Model inference method, and client, server, AF and anlf

Abstract

The present disclosure relates to the technical field of communications, and relates to a model inference method, and a client, a server, an AF and an AnLF. The model inference method in the present disclosure is executed by the server, and comprises: sending first request information to a client, wherein the first request information is used for requesting vertical federated learning (VFL) inference, and the first request information comprises at least one of a VFL association identifier, first sample indication information and first feature indication information; receiving first response information sent by the client, wherein the first response information comprises at least one of an intermediate result of the VFL inference, the VFL association identifier, second sample indication information and second feature indication information; and on the basis of the first response information and / or an intermediate result of local VFL inference, generating an inference result.

Description

Inference method of model, client, server, AF and AnLF

[0001] Cross-reference to Related Applications

[0002] This application is based on the application with CN application number 202411096481.0 and application date August 9, 2024, and claims priority thereto, the disclosure of which is hereby incorporated by reference in its entirety. TECHNICAL FIELD

[0003] The present disclosure relates to the field of communication technology, in particular to an inference method of model, client, server, AF and AnLF. BACKGROUND

[0004] The concept of federated learning has been introduced in 3GPP ((3rd Generation Partnership Project, 3rd Generation Partnership Project) R18 (Release 18, Release 18) eNA_ph3 (Study on Enablers for Network Automation for 5G-phase 3, 5G-phase 3 Network Automation Enabler Research) project, but mainly focuses on horizontal federated learning. In horizontal federated learning (Horizontal Federated Learning, HFL), the input data used for training has the same sample feature and different data samples. The 5GC (5G Core, 5G Core) allows the model to be passed between the NWDAF (Network Data Analytics Function, Network Data Analytics Function), and the NWDAF does not need to exchange local data, thereby ensuring data privacy. Moreover, in horizontal federated learning, the transmission of the model and the interaction of network elements only exist between NWDAFs, and do not involve other network entities.

[0005] However, vertical federated learning (Vertical Federated Learning, VFL) is different from horizontal federated learning. Vertical federated learning has the same data sample and different sample features. For example, in the 5GC, the data in the AF (Application Function) and the NWDAF for the QoE (Quality of Experience) information of the user has different data features and the same data sample. Vertical federated learning in the 5GC can effectively break the data isolation and obtain a better model by exchanging intermediate data such as gradients.

[0006] On the other hand, the AI function of the current 5GC is divided into two processes of training and inference, and the inference process occurs after model training. Compared with the traditional inference process, the inference process of vertical federated learning also needs the joint participation of all parties of vertical federated learning participants. Therefore, in the 5GC, the method supporting inference in VFL needs to be studied, and this problem has been further studied as one of the main problems. SUMMARY

[0007] According to some embodiments of the present disclosure, a method for inference of a model is provided, executed by a server, comprising: sending first request information to a client, wherein the first request information is used to request vertical federated learning (VFL) inference, and the first request information comprises at least one of a VFL association identifier, first sample indication information, and first feature indication information; receiving first response information sent by the client, wherein the first response information comprises at least one of an intermediate result of VFL inference, the VFL association identifier, second sample indication information, and second feature indication information; and generating an inference result according to the first response information and / or an intermediate result of local VFL inference.

[0008] In some embodiments, the VFL association identifier is used to associate the process of VFL and / or a VFL model; the first sample indication information is used to indicate a data sample; the first feature indication information is used to indicate a data feature; the second sample indication information is used to indicate a data sample; and / or the second feature indication information is used to indicate a data feature.

[0009] In some embodiments, the method further comprises performing VFL inference according to data in the server to obtain an intermediate result of local VFL inference.

[0010] In some embodiments, the method further comprises receiving second request information sent by a consuming network function (NF), wherein the second request information is used to request an analysis result, and the second request information comprises at least one of an analysis identifier and a VFL indicator, the analysis identifier is used to identify a requested analysis and / or an analysis result generated using a VFL manner, and the VFL indicator is used to indicate that a VFL trained model is used; and wherein sending the first request information to the client comprises: sending the first request information to the client according to the second request information.

[0011] In some embodiments, the second request information further comprises filter information, wherein the filter information comprises at least one of single network slice selection assistance information, an application identifier, a region of interest, and a data network access identifier.

[0012] In some embodiments, the first request information further comprises at least one of an interoperation indicator, an analysis identifier, filter information, and a VFL indicator.

[0013] In some embodiments, the method further includes: sending, to the consumer NF, second response information, wherein the second response information includes at least one of an analysis result and second validity time information, and the second validity time information is used to indicate a validity time of the analysis result.

[0014] In some embodiments, the first response information further includes: first validity time information used to indicate a validity time of the intermediate result of the VFL inference.

[0015] In some embodiments, the first sample indication information includes at least one of a user equipment identity, a generic public user identifier GPSI, a subscription permanent identifier SUPI, location information, a sample identity, and a group identity; and / or the second sample indication information includes at least one of a user equipment identity, a generic public user identifier GPSI, a subscription permanent identifier SUPI, location information, a sample identity, and a group identity.

[0016] In some embodiments, the first feature indication information is used to indicate a server-side used data feature and / or a client-side used data feature; the second feature indication information is used to indicate a server-side used data feature and / or a client-side used data feature; the first feature indication information includes at least one of service experience information, quality of experience indicator information, voice quality information, and a feature identity; and / or the second feature indication information includes at least one of service experience information, quality of experience indicator information, voice quality information, and a feature identity.

[0017] In some embodiments, the method further includes: collecting data from the corresponding network element according to the first request information.

[0018] In some embodiments, the analysis result is in the form of an analysis ID output, and includes at least one of each item of data of the analysis ID output, or includes at least one of quality of experience indicator information, latency information, and 5G quality of service indicators.

[0019] In some embodiments, the consumer NF is an analytics logic function AnLF, and the server is an application function AF.

[0020] In some embodiments, the server is a network data analytics function NWDAF, and the client is an application function AF, or the server and the client are different NWDAFs, or the server is an AF and the client is an NWDAF.

[0021] In some embodiments, in the case that the AF is a non-trusted AF, sending the first request information to the client comprises: sending the first request information to the client through a network exposure function (NEF), wherein, in the case that the server is a non-trusted AF, the VFL association identifier in the first request information is an external VFL association identifier, the first sample indication information comprises first external sample indication information, and / or the first feature indication information comprises first external feature indication information, the NEF maps the external VFL association identifier in the first request information to an internal VFL association identifier, maps the first external sample indication information to first internal sample indication information, and / or maps the first external feature indication information to first internal feature indication information, in the case that the client is a non-trusted AF, the VFL association identifier in the first request information is an internal VFL association identifier, the first sample indication information comprises first internal sample indication information, and / or the first feature indication information comprises first internal feature indication information, the NEF maps the internal VFL association identifier in the first request information to an external VFL association identifier, maps the first internal sample indication information to first external sample indication information, and / or maps the first internal feature indication information to first external feature indication information.

[0022] In some embodiments, in the case that the AF is a non-trusted AF, receiving the first response information sent by the client comprises: receiving the first response information sent by the client forwarded through a network exposure function (NEF), wherein, in the case that the client is a non-trusted AF, the intermediate result of the VFL inference in the first response information is an external intermediate result, the VFL association identifier is an external VFL association identifier, the second sample indication information comprises second external sample indication information, and / or the second feature indication information comprises second external feature indication information, the NEF maps the external intermediate result to an internal intermediate result, maps the external VFL association identifier in the first request information to an internal VFL association identifier, maps the second external sample indication information to second internal sample indication information, and / or maps the second external feature indication information to second internal feature indication information, in the case that the server is a non-trusted AF, the intermediate result of the VFL inference in the first response information is an internal intermediate result, the VFL association identifier is an internal VFL association identifier, the second sample indication information comprises second internal sample indication information, and the second feature indication information comprises second internal feature indication information, the NEF maps the internal intermediate result in the first request information to an external intermediate result, maps the internal VFL association identifier to an external VFL association identifier, maps the second internal sample indication information to second external sample indication information, and / or maps the second internal feature indication information to second external feature indication information.

[0023] In some embodiments, the method further comprises, in a case that the service end is the NWDAF and the VFL model is not stored in the NWDAF, sending first model request information to an analysis data storage function (ADRF), wherein the first model request information comprises at least one of a VFL association identifier, first sample indication information, and first feature indication information; and receiving a first model request response sent by the ADRF, wherein the ADRF retrieves the VFL model according to the first model request information, and the first model request response comprises at least one of a file of the VFL model, a storage address, and validity time information of the VFL model.

[0024] According to some embodiments of the present disclosure, a model inference method is provided, which is performed by a client and comprises: receiving first request information sent by a service end, wherein the first request information is used to request vertical federated learning (VFL) inference, and the first request information comprises at least one of a VFL association identifier, first sample indication information, and first feature indication information; and sending first response information to the service end, wherein the first response information comprises at least one of an intermediate result of the VFL inference, the VFL association identifier, second sample indication information, and second feature indication information, so that the service end generates an inference result according to the first response information and / or an intermediate result of local VFL inference.

[0025] In some embodiments, the VFL association identifier is used to associate a process of the VFL and / or a VFL model; the first sample indication information is used to indicate a data sample; the first feature indication information is used to indicate a data feature; the second sample indication information is used to indicate a data sample; and / or the second feature indication information is used to indicate a data feature.

[0026] In some embodiments, the method further comprises: performing VFL inference according to data in the client to obtain an intermediate result of the VFL inference.

[0027] In some embodiments, the first request information further comprises at least one of an interoperability indicator, an analysis identifier, filter information, and a VFL indicator, wherein the analysis identifier is used to identify an analysis requested and / or an analysis result generated in a VFL manner, the filter information comprises at least one of single network slice selection assistance information, an application identifier, a region of interest, and data network access identification, and the VFL indicator is used to indicate a model trained using VFL.

[0028] In some embodiments, the first response information further comprises first validity time information used to indicate a validity time of the intermediate result of the VFL inference.

[0029] In some embodiments, the first sample indication information comprises at least one of a user equipment identity, a generic public user identifier GPSI, a subscription permanent identifier SUPI, location information, a sample identity, a group identity; and / or the second sample indication information comprises at least one of a user equipment identity, a generic public user identifier GPSI, a subscription permanent identifier SUPI, location information, a sample identity, a group identity.

[0030] In some embodiments, the first feature indication information is used to indicate a data feature used by the server and / or a data feature used by the client; the second feature indication information is used to indicate a data feature used by the server and / or a data feature used by the client; the first feature indication information comprises at least one of service experience information, quality of experience indicator information, voice quality information, a feature identity; and / or the second feature indication information comprises at least one of service experience information, quality of experience indicator information, voice quality information, a feature identity.

[0031] In some embodiments, the method further comprises collecting the data samples from the corresponding network element according to the first request information.

[0032] In some embodiments, the server is a network data analytics function NWDAF, and the client is an application function AF, or the server and the client are different NWDAFs, or the server is an AF, and the client is a NWDAF.

[0033] In some embodiments, in a case that the AF is a non-trusted AF, the receiving the first request information sent by the server comprises: receiving the first request information sent by the server forwarded by a network exposure function NEF, wherein, in a case that the server is a non-trusted AF, the VFL association identity in the first request information is an external VFL association identity, the first sample indication information comprises first external sample indication information, and / or the first feature indication information comprises first external feature indication information, the NEF maps the external VFL association identity in the first request information to an internal VFL association identity, maps the first external sample indication information to first internal sample indication information, and / or maps the first external feature indication information to first internal feature indication information, in a case that the client is a non-trusted AF, the VFL association identity in the first request information is an internal VFL association identity, the first sample indication information comprises first internal sample indication information, and / or the first feature indication information comprises first internal feature indication information, the NEF maps the internal VFL association identity in the first request information to an external VFL association identity, maps the first internal sample indication information to first external sample indication information, and / or maps the first internal feature indication information to first external feature indication information.

[0034] In some embodiments, in the case that the AF is a non-trusted AF, sending the first response information to the service end includes: sending the first response information to the service end through a network exposure function (NEF), wherein, in the case that the client is a non-trusted AF, the intermediate result of the VFL inference in the first response information is an external intermediate result, the VFL association identifier is an external VFL association identifier, the second sample indication information includes second external sample indication information, and / or the second feature indication information includes second external feature indication information, the NEF maps the external intermediate result to an internal intermediate result, maps the external VFL association identifier in the first request information to an internal VFL association identifier, maps the second external sample indication information to second internal sample indication information, and / or maps the second external feature indication information to second internal feature indication information, in the case that the service end is a non-trusted AF, the intermediate result of the VFL inference in the first response information is an internal intermediate result, the VFL association identifier is an internal VFL association identifier, the second sample indication information includes second internal sample indication information, and the second feature indication information includes second internal feature indication information, the NEF maps the internal intermediate result in the first request information to an external intermediate result, maps the internal VFL association identifier to an external VFL association identifier, maps the second internal sample indication information to second external sample indication information, and / or maps the second internal feature indication information to second external feature indication information.

[0035] In some embodiments, the method further includes: in the case that the client is a NWDAF and the VFL model is not stored in the NWDAF, sending second model request information to an analysis data storage function (ADRF), wherein the second model request information includes at least one of the VFL association identifier, the second sample indication information, and the second feature indication information; and receiving a second model request response sent by the ADRF, wherein the ADRF retrieves the VFL model according to the second model request information, and the second model request response includes at least one of a file of the VFL model, a storage address, and validity time information of the VFL model.

[0036] According to yet some embodiments of the present disclosure, a model inference method is provided, performed by an application function (AF), and includes: receiving third request information sent by an analysis logic function (AnLF), wherein the third request information is used to request an analysis result, and the third request information includes at least one of an analysis identifier and a vertical federated learning (VFL) indicator; performing VFL inference according to the third request information; and sending third response information to the AnLF, wherein the third response information includes at least one of the analysis result and third validity time information.

[0037] In some embodiments, the analysis identifier is used to identify the requested analysis and / or the analysis result is generated using a VFL manner; the VFL indicator is used to indicate a model trained using VFL; and / or the third validity time information is used to indicate a validity time of the analysis result.

[0038] In some embodiments, the third request information further comprises filter information, wherein the filter information comprises at least one of a single network slice selection assistance information, an application identifier, a region of interest, and a data network access identifier.

[0039] In some embodiments, the analysis result is in a form of an analysis ID output, and comprises at least one of each data of the analysis ID output, or at least one of quality of experience indicator information, latency information, and 5G quality of service indicators.

[0040] In some embodiments, in a case that the AF is a non-trusted AF, the receiving the third request information sent by the analysis logic function AnLF comprises: receiving the third request information sent by the AnLF through a network exposure function NEF; and the sending the third response information to the AnLF comprises: sending the third response information to the AnLF through the NEF.

[0041] In some embodiments, in a case that the third request information comprises internal correlation information and / or internal location information of the user equipment, the NEF maps the internal correlation information to external correlation information and / or maps the internal location information to external location information.

[0042] According to still other embodiments of the present disclosure, there is provided a model inference method performed by an analysis logic function AnLF, comprising: sending third request information to an application function AF, wherein the third request information is used to request an analysis result, and the third request information comprises at least one of an analysis identifier and a vertical federated learning VFL indicator; and receiving third response information sent by the AF, wherein the third response information comprises at least one of the analysis result and third validity time information.

[0043] In some embodiments, the analysis identifier is used to identify the requested analysis and / or the analysis result is generated using a VFL manner; the VFL indicator is used to indicate a model trained using VFL; and / or the third validity time information is used to indicate a validity time of the analysis result.

[0044] In some embodiments, in the case that the AF is a non-trusted AF, sending the third request information to the application function AF comprises: sending the third request information to the AF through a network exposure function NEF, wherein, in the case that the third request information comprises internal related information and / or internal location information of the user equipment, the NEF maps the internal related information to external related information and maps the internal location information to external location information; and receiving the third response information sent by the AF comprises: receiving the third response information sent by the AF through the NEF.

[0045] According to yet some embodiments of the present disclosure, a server end is provided, comprising: a sending module configured to send first request information to a client end, wherein the first request information is used to request vertical federated learning VFL inference, and the first request information comprises at least one of a VFL association identifier, first sample indication information and first feature indication information; a receiving module configured to receive first response information sent by the client end, wherein the first response information comprises at least one of an intermediate result of the VFL inference, the VFL association identifier, second sample indication information and second feature indication information; and a generating module configured to generate an inference result according to the first response information and / or an intermediate result of local VFL inference.

[0046] According to still some embodiments of the present disclosure, a client end is provided, comprising: a receiving module configured to receive first request information sent by a server end, wherein the first request information is used to request vertical federated learning VFL inference, and the first request information comprises at least one of a VFL association identifier, first sample indication information and first feature indication information; and a sending module configured to send first response information to the server end, wherein the first response information comprises at least one of an intermediate result of the VFL inference, the VFL association identifier, second sample indication information and second feature indication information, so that the server end generates an inference result according to the first response information and / or an intermediate result of local VFL inference.

[0047] According to yet some embodiments of the present disclosure, an application function is provided, comprising: a receiving module configured to receive third request information sent by an analysis logic function AnLF, wherein the third request information is used to request an analysis result, and the third request information comprises at least one of an analysis identifier and a vertical federated learning VFL indicator; an inference module configured to perform VFL inference according to the third request information; and a sending module configured to send third response information to the AnLF, wherein the third response information comprises at least one of the analysis result and third valid time information.

[0048] According to still another embodiment of the present disclosure, an analysis logic function AnLF is provided, comprising: a sending module configured to send third request information to an application function AF, wherein the third request information is used to request an analysis result, and the third request information comprises at least one of an analysis identifier and a vertical federated learning VFL indicator; and a receiving module configured to receive third response information sent by the AF, wherein the third response information comprises at least one of the analysis result and fourth valid time information.

[0049] According to still another embodiment of the present disclosure, an electronic device is provided, comprising: a processor; and a memory coupled to the processor, configured to store instructions, which, when executed by the processor, cause the processor to perform the inference method of the model of any embodiment of the present disclosure.

[0050] According to still another embodiment of the present disclosure, a computer readable storage medium is provided, having stored thereon a computer program, wherein the program, when executed by a processor, implements the inference method of the model of any embodiment of the present disclosure.

[0051] According to still another embodiment of the present disclosure, a computer program product is provided, comprising instructions which, when executed by a processor, cause the processor to perform the inference method of the model of any embodiment of the present disclosure.

[0052] Other features of the present disclosure, and the advantages thereof over the existing techniques will become more apparent from the following detailed description of exemplary embodiments of the present disclosure with reference to the attached drawings. BRIEF DESCRIPTION OF DRAWINGS

[0053] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present disclosure, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0054] FIG. 1 shows a flowchart of the inference method of the model of some embodiments of the present disclosure.

[0055] FIG. 2 shows a flowchart of the inference method of the model of another embodiment of the present disclosure.

[0056] FIG. 3 shows a flowchart of the inference method of the model of still another embodiment of the present disclosure.

[0057] FIG. 4 shows a flowchart of the inference method of the model of still another embodiment of the present disclosure.

[0058] FIG. 5 shows a flowchart of the inference method of the model of still another embodiment of the present disclosure.

[0059] FIG. 6 shows a flowchart of an inference method of a model according to some embodiments of the present disclosure.

[0060] FIG. 7 shows a structural diagram of a server according to some embodiments of the present disclosure.

[0061] FIG. 8 shows a structural diagram of a client according to some embodiments of the present disclosure.

[0062] FIG. 9 shows a structural diagram of an AF according to some embodiments of the present disclosure.

[0063] FIG. 10 shows a structural diagram of an AnLF according to some embodiments of the present disclosure.

[0064] FIG. 11 shows a structural diagram of an electronic device according to some embodiments of the present disclosure.

[0065] FIG. 12 shows a structural diagram of an electronic device according to some other embodiments of the present disclosure.

[0066] FIG. 13 shows a structural diagram of an inference system of a model according to some embodiments of the present disclosure. DETAILED DESCRIPTION

[0067] The technical solutions in the embodiments of the present disclosure will be described clearly and completely below with reference to the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all the embodiments of the present disclosure. The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way limiting on the present disclosure and its applications or uses. Based on the embodiments in the present disclosure, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present disclosure.

[0068] The present disclosure proposes a signaling interaction process for an inference process of vertical federated learning, which can realize cooperation between a client and a server to complete the inference process of vertical federated learning.

[0069] The present disclosure proposes an inference method of a model, which will be described below with reference to FIGS. 1-6.

[0070] FIG. 1 is a flowchart of an inference method of a model according to some embodiments of the present disclosure. As shown in FIG. 1, the method of this embodiment includes steps S102-S106.

[0071] In step S102, the server sends first request information to the client, and correspondingly, the client receives the first request information sent by the server.

[0072] In some embodiments, the first request information is used to request VFL inference, and the first request information includes at least one of a VFL association identifier, first sample indication information, and first feature indication information.

[0073] The VFL correlation ID is used to associate the process of VFL and / or the VFL model, so that the server and the client know which VFL task and / or VFL model. The VFL correlation ID in the inference process can also be associated with the VFL correlation ID in the training process.

[0074] In some embodiments, the first sample indication information is used to indicate the data sample. For example, the first sample indication information is denoted as Sample, which is not limited to the example, and is used to indicate the data sample used for VFL inference in the server and / or the client. In some embodiments, the first sample indication information includes at least one of a user equipment identifier (UE ID), a generic public subscription identifier (GPSI), a subscription permanent identifier (SUPI), location information, a sample identifier (Sample ID), and a group identifier (Group ID).

[0075] The server and the client can use the same data sample, for example, the data corresponding to a certain UE ID is used as the data for VFL inference. The sample identifier can be the sample identifier determined after the sample alignment in the previous process.

[0076] In some embodiments, the first feature indication information is used to indicate the data feature. For example, the first feature indication information is denoted as Feature, which is not limited to the example, and is used to indicate the data feature used by the server and / or the data feature used by the client. In some embodiments, the first feature indication information includes at least one of service experience information, quality of experience metrics (QoE metrics), voice quality information (for example, MOS (Mean Opinion Score)), and a feature identifier (Feature ID). The client can determine the data feature to be used according to the first feature indication information.

[0077] In some embodiments, the first request information further includes at least one of an interoperation indicator, an analysis identifier, filter information, and a VFL indicator.

[0078] In some embodiments, an analytics ID is used to identify the requested analytics and / or the analytics result is generated using VFL. The analytics ID can indicate the analytics result is generated using VFL while indicating the requested analytics. For example, the analytics ID can indicate the requested analytics service experience.

[0079] In some embodiments, the filter information includes at least one of a single network slice selection assistance information (S-NSSAI), an application ID, an area of interest (AOI), and a data network access identifier (DNAI).

[0080] In some embodiments, the VFL indicator is used to indicate a VFL model to be used.

[0081] In step S104, the client sends first response information to the server, and correspondingly, the server receives the first response information sent by the client.

[0082] In some embodiments, the first response information includes at least one of a VFL inference intermediate result, a VFL association ID, second sample indication information, and second feature indication information.

[0083] The client can determine the VFL-related model, data sample, data feature, and the like according to the first request information, and then perform VFL inference to obtain an intermediate result, which is sent to the server through the first response information. If the client needs to obtain data from other network elements for VFL inference, the client can collect data from the corresponding network elements according to the first request information for VFL inference.

[0084] The server can also perform VFL inference. In some embodiments, the server performs VFL inference according to data in the server to obtain a local VFL inference intermediate result. If the server needs to obtain data from other network elements for VFL inference, the server can collect data from the corresponding network elements for VFL inference.

[0085] The second sample indication information can be the same as or different from the first sample indication information, and the second feature indication information can be the same as or different from the first feature indication information.

[0086] In some embodiments, the second sample indication information is used to indicate the data sample. For example, the second sample indication information is denoted as Sample, which is not limited to the example, and is used to indicate the data sample used by the server and / or the client for the VFL inference. In some embodiments, the second sample indication information includes at least one of a user equipment identifier (UE ID), a generic public subscription identifier (GPSI), a subscription permanent identifier (SUPI), location information, a sample identifier (Sample ID), and a group identifier (Group ID).

[0087] In some embodiments, the second feature indication information is used to indicate the data feature. For example, the second feature indication information is denoted as Feature, which is not limited to the example, and is used to indicate the data feature used by the server and / or the data feature used by the client. In some embodiments, the second feature indication information includes at least one of service experience information, quality of experience metrics (QoE metrics), voice quality information (e.g., a mean opinion score (MOS)), and a feature identifier (Feature ID).

[0088] If the data indicated by the first sample indication information and the first feature indication information does not exist in the client, the second sample indication information different from the first sample indication information and / or the second feature indication information different from the first feature indication information can be returned to the server, to indicate the available data sample and / or data feature. The server can determine the data sample and / or data feature used by the VFL inference according to the second sample indication information and / or the second feature indication information, and then instruct the client to perform the VFL inference.

[0089] In step S106, the server generates the inference result according to the first response information and / or the intermediate result of the local VFL inference.

[0090] The server can generate the final inference result (or analysis result) according to the intermediate result of the VFL inference of the client, and also in combination with the intermediate result of the local VFL inference. The intermediate result can be an intermediate value output by the VFL model, which is not limited to the example.

[0091] In some embodiments, the first response information further includes first validity time information, which is used to indicate the validity time of the intermediate result of the VFL inference of the client.

[0092] In some embodiments, the server is a NWDAF, the client is an AF, or the server and the client are different NWDAFs, or the server is an AF, and the client is a NWDAF.

[0093] In the method of the above embodiments, the server sends first request information to the client, for requesting to perform VFL inference, and the first request information includes at least one of a VFL association identifier, first sample indication information, and first feature indication information, the first sample indication information and the first feature indication information can be used to indicate data samples and data features that need to be used respectively, the client sends first response information to the server after performing VFL inference, the first response information includes at least one of an intermediate result of VFL inference, the VFL association identifier, second sample indication information, and second feature indication information, the second sample indication information and the second feature indication information can be used to indicate data samples and data features that are used respectively, and then the server generates an inference result according to the first response information and / or an intermediate result of local VFL inference. By initiating VFL inference by the server, the signaling interaction between the server and the client can realize accurate transmission of information required in the VFL inference process, so that multiple participants of the VFL complete the VFL inference together, and the accuracy and effectiveness of the VFL inference are improved.

[0094] The following describes another embodiment of the inference method of the model of the present disclosure in combination with FIG. 2.

[0095] FIG. 2 is a flowchart of another embodiment of the inference method of the model of the present disclosure. As shown in FIG. 2, the method of this embodiment includes steps S202-S214.

[0096] In step S202, a Consumer NF sends second request information to a server, and correspondingly, the server receives the second request information sent by the Consumer NF.

[0097] In some embodiments, the second request information is used to request an analysis result, and the second request information includes at least one of an analysis identifier and a VFL indicator, the analysis identifier is used to identify that the requested analysis and / or the analysis result is generated in a VFL manner, and the VFL indicator is used to indicate that a VFL training model is used. The analysis identifier and the VFL indicator can refer to the foregoing embodiments, and will not be described here again.

[0098] In some embodiments, the second request information further includes filter information, which can refer to the foregoing embodiments, and will not be described here again.

[0099] In step S204, the server sends first request information to the client according to the second request information, and correspondingly, the client receives the first request information sent by the server.

[0100] In some embodiments, the first request information comprises at least one of a VFL association identifier, first sample indication information, first feature indication information, an interoperation indicator, an analysis identifier, filter information, and a VFL indicator. Each item of information can refer to the foregoing embodiments, and will not be described again here.

[0101] Optionally, in step S205, the client collects data from other network elements.

[0102] Optionally, in step S206, the server collects data from other network elements.

[0103] If the data for VFL inference is stored in other network elements, the client and the server can collect data from other network elements. The client and the server can collect data from the same or different network elements (NFs).

[0104] In step S207, the client performs VFL inference to obtain a VFL inference intermediate result corresponding to the client.

[0105] The client performs VFL inference using the model indicated in the first request information, the data sample, the data feature, and the corresponding data.

[0106] In step S208, the server performs VFL inference to obtain a VFL inference intermediate result corresponding to the server.

[0107] The server performs VFL inference using the VFL model and the corresponding data.

[0108] In step S210, the client sends first response information to the server, and correspondingly, the server receives the first response information sent by the client.

[0109] In some embodiments, the first response information comprises at least one of a VFL inference intermediate result, a VFL association identifier, second sample indication information, second feature indication information, and first validity time information.

[0110] In step S212, the server generates an inference result (or an analysis result) according to the first response information and / or the local VFL inference intermediate result.

[0111] In step S214, the server sends second response information to a consuming NF, and correspondingly, the consuming NF receives the second response information sent by the server.

[0112] In some embodiments, the second response information comprises at least one of an analysis result and second validity time information, and the second validity time information is used to indicate the validity time of the analysis result.

[0113] The analysis result can be a reasoning result or generated according to the reasoning result. In some embodiments, the analysis result is in the form of an Analytics ID output, includes at least one of data in the Analytics ID output, or includes at least one of quality of experience indicator information, latency information, and a 5G quality of service identifier (5G QoS Identifier, 5QI). The data in the Analytics ID output is defined in a standard, and details are not described again.

[0114] In some embodiments, the consumer NF is an Analytics Logical Function (AnLF), and the server is an AF.

[0115] The server can be an AF, and the client can also be an AF. If the server or the client is a non-trusted AF, in order to improve the security of the VFL reasoning process, the information exchanged between the server and the client needs to be processed.

[0116] In some embodiments, the server sends the first request information to the client through a Network Exposure Function (NEF), and correspondingly, the client receives the first request information sent by the server forwarded through the NEF.

[0117] In some embodiments, in the case where the server is a non-trusted AF, the VFL association identifier in the first request information is an external VFL association identifier, the first sample indication information includes first external sample indication information, and / or the first feature indication information includes first external feature indication information. The NEF maps the external VFL association identifier in the first request information to an internal VFL association identifier, maps the first external sample indication information to first internal sample indication information, and / or maps the first external feature indication information to first internal feature indication information.

[0118] In some embodiments, in the case where the client is a non-trusted AF, the VFL association identifier in the first request information is an internal VFL association identifier, the first sample indication information includes first internal sample indication information, and / or the first feature indication information includes first internal feature indication information. The NEF maps the internal VFL association identifier in the first request information to an external VFL association identifier, maps the first internal sample indication information to first external sample indication information, and / or maps the first internal feature indication information to first external feature indication information.

[0119] For example, the NEF stores a mapping relationship between an internal VFL correlation ID and an external VFL correlation ID, and when it receives a message from the NWDAF, it maps the internal VFL correlation ID to the external VFL correlation ID, and when it receives a message from the untrusted AF, it maps the external VFL correlation ID to the internal VFL correlation ID.

[0120] For example, the first sample indication information includes a Sample ID, and the NEF stores a mapping relationship between an internal Sample ID and an external Sample ID, and when it receives a message from the NWDAF, it maps the internal Sample ID to the external Sample ID, and when it receives a message from the untrusted AF, it maps the external Sample ID to the internal Sample ID.

[0121] For example, the first sample indication information includes a UE ID, and the NEF stores a mapping relationship between an external UE ID and an internal UE ID, and when it receives a message from the NWDAF, it maps the internal UE ID to the external UE ID, and when it receives a message from the untrusted AF, it maps the external UE ID to the internal UE ID.

[0122] For example, the first feature indication information includes a Feature ID and / or other Feature parameters, and the NEF stores a mapping relationship between an internal Feature ID and / or other Feature parameters and an external Feature ID and / or other Feature parameters, and when it receives a message from the NWDAF, it maps the internal Feature ID and / or other Feature parameters to the external Feature ID and / or other Feature parameters, and when it receives a message from the untrusted AF, it maps the external Feature ID and / or other Feature parameters to the internal Feature ID and / or other Feature parameters.

[0123] If the first request information does not include the first sample indication information or the first feature indication information, or the first sample indication information does not include information that needs to be converted, or the first feature indication information does not include information that needs to be converted, the NEF can not convert the first sample indication information or the first feature indication information.

[0124] In the case that the client is a non-trusted AF, the first response information sent by the client to the server also needs to be processed by the NEF to improve the security of the VFL inference.

[0125] In some embodiments, the client sends the first response information to the server through the NEF, and correspondingly, the server receives the first response information sent by the client forwarded through the NEF.

[0126] In some embodiments, in the case that the client is a non-trusted AF, the intermediate result of the VFL inference in the first response information is an external intermediate result, the VFL association identifier is an external VFL association identifier, the second sample indication information includes second external sample indication information, and / or the second feature indication information includes second external feature indication information, the NEF maps the external intermediate result to an internal intermediate result, maps the external VFL association identifier in the first request information to an internal VFL association identifier, maps the second external sample indication information to second internal sample indication information, and / or maps the second external feature indication information to second internal feature indication information.

[0127] The mapping of the second external sample indication information and the second external feature indication information can refer to the mapping of the first external sample indication information and the first external feature indication information, which will not be described again.

[0128] For example, the NEF stores the mapping relationship between the internal intermediate result parameter and the external intermediate result parameter, and when it receives information from the NWDAF, it maps the internal intermediate result parameter to the external intermediate result parameter, and when it receives a message from a non-trusted AF, it maps the external intermediate result to the internal intermediate result parameter.

[0129] The method of the above embodiments can guarantee the security of the information exchanged between the server and the client in the case that the client or the server is a non-trusted AF, thereby improving the security of the VFL inference, through the processing of the information exchanged between the server and the client by the NEF.

[0130] Based on the above scheme of processing the information exchanged between the server and the client by the NEF, some other embodiments of the inference method of the model of the present disclosure are described below in conjunction with FIG. 3.

[0131] FIG. 3 is a flowchart of some other embodiments of the inference method of the model of the present disclosure. As shown in FIG. 3, the method of the embodiments includes steps S302-S316.

[0132] In step S302, the server sends first request information to the NEF.

[0133] In step S304, the NEF maps the information in the first request information.

[0134] The specific mapping method can refer to the foregoing embodiments, which will not be described here again.

[0135] In step S306, the NEF sends the first request information to the client.

[0136] The first request information sent by the NEF is the first request information after the mapping is completed.

[0137] Optionally, in step S307, the client collects data from other network elements.

[0138] Optionally, in step S308, the server collects data from other network elements.

[0139] In step S309, the client performs VFL inference to obtain the intermediate result of the VFL inference corresponding to the client.

[0140] In step S310, the server performs VFL inference to obtain the intermediate result of the VFL inference corresponding to the server.

[0141] In step S312, the client sends the first response information to the NEF.

[0142] In step S314, the NEF maps the information in the first response information.

[0143] The specific mapping method can refer to the foregoing embodiments, which will not be described here again.

[0144] In step S316, the NEF sends the first response information to the server.

[0145] The first response information sent by the NEF is the first response information after the mapping is completed.

[0146] The VFL model can be stored in different network elements. For the server or the client, if the VFL model is not stored locally, it needs to be obtained from other network elements.

[0147] The method of the server or the client obtaining the VFL model from other network elements will be described below in combination with some embodiments.

[0148] FIG. 4 is a flowchart of still some embodiments of the inference method of the model of the disclosure. As shown in FIG. 4, the method of the embodiment includes steps S402-S404. For example, the server is the NWDAF, and the VFL model is not stored in the NWDAF.

[0149] In step S402, the server sends first model request information to the ADRF (Analytics Data Repository Function, analytics data storage function), and correspondingly, the ADRF receives the first model request information sent by the server.

[0150] In some embodiments, the first model request information comprises at least one of a VFL association identifier, first sample indication information, first feature indication information, a machine learning model identifier (ML model ID), an interoperation indicator, an analytics identifier, and filter information. Each item of information can refer to the foregoing embodiments and will not be described again here.

[0151] In step S404, the ADRF sends a first model request response to the service end, and correspondingly, the service end receives the first model request response sent by the ADRF.

[0152] In some embodiments, the ADRF retrieves the VFL model according to the first model request information, and the first model request response comprises at least one of a file of the VFL model, a storage address, and valid time information of the VFL model.

[0153] The valid time information of the VFL model is used to indicate a valid period or a valid time of the VFL model.

[0154] The service end can obtain information of the VFL model according to the first model request response, for VFL inference.

[0155] FIG. 5 is a flowchart of still another embodiment of a model inference method of the present disclosure. As shown in FIG. 5, the method of this embodiment comprises steps S502-S504. For example, the client is a NWDAF, and the VFL model is not stored in the NWDAF.

[0156] In step S502, the client sends second model request information to the ADRF, and correspondingly, the ADRF receives the second model request information sent by the client.

[0157] In some embodiments, the second model request information comprises at least one of a VFL association identifier, second sample indication information, second feature indication information, a machine learning model identifier (ML model ID), an interoperation indicator, an analytics identifier, and filter information. Each item of information can refer to the foregoing embodiments and will not be described again here.

[0158] In step S504, the ADRF sends a second model request response to the client, and correspondingly, the client receives the second model request response sent by the ADRF.

[0159] In some embodiments, the ADRF retrieves the VFL model according to the second model request information, and the second model request response comprises at least one of a file of the VFL model, a storage address, and valid time information of the VFL model.

[0160] The client can obtain information of the VFL model according to the second model request response, for VFL inference.

[0161] The following describes how the consumer NF initiates a request for an analysis result in the case of the consumer NF being an AnLF and the server being an AF, in combination with some embodiments.

[0162] FIG. 6 is a flowchart of still other embodiments of the inference method of the model of the present disclosure. As shown in FIG. 6, the method of this embodiment includes steps S602-S604.

[0163] In step S602, the AnLF sends third request information to the AF, and correspondingly, the AF receives the third request information sent by the AnLF.

[0164] In some embodiments, the third request information is used to request an analysis result, and the third request information includes at least one of an analysis identifier, a VFL indicator, and filter information. In some embodiments, the analysis identifier is used to identify the requested analysis and / or to identify that the analysis result is generated using a VFL method. The VFL indicator is used to indicate that the model is trained using a VFL. The filter information includes at least one of single network slice selection assistance information, an application identifier, a region of interest, and data network access identification.

[0165] In step S604, the AF sends third response information to the AnLF.

[0166] The AF performs VFL inference according to the third request information. The specific inference process can refer to the foregoing embodiments of the interaction between the server and the client, and will not be described here again.

[0167] In some embodiments, the third response information includes at least one of an analysis result and third validity time information. The third validity time information is used to indicate the validity time of the analysis result.

[0168] In some embodiments, the analysis result is in the form of an analysis ID output, includes at least one of the data of the analysis ID output, or includes at least one of quality of experience indicator information (QoE), delay information (Delay), and 5G quality of service indicator (5QI).

[0169] In some embodiments, in the case of the AF being a non-trusted AF, the AF receives the third request information sent by the AnLF through the NEF; and the AF sends the third response information to the AnLF through the NEF.

[0170] In some embodiments, in the case that the third request information includes internal related information and / or internal location information of a user equipment, the NEF maps the internal related information to external related information and / or maps the internal location information to external location information.

[0171] Internal related information of the user equipment is, for example, an identifier such as an internal UE ID, and internal location information such as area information. By mapping the information sent between the AF and the AnLF through the NEF, the security of the information can be improved, and the security of the VFL inference can be improved.

[0172] The method of the above embodiment can realize that the AnLF initiates an analysis result request, triggers the process of the VFL inference of the AF, and through the interaction between the AnLF and the AF, the inference process of the VFL can be realized, and the accuracy of the inference is improved.

[0173] The present disclosure also provides a server, which is described below in conjunction with FIG. 7.

[0174] FIG. 7 is a structural diagram of some embodiments of a server of the present disclosure. As shown in FIG. 7, the server 70 of the embodiment includes a sending module 710, a receiving module 720, and a generating module 730.

[0175] The sending module 710 is configured to send first request information to the client, where the first request information is used to request vertical federated learning (VFL) inference, and the first request information includes at least one of a VFL association identifier, first sample indication information, and first feature indication information.

[0176] The receiving module 720 is configured to receive first response information sent by the client, where the first response information includes at least one of an intermediate result of the VFL inference, the VFL association identifier, second sample indication information, and second feature indication information.

[0177] The generating module 730 is configured to generate an inference result according to the first response information and / or an intermediate result of local VFL inference.

[0178] In some embodiments, the VFL association identifier is used to associate the process of the VFL and / or a VFL model; the first sample indication information is used to indicate a data sample; the first feature indication information is used to indicate a data feature; the second sample indication information is used to indicate a data sample; and / or the second feature indication information is used to indicate a data feature.

[0179] In some embodiments, the server 70 further includes an inference module 740 configured to perform VFL inference according to data in the server to obtain an intermediate result of local VFL inference.

[0180] In some embodiments, the receiving module 720 is further configured to receive second request information sent by a consuming network function (NF), wherein the second request information is used to request an analysis result, and the second request information comprises at least one of an analysis identifier and a VFL indicator, the analysis identifier is used to identify a requested analysis and / or the analysis result is generated using a VFL manner, and the VFL indicator is used to indicate that a VFL training model is used. The sending module 710 is configured to send the first request information to the client according to the second request information.

[0181] In some embodiments, the second request information further comprises filter information, wherein the filter information comprises at least one of a single network slice selection assistance information, an application identifier, a region of interest, and a data network access identifier.

[0182] In some embodiments, the first request information further comprises at least one of an interoperability indicator, an analysis identifier, filter information, and a VFL indicator.

[0183] In some embodiments, the sending module 710 is configured to send second response information to the consuming NF, wherein the second response information comprises at least one of an analysis result and second validity time information, and the second validity time information is used to indicate a validity time of the analysis result.

[0184] In some embodiments, the first response information further comprises first validity time information used to indicate a validity time of the intermediate result of the VFL inference.

[0185] In some embodiments, the first sample indication information comprises at least one of a user equipment identifier, a generic public user identifier (GPSI), a subscription permanent identifier (SUPI), location information, a sample identifier, and a group identifier; and / or the second sample indication information comprises at least one of a user equipment identifier, a generic public user identifier (GPSI), a subscription permanent identifier (SUPI), location information, a sample identifier, and a group identifier.

[0186] In some embodiments, the first feature indication information is used to indicate a server-side used data feature and / or a client-side used data feature; the second feature indication information is used to indicate a server-side used data feature and / or a client-side used data feature; the first feature indication information comprises at least one of service experience information, quality of experience indicator information, voice quality information, and a feature identifier; and / or the second feature indication information comprises at least one of service experience information, quality of experience indicator information, voice quality information, and a feature identifier.

[0187] In some embodiments, the data collection module 750 is configured to collect data from corresponding network elements according to the first request information.

[0188] In some embodiments, the analysis result is in the form of an analysis ID output, including at least one of various data of the analysis ID output, or including at least one of quality of experience information, latency information, and 5G quality of service indicators.

[0189] In some embodiments, the consumer NF is an analytics logic function AnLF, and the server is an application function AF.

[0190] In some embodiments, the server is a network data analytics function NWDAF, and the client is an application function AF, or the server and the client are different NWDAFs, or the server is an AF and the client is a NWDAF.

[0191] In some embodiments, when the AF is a non-trusted AF, the sending module 710 is configured to send the first request information to the client through a network exposure function NEF, wherein, when the server is a non-trusted AF, the VFL association identifier in the first request information is an external VFL association identifier, the first sample indication information includes first external sample indication information, and / or the first feature indication information includes first external feature indication information, the NEF maps the external VFL association identifier in the first request information to an internal VFL association identifier, maps the first external sample indication information to first internal sample indication information, and / or maps the first external feature indication information to first internal feature indication information, and / or, when the client is a non-trusted AF, the VFL association identifier in the first request information is an internal VFL association identifier, the first sample indication information includes first internal sample indication information, and / or the first feature indication information includes first internal feature indication information, the NEF maps the internal VFL association identifier in the first request information to an external VFL association identifier, maps the first internal sample indication information to first external sample indication information, and / or maps the first internal feature indication information to first external feature indication information.

[0192] In some embodiments, in the case that the AF is a non-trusted AF, the receiving module 720 is configured to receive first response information sent by the client and forwarded by a network exposure function (NEF), wherein, in the case that the client is a non-trusted AF, the intermediate result of the VFL inference in the first response information is an external intermediate result, the VFL association identifier is an external VFL association identifier, the second sample indication information includes second external sample indication information, and / or the second feature indication information includes second external feature indication information, the NEF maps the external intermediate result to an internal intermediate result, maps the external VFL association identifier in the first request information to an internal VFL association identifier, maps the second external sample indication information to second internal sample indication information, and / or maps the second external feature indication information to second internal feature indication information, and / or, in the case that the server is a non-trusted AF, the intermediate result of the VFL inference in the first response information is an internal intermediate result, the VFL association identifier is an internal VFL association identifier, the second sample indication information includes second internal sample indication information, and the second feature indication information includes second internal feature indication information, the NEF maps the internal intermediate result in the first request information to an external intermediate result, maps the internal VFL association identifier to an external VFL association identifier, maps the second internal sample indication information to second external sample indication information, and / or maps the second internal feature indication information to second external feature indication information.

[0193] In some embodiments, in the case that the server is a NWDAF and the VFL model is not stored in the NWDAF, the sending module 710 is configured to send first model request information to an analytics data storage function (ADRF), wherein the first model request information includes at least one of the VFL association identifier, the first sample indication information, and the first feature indication information; and the receiving module 720 is configured to receive a first model request response sent by the ADRF, wherein the ADRF retrieves the VFL model according to the first model request information, and the first model request response includes at least one of a file of the VFL model, a storage address, and validity time information of the VFL model.

[0194] The present disclosure also provides a server, which is described below in conjunction with FIG. 7.

[0195] FIG. 7 is a structural diagram of some embodiments of a server of the present disclosure. As shown in FIG. 7, the server 70 of this embodiment includes a receiving module 710 and a sending module 720.

[0196] The receiving module 710 is configured to receive first request information sent by the client, wherein the first request information is used to request vertical federated learning (VFL) inference, and the first request information includes at least one of a VFL association identifier, first sample indication information, and first feature indication information.

[0197] The sending module 820 is configured to send first response information to the server, wherein the first response information comprises at least one of an intermediate result of VFL inference, a VFL association identifier, second sample indication information, and second feature indication information, so that the server generates an inference result according to the first response information and / or the intermediate result of local VFL inference.

[0198] In some embodiments, the VFL association identifier is used to associate a process of the VFL and / or a VFL model; the first sample indication information is used to indicate a data sample; the first feature indication information is used to indicate a data feature; the second sample indication information is used to indicate a data sample; and / or the second feature indication information is used to indicate a data feature.

[0199] In some embodiments, the client 80 further comprises an inference module 830 configured to perform VFL inference according to data in the client to obtain an intermediate result of VFL inference.

[0200] In some embodiments, the first request information further comprises at least one of an interoperation indicator, an analysis identifier, filter information, and a VFL indicator, wherein the analysis identifier is used to identify an analysis of the request and / or an analysis result is generated in a VFL manner, the filter information comprises at least one of single network slice selection assistance information, application identifier, area of interest, data network access identifier, and the VFL indicator is used to indicate a model trained using VFL.

[0201] In some embodiments, the first response information further comprises first validity time information used to indicate a validity time of the intermediate result of VFL inference.

[0202] In some embodiments, the first sample indication information comprises at least one of a user equipment identifier, a generic public subscription identifier GPSI, a subscription permanent identifier SUPI, location information, a sample identifier, and a group identifier; and / or the second sample indication information comprises at least one of a user equipment identifier, a generic public subscription identifier GPSI, a subscription permanent identifier SUPI, location information, a sample identifier, and a group identifier.

[0203] In some embodiments, the first feature indication information is used to indicate a data feature used by the server and / or a data feature used by the client; the second feature indication information is used to indicate a data feature used by the server and / or a data feature used by the client; the first feature indication information comprises at least one of service experience information, quality of experience indicator information, voice quality information, and a feature identifier; and / or the second feature indication information comprises at least one of service experience information, quality of experience indicator information, voice quality information, and a feature identifier.

[0204] In some embodiments, the data collection module 840 is configured to collect the data sample from the corresponding network element according to the first request information.

[0205] In some embodiments, the server is a network data analytics function (NWDAF), and the client is an application function (AF), or the server and the client are different NWDAFs, or the server is an AF, and the client is an NWDAF.

[0206] In some embodiments, in a case where the AF is a non-trusted AF, the receiving module 810 is configured to receive the first request information sent by the server and forwarded by a network exposure function (NEF), wherein, in a case where the server is a non-trusted AF, the VFL association identifier in the first request information is an external VFL association identifier, the first sample indication information includes first external sample indication information, and / or the first feature indication information includes first external feature indication information, the NEF maps the external VFL association identifier in the first request information to an internal VFL association identifier, maps the first external sample indication information to first internal sample indication information, and / or maps the first external feature indication information to first internal feature indication information, and / or, in a case where the client is a non-trusted AF, the VFL association identifier in the first request information is an internal VFL association identifier, the first sample indication information includes first internal sample indication information, and / or the first feature indication information includes first internal feature indication information, the NEF maps the internal VFL association identifier in the first request information to an external VFL association identifier, maps the first internal sample indication information to first external sample indication information, and / or maps the first internal feature indication information to first external feature indication information.

[0207] In some embodiments, in a case where the AF is a non-trusted AF, the sending module 820 is configured to send the first response information to the server through a network exposure function (NEF),

[0208] In a case where the client is a non-trusted AF, the intermediate result of the VFL inference in the first response information is an external intermediate result, the VFL association identifier is an external VFL association identifier, the second sample indication information includes second external sample indication information, and / or the second feature indication information includes second external feature indication information, the NEF maps the external intermediate result to an internal intermediate result, maps the external VFL association identifier in the first request information to an internal VFL association identifier, maps the second external sample indication information to second internal sample indication information, and / or maps the second external feature indication information to second internal feature indication information, and / or in a case where the server is a non-trusted AF, the intermediate result of the VFL inference in the first response information is an internal intermediate result, the VFL association identifier is an internal VFL association identifier, the second sample indication information includes second internal sample indication information, and the second feature indication information includes second internal feature indication information, the NEF maps the internal intermediate result in the first request information to an external intermediate result, maps the internal VFL association identifier to an external VFL association identifier, maps the second internal sample indication information to second external sample indication information, and / or maps the second internal feature indication information to second external feature indication information.

[0209] In some embodiments, in a case where the client is a NWDAF and the VFL model is not stored in the NWDAF, the sending module 820 is configured to send second model request information to an analytics data storage function (ADRF), wherein the second model request information includes at least one of the VFL association identifier, the second sample indication information, and the second feature indication information; and the receiving module 810 is configured to receive a second model request response sent by the ADRF, wherein the ADRF retrieves the VFL model according to the second model request information, and the second model request response includes at least one of a file of the VFL model, a storage address, and validity time information of the VFL model.

[0210] The present disclosure also provides an application function, which is described below in conjunction with FIG. 9.

[0211] FIG. 9 is a structural diagram of some embodiments of the application function of the present disclosure. As shown in FIG. 9, the AF 90 of this embodiment includes a receiving module 910, an inference module 920, and a sending module 930.

[0212] The receiving module 910 is configured to receive third request information sent by an analytics logic function (AnLF), wherein the third request information is used to request an analysis result, and the third request information includes at least one of an analysis identifier and a vertical federated learning (VFL) indicator.

[0213] The inference module 920 is configured to perform VFL inference according to the third request information.

[0214] The sending module 930 is configured to send third response information to the AnLF, wherein the third response information comprises at least one of an analysis result and third validity time information.

[0215] In some embodiments, the analysis identifier is used to identify the requested analysis and / or the analysis result is generated using a VFL manner; the VFL indicator is used to indicate a model trained using VFL; and / or the third validity time information is used to indicate a validity time of the analysis result.

[0216] In some embodiments, the third request information further comprises filter information, wherein the filter information comprises at least one of a single network slice selection assistance information, an application identifier, a region of interest, and a data network access identifier.

[0217] In some embodiments, the analysis result is in the form of an analysis ID output, and comprises at least one of various data of the analysis ID output, or comprises at least one of quality of experience indicator information, latency information, and 5G quality of service indicators.

[0218] In some embodiments, in the case that the AF is a non-trusted AF, the receiving module 910 is configured to receive the third request information sent by the AnLF through a network exposure function (NEF); and the sending module 920 is configured to send the third response information to the AnLF through the NEF.

[0219] In some embodiments, in the case that the third request information comprises internal related information and / or internal location information of a user equipment, the NEF maps the internal related information to external related information and / or maps the internal location information to external location information.

[0220] The present disclosure also provides an analysis logic function (AnLF), which is described below in conjunction with FIG. 10.

[0221] FIG. 10 is a structural diagram of some embodiments of the AnLF of the present disclosure. As shown in FIG. 10, the AnLF 100 of this embodiment comprises a sending module 1010 and a receiving module 1020.

[0222] The sending module 1010 is configured to send third request information to an application function (AF), wherein the third request information is used to request an analysis result, and the third request information comprises at least one of an analysis identifier and a vertical federated learning (VFL) indicator.

[0223] The receiving module 1020 is configured to receive third response information sent by the AF, wherein the third response information comprises at least one of an analysis result and fourth validity time information.

[0224] In some embodiments, the analysis identifier is used to identify the analysis of the request and / or the analysis result is generated in a VFL manner; the VFL indicator is used to indicate a model trained using VFL; and / or the third validity time information is used to indicate a validity time of the analysis result.

[0225] In some embodiments, in the case that the AF is a non-trusted AF, the sending module 1010 is configured to send third request information to the AF through a network exposure function (NEF), wherein, in the case that the third request information includes internal correlation information and / or internal location information of the user equipment, the NEF maps the internal correlation information to external correlation information and maps the internal location information to external location information; and the receiving module 1020 is configured to receive third response information sent by the AF through the NEF.

[0226] The present disclosure also provides an electronic device, which is described below in conjunction with FIGS. 11 and 12.

[0227] FIG. 11 is a structural diagram of some embodiments of the electronic device of the present disclosure. As shown in FIG. 11, the electronic device 110 of this embodiment includes a memory 1110 and a processor 1120 coupled to the memory 1110, and the processor 1120 is configured to perform the inference method of the model in any of some embodiments of the present disclosure based on instructions stored in the memory 1110.

[0228] The memory 510 may, for example, include system memory, fixed non-volatile storage media, etc. The system memory may, for example, store an operating system, application programs, a boot loader, a database, and other programs, etc.

[0229] FIG. 12 is a structural diagram of another embodiment of the electronic device of the present disclosure. As shown in FIG. 12, the electronic device 120 of this embodiment includes a memory 1210 and a processor 1220, which are similar to the memory 510 and the processor 520, respectively. It may also include an input / output interface 1230, a network interface 1240, a storage interface 1250, etc. These interfaces 1230, 1240, 1250 and the memory 1210 and the processor 1220 may, for example, be connected through a bus 1260. The input / output interface 1230 provides a connection interface for display, mouse, keyboard, touch screen, and other input / output devices. The network interface 1240 provides a connection interface for various networking devices, such as a database server or a cloud storage server, etc. The storage interface 1250 provides a connection interface for external storage devices such as SD cards, USB flash drives, etc.

[0230] The present disclosure also provides an inference system of a model, which is described below in conjunction with FIG. 13.

[0231] FIG. 13 is a structural diagram of some embodiments of the inference system of the model of the present disclosure. As shown in FIG. 13, the system 13 of this embodiment includes the server 70 and the client 80 of any embodiment of the present disclosure.

[0232] In some embodiments, the server is a network data analytics function (NWDAF), and the client is an application function (AF), or the server and the client are different NWDAFs, or the server is an AF, and the client is an NWDAF.

[0233] In some embodiments, the system 13 further includes a consumer NF 132 configured to send second request information to the server 70 and receive second response information sent by the server 70.

[0234] In some embodiments, the consumer NF is an analytics logic function (AnLF), and the server is an application function (AF).

[0235] In some embodiments, the system 13 further includes an NEF 134 configured to receive the first request information sent by the server 70 and send to the client 80, and receive the first response information sent by the client 80 and send to the server 70.

[0236] In some embodiments, in the case where the server is a non-trusted AF, the VFL association identifier in the first request information is an external VFL association identifier, the first sample indication information includes first external sample indication information, and / or the first feature indication information includes first external feature indication information, the NEF 134 is configured to map the external VFL association identifier in the first request information to an internal VFL association identifier, map the first external sample indication information to first internal sample indication information, and / or map the first external feature indication information to first internal feature indication information.

[0237] In some embodiments, in the case where the client is a non-trusted AF, the VFL association identifier in the first request information is an internal VFL association identifier, the first sample indication information includes first internal sample indication information, and / or the first feature indication information includes first internal feature indication information, the NEF 134 is configured to map the internal VFL association identifier in the first request information to an external VFL association identifier, map the first internal sample indication information to first external sample indication information, and / or map the first internal feature indication information to first external feature indication information.

[0238] In some embodiments, in the case that the client is a non-trusted AF, the intermediate result of the VFL inference in the first response information is an external intermediate result, the VFL association identifier is an external VFL association identifier, the second sample indication information comprises second external sample indication information, and / or the second feature indication information comprises second external feature indication information, the NEF 134 is configured to map the external intermediate result to an internal intermediate result, map the external VFL association identifier in the first request information to an internal VFL association identifier, map the second external sample indication information to second internal sample indication information, and / or map the second external feature indication information to second internal feature indication information.

[0239] In some embodiments, in the case that the server is a non-trusted AF, the intermediate result of the VFL inference in the first response information is an internal intermediate result, the VFL association identifier is an internal VFL association identifier, the second sample indication information comprises second internal sample indication information, and the second feature indication information comprises second internal feature indication information, the NEF 134 is configured to map the internal intermediate result in the first request information to an external intermediate result, map the internal VFL association identifier to an external VFL association identifier, map the second internal sample indication information to second external sample indication information, and / or map the second internal feature indication information to second external feature indication information.

[0240] In some embodiments, the system 13 further comprises an ADRF 136, configured to, in the case that the server is a NWDAF and the VFL model is not stored in the NWDAF, receive the first model request information sent by the server 70, and send a first model request response to the server 70.

[0241] In some embodiments, the ADRF 136 is configured to receive the second model request information sent by the client 80, and send a second model request response to the client 80.

[0242] The information and the functions of the various network elements can refer to the descriptions of the foregoing embodiments, and will not be described here.

[0243] The present disclosure also provides a computer readable storage medium having stored thereon a computer program, wherein the program, when executed by a processor, implements the inference method of the model according to any of the foregoing embodiments.

[0244] The present disclosure also provides a computer program product, comprising instructions, wherein the instructions, when executed by a processor, implement the inference method of the model according to any of the foregoing embodiments.

[0245] The present disclosure also provides a computer program, comprising instructions, wherein the instructions, when executed by a processor, implement the inference method of the model according to any of the foregoing embodiments.

[0246] Those skilled in the art will appreciate that embodiments of the disclosure can be supplied as a method, a system, or a computer program product. Accordingly, the present disclosure can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the disclosure can take the form of a computer program product on one or more computer-usable non-transitory storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, and the like) having computer-usable program code embodied in the medium.

[0247] The present disclosure is described in reference to the flowchart and / or block diagrams of the method, apparatus (system) and computer program product according to embodiments of the present disclosure. It should be understood that each flow and / or block in the flowchart and / or block diagrams, and a combination of flows and / or blocks in the flowchart and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, a special purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions, which are executed via the processor of the computer or other programmable data processing apparatus, generate means for implementing the functions specified in the flowchart one or more flows and / or block diagram one or more blocks.

[0248] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to work in a particular manner, so that the instructions stored in the computer readable memory produce an article of manufacture including instruction means, which implement the functions specified in the flowchart one or more flows and / or block diagram one or more blocks.

[0249] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operational steps are performed on the computer or other programmable data processing apparatus to generate a computer implemented process, so that the instructions executed on the computer or other programmable data processing apparatus provide steps for implementing the functions specified in the flowchart one or more flows and / or block diagram one or more blocks.

[0250] The above description is only the preferred embodiment of the present disclosure, and is not intended to limit the present disclosure. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A model inference method, executed by a server, comprising: sending, to a client, first request information, wherein the first request information is used to request vertical federated learning (VFL) inference, and the first request information comprises at least one of a VFL association identifier, first sample indication information, and first feature indication information; receiving first response information sent by the client, wherein the first response information comprises at least one of an intermediate result of the VFL inference, the VFL association identifier, second sample indication information, and second feature indication information; generating an inference result according to the first response information and / or an intermediate result of local VFL inference.

2. The inference method of claim 1, wherein: the VFL association identifier is used to associate a process of the VFL and / or a VFL model; the first sample indication information is used to indicate a data sample; the first feature indication information is used to indicate a data feature; the second sample indication information is used to indicate a data sample; and / or the second feature indication information is used to indicate a data feature.

3. The inference method of claim 1 or 2, further comprising: performing VFL inference according to data in the server to obtain the intermediate result of the local VFL inference.

4. The inference method of any one of claims 1-3, further comprising: receiving second request information sent by a consuming network function (NF), wherein the second request information is used to request an analysis result, and the second request information comprises at least one of an analysis identifier and a VFL indicator, the analysis identifier is used to identify a requested analysis and / or an analysis result generated in a VFL manner, and the VFL indicator is used to indicate a VFL trained model; wherein the sending, to the client, first request information comprises: sending, to the client, the first request information according to the second request information.

5. The inference method of claim 4, wherein: the second request information further comprises filter information, wherein the filter information comprises at least one of a single network slice selection assistance information, an application identifier, a region of interest, and a data network access identifier.

6. The inference method of claim 5, wherein, the first request information further comprises at least one of an interoperation indicator, the analysis identifier, the filter information, and the VFL indicator.

7. The inference method of any one of claims 4-6, further comprising: sending, to the consuming NF, second response information, wherein the second response information comprises at least one of the analysis result and second validity time information, and the second validity time information is used to indicate a validity time of the analysis result.

8. The inference method of any one of claims 1-7, wherein, the first response information further comprises first validity time information, used to indicate a validity time of the intermediate result of the VFL inference.

9. The inference method of any one of claims 1-8, wherein: the first sample indication information comprises at least one of a user equipment identifier, a generic public subscription identifier (GPSI), a subscription permanent identifier (SUPI), location information, a sample identifier, and a group identifier; and / or The second sample indication information includes at least one of a user equipment identifier, a general public user identifier (GPSI), a subscription permanent identifier (SUPI), location information, a sample identifier, and a group identifier.

10. The reasoning method of any of claims 1-9, wherein, The first feature indication information is used to indicate a data feature used by the service end and / or a data feature used by the client end. The second feature indication information is used to indicate a data feature used by the service end and / or a data feature used by the client end. The first feature indication information includes at least one of service experience information, quality of experience indicator information, voice quality information, and a feature identifier; and / or The second feature indication information includes at least one of service experience information, quality of experience indicator information, voice quality information, and a feature identifier.

11. The reasoning method of any of claims 1-10, further comprising: collecting data from a corresponding network element according to the first request information.

12. The reasoning method of any of claims 1-11, wherein, The analysis result is in the form of an analysis ID output, includes at least one of various data of the analysis ID output, or includes at least one of quality of experience indicator information, latency information, and 5G quality of service indicators.

13. The inference method of any one of claims 4-12, wherein, The consuming NF is an analysis logic function (AnLF), and the service end is an application function (AF).

14. The reasoning method of any of claims 1-13, wherein, The service end is a network data analysis function (NWDAF), and the client end is an application function (AF), or the service end and the client end are different NWDAFs, or the service end is an AF and the client end is an NWDAF.

15. The inference method of claim 14, wherein, In a case where the AF is a non-trusted AF, the sending of the first request information to the client end includes: sending the first request information to the client end through a network exposure function (NEF), In a case where the service end is the non-trusted AF, the VFL association identifier in the first request information is an external VFL association identifier, the first sample indication information includes first external sample indication information, and / or the first feature indication information includes first external feature indication information, the NEF maps the external VFL association identifier in the first request information to an internal VFL association identifier, maps the first external sample indication information to first internal sample indication information, and / or maps the first external feature indication information to first internal feature indication information, In a case where the client is the untrusted AF, the VFL association identifier in the first request information is an internal VFL association identifier, the first sample indication information includes first internal sample indication information, and / or the first feature indication information includes first internal feature indication information, the NEF maps the internal VFL association identifier in the first request information to an external VFL association identifier, maps the first internal sample indication information to first external sample indication information, and / or maps the first internal feature indication information to first external feature indication information.

16. The inference method of claim 14 or 15, wherein, In a case where the AF is an untrusted AF, the receiving the first response information sent by the client includes: receiving the first response information sent by the client forwarded by a network exposure function (NEF), In a case where the client is the untrusted AF, the intermediate result of the VFL inference in the first response information is an external intermediate result, the VFL association identifier is an external VFL association identifier, the second sample indication information includes second external sample indication information, and / or the second feature indication information includes second external feature indication information, the NEF maps the external intermediate result to an internal intermediate result, maps the external VFL association identifier in the first request information to an internal VFL association identifier, maps the second external sample indication information to second internal sample indication information, and / or maps the second external feature indication information to second internal feature indication information, In a case where the client is the untrusted AF, the intermediate result of the VFL inference in the first response information is an external intermediate result, the VFL association identifier is an external VFL association identifier, the second sample indication information includes second external sample indication information, and / or the second feature indication information includes second external feature indication information, the NEF maps the external intermediate result to an internal intermediate result, maps the external VFL association identifier in the first request information to an internal VFL association identifier, maps the second external sample indication information to second internal sample indication information, and / or maps the second external feature indication information to second internal feature indication information.

17. The reasoning method of any of claims 14-16, further comprising: In a case where the server is the NWDAF and the VFL model is not stored in the NWDAF, sending first model request information to an analytics data repository function (ADRF), wherein the first model request information includes at least one of the VFL association identifier, the first sample indication information, and the first feature indication information; receiving a first model request response sent by the ADRF, wherein the ADRF retrieves the VFL model according to the first model request information, and the first model request response includes at least one of a file of the VFL model, a storage address, and validity time information of the VFL model.

18. A model inference method performed by a client, comprising: receive first request information sent by a server, wherein the first request information is used to request vertical federated learning (VFL) inference, and the first request information comprises at least one of a VFL association identifier, first sample indication information, and first feature indication information; send first response information to the server, wherein the first response information comprises at least one of an intermediate result of the VFL inference, the VFL association identifier, second sample indication information, and second feature indication information, so that the server generates an inference result according to the first response information and / or an intermediate result of local VFL inference.

19. The inference method of claim 18, wherein the VFL association identifier is used to associate a process of the VFL and / or a VFL model; the first sample indication information is used to indicate a data sample; the first feature indication information is used to indicate a data feature; the second sample indication information is used to indicate a data sample; and / or the second feature indication information is used to indicate a data feature.

20. The inference method of claim 18 or 19, further comprising: performing VFL inference according to data in the client to obtain the intermediate result of the VFL inference.

21. The reasoning method according to any one of claims 18-20, wherein, the first request information further comprises at least one of an interoperation indicator, an analysis identifier, filter information, and a VFL indicator, wherein the analysis identifier is used to identify that requested analysis and / or an analysis result is generated in a VFL manner, the filter information comprises at least one of a single network slice selection assistance information, an application identifier, a region of interest, and a data network access identifier, and the VFL indicator is used to indicate a model trained using VFL.

22. The inference method of any one of claims 18-21, wherein, the first response information further comprises first validity time information used to indicate a validity time of the intermediate result of the VFL inference.

23. The inference method of any of claims 18-22, wherein the first sample indication information comprises at least one of a user equipment identifier, a generic public subscription identifier (GPSI), a subscription permanent identifier (SUPI), location information, a sample identifier, and a group identifier; and / or the second sample indication information comprises at least one of a user equipment identifier, a generic public subscription identifier (GPSI), a subscription permanent identifier (SUPI), location information, a sample identifier, and a group identifier.

24. The inference method of any of claims 18-23, wherein the first feature indication information is used to indicate a data feature used by the server and / or a data feature used by the client; the second feature indication information is used to indicate a data feature used by the server and / or a data feature used by the client; the first feature indication information comprises at least one of service experience information, quality of experience (QoE) indicator information, voice quality information, and a feature identifier; and / or the second feature indication information comprises at least one of service experience information, quality of experience (QoE) indicator information, voice quality information, and a feature identifier.

25. The inference method of any of claims 18-24, further comprising: collecting data samples from corresponding network elements according to the first request information.

26. The inference method of any one of claims 18-25, wherein, the server is a network data analytics function (NWDAF) and the client is an application function (AF), or the server and the client are different NWDAFs, or the server is an AF and the client is a NWDAF.

27. The inference method of claim 26, wherein, in the case that the AF is a non-trusted AF, the receiving the first request information sent by the server comprises: receiving the first request information sent by the server forwarded by a network exposure function (NEF), in the case that the server is the non-trusted AF, the VFL correlation identifier in the first request information is an external VFL correlation identifier, the first sample indication information comprises first external sample indication information, and / or the first feature indication information comprises first external feature indication information, the NEF maps the external VFL correlation identifier in the first request information to an internal VFL correlation identifier, maps the first external sample indication information to first internal sample indication information, and / or maps the first external feature indication information to first internal feature indication information, in the case that the client is the non-trusted AF, the VFL correlation identifier in the first request information is an internal VFL correlation identifier, the first sample indication information comprises first internal sample indication information, and / or the first feature indication information comprises first internal feature indication information, the NEF maps the internal VFL correlation identifier in the first request information to an external VFL correlation identifier, maps the first internal sample indication information to first external sample indication information, and / or maps the first internal feature indication information to first external feature indication information.

28. The reasoning method of claim 26 or 27, wherein, in the case that the AF is a non-trusted AF, the sending the first response information to the server comprises: sending the first response information to the server through a network exposure function (NEF), in the case that the client is the non-trusted AF, the intermediate result of the VFL inference in the first response information is an external intermediate result, the VFL correlation identifier is an external VFL correlation identifier, the second sample indication information comprises second external sample indication information, and / or the second feature indication information comprises second external feature indication information, the NEF maps the external intermediate result to an internal intermediate result, maps the external VFL correlation identifier in the first request information to an internal VFL correlation identifier, maps the second external sample indication information to second internal sample indication information, and / or maps the second external feature indication information to second internal feature indication information, In a case where the service end is the untrusted AF, the intermediate result of the VFL inference in the first response information is an internal intermediate result, the VFL association identifier is an internal VFL association identifier, the second sample indication information includes second internal sample indication information, and the second feature indication information includes second internal feature indication information. The NEF maps the internal intermediate result in the first request information to the external intermediate result, maps the internal VFL association identifier to an external VFL association identifier, maps the second internal sample indication information to second external sample indication information, and / or maps the second internal feature indication information to second external feature indication information.

29. The inferential method of any of claims 26-28, further comprising: In a case where the client is the NWDAF and the VFL model is not stored in the NWDAF, sending second model request information to an analysis data storage function (ADRF), wherein the second model request information includes at least one of the VFL association identifier, the second sample indication information, and the second feature indication information; receiving second model request response sent by the ADRF, wherein the ADRF retrieves the VFL model according to the second model request information, and the second model request response includes at least one of a file of the VFL model, a storage address, and validity time information of the VFL model.

30. A model inference method, performed by an application function (AF), comprising: receiving third request information sent by an analysis logic function (AnLF), wherein the third request information is used to request an analysis result, and the third request information includes at least one of an analysis identifier and a vertical federated learning (VFL) indicator; performing VFL inference according to the third request information; sending third response information to the AnLF, wherein the third response information includes at least one of the analysis result and third validity time information.

31. The inference method of claim 30, wherein the analysis identifier is used to identify a requested analysis and / or to identify that the analysis result is generated in a VFL manner; the VFL indicator is used to indicate a model trained using VFL; and / or the third validity time information is used to indicate a validity time of the analysis result.

32. The inference method of claim 30, wherein the third request information further includes filter information, wherein the filter information includes at least one of single network slice selection assistance information, an application identifier, a region of interest, and data network access identification.

33. The inference method of claim 30, wherein the analysis result is in a form of an analysis ID output, includes at least one of various data of the analysis ID output, or includes at least one of quality of experience (QoE) indicator information, latency information, and 5G quality of service (QoS) indicators.

34. The inference method of claim 30, wherein, In a case where the AF is an untrusted AF, the receiving third request information sent by an analysis logic function (AnLF) includes: receiving, by a network exposure function (NEF), the third request information sent by the AnLF; the sending third response information to the AnLF includes: sending third response information to the AnLF via the NEF.

35. The inferential method of claim 34 wherein, In a case where the third request information includes internal correlation information and / or internal location information of a user equipment, the NEF maps the internal correlation information to external correlation information and / or maps the internal location information to external location information.

36. A reasoning method of a model, performed by an analytics logic function (AnLF), comprising: sending, to an application function (AF), third request information, wherein the third request information is used to request an analysis result, and the third request information includes at least one of an analysis identifier and a vertical federated learning (VFL) indicator; receiving third response information sent by the AF, wherein the third response information includes at least one of the analysis result and third validity time information.

37. The reasoning method of claim 36, wherein: the analysis identifier is used to identify a requested analysis and / or identify that the analysis result is generated in a VFL manner; the VFL indicator is used to indicate a model trained using VFL; and / or the third validity time information is used to indicate a validity time of the analysis result.

38. The reasoning method of claim 36, wherein: in a case where the AF is a non-trusted AF, the sending, to an application function (AF), third request information includes: sending, to the AF via a network exposure function (NEF), the third request information, wherein, in a case where the third request information includes internal correlation information and / or internal location information of a user equipment, the NEF maps the internal correlation information to external correlation information and maps the internal location information to external location information; the receiving third response information sent by the AF includes: receiving, via the NEF, third response information sent by the AF.

39. A server, comprising: a sending module configured to send, to a client, first request information, wherein the first request information is used to request vertical federated learning (VFL) reasoning, and the first request information includes at least one of a VFL correlation identifier, first sample indication information, and first feature indication information; a receiving module configured to receive first response information sent by the client, wherein the first response information includes at least one of an intermediate result of the VFL reasoning, the VFL correlation identifier, second sample indication information, and second feature indication information; a generating module configured to generate a reasoning result according to the first response information and / or an intermediate result of local VFL reasoning.

40. A client, comprising: a receiving module configured to receive first request information sent by a server, wherein the first request information is used to request vertical federated learning (VFL) reasoning, and the first request information includes at least one of a VFL correlation identifier, first sample indication information, and first feature indication information; The sending module is configured to send first response information to the service end, wherein the first response information comprises at least one of the intermediate result of the VFL inference, the VFL association identifier, second sample indication information and second feature indication information, so that the service end generates an inference result according to the first response information and / or the intermediate result of the local VFL inference.

41. An application function comprising: a receiving module configured to receive third request information sent by an analysis logic function AnLF, wherein the third request information is used to request an analysis result, and the third request information comprises at least one of an analysis identifier and a vertical federated learning VFL indicator; an inference module configured to perform VFL inference according to the third request information; a sending module configured to send third response information to the AnLF, wherein the third response information comprises at least one of the analysis result and third valid time information.

42. An analysis logic function AnLF comprising: a sending module configured to send third request information to an application function AF, wherein the third request information is used to request an analysis result, and the third request information comprises at least one of an analysis identifier and a vertical federated learning VFL indicator; a receiving module configured to receive third response information sent by the AF, wherein the third response information comprises at least one of the analysis result and fourth valid time information.

43. An electronic device comprising: a processor; and a memory coupled to the processor for storing instructions, which, when executed by the processor, cause the processor to perform the inference method of any one of claims 1 to 38.

44. A computer readable storage medium having stored thereon a computer program, wherein, The program is executed by the processor to implement the inference method of any one of claims 1 to 38. The program is executed by the processor to implement the inference method of any one of claims 1 to 38.

Images