Model sharing authorization method, model inference method, electronic device, and storage medium

Abstract

Provided in the present application are a model sharing authorization method, a model inference method, an electronic device, and a storage medium. The model sharing authorization method comprises: acquiring a model sharing acquisition request sent by a network function consumer, wherein the model sharing acquisition request at least carries consumer information of the network function consumer; sending the consumer information and a consumer authentication request to a security main body; and in response to authentication success information, which is fed back by the security main body on the basis of the consumer authentication request and the consumer information, sharing a model with the network function consumer.

Description

Model sharing authorization, model reasoning method, electronic device and storage medium Technical Field

[0001] The present application relates to the field of wireless communication technology, and in particular to model sharing authorization, model reasoning methods, electronic devices and storage media. Background Art

[0002] The Network Data Analytics Function (NWDAF) is a 5G Core Network (5GC) Network Function (NF) located in the control plane. It performs statistical data and machine learning tasks in the 5G system. The NWDAF interacts with different entities for different purposes: collecting data based on event subscriptions provided by the Access and Mobility Management Function (AMF), Session Management Function (SMF), User Plane Function (UPF), Policy Control Function (PCF), Unified Data Management (UDM), NSACF, Application Function (AF), and Operation Administration and Maintenance (OAM); retrieving information from data repositories; collecting location information data from the Location Services (LCS) system; retrieving information about NFs; delivering analytics to consumers on demand; providing batch data associated with analytics IDs; and providing accuracy information for analytics IDs. It also provides information about machine learning (ML) models, including accuracy or degradation of ML model accuracy. A single instance or multiple instances of NWDAF can also be deployed in a public land mobile network (PLMN). NWDAF can include the following logical functions: analysis logic function and model training logic function. The Data Collection Coordination Function (DCCF) is also an NF on the 5GC control plane. DCCF is responsible for coordinating the collection and distribution of data requested by NF consumers. It prevents the data source from processing multiple subscriptions to the same data and prevents multiple notifications containing the same information from being sent due to uncoordinated requests from data consumers. DCCF is applicable to NWDAFs that request data from data sources (such as for computational analysis); NF consumers that request analysis from NWDAF data sources; NF consumers that request data from the Analytics Data Repository Function (ADRF) data source; and ADRFs that receive data from NF data sources.

[0003] Supports ML model sharing between NWDAFs. For example, in Federated Learning (FL), the FL server distributes the initial model to the FL client and transmits the aggregated model to the service client after the federated learning process is completed. For example, an NWDAF containing the Model Training Logical Function (MTLF) transmits the model to an NWDAF containing the Analytics Logical Function (AnLF) for model inference. Another example is that after training the machine learning model, the NWDAF containing the MTLF sends it to the ADRF for storage. These methods all involve model sharing, but do not provide solutions to ensure the security of the model sharing process, which poses security risks.

[0004] Summary of the Invention

[0005] The embodiments of the present application aim to provide model sharing authorization, model reasoning methods, electronic devices and storage media to achieve the security of model sharing within a communication network and reduce the risk of privacy leakage caused by model sharing.

[0006] The present application provides a model sharing authorization method, which is applied to a network data analysis function unit and includes:

[0007] Obtain a model sharing acquisition request sent by a network function consumer, wherein the model sharing acquisition request carries at least consumer information of the network function consumer; send the consumer information and a consumer authentication request to a security subject; and share a model with the network function consumer in response to authentication success information fed back by the security subject based on the consumer authentication request and the consumer information.

[0008] The present application also provides a model sharing authorization method, which is applied to a security subject and includes:

[0009] Obtain consumer information and consumer authentication request sent by the network data analysis function unit; generate an authentication token based on a preset encryption algorithm and the consumer information; authenticate the network function consumer corresponding to the consumer information based on the authentication token; and feed back the authentication success information of the authentication to the network data analysis function unit so that the network data analysis function unit shares the model with the network function consumer.

[0010] The present application also provides a model sharing authorization method, which is applied to a network function consumer and includes:

[0011] Send a model sharing acquisition request to the network data analysis functional unit, wherein the model sharing acquisition request carries at least consumer information; perform authentication based on the authentication token sent by the security subject; obtain an authentication success response fed back by the network data analysis functional unit to obtain the model shared by the network data analysis functional unit.

[0012] The present application also provides a model reasoning method, which is applied to a network data analysis functional unit and includes:

[0013] Obtain an analysis acquisition request sent by a network function consumer; feed back model call information of the inference model corresponding to the analysis acquisition request to the network function consumer; obtain inference data collected by the network function consumer based on the model call information; generate an inference result based on the inference data and the inference model, and feed back the inference result to the network function consumer.

[0014] The present application also provides a model reasoning method, which is applied to a network function consumer and includes:

[0015] Send an analysis acquisition request to the network data analysis functional unit; receive the model call information fed back by the network data analysis functional unit, and collect inference data based on the model call information; send the inference data to the network data analysis functional unit so that the network data analysis functional unit generates an inference result based on the inference model and the inference data; receive the inference result fed back by the network data analysis functional unit.

[0016] An embodiment of the present application further provides an electronic device, comprising:

[0017] One or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the model sharing authorization method or the model reasoning method as described in any one of the embodiments of the present application.

[0018] An embodiment of the present application also provides a computer-readable storage medium, wherein the computer-readable storage medium stores one or more programs, and the one or more programs are executed by one or more processors to implement the model sharing authorization method or the model reasoning method as described in any one of the embodiments of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] FIG1 is a structural diagram of a 5GC architecture provided in an embodiment of the present application;

[0020] FIG2 is an example diagram of an NWDAF data collection architecture provided in an embodiment of the present application;

[0021] FIG3 is an exemplary diagram of an architecture for implementing NWDAF functions provided in an embodiment of the present application;

[0022] FIG4 is a structural example diagram of a DCCF data collection architecture provided in an embodiment of the present application;

[0023] FIG5 is an example diagram of an architecture in which a DCCF helps an NWDAF function to be implemented, provided in an embodiment of the present application;

[0024] FIG6 is a schematic diagram of a structure of a NWDAF model sharing provided in an embodiment of the present application;

[0025] FIG7 is a flowchart of a model sharing authorization method provided in an embodiment of the present application;

[0026] FIG8 is a flowchart of another model sharing authorization method provided in an embodiment of the present application;

[0027] FIG9 is a flowchart of another model sharing authorization method provided in an embodiment of the present application;

[0028] FIG10 is a flowchart of another model sharing authorization method provided in an embodiment of the present application;

[0029] FIG11 is an example diagram of a model sharing authorization method provided in an embodiment of the present application;

[0030] FIG12 is a flow chart of a model reasoning method provided in an embodiment of the present application;

[0031] FIG13 is a flowchart of another model reasoning method provided in an embodiment of the present application;

[0032] FIG14 is an example diagram of a model reasoning method provided in an embodiment of the present application;

[0033] FIG15 is a schematic structural diagram of a model sharing authorization device provided in an embodiment of the present application;

[0034] FIG16 is a schematic structural diagram of another model sharing authorization device provided in an embodiment of the present application;

[0035] FIG17 is a schematic structural diagram of another model sharing authorization device provided in an embodiment of the present application;

[0036] FIG18 is a schematic diagram of the structure of a model reasoning device provided in an embodiment of the present application;

[0037] FIG19 is a schematic diagram of the structure of another model reasoning device provided in an embodiment of the present application;

[0038] FIG20 is a schematic structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0039] In the subsequent description, suffixes such as "module", "component" or "unit" used to represent elements are only used to facilitate the description of this application and have no specific meaning in themselves. Therefore, "module", "component" or "unit" can be used interchangeably.

[0040] The model sharing authorization method and model reasoning method provided in the embodiments of the present application are applied to the 5G core network architecture, see Figure 1, which includes user equipment, radio access network (Radio Access Network, RAN), AMF, SMF, UPF, PCF, and UDM. Among them, RAN manages wireless resources, transmits user data received through N3 to UE, and transmits user data from UE through N3 interface. RAN maps between DRB (dedicated radio bearer) and QoS traffic in PDU session. AMF includes the following functions: registration management, connection management, reachability management and mobility management. This function also performs access verification and access authorization. AMF is a non-access stratum (NAS) security terminal, forwarding SM NAS between UE and SMF, etc. SMF functions include the following functions: session establishment, modification and release, UE IP address allocation and management (including optional authorization function), UP function selection and control, downlink data notification, etc. SMF controls UPF through N4 association. The SMF provides the UPF with Packet Detection Rules (PDRs), instructing it on how to detect user data traffic. It also provides Forwarding Action Rules (FARs), QoS Enforcement Rules (QERs), and URRs (Usage Reporting Rules) to instruct the UPF on how to forward user data traffic, perform QoS processing, and report usage for user data traffic detected using the PDRs. The UPF includes functions such as serving as an anchor point for intra- and inter-Radio Access Type (RAT) mobility, packet routing and forwarding, traffic usage reporting, user plane QoS processing, downlink packet buffering, and downlink data notification triggering. A GTP-U tunnel is used for the N3 interface between the RAN and the UPF. GTP-U tunnels are implemented per PDU session. For downlink traffic, the UPF uses the FARs received from the SMF to bind downlink traffic to QoS traffic within the PDU session GTP-U tunnel. For uplink traffic, the RAN transfers user plane traffic to the QoS flow identified by the UE. The PCF control plane function provides QoS policy rules to enforce these rules. The PCF converts AF requests into PCC rules applicable to the PDU session. The UDM performs the generation of 3GPP AKA authentication credentials, access authorization based on subscription data, UE registration with the serving NF (e.g., storing the service AMF for the UE and storing the PDU session storage SMF for the UE), and subscription management. The UDM accesses the UDR to retrieve UE subscription data and stores the UE context in the UDR. The UDM and UDR can be deployed together.

[0041] A single or multiple instances of the NWDAF can be deployed within a PLMN. The NWDAF can include the following logical functions: the ANLF and the MTLF. The ANLF is a logical function within the NWDAF responsible for reasoning, deriving analytical information, and exposing analytical services; the MTLF is used to train ML models and expose new training services. An NWDAF can include one ANLF, one MTLF, or both. As shown in Figure 2, the NWDAF can collect data from other network functional units (NFs). As shown in Figure 3, the NWDAF can also provide analysis, statistics, and prediction result sharing to other NFs. Furthermore, the DCCF can also be an NF on the 5GC control plane. The DCCF coordinates the collection and distribution of data requested by NF consumers, preventing data sources from processing multiple subscriptions to the same data and preventing multiple notifications containing the same information from being sent due to uncoordinated requests from data consumers. As shown in Figure 4, the NWDAF can leverage the DCCF to collect data from other NFs. Through the architecture shown in Figure 5, the DCCF facilitates the distribution of the analysis and prediction results provided by the NWDAF. As shown in Figure 6, the model sharing process can be performed via URL or FQDN. The process includes:

[0042] 1. NWDAF service consumers (i.e., NWDAFs that include analysis logic) subscribe, modify, or unsubscribe to a trained machine learning model associated with one or a set of analysis IDs by calling the Nnwdaf_MLModelProvision_Subscribe / Nnwdaf_MLModelProvision_Unsubscribe service operations.

[0043] 2. If the NWDAF service consumer subscribes to trained machine learning models associated with one or a set of analysis IDs, then the NWDAF, which contains the model training logic, will send a notification to the NWDAF service consumer by calling the Nnwdaf_MLModelProvision_Notify service operation. The notification includes: a set of unique machine learning model identifiers associated with each analysis ID requested by the service consumer and machine learning model information. The machine learning model information includes the machine learning model file address (such as a URL or FQDN).

[0044] However, the model sharing process currently does not provide a solution for how to securely share and use the model. To address this problem, an embodiment of the present application provides a model sharing authorization and model reasoning method.

[0045] FIG7 is a flowchart of a model sharing authorization method provided in an embodiment of the present application. The embodiment of the present application is applied to the case where model security authorization sharing is supported within a 5G system, and is generally applied to a network data analysis functional unit. Referring to FIG7 , the method provided in the embodiment of the present application includes:

[0046] 110. Obtain a model sharing acquisition request sent by a network function consumer, where the model sharing acquisition request carries at least consumer information of the network function consumer.

[0047] A network function consumer may be a unit that requests the use of a shared model. A network function consumer may include a network function unit, an OAM unit, etc., which has an NWDAF and 5GC with AnLF. A model sharing acquisition request may be a request from a network function consumer to a network data analysis function unit to obtain a model. The network data analysis function unit generally has an MTLF. Consumer information may be information that identifies attributes related to the network function consumer. Consumer information may include but is not limited to the consumer's number and the consumer's device type.

[0048] In an embodiment of the present application, a model sharing acquisition request sent by a network function consumer can be received, and the model sharing acquisition request carries at least consumer information of the network function consumer. In some embodiments, the model sharing acquisition request can also carry identification information of the model requested to be shared.

[0049] 120. Send the consumer information and the consumer authentication request to the security subject.

[0050] The security subject may be a security function unit within a wireless communication system, which may provide a security authentication service; the consumer authentication request may be a request for the security subject to authenticate the identity information of the network function consumer.

[0051] In an embodiment of the present application, the network data analysis function unit can obtain consumer information in the model sharing acquisition request and generate a consumer authentication request corresponding to the consumer information. The network data analysis function unit can transmit the consumer authentication request and the corresponding consumer information to the security subject so that the security subject can authenticate and verify the network function consumer.

[0052] 130. In response to the authentication success information fed back by the security subject based on the consumer authentication request and consumer information, share the model with the network function consumer.

[0053] The authentication success information may be the authentication verification result fed back by the security subject to the network data analysis function unit, and subtracting the success information may indicate that the authentication verification of the network function consumer is successful.

[0054] In an embodiment of the present application, after the security subject successfully verifies the network function consumer through the consumer authentication request and consumer information, it can generate authentication success information. The security subject can send the authentication success information to the network data analysis function unit. After receiving the authentication success information, the network data analysis function unit can share the model with the network function consumer.

[0055] In some application embodiments, the consumer information includes at least one of the following: network function consumer identification information and network function consumer device vendor information.

[0056] In an embodiment of the present application, the consumer information transmitted by the network function consumer includes at least network function consumer identification information and network function consumer device vendor information. The network function consumer identification information may be the unique identification information of the network function consumer, and may include a number or a function unit name. The network function consumer device vendor information may be information indicating the device vendor, and may indicate the hardware device type or hardware device category of the network function consumer.

[0057] In some application embodiments, in response to the security subject providing authentication success information based on the consumer authentication request and consumer information feedback, sharing the model with the network function consumer includes:

[0058] Receive the authentication success information fed back by the security subject, and send an authentication success response to the network function consumer, wherein the authentication success response carries at least the model sharing information of the model.

[0059] The successful authentication response can be the information that the network data analysis function unit feeds back to the network function consumer that the authentication is successful. The successful authentication response can also carry the model sharing information of the shared model requested by the network function consumer. The model sharing information can include information indicating the model, such as the model name and model number.

[0060] In an embodiment of the present application, the network data analysis function unit can generate an authentication success response after receiving the authentication success information fed back by the security subject. The authentication success response can carry the model sharing information of the network function consumer request model, so that the network function consumer can share the model based on the model sharing information. It can be understood that in some application embodiments, the model sharing information may not be carried by the authentication success response, and the network data analysis function unit can send the authentication success response and the model sharing information to the network function consumer separately.

[0061] In the above application embodiment, the model sharing information includes at least one of a model file, a model uniform resource locator, and a fully qualified domain name.

[0062] In an embodiment of the present application, the model sharing information may exist in the form of a model file, a model uniform resource locator, or a fully qualified domain name, and the network function consumer may use the shared model through the model sharing information.

[0063] FIG8 is a flowchart of another model sharing authorization method provided in an embodiment of the present application. The embodiment of the present application is applied to the case where model security authorization sharing is supported within a 5G system, and is generally applied to a security subject. Referring to FIG8 , the method provided in the embodiment of the present application includes:

[0064] 210. Obtain the consumer information and consumer authentication request sent by the network data analysis function unit.

[0065] In an embodiment of the present application, the security subject can receive consumer information and consumer authentication requests sent by the network data analysis function unit.

[0066] 220. Generate an authentication token based on a preset encryption algorithm and consumer information.

[0067] The preset encryption algorithm can be an algorithm for processing consumer information to generate an authentication token. The preset encryption algorithm includes but is not limited to the HMAC encryption algorithm, the JWT encryption algorithm, etc. The authentication token can be information generated by the preset encryption algorithm. The authentication token information can be generated by encrypting information including at least consumer information according to the preset encryption algorithm.

[0068] In an embodiment of the present application, the security subject can call a preset encryption algorithm to encrypt information including at least consumer information, and can use the generated ciphertext as an authentication token.

[0069] 230. Authenticate the network function consumer corresponding to the consumer information based on the authentication token.

[0070] In embodiments of the present application, the network function consumer corresponding to the consumer information can be determined, and the generated authentication token can be used to authenticate the network function consumer to determine whether the network function consumer is legitimate. This authentication process can include the security entity comparing the authentication token with the network function consumer's historical authentication tokens, or sending the authentication token to the network function consumer, having the network function consumer re-upload the consumer information, and the security entity performing authentication by verifying the re-uploaded consumer information.

[0071] 240. Feedback authentication success information to the network data analysis function unit so that the network data analysis function unit shares the model with the network function consumer.

[0072] In an embodiment of the present application, when the network function consumer is successfully authenticated, authentication success information can be generated and sent to the network data analysis function unit, so that the network data analysis function unit can share the model with the network function consumer.

[0073] FIG9 is a flowchart of another model sharing authorization method provided in an embodiment of the present application. The present embodiment of the present application is based on the above-mentioned embodiment and describes the authentication process in detail. Referring to FIG9 , the method provided in the embodiment of the present application includes:

[0074] 310. Obtain the consumer information and consumer authentication request sent by the network data analysis function unit.

[0075] 320. Generate an authentication token based on a preset encryption algorithm and consumer information.

[0076] 330. Send an authentication link carrying an authentication token to the network function consumer.

[0077] The authentication link may be a uniform resource locator used to authenticate the network function consumer, and the consumer information of the network function consumer may be obtained in the authentication link.

[0078] In an embodiment of the present application, the authentication token can be written into the authentication link, and the authentication link can be sent to the network function consumer, so that the network function consumer can perform authorization verification through the authentication link.

[0079] 340. Receive an authentication request sent by the network function consumer based on the authentication link.

[0080] The authentication request may be a regenerated authentication request by the network function consumer, which may carry the collected consumer information of the network function consumer. The consumer information carried in the authentication request may be actively uploaded by the network function consumer or automatically collected by the page corresponding to the authentication link.

[0081] In an embodiment of the present application, when a network function consumer receives an authentication link, it can generate an authentication request by accessing the authentication link, and the authentication request can be sent by the network function consumer to the security subject.

[0082] 350. Extract the authentication token carried in the authentication request and verify the authentication token according to a preset encryption algorithm.

[0083] In an embodiment of the present application, the authentication token can be extracted from the authentication request, and the consumer information carried in the authentication request can be encrypted according to a preset encryption algorithm. The encrypted ciphertext can be compared with the authentication token. When the ciphertext and the authentication token are the same, the verification is successful, otherwise the verification fails.

[0084] 360. Feedback the authentication success information to the network data analysis function unit so that the network data analysis function unit shares the model with the network function consumer.

[0085] In some application embodiments, it also includes: feeding back authentication failure information of the authentication to the network data analysis function unit, so that the network data analysis function unit sends an authentication failure response to the network function consumer, wherein the authentication failure information at least carries the reason for the authentication failure.

[0086] In an embodiment of the present application, the security subject may generate authentication failure information when determining that the authentication has failed, and may send the authentication failure information to the network data analysis function unit. When the network data analysis function unit receives the authentication failure information, it may send an authentication failure response to the network function consumer to indicate that the authentication verification has failed. The authentication failure information and the authentication failure response may carry at least the reason for the authentication failure.

[0087] In some other application embodiments, the method further includes: sending an authentication failure response to the network function consumer, wherein the authentication failure response carries at least the reason for the authentication failure.

[0088] In an embodiment of the present application, when the security subject determines that the authentication has failed, it can directly send an authentication failure response to the network function consumer, indicating that the model sharing authentication of the network function consumer has failed. It can be understood that the authentication failure response carries at least the reason for the authentication failure, so that the network function consumer knows the reason for the failure of the model sharing authentication.

[0089] FIG10 is a flowchart of another model sharing authorization method provided in an embodiment of the present application. The embodiment of the present application is applied to the case where model security authorization sharing is supported within a 5G system, and is generally applied to network function consumers. Referring to FIG10 , the method provided in the embodiment of the present application includes:

[0090] 410. Send a model sharing acquisition request to the network data analysis function unit, where the model sharing acquisition request carries at least consumer information.

[0091] In an embodiment of the present application, when a network function consumer needs to share a model, he or she can generate a model sharing acquisition request according to his or her own consumer information, and send the model sharing acquisition request to the network data analysis function unit to request the shared model.

[0092] 420. Perform authentication based on the authentication token sent by the security subject.

[0093] In an embodiment of the present application, a network function consumer can receive an authentication token sent by a security entity and can perform authentication through the authentication token. The authentication method may include but is not limited to: the security entity compares the authentication token with the historical authentication token of the network function consumer, or the authentication token is sent to the network function consumer, and the network function consumer re-uploads the consumer information, and the security entity performs authentication by verifying the re-uploaded consumer information, etc.

[0094] 430. Obtain the authentication success response fed back by the network data analysis functional unit to obtain the model shared by the network data analysis functional unit.

[0095] In an embodiment of the present application, when the network function consumer is successfully authenticated, authentication success information can be generated. The network function consumer can receive the authentication success information and share the model through the model sharing information carried in the authentication success information.

[0096] In some application embodiments, the consumer information includes at least one of the following: network function consumer identification information and network function consumer device vendor information.

[0097] Based on the above application embodiment, authentication is performed according to the authentication token sent by the security subject, including: obtaining the authentication link sent by the security subject, wherein the authentication link carries at least the authentication token; and sending an authentication request to the security subject according to the authentication link.

[0098] In an embodiment of the present application, the network function consumer can receive an authentication link sent by the security entity, and the authentication link can carry an authentication token produced by the security entity. The network function consumer can feedback an authentication request to the security entity by accessing the authentication link. It can be understood that the authentication request can carry consumer information used for authorization verification of the component sharing model. The consumer information can be uploaded by the network function consumer or automatically collected and generated by the web page corresponding to the authentication link.

[0099] Based on the above application embodiment, obtaining the authentication success response fed back by the network data analysis functional unit to obtain the model shared by the network data analysis functional unit includes:

[0100] Receive an authentication success response sent by the network data analysis functional unit, wherein the authentication success response carries at least model sharing information of the model.

[0101] In an embodiment of the present application, the network data analysis function unit can send an authentication success response to the network function consumer. The authentication success response can be generated by the network data analysis function unit, and the authentication success response can carry model sharing information. The network function consumer can share the model of the network data analysis function unit through the model sharing information.

[0102] In some application embodiments, the model sharing information includes at least one of a model file, a model uniform resource locator, and a fully qualified domain name.

[0103] In some application embodiments, the method further includes: receiving an authentication failure response sent by the network data analysis functional unit, wherein the authentication failure response carries at least an authentication identification reason.

[0104] In an embodiment of the present application, the authentication failure response may be sent by the network data analysis function unit, and the network function consumer may receive the authentication failure response sent by the network data analysis function unit that carries at least the reason for the authentication failure.

[0105] In some other application embodiments, the method further includes: receiving an authentication failure response sent by a security subject, wherein the authentication failure response carries at least a reason for the authentication failure.

[0106] The authentication failure response can also be sent directly by the security subject. The authentication failure response can carry the reason for the authentication failure. The network function consumer can receive the authentication failure response sent by the security subject.

[0107] In an exemplary embodiment, FIG11 is an example diagram of a model sharing authorization method provided in an embodiment of the present application. Before the 5GC system shares the model, the consumer network element can be authenticated to ensure the security of the model receiving entity. Referring to FIG11 , the model sharing authorization method includes:

[0108] 1. The NF consumer requests a model from the NWDAF that includes the MTLF. The NF consumer is the NWDAF that includes the AnLF, 5GC NF, or OAM. The request for the model can carry the Analytic ID and its own information, including but not limited to identification information and equipment vendor information.

[0109] 2. The model owner sends the NF consumer's information (identification information, equipment vendor information) and the consumer authentication request to the security subject (another network element).

[0110] 3. The security subject generates an authentication token based on the user information and encryption algorithm. The encryption algorithm includes but is not limited to HMAC, JWT, etc.

[0111] 4. The security subject sends an authentication URL link with a token to the NF consumer.

[0112] 5.NF Consumer sends the request again using the authentication URL.

[0113] 6. After receiving the request, the security subject parses the authentication token in the URL and verifies it using the same encryption algorithm.

[0114] 7a / 7b. The security subject sends an NF consumer authentication success response to the model owner NWDAF, and then the model owner NWDAF sends an authentication success response to the NF consumer. The authentication success response can carry the model file, the URL or FQDN of the model file, etc.

[0115] 7c / 7d. The security subject sends an authentication failure response to the model owner NWDAF, including the authentication failure reason. The model owner NWDAF passes this authentication failure message to the NF consumer.

[0116] Alternatively, the security subject directly sends an authentication failure response to the consumer, including the reason for the authentication failure.

[0117] FIG12 is a flow chart of a model reasoning method provided in an embodiment of the present application. The embodiment of the present application is applicable to the case of model sharing and is generally applied to a network data analysis functional unit. Referring to FIG12 , the method provided in the embodiment of the present application includes:

[0118] 510. Obtain the analysis acquisition request sent by the network function consumer.

[0119] The analysis acquisition request may be a request to call an inference model in the network data analysis functional unit, and the sharing acquisition request may carry indication information of the inference model.

[0120] In an embodiment of the present application, the network data analysis function unit can receive an analysis acquisition request sent by a network function consumer.

[0121] 520. Feedback the model call information of the inference model corresponding to the analysis request to the network function consumer.

[0122] The inference model can be a machine learning model that has been trained within the network data analysis model. The model call information can be information for calling the inference model. The model call information can include a uniform resource locator, interface or encrypted interface for calling the inference model. The inference call information can include inference input information for using the inference model. The inference input information can be the input requirement of the inference model.

[0123] In an embodiment of the present application, the network data analysis function unit can generate model call information in response to the analysis acquisition request and send the model call information to the network function consumer.

[0124] 530. Obtain inference data collected by network function consumers based on model call information.

[0125] The inference data may be input information used for inference in the inference model, and the inference data may be collected and generated by the network function consumer according to the model call information.

[0126] In an embodiment of the present application, the network function consumer can collect inference data according to the model call information, and the network function consumer can feed back the collected inference data to the network data analysis function unit, and the network data analysis function unit receives the inference data.

[0127] 540. Generate inference results based on the inference data and the inference model, and feed back the inference results to the network function consumer.

[0128] In an embodiment of the present application, the network data analysis functional unit may invoke an inference model and input inference data into the inference model for inference, thereby generating an inference result. The network data analysis functional unit may then feed the generated inference result back to the network function consumer. It is understood that the inference result may be returned to the network function consumer via an interface, a uniform resource locator, an encrypted interface, or the like, or the inference result may be fed back directly to the network function consumer.

[0129] Based on the above application embodiment, the model call information of the inference model corresponding to the analysis request is fed back to the network function consumer, including:

[0130] Extract the analysis identifier carried in the analysis acquisition request; and feed back model call information to the network function consumer based on the analysis identifier; wherein the model call information at least includes the inference data requirements of the inference model.

[0131] The analysis identifier may be information indicating the reasoning model, and the analysis identifier may include information such as the model name and model number of the reasoning model.

[0132] In an embodiment of the present application, the analysis identifier can be extracted from the analysis acquisition request, and the corresponding reasoning model can be determined according to the analysis identifier. Model call information corresponding to the reasoning model can be generated. The model call information can carry the reasoning data requirements of the reasoning model. The reasoning data requirements can include data input requirements for the reasoning model. The reasoning data requirements can be used by network function consumers to collect reasoning data.

[0133] In some application embodiments, the type of model call information includes at least one of the following: an encrypted interface, a uniform resource locator link, and a signaling service.

[0134] In an embodiment of the present application, the model call information can exist in the form of an encrypted interface, a uniform resource locator link, a signaling service, etc., and the carried inference data requirements can be transmitted to the network function consumer.

[0135] FIG13 is a flowchart of another model reasoning method provided by an embodiment of the present application. The embodiment of the present application is applicable to the case of model sharing, and is generally applied to network function consumers. Referring to FIG13 , the method provided by the embodiment of the present application includes:

[0136] 610. Send an analysis acquisition request to the network data analysis function unit.

[0137] In an embodiment of the present application, the network function consumer may send an analysis acquisition request to the network data analysis function unit to use the inference model within the network data analysis function unit.

[0138] 620. Receive model call information fed back by the network data analysis functional unit, and collect inference data based on the model call information.

[0139] In an embodiment of the present application, the network function consumer can receive model call information sent by the network data analysis function unit and collect inference data according to the model call information.

[0140] 630. Send the inference data to the network data analysis functional unit so that the network data analysis functional unit generates an inference result according to the inference model and the inference data.

[0141] The network function consumer can send the collected inference data to the network data analysis function unit, and the network data analysis function unit can generate inference results according to the inference data and the inference model. It can be understood that the inference model can be the inference model indicated by the analysis acquisition request sent by the network function consumer.

[0142] 640. Receive the inference result fed back by the network data analysis functional unit.

[0143] In an embodiment of the present application, the network data analysis function unit can feed back the generated inference results to the network function consumer.

[0144] Based on the above application embodiment, the inference data is collected according to the model call information, including:

[0145] Extract the inference data requirements included in the model call information; and collect inference data according to the inference data requirements.

[0146] In an embodiment of the present application, the model call information may include inference data requirements. When the network function consumer receives the model call information fed back by the network data analysis function unit, it can extract the inference data requirements in the model call information and collect the inference data according to the inference data requirements in order to use the inference model.

[0147] Based on the above application embodiment, the type of model call information includes at least one of the following: an encrypted interface, a uniform resource locator link, and a signaling service.

[0148] In an exemplary embodiment, FIG14 is an example diagram of a model inference method provided by an embodiment of the present application. The network function consumer directly uses the encrypted interface of the model to call the inference function of the inference model. It can be understood that the call of the inference model can also be implemented using a uniform resource locator, a signaling service, or an unencrypted interface. The network function consumer can call the inference model instead of directly sharing the inference model. In addition, the service area of ​​the NWDAF that owns the model may not have the conditions for collecting certain model inference data, but the inference data can be collected within the consumer's service area. Referring to FIG14, the model inference method provided by the embodiment of the present application includes:

[0149] 1. The network function consumer sends an analysis acquisition request to the NWDAF. The analysis acquisition request includes an analysis ID. The network function consumer may include a 5GC NF or OAM.

[0150] 2. NWDAF provides network function consumers with an encrypted interface or Uniform Resource Locator (URL) link containing model inference input information. The encrypted interface can also be used here, and the signaling service within the 5GC can be used for transmission.

[0151] 3. The network function consumer reads the inference input information provided in the interface and collects corresponding inference data according to the inference input information.

[0152] 4. If the NWDAF sends an encrypted interface to the consumer in operation 2, the consumer returns the collected inference data to the NWDAF via the encrypted interface. If the encrypted interface is not used, the inference input data is returned using traditional 5GC signaling.

[0153] 5. NWDAF uses the trained model for inference.

[0154] 6. NWDAF sends the inference results to the network function consumer.

[0155] Figure 15 is a schematic diagram of the structure of a model sharing authorization device provided in an embodiment of the present application. The device can execute the model sharing authorization method provided in any embodiment of the present application and has the corresponding functional modules and beneficial effects of the execution method. The device can be implemented by software and / or hardware and is generally applied to the network data analysis function unit. As shown in Figure 15, the device provided in an embodiment of the present application includes:

[0156] The request acquisition module 710 is configured to acquire a model sharing acquisition request sent by a network function consumer, wherein the model sharing acquisition request at least carries the consumer information of the network function consumer.

[0157] The security authentication module 720 is used to send the consumer information and the consumer authentication request to the security subject.

[0158] The success response module 730 is used to respond to the authentication success information fed back by the security subject according to the consumer authentication request and the consumer information, and share the model with the network function consumer.

[0159] Based on the above application embodiment, the consumer information in the device includes at least one of the following: network function consumer identification information and network function consumer equipment vendor information.

[0160] Based on the above-mentioned application embodiment, the device success response module 730 is used to: receive the authentication success information fed back by the security subject, and send an authentication success response to the network function consumer, wherein the authentication success response carries at least the model sharing information of the model.

[0161] Based on the above application embodiment, the model sharing information includes at least one of a model file, a model uniform resource locator, and a fully qualified domain name.

[0162] Figure 16 is a schematic diagram of the structure of another model sharing authorization device provided in an embodiment of the present application. The device can execute the model sharing authorization method provided in any embodiment of the present application and has the corresponding functional modules and beneficial effects of the execution method. The device can be implemented by software and / or hardware and is generally applied to security subjects. As shown in Figure 16, the device provided in an embodiment of the present application includes:

[0163] The authentication acquisition module 810 is used to obtain the consumer information and the consumer authentication request sent by the network data analysis function unit.

[0164] The token generation module 820 is used to generate an authentication token according to a preset encryption algorithm and the consumer information.

[0165] The authentication execution module 830 is used to authenticate the network function consumer corresponding to the consumer information according to the authentication token.

[0166] The success feedback module 840 is used to feed back the authentication success information of the authentication to the network data analysis function unit so that the network data analysis function unit shares the model with the network function consumer.

[0167] In some application embodiments, the authentication execution module 830 includes:

[0168] A link sending unit is used to send an authentication link carrying the authentication token to the network function consumer.

[0169] A request receiving unit is used to receive an authentication request sent by the network function consumer based on the authentication link.

[0170] The verification and comparison unit is used to extract the authentication token carried in the authentication request and verify the authentication token according to the preset encryption algorithm.

[0171] In other application embodiments, it also includes: a first failure feedback module, which is used to feedback the authentication failure information of the authentication to the network data analysis function unit, so that the network data analysis function unit sends an authentication failure response to the network function consumer, wherein the authentication failure information at least carries the reason for the authentication failure.

[0172] In some other application embodiments, it also includes: a second failure feedback module, used to send an authentication failure response to the network function consumer, wherein the authentication failure response at least carries the reason for the authentication failure.

[0173] Figure 17 is a schematic diagram of the structure of another model sharing authorization device provided in an embodiment of the present application. The device can execute the model sharing authorization method provided in any embodiment of the present application and has the corresponding functional modules and beneficial effects of the execution method. The device can be implemented by software and / or hardware and is generally used for network function consumers. As shown in Figure 17, the device provided in an embodiment of the present application includes:

[0174] The sharing request module 910 is used to send a model sharing acquisition request to the network data analysis function unit, wherein the model sharing acquisition request at least carries consumer information.

[0175] The authentication execution module 920 is used to perform authentication according to the authentication token sent by the security subject.

[0176] The model acquisition module 930 is used to obtain the authentication success response fed back by the network data analysis functional unit to obtain the model shared by the network data analysis functional unit.

[0177] In some application embodiments, the in-device consumer information includes at least one of the following: network function consumer identification information and network function consumer device vendor information.

[0178] In some application embodiments, the in-device authentication execution module 920 is used to: obtain an authentication link sent by the security subject, wherein the authentication link carries at least the authentication token; and send an authentication request to the security subject according to the authentication link.

[0179] In some application embodiments, the in-device model acquisition module 930 is used to: receive the authentication success response sent by the network data analysis functional unit, wherein the authentication success response carries at least the model sharing information of the model.

[0180] In some other application embodiments, the intra-device model sharing information includes at least one of a model file, a model uniform resource locator, and a fully qualified domain name.

[0181] In some other application embodiments, the device also includes a first failure module for receiving an authentication failure response sent by the network data analysis functional unit, wherein the authentication failure response carries at least a reason for the authentication failure.

[0182] In some other application embodiments, the apparatus further includes a second failure module configured to receive an authentication failure response sent by the security subject, wherein the authentication failure response carries at least a reason for the authentication failure.

[0183] Figure 18 is a schematic diagram of the structure of a model inference device provided in an embodiment of the present application. The device can execute the model inference method provided in any embodiment of the present application and has the corresponding functional modules and beneficial effects of the execution method. The device can be implemented by software and / or hardware and is generally used in a network data analysis functional unit. As shown in Figure 18, the device provided in an embodiment of the present application includes:

[0184] The analysis acquisition request module 1010 is used to acquire the analysis acquisition request sent by the network function consumer.

[0185] The model calling module 1020 is used to feed back the model calling information of the inference model corresponding to the analysis acquisition request to the network function consumer.

[0186] The inference data module 1030 is used to obtain the inference data collected by the network function consumer based on the model call information.

[0187] The result reasoning module 1040 is used to generate a reasoning result based on the reasoning data and the reasoning model, and feed back the reasoning result to the network function consumer.

[0188] Based on the above application embodiment, the in-device model calling module 1020 is used to:

[0189] Extracting the analysis identifier carried in the analysis acquisition request; feeding back the model call information to the network function consumer according to the analysis identifier; wherein the model call information at least includes the inference data requirements of the inference model.

[0190] Based on the above application embodiment, the type of the model calling information within the device includes at least one of the following: an encrypted interface, a uniform resource locator link, and a signaling service.

[0191] Figure 19 is a schematic diagram of the structure of another model reasoning device provided in an embodiment of the present application. The device can execute the model reasoning method provided in any embodiment of the present application and has the corresponding functional modules and beneficial effects of the execution method. The device can be implemented by software and / or hardware and is generally used for network function consumers. As shown in Figure 19, the device provided in an embodiment of the present application includes:

[0192] The request sending module 1110 is used to send an analysis acquisition request to the network data analysis functional unit.

[0193] The data collection module 1120 is used to receive the model call information fed back by the network data analysis functional unit and collect inference data based on the model call information.

[0194] The data uploading module 1130 is used to send the inference data to the network data analysis functional unit, so that the network data analysis functional unit generates an inference result according to the inference model and the inference data.

[0195] The inference result module 1140 is used to receive the inference result fed back by the network data analysis functional unit.

[0196] In an embodiment of the present application, the data collection module 1120 is used to: extract the inference data requirements included in the model call information; and collect the inference data according to the inference data requirements.

[0197] In the embodiment of the present application, the type of the model calling information within the device includes at least one of the following: an encrypted interface, a uniform resource locator link, and a signaling service.

[0198] Figure 20 is a structural diagram of an electronic device provided in an embodiment of the present application, which electronic device includes a processor 10 and a memory 11; the number of processors 10 in the electronic device can be one or more, and Figure 20 takes one processor 10 as an example; the processor 10 and the memory 11 in the electronic device can be connected via a bus or other means, and Figure 20 takes connection via a bus as an example.

[0199] The memory 11, as a computer-readable storage medium, can be used to store software programs, computer executable programs and modules, such as the module corresponding to the model sharing authorization device in the embodiment of the present application. The processor 10 executes various functional applications and data processing of the electronic device by running the software programs, instructions and modules stored in the memory 11, that is, implements the above-mentioned model sharing authorization method. Or, such as the module corresponding to the model reasoning device in the embodiment of the present application. The processor 10 executes various functional applications and data processing of the electronic device by running the software programs, instructions and modules stored in the memory 11, that is, implements the above-mentioned model reasoning method.

[0200] The memory 11 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created based on the use of the electronic device, etc. In addition, the memory 11 may include a high-speed random access memory and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 11 may further include a memory remotely located relative to the processor 10, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0201] The present application also provides a storage medium containing computer-executable instructions. When the computer-executable instructions are executed by a computer processor, they are used to perform a model sharing authorization method. The method is applied to a network data analysis functional unit, and the method includes:

[0202] Obtain a model sharing acquisition request sent by a network function consumer, wherein the model sharing acquisition request carries at least consumer information of the network function consumer; send the consumer information and a consumer authentication request to a security subject; and in response to authentication success information fed back by the security subject based on the consumer authentication request and the consumer information, share the model with the network function consumer. Alternatively, the computer-executable instructions, when executed by a computer processor, are used to execute a model sharing authorization method, which is applied to a security subject and includes: obtaining consumer information and a consumer authentication request sent by a network data analysis function unit; generating an authentication token based on a preset encryption algorithm and the consumer information; authenticating the network function consumer corresponding to the consumer information based on the authentication token; and feeding back authentication success information to the network data analysis function unit so that the network data analysis function unit shares the model with the network function consumer.

[0203] Alternatively, the computer executable instructions, when executed by a computer processor, are used to execute a model sharing authorization method, which is applied to a network function consumer, and includes: sending a model sharing acquisition request to a network data analysis function unit, wherein the model sharing acquisition request carries at least consumer information; performing authentication based on an authentication token sent by a security subject; and obtaining an authentication success response fed back by the network data analysis function unit to obtain the model shared by the network data analysis function unit.

[0204] Alternatively, the computer executable instructions, when executed by a computer processor, are used to perform a model inference method, which is applied to a network data analysis function unit, and the method includes: obtaining an analysis acquisition request sent by a network function consumer; feeding back model call information of the inference model corresponding to the analysis acquisition request to the network function consumer; obtaining inference data collected by the network function consumer based on the model call information; generating an inference result based on the inference data and the inference model, and feeding back the inference result to the network function consumer.

[0205] Alternatively, the computer executable instructions are used to perform a model inference method when executed by a computer processor, and the method is applied to a network function consumer, and the method includes: sending an analysis acquisition request to a network data analysis function unit; receiving model call information fed back by the network data analysis function unit, and collecting inference data based on the model call information; sending the inference data to the network data analysis function unit so that the network data analysis function unit generates an inference result based on the inference model and the inference data; and receiving the inference result fed back by the network data analysis function unit.

[0206] Through the above description of the implementation methods, those skilled in the art can clearly understand that the present application can be implemented with the help of software and necessary general-purpose hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application can essentially be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0207] In the embodiment of the above-mentioned device, the various units and modules included are divided only according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be realized; in addition, the names of the various functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application.

[0208] Those skilled in the art will appreciate that all or some of the operations, devices, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, or appropriate combinations thereof.

[0209] In a hardware implementation, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or operation may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. The corresponding software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or temporary medium). As known to those of ordinary skill in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, as is well known to those skilled in the art, communication media typically embodies computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

Claims

1. A model sharing authorization method, applied to a network data analysis functional unit, comprising: Obtaining a model sharing acquisition request sent by a network function consumer, wherein the model sharing acquisition request carries at least consumer information of the network function consumer; Sending the consumer information and the consumer authentication request to the security subject; In response to authentication success information fed back by the security subject according to the consumer authentication request and the consumer information, a model is shared with the network function consumer.

2. The method according to claim 1, wherein: The consumer information includes at least one of the following: network function consumer identification information and network function consumer device vendor information.

3. The method according to claim 1, wherein: The step of sharing the model with the network function consumer in response to authentication success information fed back by the security subject according to the consumer authentication request and the consumer information includes: Receive the authentication success information fed back by the security subject, and send an authentication success response to the network function consumer, wherein the authentication success response carries at least the model sharing information of the model.

4. The method according to claim 3, wherein: The model sharing information includes at least one of a model file, a model uniform resource locator, and a fully qualified domain name.

5. A model sharing authorization method, applied to a security subject, comprising: Obtaining consumer information and consumer authentication requests sent by the network data analysis function unit; Generate an authentication token based on a preset encryption algorithm and the consumer information; authenticating the network function consumer corresponding to the consumer information according to the authentication token; Feedback the authentication success information of the authentication to the network data analysis function unit so that the network data analysis function unit shares the model with the network function consumer.

6. The method according to claim 5, wherein: authenticating the network function consumer corresponding to the consumer information according to the authentication token, including: Sending an authentication link carrying the authentication token to the network function consumer; receiving an authentication request sent by the network function consumer based on the authentication link; Extract the authentication token carried in the authentication request, and verify the authentication token according to the preset encryption algorithm.

7. The method according to claim 5, further comprising: Feedback the authentication failure information of the authentication to the network data analysis function unit, so that the network data analysis function unit sends an authentication failure response to the network function consumer, wherein the authentication failure information at least carries the reason for the authentication failure.

8. The method according to claim 5, further comprising: An authentication failure response is sent to the network function consumer, wherein the authentication failure response carries at least an authentication failure reason.

9. A model sharing authorization method, applied to a network function consumer, comprising: Sending a model sharing acquisition request to the network data analysis function unit, wherein the model sharing acquisition request carries at least consumer information; Authentication is performed based on the authentication token sent by the security subject; Obtain the authentication success response fed back by the network data analysis functional unit to obtain the model shared by the network data analysis functional unit.

10. The method according to claim 9, wherein: The consumer information includes at least one of the following: network function consumer identification information and network function consumer device vendor information.

11. The method according to claim 9, wherein: The authentication according to the authentication token sent by the security subject includes: Obtaining an authentication link sent by the security subject, wherein the authentication link carries at least the authentication token; An authentication request is sent to the security subject according to the authentication link.

12. The method according to claim 9, wherein: The obtaining of the authentication success response fed back by the network data analysis functional unit to obtain the model shared by the network data analysis functional unit includes: Receive the authentication success response sent by the network data analysis functional unit, wherein the authentication success response carries at least the model sharing information of the model.

13. The method according to claim 12, wherein: The model sharing information includes at least one of a model file, a model uniform resource locator, and a fully qualified domain name.

14. The method according to claim 9, further comprising: Receive an authentication failure response sent by the network data analysis function unit, wherein the authentication failure response at least carries a reason for the authentication failure.

15. The method according to claim 9, further comprising: An authentication failure response sent by the security subject is received, wherein the authentication failure response carries at least an authentication failure reason.

16. A model reasoning method, applied to a network data analysis functional unit, comprising: Get analytics acquisition requests sent by network function consumers; Feedback to the network function consumer the model call information of the inference model corresponding to the analysis acquisition request interest; Obtaining inference data collected by the network function consumer based on the model call information; An inference result is generated according to the inference data and the inference model, and the inference result is fed back to the network function consumer.

17. The method according to claim 16, wherein: Feedback of model call information of the inference model corresponding to the analysis acquisition request to the network function consumer includes: Extracting the analysis identifier carried in the analysis acquisition request; Feedback the model calling information to the network function consumer according to the analysis identifier; The model calling information at least includes the inference data requirements of the inference model.

18. The method according to claim 16 or 17, wherein: The type of the model call information includes at least one of the following: Encrypted interface, Uniform Resource Locator link, signaling service.

19. A model reasoning method, applied to a network function consumer, comprising: Sending an analysis acquisition request to the network data analysis function unit; receiving model call information fed back by the network data analysis functional unit, and collecting inference data according to the model call information; Sending the inference data to the network data analysis functional unit, so that the network data analysis functional unit generates an inference result according to the inference model and the inference data; Receive the inference result fed back by the network data analysis functional unit.

20. The method according to claim 19, wherein: The collecting of inference data according to the model calling information includes: Extracting the inference data requirements included in the model call information; The inference data is collected according to the inference data requirement.

21. The method according to claim 19, wherein: The type of the model call information includes at least one of the following: Encrypted interface, Uniform Resource Locator link, signaling service.

22. An electronic device comprising: at least one processor; a memory configured to store at least one program; When the at least one program is executed by the at least one processor, the at least one processor The device implements the model sharing authorization method or the model reasoning method as described in any one of claims 1-21.

23. A computer-readable storage medium storing at least one program, wherein the at least one program is executed by at least one processor to implement the model sharing authorization method or the model reasoning method as described in any one of claims 1 to 21.

Images