs based on at least one of load status, service capacity, and network reachability in the service parameter information to obtain the target NWDAF; and assigning a temporary external identifier to the target NWDAF; and sending a client discovery response carrying the temporary external identifier to the untrusted AF.

[0110] In one embodiment of this disclosure, candidate NWDAFs are filtered based on at least one of load status, service capacity, and network reachability in the service parameter information to obtain target NWDAFs. The method further includes: if the VFL information of multiple target NWDAFs includes a VFL client aggregation capability identifier, then one of the multiple target NWDAFs is designated as an aggregation node and fed back to the untrusted AF.

[0111] In some embodiments, VFL information includes analysis task ID, VFL capability information, VFL interoperability indicators, and feature identifiers.

[0112] In some embodiments, service parameter information includes NWDAF load status, service capacity, and network reachability.

[0113] In some embodiments, NEF parses the NF discovery response returned by NRF and extracts information on all NWDAF clients that meet the filtering criteria, including: Each NWDAF's NF Profile includes the analysis task ID, VFL capability information, VFL interoperability indicators, and feature identifiers. Additional information includes each NWDAF's load status, service capacity, network reachability, etc., and each NWDAF's VFL client aggregation capability identifier (vflClientAggrCap).

[0114] Furthermore, after confirming that the NWDAF's analysis task ID, VFL interoperability indicator, and feature identifier match the AF's requirements, NWDAFs with excessive load, insufficient capacity, or abnormal network status are excluded, while a candidate list of healthy and capable NWDAFs is retained.

[0115] Optionally, check if there is a node in the candidate NWDAF that is marked as supporting aggregation (vflClientAggrCap). If it exists, NEF can select one from the NWDAFs with aggregation capabilities according to the load balancing strategy and designate it to serve as the aggregation node for this VFL task, responsible for aggregating the intermediate calculation results of other clients. If it does not exist or aggregation is not required, this step can be skipped, and only ordinary client nodes can be retained.

[0116] In some embodiments, NEF assigns temporary external NWDAF identifiers to all selected NWDAFs and establishes the following mapping: External NWDAF identifier and internal identifier of real NWDAF; the association between external NWDAF identifier and its VFL task and untrusted AF. The external identifier is only valid in the interaction of this VFL task. When the untrusted AF communicates with NEF in the future, it will only use the identifier to refer to the target NWDAF and hide the real network element information.

[0117] In some embodiments, if an aggregation node has been specified, the result only includes the external identifier of the aggregation node and the VFL interoperability indicator. The aggregation node is responsible for subsequent collaboration with other clients, and the result does not carry any real internal identifier or sensitive configuration information of NWDAF.

[0118] In this embodiment, based on the client list initially screened by NRF, NEF first performs a secondary screening based on the task ID, interoperability indicator, and feature identifier to select candidate NWDAFs that are fully compatible with the business capabilities. Then, it performs status filtering by combining operational status parameters such as load status, service capacity, and network reachability to eliminate nodes that are abnormal, lack resources, or have broken links. Finally, it determines the target NWDAF that is compatible and assigns a temporary external identifier to the target NWDAF to mask the real internal network identifier of the network element. This helps to improve the rationality and availability of the selection of vertical federation learning nodes. The use of temporary identifiers can also hide internal network element information, which helps to reduce the risk of internal network information exposure.

[0119] Figure 5 The process of registering and discovering VFLs using NWDAF as a server / client is illustrated. In order to support VFL training and inference, when an NWDAF that supports VFLs registers with the NRF, it must include a list of relevant VFL information. When the NRF attempts to discover an NWDAF that supports VFL training and inference, it must return the list of relevant VFL information upon discovery.

[0120] NWDAF acts as both a VFL server and client in the VFL information registration and discovery process, including: In step S502, NWDAF, acting as the VFL server, queries the NWDAFFunction configuration file, which includes parameters such as VFL capability information corresponding to each analysis task ID, VFL interoperability indicators, and optional supported feature identifiers.

[0121] In step S504, NWDAF, acting as a VFL client, queries the NWDAFFunction configuration file, which includes parameters such as VFL capability information corresponding to each analysis task ID, VFL interoperability indicators, and optional supported feature identifiers.

[0122] In step S506, the VFL server and VFL client send a VFL registration request to the NRF, including a list of VFL information.

[0123] The VFL information list includes parameters such as VFL capability information corresponding to each analysis task ID, VFL interoperability indicators, and optional supported feature identifiers.

[0124] In step S508, the trusted AF, acting as a VFL client, sends a VFL registration request to the NRF to directly register its NF configuration file with the NRF, including parameters such as NF type, Analytics ID, corresponding VFL capability information, VFL interoperability indicator, and optional supported feature identifiers.

[0125] In step S510, the untrusted AF, acting as a VFL client, needs to send a VFL registration request to NEF through OAM configuration to register its analysis task ID, corresponding VFL capability information, VFL interoperability indicator, optional supported feature identifiers, and other parameters.

[0126] In step S512, NEF sends a VFL registration request to NRF to update the NEF configuration file in NRF.

[0127] In step S514, the NRF receives registration requests from the VFL server and client, and stores the NF configuration file.

[0128] In step S516, the NRF sends a registration response to the NWDAF, which acts as the VFL server.

[0129] In step S518, NRF sends a registration response to NWDAF, which is acting as a VFL client.

[0130] In step S520, NRF sends a registration response to NEF.

[0131] In step S522, the NRF sends a registration response to the trusted AF.

[0132] The VFL discovery process can be triggered by the VFL server or by the consumer. Once the VFL server is identified, it determines whether the ML model needs to adopt VFL technology based on information such as carrier policies, analytics task ID, VFL interoperability indicators, and service areas.

[0133] In step S524, the VFL server sends an NF discovery request to the NRF to discover other NWDAFs and / or AFs that are VFL clients from the NRF. The selection criteria may include NF type (NWDAF, AF, or NEF), analysis task ID, VFL client capability type, VFL interoperability indicator, etc.

[0134] In step S526, the NRF authenticates and authorizes the NF that initiated the request, ensuring that the NF has the right to obtain information about the target network function.

[0135] In step S528, the NRF returns an NF discovery response to the VFL server, which includes details of one or more eligible VFL clients (such as NWDAF or AF).

[0136] To support VFL training and inference, NWDAFs that support VFL must include a list of relevant VFL information when registering with the NRF. When the NRF attempts to discover an NWDAF that supports VFL training and inference, it must return the relevant VFL information list upon discovery. When an untrusted Application Function (AF) acts as a VFL server, it can only select an NWDAF as a VFL client. Figure 6 The process of NWDAF as a VFL client for VFL information registration and discovery is illustrated, including: Step S602: The candidate VFL client (taking NWDAF as an example) queries the NWDAFFunction configuration file, including parameters such as VFL capability information corresponding to each analysis task ID, VFL interoperability indicators, and optional supported feature identifiers.

[0137] In step S604, the candidate VFL client (taking NWDAF as an example) sends a VFL registration request to the NRF to register the NWDAFFunction configuration file, including parameters such as VFL capability information corresponding to each analysis task ID, VFL interoperability indicators, and optional supported feature identifiers.

[0138] In step S606, the untrusted AF, acting as a VFL server, needs to send a VFL registration request to NEF through OAM configuration, registering its analysis task ID, corresponding VFL capability information, VFL interoperability indicator, optional supported feature identifiers, and other parameters.

[0139] In step S608, the NEF then sends a VFL registration request for the untrusted AF to the NRF to update the NEF profile in the NRF.

[0140] Step S610: NRF stores the NF configuration files for the VFL server and VFL client.

[0141] In step S612, NRF sends a registration response to the VFL client.

[0142] In step S614, NRF sends a registration response to the VFL server.

[0143] In step S616, when VFL operation is required, the untrusted AF sends a VFL client discovery request to the NEF and provides the conditions for selecting a VFL client (NWDAF), including the analysis task ID, the corresponding VFL capability information, the VFL interoperability indicator, and optional supported feature identifiers.

[0144] In step S618, NEF checks whether AF has the right to request the VFL client for the analysis task ID according to the configured policy.

[0145] In step S620, NEF uses the conditions provided by AF to send a VFL client discovery request to NRF on behalf of AF to invoke the discovery service.

[0146] In step S622, NRF queries the local NWDAFFunction configuration file, which includes parameters such as VFL capability information corresponding to each analysis task ID, VFL interoperability indicators, and optional supported feature identifiers.

[0147] In step S624, the NRF returns an NF discovery response to the NEF, which includes details of one or more eligible VFL clients (NWDAF).

[0148] In step S626, NEF selects a suitable NWDAF as the VFL client based on information such as NF Profile, load, and capacity, and assigns it a temporary external NWDAF identifier.

[0149] NEF can also specify a particular NWDAF to act as an aggregation node based on the VFL client intermediate result aggregation capability (vflClientAggrCap).

[0150] The NF Profile includes parameters such as the analysis task ID, the corresponding VFL capability information, VFL interoperability indicators, and optional supported feature identifiers.

[0151] In step S628, NEF returns the discovery results to the untrusted AF. This only includes information such as the external NWDAF identifier and AF correspondence, VFL interoperability indicator, etc. If an aggregation node is specified, NEF will only return information about that aggregation node.

[0152] In step S630, the AF stores the received external NWDAF identifier, which is used to specify the target VFL client when interacting with NEF later.

[0153] Those skilled in the art will understand that various aspects of the present invention can be implemented as systems, methods, or program products. Therefore, various aspects of the present invention can be specifically implemented in the following forms: entirely in hardware, entirely in software (including firmware, microcode, etc.), or in a combination of hardware and software, collectively referred to herein as “circuit,” “module,” or “system.”

[0154] The following reference Figure 7 This describes a configuration device 700 for vertical federated learning information according to this embodiment of the invention. Figure 7 The configuration device 700 for vertical federated learning information shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0155] The configuration device 700 for vertical federated learning information is manifested as a hardware module. Components of the configuration device 700 may include, but are not limited to: a calling module 702, used to call the recognizable name (DN) of the machine learning analysis information MlAnalyticsInfo, query the NWDAFFunction configuration file, and obtain a list of vertical federated learning VFL information. The VFL information list includes a machine learning analysis task ID and corresponding VFL capability information, a VFL interoperability indicator, and a feature identifier. The machine learning analysis task ID represents the network data analysis and / or machine learning task to be vertically federated; the VFL capability information represents the NWDAF as a VFL server and / or VFL client; the VFL interoperability indicator is used to match network functions (NFs) with the same collaborative capabilities; and the feature identifier represents the data characteristics of the machine learning analysis task. A first sending module 704 is used to send a first VFL registration request to the network storage function (NRF), the first VFL registration request carrying the VFL information list. A first receiving module 706 is used to receive a first registration success response from the NRF.

[0156] The following reference Figure 8 This describes a configuration device 800 for vertical federated learning information according to this embodiment of the invention. Figure 8 The configuration device 800 for vertical federated learning information shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0157] The configuration device 800 for vertical federated learning information is manifested in the form of a hardware module. Components of the configuration device 800 for vertical federated learning information may include, but are not limited to: The second receiving module 802 is used to receive a first VFL registration request sent by NWDAF. The first VFL registration request carries a VFL information list, which includes a machine learning analysis task ID and corresponding VFL capability information, a VFL interoperability indicator, and a feature identifier. The machine learning analysis task ID represents the network data analysis and / or machine learning task to be performed for vertical federated learning. The VFL capability information represents NWDAF as a VFL server and / or VFL client. The VFL interoperability indicator is used to match network functions (NFs) with the same collaborative capabilities. The feature identifier represents the data characteristics of the machine learning analysis task. The registration module 804 is used to register NWDAF with VFL based on the VFL information list. The second sending module 806 is used to send a first registration success response to NWDAF.

[0158] The following reference Figure 9 This describes a configuration device 900 for vertical federated learning information according to this embodiment of the invention. Figure 9The configuration device 900 for vertical federated learning information shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0159] The configuration device 900 for vertical federated learning information is manifested in the form of a hardware module. Components of the configuration device 900 may include, but are not limited to: a third receiving module 902, used to receive a third VFL registration request sent by an untrusted AF; a third sending module 904, used to send a third VFL registration request to the NRF so that the NRF registers the untrusted AF, the NRF also storing a list of VFL information sent by the NWDAF, the VFL information list including a machine learning analysis task ID and corresponding VFL capability information, a VFL interoperability indicator, and a feature identifier, the machine learning analysis task ID representing the network data analysis and / or machine learning task to be performed in vertical federated learning, the VFL capability information representing the NWDAF as a VFL server and / or VFL client, the VFL interoperability indicator used to match network functions (NFs) with the same cooperative capabilities, and the feature identifier representing the data characteristics of the machine learning analysis task; and a fourth receiving module 906, used to receive a third registration success response from the NRF, wherein one of the untrusted AF and the NWDAF acts as a VFL server and the other as a VFL client.

[0160] The following reference Figure 10 This describes an electronic device 1000 according to this embodiment of the present invention. It may be a network device. Figure 10 The electronic device 1000 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0161] like Figure 10 As shown, the electronic device 1000 is manifested in the form of a general-purpose computing device. The components of the electronic device 1000 may include, but are not limited to: at least one processing unit 1010, at least one storage unit 1020, and a bus 1030 connecting different system components (including storage unit 1020 and processing unit 1010).

[0162] The storage unit stores program code that can be executed by the processing unit 1010, causing the processing unit 1010 to perform the steps described in the "Exemplary Methods" section of this specification according to various exemplary embodiments of the present invention. For example, the processing unit 1010 can perform, as follows: Figure 1 The described solution.

[0163] Storage unit 1020 may include readable media in the form of volatile storage units, such as random access memory (RAM) 10201 and / or cache memory 10202, and may further include read-only memory (ROM) 10203.

[0164] Storage unit 1020 may also include a program / utility 10204 having a set (at least one) program module 10205, such program module 10205 including but not limited to: operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0165] Bus 1030 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the multiple bus structures.

[0166] Electronic device 1000 can also communicate with one or more external devices 1070 (e.g., keyboard, pointing device, Bluetooth device, etc.), one or more devices that enable a user to interact with electronic device 1000, and / or any device that enables electronic device 1000 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 1050. Furthermore, electronic device 1000 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 1060. As shown, network adapter 1060 communicates with other modules of electronic device 1000 via bus 1030. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 1000, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0167] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, terminal device, or network device, etc.) to execute the methods according to the embodiments of this disclosure.

[0168] In exemplary embodiments of this disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the methods described above is stored. In some possible embodiments, various aspects of the invention may also be implemented as a program product comprising program code that, when the program product is run on an electronic device, causes the electronic device to perform the steps of the various exemplary embodiments of the invention described in the "Exemplary Methods" section above.

[0169] According to embodiments of the present invention, a program product for implementing the above-described method may employ a portable compact disc read-only memory (CD-ROM) and include program code, and may run on an electronic device, such as a personal computer. However, the program product of the present invention is not limited thereto. In this document, a readable storage medium may be any tangible medium that includes or stores a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0170] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0171] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying 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 readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting programs for use by or in conjunction with an instruction execution system, apparatus, or device.

[0172] The program code included on the readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0173] Program code for performing the operations of this invention can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0174] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0175] Furthermore, although the steps of the method in this disclosure are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or a step may be broken down into multiple steps.

[0176] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, mobile terminal, or network device, etc.) to execute the methods according to the embodiments of this disclosure.

[0177] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the appended claims.

Claims

1. A method for configuring information in vertical federated learning, characterized in that, The network data analysis function NWDAF includes: The machine learning analysis information MlAnalyticsInfo is invoked to retrieve the recognizable name DN, and the NWDAFFunction configuration file is queried to obtain the vertical federated learning VFL information list. The VFL information list includes the machine learning analysis task ID and the corresponding VFL capability information, VFL interoperability indicator, and feature identifier. The machine learning analysis task ID represents the network data analysis and / or machine learning task to be performed in vertical federated learning. The VFL capability information represents the NWDAF as a VFL server and / or VFL client. The VFL interoperability indicator is used to match network functions NF with the same collaborative capabilities. The feature identifier represents the data characteristics of the machine learning analysis task. Send a first VFL registration request to the Network Storage Function (NRF), the first VFL registration request carrying the VFL information list; Receive the first registration success response from the NRF feedback.

2. The method for configuring vertical federated learning information according to claim 1, characterized in that, The system retrieves the identifiable name (DN) of the machine learning analytics information MlAnalyticsInfo, queries the NWDAFFunction configuration file, and obtains a list of vertical federated learning (VFL) information, including: Based on the identifiable name DN of the MlAnalyticsInfo, locate the corresponding configuration node in the NWDAFFunction configuration file; Read the attribute parameters of all VFL-related information object class (IOC) instances mounted under the configuration node. The attribute parameters include the VFL capability information, the VFL interoperability indicator, and the feature identifier. Using the machine learning analysis task ID as a group index, the attribute parameters belonging to the same group are read to obtain the VFL information list.

3. The method for configuring vertical federated learning information according to claim 2, characterized in that, When the NWDAF acts as the VFL server, the feature identifier includes at least one of the following: network traffic to be analyzed, user behavior, network slicing, service latency, and terminal access characteristics. When the NWDAF acts as the VFL client, the feature identifier includes at least one of the following: locally stored network traffic, user behavior, network slicing, service latency, and terminal access features.

4. The method for configuring vertical federated learning information according to claim 2, characterized in that, When the NWDAF acts as the VFL client, the attribute parameters read also include a client aggregation capability identifier, which indicates whether the VFL client has the ability to aggregate intermediate computation results from other VFL clients.

5. The method for configuring vertical federated learning information according to claim 1, characterized in that, When the NWDAF acts as the VFL server, it also includes: Send a Network Function (NF) Discovery Request to the NRF, the NF Discovery Request being used to request the NRF to filter VFL clients that meet the conditions for vertical federated learning collaboration; The list of clients receiving the NRF feedback includes at least one of NWDAF clients, Trusted Application Function (AF) clients, and Untrusted AF clients.

6. The method for configuring vertical federated learning information according to claim 5, characterized in that, Sending a Network Function (NF) Discovery Request to the NRF, including: The NWDAF triggers or receives a VFL client discovery command sent by the consumer, and performs local access verification based on the operator policy, the target analytics task ID, the VFL interoperability indicator, and the service area to determine whether to enable VFL for the target analytics task. If it is determined that the VFL is enabled, the network element type, the target analysis task ID, the VFL client capability, and the VFL interoperability indicator are used as filtering conditions, and the NF discovery request is generated based on the filtering conditions. Send the NF discovery request to the NRF.

7. A method for configuring information in vertical federated learning, characterized in that, Applied to NRF, including: The system receives a first VFL registration request sent by the NWDAF. The first VFL registration request carries a VFL information list, which includes a machine learning analysis task ID and corresponding VFL capability information, a VFL interoperability indicator, and a feature identifier. The machine learning analysis task ID represents the network data analysis and / or machine learning task to be performed for vertical federated learning. The VFL capability information represents the NWDAF as a VFL server and / or VFL client. The VFL interoperability indicator is used to match network functions (NFs) with the same collaborative capabilities. The feature identifier represents the data characteristics of the machine learning analysis task. Based on the VFL information list, the NWDAF is registered and filed as a VFL. Send a first registration success response to the NWDAF.

8. The method for configuring vertical federated learning information according to claim 7, characterized in that, Also includes: Receive a second VFL registration request sent by a trusted AF and / or a third VFL registration request sent by a Network Open Function (NEF), wherein the third VFL registration request is sent to the NEF by an untrusted AF; Based on the second VFL registration request, the trusted AF is registered and filed using VFL, and / or based on the third VFL registration request, the untrusted AF is registered and filed using VFL; Send a second registration success response to the trusted AF and / or send a third registration success response to the NEF.

9. The method for configuring vertical federated learning information according to claim 7, characterized in that, Also includes: Receive a Network Function (NF) Discovery Request sent by a VFL server. The NF Discovery Request is used to request NRF to filter VFL clients that meet the conditions for vertical federated learning collaboration. The VFL server is any one of NWDAF server, Trusted AF, or NEF. Perform authentication on the NF discovery request; If authentication is successful, the registered NF files are filtered to obtain a list of clients that meet the conditions for vertical federated learning collaboration. The VFL server is fed back the client list, which includes at least one of NWDAF clients, trusted AFs, and untrusted AFs.

10. The method for configuring vertical federated learning information according to claim 9, characterized in that, The authentication operation for the NF discovery request includes: The network element identity and network access qualifications of the VFL server are verified; and Perform discovery service authorization verification on the VFL server.

11. The method for configuring vertical federated learning information according to claim 9, characterized in that, The registered VFL clients are filtered to obtain a list of clients that meet the aforementioned vertical federated learning collaboration conditions, including: Based on the network element type, target analysis task ID, VFL client capabilities, and VFL interoperability indicator carried in the NF discovery request, all the registered NF files are matched to filter out clients that meet the vertical federated learning collaboration conditions, thus obtaining the client list.

12. A method for configuring information in vertical federated learning, characterized in that, Applied to NEF, including: Receive a third VFL registration request sent by an untrusted AF; The third VFL registration request is sent to the NRF so that the NRF can register the untrusted AF. The NRF also stores a list of VFL information sent by the NWDAF. The VFL information list includes a machine learning analysis task ID and the corresponding VFL capability information, a VFL interoperability indicator, and a feature identifier. The machine learning analysis task ID represents the network data analysis and / or machine learning task to be performed for vertical federated learning. The VFL capability information represents the NWDAF as a VFL server and / or VFL client. The VFL interoperability indicator is used to match network functions (NFs) with the same cooperative capabilities. The feature identifier represents the data characteristics of the machine learning analysis task. Receive the third registration success response from the NRF feedback, wherein one of the untrusted AF and the NWDAF acts as a VFL server and the other acts as a VFL client.

13. The method for configuring vertical federated learning information according to claim 12, characterized in that, Also includes: Receive the NF discovery request sent by the untrusted AF, the NF discovery request being used to request the NRF to filter VFL clients that meet the vertical federated learning collaboration conditions; The NF discovery request is authenticated based on the operator's policy to verify whether the untrusted AF has permission to access the VFL client corresponding to the machine learning analysis task ID. The NF discovery request is authenticated and successfully sent to the NRF; A list of clients receiving the NRF feedback, the list of clients including the NWDAF that satisfies the vertical federated learning collaboration conditions.

14. The method for configuring vertical federated learning information according to claim 13, characterized in that, The NF discovery request includes the target analysis task ID, the requirement for VFL capability information, the VFL interoperability indicator, and optional supported feature identifiers.

15. The method for configuring vertical federated learning information according to claim 13, characterized in that, Also includes: Extract the VFL information and service parameter information for each NWDAF from the client list; Based on the VFL information, the NWDAF with the same target analysis task ID, the same VFL interoperability indicator, and the feature identifier that matches the requirements is identified as a candidate NWDAF; Based on at least one of the load status, service capacity, and network reachability in the service parameter information, the candidate NWDAFs are filtered to obtain the target NWDAF, and a temporary external identifier is assigned to the target NWDAF. The client discovery response carrying the temporary external identifier is sent to the untrusted AF.

16. The method for configuring vertical federated learning information according to claim 15, characterized in that, The process of filtering the candidate NWDAFs based on at least one of the load status, service capacity, and network reachability in the service parameter information to obtain the target NWDAF further includes: If the VFL information of multiple target NWDAFs includes a VFL client aggregation capability identifier, then one of the multiple target NWDAFs is designated as the aggregation node and fed back to the untrusted AF.

17. A configuration device for vertical federated learning information, characterized in that, The network data analysis function NWDAF includes: The calling module is used to call the recognizable name (DN) of the machine learning analysis information MlAnalyticsInfo, query the NWDAFFunction configuration file, and obtain the vertical federated learning VFL information list. The VFL information list includes the machine learning analysis task ID and the corresponding VFL capability information, VFL interoperability indicator, and feature identifier. The machine learning analysis task ID represents the network data analysis and / or machine learning task to be performed in vertical federated learning. The VFL capability information represents the NWDAF as a VFL server and / or VFL client. The VFL interoperability indicator is used to match network functions (NFs) with the same collaborative capabilities. The feature identifier represents the data characteristics of the machine learning analysis task. The first sending module is used to send a first VFL registration request to the Network Storage Function (NRF), wherein the first VFL registration request carries the VFL information list; The first receiving module is used to receive the first registration success response from the NRF.

18. A configuration device for vertical federated learning information, characterized in that, Applied to NRF, including: The second receiving module is used to receive a first VFL registration request sent by the NWDAF. The first VFL registration request carries a VFL information list. The VFL information list includes a machine learning analysis task ID and the corresponding VFL capability information, a VFL interoperability indicator, and a feature identifier. The machine learning analysis task ID represents the network data analysis and / or machine learning task to be performed for vertical federated learning. The VFL capability information represents the NWDAF as a VFL server and / or a VFL client. The VFL interoperability indicator is used to match network functions (NFs) with the same cooperative capabilities. The feature identifier represents the data characteristics of the machine learning analysis task. The registration module is used to register and file the NWDAF as a VFL based on the VFL information list; The second sending module is used to send a first registration success response to the NWDAF.

19. A configuration device for vertical federated learning information, characterized in that, Applied to NEF, including: The third receiving module is used to receive the third VFL registration request sent by the untrusted AF; The third sending module is used to send the third VFL registration request to the NRF so that the NRF can register the untrusted AF. The NRF also stores a list of VFL information sent by the NWDAF. The VFL information list includes a machine learning analysis task ID and the corresponding VFL capability information, a VFL interoperability indicator, and a feature identifier. The machine learning analysis task ID represents the network data analysis and / or machine learning task to be performed for vertical federated learning. The VFL capability information represents the NWDAF as a VFL server and / or VFL client. The VFL interoperability indicator is used to match network functions (NFs) with the same cooperative capabilities. The feature identifier represents the data characteristics of the machine learning analysis task. The fourth receiving module is used to receive the third registration success response from the NRF, wherein one of the untrusted AF and the NWDAF acts as a VFL server and the other acts as a VFL client.

20. A network device, characterized in that, include: processor; as well as Memory for storing the executable instructions of the processor; The processor is configured to execute the configuration method of vertical federated learning information according to any one of claims 1 to 8 by executing the executable instructions.

21. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the configuration method of vertical federated learning information as described in any one of claims 1 to 6, or the configuration method of vertical federated learning information as described in any one of claims 7 to 11, or the configuration method of vertical federated learning information as described in any one of claims 12 to 16.

22. A computer program product having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the configuration method of vertical federated learning information as described in any one of claims 1 to 6, or the configuration method of vertical federated learning information as described in any one of claims 7 to 11, or the configuration method of vertical federated learning information as described in any one of claims 12 to 16.

Images