A communication method, a communication device, and a communication system

Abstract

This application provides a communication method, communication device, and communication system. The method includes: an inference network element sending a first request message to a training network element, the first request message including analysis type identification information, wherein the training network element and the inference network element are of different vendor types but the model deployment platform types are the same; receiving a first response message from the training network element, the first response message including an encrypted model or the address information of the encrypted model; obtaining an encrypted analysis result based on the encrypted model; and obtaining a decrypted analysis result based on the encrypted analysis result. This scheme allows the inference network element and the training network element to be deployed by different vendors, breaking the limitation in existing schemes where models can only be deployed by the same vendor.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a communication method, communication device and communication system. Background Technology

[0002] The training network element can train the model and provide the trained model to the inference network element. The inference network element inputs the data to be analyzed into the model to perform inference and obtain the analysis results.

[0003] To support the selection of appropriate training elements by the inference network element, the current practice is to configure the address information of one or more training elements and the identifier information of the analysis types supported by each training element locally on the inference network element. The inference network element can then select a training element from these one or more training elements that can provide the model based on the analysis type corresponding to the data to be analyzed. Furthermore, the inference network element and each training element are from the same vendor and use the same model deployment platform.

[0004] However, limiting inference and training network elements to the same vendor makes it impossible to share models across vendors. Summary of the Invention

[0005] This application provides a communication method, communication device, and communication system to enable cross-vendor model sharing.

[0006] In a first aspect, embodiments of this application provide a communication method, which can be executed by an inference network element or a module (such as a chip) applied in the inference network element. Taking the execution of the communication method by an inference network element as an example, the method includes: the inference network element sending a first request message to a training network element, the first request message including identification information of the analysis type, the first request message being used to request a model that supports the analysis type, the training network element and the inference network element having different vendor types, and the inference network element and the training network element having the same model deployment platform type; the inference network element receiving a first response message from the training network element, the first response message including an encrypted model or the address information of the encrypted model, the encrypted model supporting the analysis type; the inference network element obtaining an encrypted analysis result based on the encrypted model; and the inference network element obtaining a decrypted analysis result based on the encrypted analysis result.

[0007] This solution deploys inference and training network elements from different vendors, but they use the same model deployment platform, breaking the limitation of existing solutions that restrict model sharing to the same vendor. It provides a process for encrypted cross-vendor model distribution, enhancing the ability of training network elements to distribute models encryptedly and mitigating the risk of inference network element deployment vendors stealing model framework and parameter information.

[0008] In one possible implementation, the inference network element sends the encrypted analysis result to the training network element; the inference network element receives the decrypted analysis result from the training network element.

[0009] This scheme, since the training network elements are the encrypted network elements of the model, allows the training network elements to decrypt the encrypted analysis results, thus achieving accurate decryption of the encrypted analysis results.

[0010] In one possible implementation, the first response message further includes a first instruction message that instructs the training network element to decrypt the encrypted analysis result.

[0011] According to this scheme, the inference network element can accurately determine that the network element encrypting the encrypted analysis result is the training network element based on the first instruction information.

[0012] In one possible implementation, the inference network element sends the encrypted analysis result and association identifier to the training network element. The association identifier is used by the training network element to determine the encryption algorithm corresponding to the encrypted model.

[0013] This scheme, through association identifiers, enables training network elements to accurately obtain the encryption algorithm corresponding to the encrypted model, and thus accurately know the decryption algorithm to be used for decryption analysis, thereby improving decryption efficiency.

[0014] In one possible implementation, the first response message further includes the address information of the first network element; the inference network element sends the encrypted analysis result to the first network element based on the address information of the first network element; the inference network element receives the decrypted analysis result from the first network element.

[0015] In this scheme, when the training network element is unable to decrypt the encrypted analysis result, the first network element can decrypt the encrypted analysis result, thus ensuring that the inference network element can obtain the decrypted analysis result.

[0016] In one possible implementation, the inference network element sends the encrypted analysis result and association identifier to the first network element based on the address information of the first network element. The association identifier is used by the first network element to determine the encryption algorithm corresponding to the encrypted model.

[0017] This scheme, through association identification, enables the first network element to accurately obtain the encryption algorithm corresponding to the encrypted model, and thus accurately know the decryption algorithm to be used for decryption analysis, thereby improving decryption efficiency.

[0018] In one possible implementation, the first response message further includes second indication information, which indicates the data type of the input data corresponding to the encrypted model.

[0019] This scheme, through the second instruction information, enables the inference network elements to preprocess the input data accordingly, thereby obtaining data to be analyzed that meets the requirements, which can improve the efficiency of data inference.

[0020] In one possible implementation, the first request message also includes the vendor type of the inference network element and the type of the model deployment platform of the inference network element.

[0021] This scheme, by carrying the vendor type and model deployment platform type of the inference network element in the first request message, enables the training network element to determine whether the vendor types of the training network element and the inference network element are the same, and whether the model deployment platform types of the training network element and the inference network element are the same. This facilitates the training network element in selecting an appropriate method to provide data inference functionality to the inference network element, thereby improving the efficiency of data inference.

[0022] In one possible implementation, before sending the first request message to the training network element, the inference network element sends a second request message to the data management network element. The second request message includes identification information of the analysis type and is used to request a network element that supports the analysis type. The inference network element receives a second response message from the data management network element, which includes the address information of the training network element.

[0023] This scheme allows inference network elements to request and discover training network elements from data management network elements, enabling accurate discovery of training network elements that can provide models.

[0024] Secondly, embodiments of this application provide a communication method, which can be executed by a training network element or a module (such as a chip) applied in the training network element. Taking the execution of the communication method by a training network element as an example, the method includes: the training network element receiving a first request message from an inference network element, the first request message including identification information of an analysis type, the first request message being used to request a model that supports the analysis type, the training network element and the inference network element having different vendor types, and the inference network element and the training network element having the same model deployment platform type; the training network element sending a first response message to the inference network element, the first response message including an encrypted model or the address information of the encrypted model; the training network element receiving an encrypted analysis result from the inference network element, the encrypted analysis result being obtained based on the encrypted model; the training network element decrypting the encrypted analysis result to obtain a decrypted analysis result; and the training network element sending the decrypted analysis result to the inference network element.

[0025] This solution deploys inference and training network elements from different vendors, but they use the same model deployment platform, breaking the limitation of existing solutions that restrict model sharing to the same vendor. It provides a process for encrypted cross-vendor model distribution, enhancing the ability of training network elements to distribute models encryptedly and mitigating the risk of inference network element deployment vendors stealing model framework and parameter information.

[0026] In one possible implementation, the first request message further includes the vendor type of the inference network element and the type of the model deployment platform of the inference network element; before the training network element sends the first response message to the inference network element, it is determined that the vendor types of the training network element and the inference network element are different, and the types of the model deployment platforms of the inference network element and the training network element are the same.

[0027] This scheme, by carrying the vendor type and model deployment platform type of the inference network element in the first request message, allows the training network element to determine whether the vendor type and model deployment platform type of the training network element and the inference network element are the same. This facilitates the training network element in selecting an appropriate method to provide data inference functionality to the inference network element, thereby improving the efficiency of data inference.

[0028] In one possible implementation, the first response message further includes a first instruction message that instructs the training network element to decrypt the encrypted analysis result.

[0029] According to this scheme, the inference network element can accurately determine that the network element encrypting the encrypted analysis result is the training network element based on the first instruction information.

[0030] In one possible implementation, the first response message further includes second indication information, which indicates the data type of the input data corresponding to the encrypted model.

[0031] This scheme, through the second instruction information, enables the inference network elements to preprocess the input data accordingly, thereby obtaining data to be analyzed that meets the requirements, which can improve the efficiency of data inference.

[0032] In one possible implementation, before receiving the first request message from the inference network element, the training network element sends a registration request message to the data management network element. The registration request message includes identification information of the analysis type and model information of the training network element. The model information includes the vendor type of the training network element and the type of model deployment platform of the training network element.

[0033] In one possible implementation, the model information in the registration request message also includes the aforementioned second instruction information.

[0034] In one possible implementation, the model information in the registration request message also includes the identification information of the second network element.

[0035] In one possible implementation, the model information in the registration request message also includes the identification information of the first network element.

[0036] In one possible implementation, the training network element receives the encrypted analysis result and association identifier from the inference network element; the training network element determines the encryption algorithm corresponding to the encrypted model based on the association identifier; the training network element determines the decryption algorithm based on the encryption algorithm; and the training network element decrypts the encrypted analysis result based on the decryption algorithm to obtain the decrypted analysis result.

[0037] This scheme, through association identifiers, allows training network elements to accurately obtain the encryption algorithm corresponding to the encrypted model, and thus accurately determine the decryption algorithm to be used for decryption analysis, thereby improving decryption efficiency.

[0038] Thirdly, embodiments of this application provide a communication method that can be executed by an inference network element or a module (such as a chip) applied in the inference network element. Taking the communication method executed by the inference network element as an example, the method includes: the inference network element sending a request message to the training network element, the request message including identification information of the analysis type, the request message being used to request a model that supports the analysis type, the training network element and the inference network element having different vendor types, and the inference network element and the training network element having different model deployment platform types; the inference network element receiving a response message from the training network element, the response message including first indication information and address information of the second network element, the first indication information indicating rejection of the request to support the model for the analysis type, the type of model deployment platform supported by the second network element including the type of model deployment platform of the training network element; the inference network element sending data to be analyzed to the second network element according to the address information of the second network element, the data to be analyzed being used by the second network element to generate an encrypted analysis result according to the encrypted model corresponding to the analysis type; the inference network element receiving a decrypted analysis result from the training network element or the first network element, the decrypted analysis result being obtained by the training network element or the first network element based on the encrypted analysis result.

[0039] This solution deploys inference and training network elements from different vendors, using different model deployment platforms, thus breaking the limitation of existing solutions that restrict model sharing to the same vendor. It provides a process for encrypted cross-vendor model distribution, enhancing the ability of training network elements to distribute models encryptedly and mitigating the risk of inference network element deployment vendors stealing model framework and parameter information.

[0040] In one possible implementation, the response message also includes a rejection reason value, which is that the training network element and the inference network element are of different vendor types and that the inference network element and the training network element are of different model deployment platform types.

[0041] This scheme uses a rejection reason value to inform the inference network element of the reason for rejection, so that the inference network element no longer sends the model to the training network element to request support for the analysis type, which can reduce the overhead of inference.

[0042] In one possible implementation, the request message also includes the vendor type of the inference network element and the type of the model deployment platform for the inference network element.

[0043] This scheme, by including the vendor type and model deployment platform type of the inference network element in the request message, enables the training network element to determine whether the vendor types and model deployment platform types of the training network element and the inference network element are the same. This facilitates the training network element in selecting an appropriate method to provide data inference functionality to the inference network element, thereby improving the efficiency of data inference.

[0044] In one possible implementation, the response message also includes second indication information, which indicates the data type of the input data corresponding to the encrypted model.

[0045] This scheme, through the second instruction information, enables the inference network elements to preprocess the input data accordingly, thereby obtaining data to be analyzed that meets the requirements, which can improve the efficiency of data inference.

[0046] In one possible implementation, the inference network element sends the data to be analyzed and the association identifier to the second network element based on the address information of the second network element. The association identifier is used by the first network element or the training network element to determine the encryption algorithm corresponding to the encryption model.

[0047] This scheme, through association identifiers, enables training network elements or the first network element to accurately obtain the encryption algorithm corresponding to the encrypted model, and thus accurately know the decryption algorithm to be used for decryption analysis, thereby improving decryption efficiency.

[0048] Fourthly, embodiments of this application provide a communication method that can be executed by a training network element or a module (such as a chip) applied in the training network element. Taking the communication method executed by the training network element as an example, the method includes: the training network element receiving a request message from the inference network element, the request message including identification information of the analysis type, the request message being used to request a model that supports the analysis type, the training network element and the inference network element having different vendor types, and the inference network element and the training network element having different model deployment platform types; the training network element sending a response message to the inference network element, the response message including first indication information and address information of the second network element, the first indication information indicating rejection of the request to support the model for the analysis type, the type of model deployment platform supported by the second network element including the type of model deployment platform of the training network element; the training network element receiving an encrypted analysis result from the second network element, the encrypted analysis result being obtained by the second network element based on the data to be analyzed by the inference network element and the encrypted model corresponding to the analysis type; the training network element decrypting the encrypted analysis result to obtain a decrypted analysis result; and the training network element sending the decrypted analysis result to the inference network element.

[0049] This solution deploys inference and training network elements from different vendors, using different model deployment platforms, thus breaking the limitation of existing solutions that restrict model sharing to the same vendor. It provides a process for encrypted cross-vendor model distribution, enhancing the ability of training network elements to distribute models encryptedly and mitigating the risk of inference network element deployment vendors stealing model framework and parameter information.

[0050] In one possible implementation, the request message further includes the vendor type of the inference network element and the type of the model deployment platform of the inference network element; before the training network element sends a response message to the inference network element, it is determined that the vendor types of the training network element and the inference network element are different, and that the types of the model deployment platforms of the inference network element and the training network element are different.

[0051] This scheme, by including the vendor type and model deployment platform type of the inference network element in the request message, allows the training network element to determine whether the vendor types and model deployment platform types of the training network element and the inference network element are the same. This facilitates the training network element in selecting an appropriate method to provide data inference functionality to the inference network element, thereby improving the efficiency of data inference.

[0052] In one possible implementation, the response message also includes a rejection reason value, which is that the training network element and the inference network element are of different vendor types and that the inference network element and the training network element are of different model deployment platform types.

[0053] This scheme uses a rejection reason value to inform the inference network element of the reason for rejection, so that the inference network element no longer sends the model to the training network element to request support for the analysis type, which can reduce the overhead of inference.

[0054] In one possible implementation, before receiving a request message from the inference network element, the training network element sends the identification information of the analysis type and the encrypted model corresponding to the analysis type to the second network element.

[0055] In one possible implementation, the response message also includes second indication information, which indicates the data type of the input data corresponding to the encrypted model.

[0056] This scheme, through the second instruction information, enables the inference network elements to preprocess the input data accordingly, thereby obtaining data to be analyzed that meets the requirements, which can improve the efficiency of data inference.

[0057] In one possible implementation, the training network element receives the encrypted analysis result and association identifier from the second network element; the training network element determines the encryption algorithm corresponding to the encrypted model based on the association identifier; the training network element determines the decryption algorithm based on the encryption algorithm; and the training network element decrypts the encrypted analysis result based on the decryption algorithm to obtain the decrypted analysis result.

[0058] This scheme, through association identifiers, allows training network elements to accurately obtain the encryption algorithm corresponding to the encrypted model, and thus accurately determine the decryption algorithm to be used for decryption analysis, thereby improving decryption efficiency.

[0059] Fifthly, embodiments of this application provide a communication method, which can be executed by a first network element or a module (such as a chip) applied in the first network element. Taking the execution of the communication method by the first network element as an example, the method includes: the first network element receiving an encrypted analysis result; the first network element decrypting the encrypted analysis result to obtain a decrypted analysis result; and the first network element sending the decrypted analysis result to an inference network element.

[0060] In one possible implementation, the first network element receives the encrypted analysis result from the inference network element.

[0061] In one possible implementation, the first network element receives the encrypted analysis result and the address information of the inference network element from the second network element; the first network element sends the decrypted analysis result to the inference network element based on the address information of the inference network element.

[0062] In one possible implementation, before the first network element receives the encrypted analysis result, the first network element receives an association identifier from the training network element and an identifier of the decryption algorithm corresponding to the association identifier; the first network element receives the encrypted analysis result and the association identifier; the first network element determines the decryption algorithm based on the association identifier; the first network element decrypts the encrypted analysis result according to the decryption algorithm to obtain the decrypted analysis result.

[0063] Sixthly, embodiments of this application provide a communication method, which can be executed by a second network element or a module (such as a chip) applied in the second network element. Taking the second network element executing the communication method as an example, the method includes: the second network element receiving identification information of an analysis type and an encrypted model supporting the analysis type from a training network element, wherein the type of model deployment platform supported by the second network element includes the type of model deployment platform of the training network element; the second network element receiving data to be analyzed from an inference network element; the second network element obtaining an encrypted analysis result based on the encrypted model and the data to be analyzed; and the second network element sending the encrypted analysis result and address information of the inference network element for receiving the decrypted analysis result to a training network element or a first network element, wherein the decrypted analysis result is obtained by the training network element or the first network element based on the encrypted analysis result.

[0064] Seventhly, embodiments of this application provide a communication method, which can be executed by an inference network element or a module (such as a chip) applied in the inference network element. Taking the execution of the communication method by an inference network element as an example, the method includes: the inference network element sending a request message to a data management network element, the request message including identification information of an analysis type, the request message being used to request a network element that supports the analysis type; the inference network element receiving a response message from the data management network element, the response message including at least one set of information, each set of information including address information of a candidate training network element and model information of the candidate training network element, the candidate training network element supporting the analysis type, the model information of the candidate training network element including the vendor type of the candidate training network element and the type of the model deployment platform of the candidate training network element; when there are one or more candidate training network elements among the at least one set of information corresponding to one or more candidate training network elements that have a different vendor type than the inference network element but the same model deployment platform type, the inference network element selects one candidate training network element from the one or more candidate training network elements as the training network element.

[0065] This solution enhances the functionality of the data management network element. The training network element first registers / updates the identification information of supported analysis types and the corresponding model information to the data management network element. Then, the inference network element discovers available training network elements or third-party network elements from the data management network element. Since the inference network element and the training network element are deployed by different vendors, and the types of model deployment platforms they use may be the same or different, this solution provides a process for encrypted cross-vendor model distribution. This enhances the ability of the training network element to distribute models encryptedly, avoids the risk of the inference network element's deployment vendor stealing the model's framework and parameters, and breaks the limitation of existing solutions where models can only be shared with one vendor.

[0066] In one possible implementation, when there is no candidate training network element among at least one candidate training network element corresponding to the at least one set of information that has a different vendor type than the inference network element but the same model deployment platform type, the inference network element determines the address information of the second network element based on the at least one set of information.

[0067] In one possible implementation, the model information of the candidate training network element includes the address information of the second network element; the inference network element obtains the address information of the second network element from the model information of the candidate training network element.

[0068] In one possible implementation, the encryption model in any of the above implementations is encrypted using one or more of a fully homomorphic encryption algorithm, a random secure averaging algorithm, or a differential privacy algorithm.

[0069] In one possible implementation, the inference network element in any of the above implementation methods can be an independent core network element or a functional module of a core network element.

[0070] In one possible implementation, the training network element in any of the above implementation methods can be an independent core network element or a functional module of a core network element.

[0071] In one possible implementation, the first network element in any of the above implementation methods can be an analysis result decryption network element, which can be used to decrypt the encrypted analysis result.

[0072] In one possible implementation, the second network element in any of the above implementations can be a model deployment and inference network element, which can be used to perform inference on the data to be analyzed based on the model to obtain analysis results. If the model used is an encrypted model, inference can be performed on the data to be analyzed based on the encrypted model to obtain encrypted analysis results.

[0073] Eighthly, embodiments of this application provide a communication device, which may be an inference network element or a module (such as a chip) applied in an inference network element. The device has the function of implementing any implementation method of the first aspect, any implementation method of the second aspect, or any implementation method of the seventh aspect. This function can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-described functions.

[0074] Ninthly, embodiments of this application provide a communication device, which may be an inference network element or a module (such as a chip) applied in an inference network element. The device has the function of implementing any implementation method of the second aspect or any implementation method of the fourth aspect described above. This function can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-described functions.

[0075] Tenthly, embodiments of this application provide a communication device, which may be a first network element or a module (such as a chip) applied in the first network element. The device has the function of implementing any of the implementation methods of the fifth aspect described above. This function can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-described functions.

[0076] Eleventhly, embodiments of this application provide a communication device, which may be a second network element or a module (such as a chip) applied in a second network element. The device has the function of implementing any of the implementation methods of the sixth aspect described above. This function can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-described functions.

[0077] In a twelfth aspect, embodiments of this application provide a communication device, including a processor and a memory; the memory is used to store computer instructions, and when the device is running, the processor executes the computer instructions stored in the memory to cause the device to perform any of the implementation methods in the first to seventh aspects described above.

[0078] In a thirteenth aspect, embodiments of this application provide a communication apparatus including units or means for performing various steps of any of the implementation methods in the first to seventh aspects described above.

[0079] In a fourteenth aspect, embodiments of this application provide a communication device, including a processor and an interface circuit. The processor is configured to communicate with other devices via the interface circuit and execute any of the implementation methods described in the first to seventh aspects. The processor may include one or more devices.

[0080] In a fifteenth aspect, embodiments of this application provide a communication device including a processor coupled to a memory, the processor being configured to invoke a program stored in the memory to execute any of the implementation methods described in the first to seventh aspects. The memory may be located within or outside the device. Furthermore, the processor may be one or more.

[0081] In a sixteenth aspect, embodiments of this application also provide a computer-readable storage medium storing instructions that, when executed on a communication device, cause any of the implementation methods in the first to seventh aspects to be performed.

[0082] In a seventeenth aspect, embodiments of this application also provide a computer program product, which includes a computer program or instructions that, when executed by a communication device, cause any of the implementation methods in the first to seventh aspects to be performed.

[0083] In an eighteenth aspect, embodiments of this application also provide a chip system, including: a processor for executing any of the implementation methods in the first to third aspects described above.

[0084] In a nineteenth aspect, embodiments of this application also provide a communication system, including an inference network element for implementing any implementation method of the first aspect described above and a training network element for implementing any implementation method of the second aspect described above.

[0085] In a twentieth aspect, embodiments of this application also provide a communication system, including an inference network element for implementing any implementation method of the third aspect above and a training network element for implementing any implementation method of the fourth aspect above. Attached Figure Description

[0086] Figure 1 This is a schematic diagram of a 5G network architecture based on a service-oriented architecture.

[0087] Figure 2 This is a schematic diagram of a 5G network architecture based on a point-to-point interface.

[0088] Figure 3 A flowchart illustrating a communication method provided in an embodiment of this application;

[0089] Figure 4 A flowchart illustrating a communication method provided in an embodiment of this application;

[0090] Figure 5 A flowchart illustrating a communication method provided in an embodiment of this application;

[0091] Figure 6 A schematic diagram of a communication device provided in an embodiment of this application;

[0092] Figure 7 This is a schematic diagram of a communication device provided in an embodiment of this application. Detailed Implementation

[0093] Figure 1 This is a schematic diagram of the fifth-generation (5G) network architecture based on a service-oriented architecture. Figure 1 The 5G network architecture shown may include terminal equipment, access network equipment, and core network equipment. Terminal equipment accesses the data network (DN) through the access network equipment and the core network. The core network equipment includes some or all of the following network elements: unified data management (UDM) network elements, unified data repository (UDR), network exposure function (NEF) network elements (not shown in the figure), application function (AF) network elements, policy control function (PCF) network elements, access and mobility management function (AMF) network elements, session management function (SMF) network elements, user plane function (UPF) network elements, network data analytics function (NWDAF) network elements, and network repository function (NRF) network elements (not shown in the figure).

[0094] Access network equipment can be radio access network (RAN) equipment. Examples include: base stations, evolved NodeBs (eNodeBs), transmission reception points (TRPs), next-generation NodeBs (gNBs) in 5G mobile communication systems, next-generation base stations in the 6th generation (6G) mobile communication systems, base stations in future mobile communication systems, or access nodes in wireless fidelity (WiFi) systems. It can also be a module or unit that performs some of the functions of a base station; for example, it can be a central unit (CU) or a distributed unit (DU). RAN equipment can be macro base stations, micro base stations, indoor stations, relay nodes, or donor nodes. The embodiments of this application do not limit the specific technologies or equipment forms used in the RAN equipment.

[0095] Terminal devices can be user equipment (UE), mobile stations, mobile terminals, etc. They can be widely used in various scenarios, such as device-to-device (D2D), vehicle-to-everything (V2X) communication, machine-type communication (MTC), the Internet of Things (IoT), virtual reality, augmented reality, industrial control, autonomous driving, telemedicine, smart grids, smart furniture, smart offices, smart wearables, smart transportation, and smart cities. Terminal devices can include mobile phones, tablets, computers with wireless transceiver capabilities, wearable devices, vehicles, urban air mobility vehicles (such as drones and helicopters), ships, robots, robotic arms, and smart home devices.

[0096] Access network equipment and terminal equipment can be fixed or mobile. They can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; they can also be deployed on water; and they can be deployed in the air on aircraft, balloons, and satellites. The embodiments of this application do not limit the application scenarios of the access network equipment and terminal equipment.

[0097] The AMF (Automatic Mobility Management) network element includes functions such as mobility management and access authentication / authorization. In addition, it is responsible for transmitting user policies between terminal devices and the PCF (Programmable Default Function).

[0098] The SMF network element includes functions such as performing session management, executing control policies issued by the PCF, selecting the UPF, and allocating Internet Protocol (IP) addresses to terminal devices.

[0099] UPF network elements, as interfaces with data networks, include functions such as user plane data forwarding, session / flow-based billing and statistics, and bandwidth limiting.

[0100] UDM network elements include functions such as managing contracted data and authorizing user access.

[0101] UDR includes functions for storing and retrieving data of various types, such as contract data, policy data, and application data.

[0102] NEF network elements are used to support the opening of capabilities and events.

[0103] An Application Provider (AF) element conveys the application's requests to the network, such as QoS requirements or user state event subscriptions. An AF can be a third-party functional entity or an application server deployed by the operator.

[0104] PCF network elements include policy control functions such as billing at the session and service flow levels, QoS bandwidth guarantee and mobility management, and terminal device policy decisions.

[0105] NRF network elements can be used to provide network element discovery functionality, providing network element information corresponding to the network element type based on requests from other network elements. NRF also provides network element management services, such as network element registration, updates, deregistration, and network element status subscription and push.

[0106] NWDAF (Network Data Center) elements are primarily used to collect data (including one or more of the following: terminal device data, access network device data, core network data, and third-party application data), and provide data analysis services. They can output data analysis results for network, network management, and application decision-making. NWDAFs can utilize machine learning models for data analysis. In 3GPP Release 17, the training and inference functions of NWDAFs were separated. An NWDAF can support only model training, only data inference, or both. An NWDAF supporting model training can also be called a training NWDAF, or an NWDAF supporting the model training logical function (MTLF) (abbreviated as NWDAF(MTLF)). The training NWDAF can train a model based on the acquired data to obtain the trained model. An NWDAF supporting data inference can also be called an inference NWDAF, or an NWDAF supporting the analytics logical function (AnLF) (abbreviated as NWDAF(AnLF)). An inference NWDAF can input input data into a trained model to obtain analysis results or inference data. In this embodiment, a training NWDAF refers to an NWDAF that at least supports model training functionality. As a possible implementation, a training NWDAF can also support data inference functionality. An inference NWDAF refers to an NWDAF that at least supports data inference functionality. As a possible implementation, an inference NWDAF can also support model training functionality. If an NWDAF supports both model training and data inference functionality, then the NWDAF can be called a training NWDAF, an inference NWDAF, a training-inference NWDAF, or simply an NWDAF. In this embodiment, an NWDAF can be a separate network element or can be co-located with other network elements, such as setting the NWDAF in a PCF network element or an AMF network element.

[0107] A Domain Provider (DN) is a network located outside of the carrier's network. A carrier's network can connect to multiple DNs, and various services can be deployed on a DN, providing data and / or voice services to terminal devices. For example, a DN might be the private network of a smart factory. Sensors installed in the workshop can act as terminal devices, and a control server for these sensors is deployed within the DN. The control server provides services to the sensors. Sensors can communicate with the control server, receive instructions from it, and transmit the collected sensor data back to the control server accordingly. Another example is a DN serving as an internal office network for a company. Employees' mobile phones or computers can act as terminal devices, accessing information and data resources on the company's internal office network.

[0108] Figure 1 Npcf, Nudr, Nudm, Naf, Namf, Nsmf, and Nnwdaf are the service interfaces provided by PCF, UDR, UDM, AF, AMF, SMF, and NWDAF, respectively, used to invoke the corresponding service operations. N1, N2, N3, N4, and N6 are interface sequence numbers; the meanings of these sequence numbers can be found in [link to documentation]. Figure 2 The description in the text.

[0109] Figure 2 This is a schematic diagram of a 5G network architecture based on a point-to-point interface. For a description of the functions of the network elements, please refer to [reference needed]. Figure 1 The functions of the corresponding network elements will not be described in detail here. Figure 2 and Figure 1 The main difference is: Figure 1 The interfaces between the various control plane network elements are service-oriented interfaces. Figure 2 The interfaces between the various control plane network elements are point-to-point interfaces.

[0110] exist Figure 2 In the architecture shown, the interface names and functions between the various network elements are as follows:

[0111] 1) N1: The interface between AMF and terminal devices, which can be used to transmit NAS signaling (such as QoS rules from AMF) to terminal devices.

[0112] 2) N2: The interface between AMF and RAN, which can be used to transmit radio bearer control information from the core network side to the RAN.

[0113] 3) N3: The interface between RAN and UPF, mainly used to transmit uplink and downlink user plane data between RAN and UPF.

[0114] 4) N4: The interface between SMF and UPF, which can be used to transmit information between the control plane and the user plane, including the distribution of forwarding rules, QoS control rules, traffic statistics rules, etc. from the control plane to the user plane, as well as the reporting of information from the user plane.

[0115] 5) N5: The interface between AF and PCF, which can be used for application service request distribution and network event reporting.

[0116] 6) N6: The interface between UPF and DN, used to transmit uplink and downlink user data streams between UPF and DN.

[0117] 7) N7: The interface between PCF and SMF, which can be used to issue protocol data unit (PDU) session granularity and business data flow granularity control strategies.

[0118] 8) N8: The interface between AMF and UDM, which can be used by AMF to obtain access and mobility management related subscription data and authentication data from UDM, as well as by AMF to register terminal device current mobility management related information with UDM.

[0119] 9) N9: User plane interface between UPFs, used to transmit uplink and downlink user data streams between UPFs.

[0120] 10) N10: The interface between SMF and UDM, which can be used by SMF to obtain session management-related subscription data from UDM, and by SMF to register terminal device current session-related information with UDM.

[0121] 11) N11: The interface between SMF and AMF, which can be used to transmit PDU session tunnel information between RAN and UPF, transmit control messages sent to terminal devices, and transmit radio resource control information sent to RAN, etc.

[0122] 12) N15: The interface between PCF and AMF, which can be used to issue terminal device policies and access control related policies.

[0123] 13) N23: The interface between PCF and NWDAF. NWDAF can collect data from PCF through this interface. It should be noted that NWDAF can also have interfaces with other devices (such as AMF, UPF, access network equipment, terminal equipment, etc.), which are not fully shown in the figure.

[0124] 14) N35: The interface between UDM and UDR, which can be used by UDM to obtain user subscription data information from UDR.

[0125] 15) N36: The interface between PCF and UDR, which can be used by PCF to obtain policy-related contract data and application data related information from UDR.

[0126] It is understood that the aforementioned network element or function can be a network component in a hardware device, a software function running on dedicated hardware, or a virtualization function instantiated on a platform (e.g., a cloud platform). As one possible implementation method, the aforementioned network element or function can be implemented by a single device, multiple devices working together, or a functional module within a single device; this application does not specifically limit this.

[0127] As one implementation method, the data management network element in this application embodiment can be the aforementioned NRF, UDM, or UDR, or it can be a network element in future communications such as 6G networks that has the functions of the aforementioned NRF, UDM, or UDR. The inference network element can be the aforementioned inference NWDAF or a network element in future communications such as 6G networks that has the aforementioned inference NWDAF function. The training network element can be the aforementioned training NWDAF or a network element in future communications such as 6G networks that has the aforementioned training NWDAF function.

[0128] As one implementation method, the data management network element in this application embodiment can be a network management-side model management device, a network management-side model management network element, or a network management-side model management service. The inference network element can be an access network device-side inference device. The training network element can be a network management-side training device, a network management-side training network element, or a network management-side training service.

[0129] As one implementation method, the data management network element in this embodiment can be a model management device on the access network device side. The inference network element can be an inference device on the access network device side. The training network element can be a training device on the access network device side.

[0130] To enable cross-vendor model sharing, this application provides a communication method. In this method, the vendor type of the training network element differs from that of the inference network element, but the type of the model deployment platform for the training network element is the same as that for the inference network element. The model deployment platform is a framework upon which the model relies for execution. Different model deployment platforms may differ in their dynamic computation graph, static computation graph, debugging methods, visualization, or parallelization characteristics. The type of the model deployment platform is used to distinguish between different model deployment platforms. For example, the vendor type can be represented by a Vendor ID, such as Vendor ID = 1 for vendor A and Vendor ID = 2 for vendor B. Similarly, the type of the model deployment platform can be represented by an AI Platform ID (or Platform ID), such as AI Platform ID = 1 for model deployment platform A and AI Platform ID = 2 for model deployment platform B. Here, AI is short for Artificial Intelligence. The vendor type and model deployment platform type are explained uniformly here and will not be elaborated further.

[0131] refer to Figure 3 The method includes the following steps:

[0132] Step 301: The inference network element sends a request message to the training network element. Correspondingly, the training network element receives the request message.

[0133] The request message includes analytics ID, which is used to request a model that supports the analytics type indicated by the analytics ID. Analytics ID indicates the analytics type and can be, for example, service experience or network element load information.

[0134] As one possible implementation, the request message also includes the vendor type and the model deployment platform type. Here, the vendor type and model deployment platform type refer to the vendor type of the inference network element and the model deployment platform type, respectively. The vendor type could be, for example, Huawei, Ericsson, or Nokia. The model deployment platform type could be, for example, MindSpore, Tensorflow, or PyTorch.

[0135] As one possible implementation, the request message may also include the version of the model deployment platform for the inference network element. The model deployment platform version could be, for example, V1.0 or V2.1.

[0136] As one possible implementation, the request message may also include an association identifier.

[0137] As one implementation method, the inference network element can request model information from the training network element by calling the Nnwdaf_MLModelProvision_Subscribe service operation. That is, the request message in step 301 can be the Nnwdaf_MLModelProvision_Subscribe service operation.

[0138] In this embodiment of the application, the request message in step 301 is also referred to as the first request message.

[0139] Step 302: The training network element is determined to be from a different vendor type than the inference network element, but the model deployment platform type is the same.

[0140] As one implementation method, the request message in step 301 above carries the vendor type of the inference network element and the type of the model deployment platform of the inference network element. The training network element determines whether the vendor type of the training network element is the same as that of the inference network element, and also determines whether the type of the model deployment platform of the training network element is the same as that of the inference network element. If the vendor types of the training network element and the inference network element are different but the type of the model deployment platform of the training network element and the inference network element are the same, then step 303 and subsequent steps are executed; otherwise, the process ends.

[0141] As an alternative implementation, if the inference network element knows the vendor type and model deployment platform type of each training network element deployed on it, then the request message in step 301 does not need to carry the vendor type and model deployment platform type of the inference network element. Instead, it carries an indication message indicating whether the vendor type of the training network element is the same as that of the inference network element, and whether the model deployment platform type of the training network element is the same as that of the inference network element. The training network element can then determine whether the vendor type and model deployment platform type of the training network element are the same as those of the inference network element based on this indication message. If the vendor types of the training network element and the inference network element are different, but the model deployment platform types of the training network element and the inference network element are the same, then step 303 and subsequent steps are executed; otherwise, the process ends.

[0142] As an alternative implementation, the vendor type and model deployment platform type of each inference network element can be pre-configured on the training network element. In this case, the request message in step 301 does not need to include the vendor type and model deployment platform type of the inference network element, nor does it need to include the aforementioned instruction information. The training network element can determine whether its vendor type and the inference network element's vendor type are the same, and whether their model deployment platform types are the same, based on its local configuration information. If the vendor types of the training network element and the inference network element are different, but their model deployment platform types are the same, then step 303 and subsequent steps are executed; otherwise, the process ends.

[0143] It should be noted that the end of the process mentioned above refers to the end of the process at that time. Figure 3 In the embodiments described above, the process ends, but other operations can still be performed after the process ends. For example, if the training network element and the inference network element are of the same vendor type and the model deployment platform type is the same, the training network element can provide the inference network element with an unencrypted model or the address information of an unencrypted model, and then the inference network element obtains unencrypted analysis results based on the unencrypted model. As another example, if the training network element and the inference network element are of different vendor types and the model deployment platform type is different, the following can be used: Figure 4 The implementation scheme enables the inference network element to obtain analysis results. For example, if the training network element and the inference network element are from the same vendor but their model deployment platforms are different, the inference network element can provide the data to be analyzed to a third-party network element (such as a second network element). The second network element uses an encrypted model and the data to be analyzed to obtain encrypted analysis results. Then, the first network element or the training network element decrypts the data to be analyzed to obtain decrypted analysis results, which are then sent to the inference network element.

[0144] Of course, the functions of the training network elements can also be pre-configured. For example, training network elements 1 to 10 can be pre-configured to provide models only to inference network elements of the same vendor type and platform type where the model is deployed. Taking training network element 1 as an example, if training network element 1 receives the request message from the inference network element in step 301 above, it assumes that the vendor type of training network element 1 is different from that of the inference network element, and that the platform type where the model of training network element 1 is deployed is the same as that where the model of the inference network element is deployed. In this implementation method, step 302 does not need to be executed.

[0145] Step 303: The training network element sends a response message to the inference network element. Correspondingly, the inference network element receives the response message.

[0146] The response message contains the encrypted model or the address information of the encrypted model, where the address information of the encrypted model may be, for example, a Uniform Resource Locator (URL) or a Fully Qualified Domain Name (FQDN).

[0147] Taking a neural network model as an example, this model includes model architecture information and model parameters. The model architecture information includes the number of neural network layers, the connections between layers, and the activation function used in each layer. The model parameters include the parameter values ​​for each layer of the neural network.

[0148] As one implementation, the encrypted model in the response message includes both unencrypted model architecture information and encrypted model parameters. As another implementation, the encrypted model in the response message includes both encrypted model architecture information and encrypted model parameters.

[0149] As one possible implementation, the response message also includes the address information of the first network element or the instruction information for the training network element to decrypt the encrypted analysis result (in this embodiment, the instruction information can also be called the first instruction information). The first network element is a third-party network element with the function of decrypting the analysis result, such as an NWDAF network element. The first network element can also be called the analysis result decryption network element.

[0150] As one possible implementation, the response message also includes indication information indicating the data type of the input data corresponding to the encrypted model (in this embodiment, this indication information can also be referred to as second indication information). For example, this indication information can be an event ID. Exemplarily, the data type can be one or more of the UE location or QoS Flow parameters.

[0151] As one possible implementation, the response message also includes a data format and / or processing parameters corresponding to each data type. The inference network element performs corresponding preprocessing on the input data corresponding to each data type based on the data format and / or processing parameters to obtain the data to be analyzed. For example, the data format includes one or more of the data reporting time window (i.e., when the data is reported) and the data cache size (i.e., how large the data cache is before reporting). The processing parameters include one or more of the maximum value, minimum value, average value, or variance value. This explanation of data types, data formats, and processing parameters also applies to other subsequent embodiments and will not be repeated here.

[0152] As one possible implementation, the response message may also include the aforementioned association identifier.

[0153] As one implementation method, the training network element can send the above information to the inference network element by calling the Nnwdaf_MLModelProvision_Notify service operation. That is, the response message in step 303 can be the Nnwdaf_MLModelProvision_Notify service operation.

[0154] In this embodiment of the application, the response message in step 303 is also referred to as the first response message.

[0155] Step 304: The inference network element obtains the encrypted analysis result based on the encrypted model.

[0156] If the response message in step 303 above carries the address information of the encrypted model, the inference network element also needs to obtain the encrypted model based on the address information. For example, the inference network element can download the encrypted model from the address indicated by the address information of the encrypted model according to the file transfer protocol (FTP).

[0157] The inference network element obtains encrypted analysis results based on the data to be analyzed and the encrypted model. In other words, the data to be analyzed is input into the encrypted model to obtain the encrypted analysis results. The data to be analyzed is the input data corresponding to the encrypted model, which is collected by the inference network element from one or more other network elements (such as UE, SMF, AMF, access network equipment, PCF, UPF, or AF).

[0158] After step 304, the inference network element obtains the decrypted analysis result based on the encrypted analysis result. The following describes two different implementation methods for the inference network element to obtain the decrypted analysis result.

[0159] As a first implementation method, if the response message of step 303 above carries instruction information (i.e., first instruction information) for instructing the training network element to decrypt the encrypted analysis result, then steps 305 to 307 are executed after step 304.

[0160] As a second implementation method, if the response message of step 303 carries the address information of the first network element, then steps 308 to 310 are executed after step 304.

[0161] Step 305: The inference network element sends a request message to the training network element. Correspondingly, the training network element receives the request message.

[0162] The request message contains the identification information of the analysis type and the encrypted analysis result. The request message is used to request the decryption of the analysis result. The identification information of the analysis type is the same as the identification information of the analysis type in step 301 above.

[0163] As one possible implementation, the request message may also include the aforementioned association identifier.

[0164] As one implementation method, the inference network element can send the analysis type identification information and the encrypted analysis result to the training network element by calling the Nnwdaf_AnalyticsDecryption_Request service operation. That is, the request message in step 305 can be the Nnwdaf_AnalyticsDecryption_Request service operation.

[0165] Step 306: The training network element decrypts the encrypted analysis result to obtain the decrypted analysis result.

[0166] The encrypted model can be encrypted using one or more of the following algorithms: fully homomorphic encryption, stochastic safety average, or differential privacy. The training network elements then use the decryption algorithm corresponding to the encryption algorithm used by the encrypted model to decrypt the encrypted analysis results and obtain the decrypted analysis results.

[0167] If both the request message in step 301 and the request message in step 305 carry the aforementioned association identifier, then before or after step 303, the training network element binds the encryption algorithm used by the encrypted model to the association identifier. Then, in step 306, the training network element can first determine the encryption algorithm corresponding to the encrypted model based on the association identifier in the request message in step 305, and then determine the decryption algorithm based on the encryption algorithm. In this way, the encrypted analysis result is decrypted according to the decryption algorithm to obtain the decrypted analysis result.

[0168] Step 307: The training network element sends a response message to the inference network element. Correspondingly, the inference network element receives the response message.

[0169] The response message contains the decrypted analysis results.

[0170] As one implementation method, the training network element can send the decrypted analysis result to the inference network element by calling the Nnwdaf_AnalyticsDecryption_Request Response service operation. That is, the response message in step 307 can be the Nnwdaf_AnalyticsDecryption_Request Response service operation.

[0171] Step 308: The inference network element sends a request message to the first network element. Correspondingly, the first network element receives the request message.

[0172] The request message contains the identification information of the analysis type and the encrypted analysis result. The request message is used to request the decryption of the analysis result. The identification information of the analysis type is the same as the identification information of the analysis type in step 301 above.

[0173] As one possible implementation, the request message may also include the aforementioned association identifier.

[0174] As one implementation method, the inference network element can send the analysis type identification information and the encrypted analysis result to the first network element by calling the Nnf_AnalyticsDecryption_Request service operation. That is, the request message in step 308 can be the Nnf_AnalyticsDecryption_Request service operation.

[0175] Step 309: The first network element decrypts the encrypted analysis result to obtain the decrypted analysis result.

[0176] The encrypted model is encrypted using one or more of the following algorithms: fully homomorphic encryption, random secure averaging, or differential privacy. The training network elements then use the decryption algorithm corresponding to the encryption algorithm used by the encrypted model to decrypt the encrypted analysis results and obtain the decrypted analysis results.

[0177] As a possible implementation, if the response message of step 303 contains the address information of the first network element, then before or after step 303, the training network element also sends the aforementioned association identifier and the decryption algorithm corresponding to the encrypted model to the first network element, and the request message of step 308 also carries the aforementioned association identifier. Thus, in step 309, the first network element can first determine the decryption algorithm corresponding to the encrypted model based on the association identifier in the request message of step 308, and then decrypt the encrypted analysis result according to the decryption algorithm to obtain the decrypted analysis result.

[0178] Step 310: The first network element sends a response message to the inference network element. Correspondingly, the inference network element receives the response message.

[0179] The response message contains the decrypted analysis results.

[0180] As one implementation method, the first network element can send the decrypted analysis result to the inference network element by calling the Nnf_AnalyticsDecryption_Response service operation. That is, the response message in step 310 can be the Nnf_AnalyticsDecryption_Response service operation.

[0181] In the above scheme, the inference network element and the training network element are deployed by different vendors, but they use the same model deployment platform. This scheme provides a process for encrypted distribution of models across vendors, enhances the ability of the training network element to distribute models in encrypted form, avoids the risk of the deployment vendor of the inference network element stealing the model's framework and parameters, ensures the security of model information, and breaks the limitation in existing solutions that models can only be shared with the same vendor.

[0182] As one implementation method, each training network element can also register its model information with the data management network element. This allows the inference network element to request and discover suitable training network elements from the data management network element when it lacks the address information of the training network element locally. For example, a training network element can send a registration request message to the data management network element. This message includes identifiers of the analysis types the training network element can provide and its model information, including the vendor type and the type of the model deployment platform. As another possible implementation method, the model information may also include indications of the data type of the input data corresponding to the encrypted model. As yet another possible implementation method, the response message may also include the data format and / or processing parameters corresponding to each data type. As yet another possible implementation method, the model information may also include the address information of a first network element or an indication that the training network element will decrypt the encrypted analysis results. The meaning of the first network element can be found in the foregoing description. As yet another possible implementation method, the model information may also include the address information of a second network element. The second network element can be a trusted third-party network element, specifically a model deployment and inference network element. The second network element can perform inference on the data to be analyzed based on the model to obtain the analysis results. If the model used is encrypted, the second network element can perform inference on the data to be analyzed based on the encrypted model to obtain encrypted analysis results.

[0183] It should be noted that when different training network elements register model information with the data management network element, the first network element in the model information of different training network elements can be the same network element or different network elements. Similarly, the second network element in the model information of different training network elements can be the same network element or different network elements.

[0184] As one implementation method, before step 301 above, the inference network element can send a request message (in this embodiment, the request message is also called the second request message) to the data management network element. This request message includes the identification information of the analysis type from step 301 above. This request message is used to request a network element that supports the analysis type. Then, the data management network element sends a response message (in this embodiment, the response message is also called the second response message) to the inference network element. This response message includes the address information of the training network element described in step 301 above. If the data management network element determines that multiple training network elements support the analysis type, it can provide the address information and model information of these multiple training network elements to the inference network element, which can then select one training network element from them.

[0185] This application provides a communication method. In this method, the vendor type of the training network element is different from that of the inference network element, and the type of the model deployment platform of the training network element is different from that of the inference network element.

[0186] refer to Figure 4 The method includes the following steps:

[0187] Step 401: The training network element encrypts the existing local model and sends the encrypted model and the identification information of the analysis type corresponding to the encrypted model to the second network element.

[0188] The second network element supports a wide variety of model deployment platforms. In this embodiment, the second network element supports at least the model deployment platform type of the training network element. The meaning of the second network element can be found in the foregoing description.

[0189] As one possible implementation, the training network element also sends the address information of the first network element to the second network element, which has the function of decrypting the analysis results.

[0190] It is understandable that the model that a training element already has locally can be a model trained by that training element or a model obtained by that training element from other training elements.

[0191] Step 401 is optional. When step 401 is not executed, other network elements or operators can pre-configure the above information to the second network element, such as one or more of the following: the encrypted model, the identification information of the analysis type corresponding to the encrypted model, and the address information of the first network element.

[0192] Step 402: The inference network element sends a request message to the training network element. Correspondingly, the training network element receives the request message.

[0193] Step 402 is the same as step 301 above, and can be referred to the foregoing description.

[0194] Step 403: The training network element sends the encrypted update model and the identification information of the analysis type corresponding to the encrypted update model to the second network element.

[0195] This step is optional. After receiving the above request message from the inference network element, if the training network element confirms that the local model needs further training, it triggers other network elements to collect data and carry out the subsequent model training process, and then encrypts the updated model obtained from the training and resends it to the second network element.

[0196] Step 404: The training network element is determined to be from a different vendor and the model deployment platform is different from that of the inference network element.

[0197] Step 404 is an optional step. The implementation method of step 404, as well as various alternative implementation methods, are similar to the description of step 302 above, and can be referred to the above description.

[0198] Step 405: The training network element sends a response message to the inference network element. Correspondingly, the inference network element receives the response message.

[0199] The response message contains the address information of the second network element and indication information for rejecting the request to support the model of the above analysis type (in this embodiment, the indication information is also referred to as the first indication information).

[0200] As one possible implementation, the response message also includes indication information indicating the data type of the input data corresponding to the encrypted model (in this embodiment, this indication information can also be called second indication information). For example, this indication information can be an event ID. As another possible implementation, the response message also includes data format and / or processing parameters corresponding to each data type.

[0201] As one possible implementation, the response message also includes a rejection reason value, which is that the training network element and the inference network element are from different vendor types and that the inference network element and the training network element are deployed on different model platforms.

[0202] As one possible implementation, if the request message in step 402 above contains an association identifier, then the response message contains the association identifier.

[0203] As one implementation method, the training network element can send the above information to the inference network element by calling the Nnwdaf_MLModelProvision_Notify service operation. That is, the response message in step 405 can be the Nnwdaf_MLModelProvision_Notify service operation.

[0204] Step 406: The inference network element sends a request message to the second network element based on the address information of the second network element. Correspondingly, the second network element receives the request message.

[0205] The request message contains the data to be analyzed and identification information for the analysis type. This request message is used to request analysis of the data. The identification information for the analysis type is the same as that for the analysis type in step 402 above.

[0206] The data to be analyzed is the input data corresponding to the encrypted model, and the data to be analyzed is collected by the inference network element from one or more other network elements (such as UE, SMF, AMF, access network equipment, PCF, UPF or AF).

[0207] As one possible implementation, the request message may also include the aforementioned association identifier.

[0208] As one implementation method, the inference network element can send the above information to the second network element by calling the Nnf_AnalyticsInfo_Request service operation. That is, the request message in step 406 can be the Nnf_AnalyticsInfo_Request service operation.

[0209] Step 407: The second network element obtains the encrypted analysis results based on the encrypted model.

[0210] Specifically, the second network element uses a locally deployed encrypted model and the data to be analyzed received from the inference network element to calculate encrypted analysis results. The locally deployed encrypted model on the second network element is configured from the training network element, other network elements, or cloud operator configurations.

[0211] After the second network element obtains the encrypted analysis result, steps 408 to 410 can be performed, or steps 411 to 413 can be performed.

[0212] Step 408: The second network element sends a request message to the training network element. Correspondingly, the training network element receives the request message.

[0213] The request message contains the analysis type identification information, the encrypted analysis result, and the address information of the inference network element. The request message is used to request the decrypted analysis result and send the decrypted analysis result to the inference network element. The analysis type identification information is the same as the analysis type identification information in step 402 above.

[0214] As one possible implementation, the request message may also include the aforementioned association identifier.

[0215] As one implementation method, the second network element can send the analysis type identifier, the encrypted analysis result, and the address information of the inference network element to the training network element by calling the Nnwdaf_AnalyticsDecryption_Request service operation. That is, the request message in step 408 can be the Nnwdaf_AnalyticsDecryption_Request service operation.

[0216] Step 409: The training network element decrypts the encrypted analysis result to obtain the decrypted analysis result.

[0217] Step 409 is the same as step 306 above, and can be referred to the foregoing description.

[0218] Step 410: The training network element sends the decrypted analysis results to the inference network element. Correspondingly, the inference network element receives the decrypted analysis results.

[0219] As one implementation method, the training network element can send the decrypted analysis results to the inference network element by calling the Nnwdaf_AnalyticsDecryption_Request Response service operation.

[0220] Step 411: The second network element sends a request message to the first network element. Correspondingly, the first network element receives the request message.

[0221] The request message contains the analysis type identification information, the encrypted analysis result, and the address information of the inference network element. The request message is used to request the decrypted analysis result and send the decrypted analysis result to the inference network element. The analysis type identification information is the same as the analysis type identification information in step 402 above.

[0222] The second network element can obtain the address information of the first network element through the above step 401.

[0223] As one possible implementation, the request message may also include the aforementioned association identifier.

[0224] Step 412: The first network element decrypts the encrypted analysis result to obtain the decrypted analysis result.

[0225] Step 412 is the same as step 309 above, and can be referred to the foregoing description.

[0226] Step 413: The first network element sends the decrypted analysis result to the inference network element. Correspondingly, the inference network element receives the decrypted analysis result.

[0227] As one implementation method, the first network element can send the decrypted analysis results to the inference network element by calling the Nnwdaf_AnalyticsDecryption_Request Response service operation.

[0228] In the above solution, the inference network element and the training network element are deployed by different vendors, and the model deployment platforms used by the two are also different. This solution provides a process for encrypted distribution of models across vendors, enhances the ability of the training network element to distribute models in encrypted form, avoids the risk that the deployment vendor of the inference network element may steal the model's framework and parameters, ensures the security of model information, and breaks the limitation in existing solutions that models can only be shared with the same vendor.

[0229] refer to Figure 5 This application provides a communication method according to an embodiment of the present application. The method includes the following steps:

[0230] Step 501: The training network element sends a registration request message to the data management network element. Correspondingly, the data management network element receives the registration request message.

[0231] The registration request message contains identification information for the analysis type and model information. The model information includes the vendor type, the type of model deployment platform, the address information of the second network element, and also includes the address information of the first network element or an instruction to the training network element to decrypt the encrypted analysis results.

[0232] As one possible implementation, the registration request message may also include the version of the model deployment platform.

[0233] As one possible implementation, the registration request message also includes indication information indicating the data type of the input data corresponding to the encrypted model. For example, this indication information could be an event ID. As another possible implementation, the registration request message also includes the data format and / or processing parameters corresponding to each data type.

[0234] The meanings of the analysis type identification information, vendor type, model deployment platform type, model deployment platform version, first network element, and second network element can be referred to the above description and will not be repeated here.

[0235] As one implementation method, the training network element can request registration from the data management network element by calling the Nnrf_NFManagement_NFRegisterRequest service operation. That is, the registration request message in step 501 can be the Nnrf_NFManagement_NFRegisterRequest service operation.

[0236] Step 502: The data management network element sends a registration response message to the training network element. Correspondingly, the training network element receives the registration response message.

[0237] As one implementation method, the data management network element can return a response to the registration request message to the training network element by calling the Nnrf_NFManagement_NFRegisterResponse service operation. That is, the registration response message in step 502 can be the Nnrf_NFManagement_NFRegister Response service operation.

[0238] Step 503: The training network element sends an update request message to the data management network element. Correspondingly, the data management network element receives the update request message.

[0239] If the model information of the training network element is updated, such as if the version of the model deployment platform is updated, the training network element can send an update request message to the data management network element to re-register the updated model information with the data management network element.

[0240] The information carried in the update request message is similar to that carried in the registration request message in step 501 above, and can be referred to the foregoing description.

[0241] As one implementation method, the training network element can request registration updates from the data management network element by calling the Nnrf_NFManagement_NFUpdateRequest service operation.

[0242] Step 504: The data management network element sends an update response message to the training network element. Correspondingly, the training network element receives the update response message.

[0243] As one implementation method, the data management network element can return a response to the update request message to the training network element by calling the Nnrf_NFManagement_NFUpdateResponse service operation.

[0244] Steps 503 to 504 above are optional steps.

[0245] Step 505: The inference network element sends a request message to the data management network element. Correspondingly, the data management network element receives the request message.

[0246] The request message includes identification information for the analysis type. As one possible implementation, the request message may also include the vendor type of the inference network element and the type of the model deployment platform.

[0247] This request message is used to request network elements that support this analysis type; specifically, it is used to request training network elements or third-party network elements that support this analysis type.

[0248] As one implementation method, the inference network element can request the data management network element to discover available training network elements or third-party network elements by calling the Nnrf_NFDiscovery_Request service operation. That is, the request message in step 505 can be the Nnrf_NFDiscovery_Request service operation.

[0249] Step 506: The data management network element sends a response message to the inference network element. Correspondingly, the inference network element receives the response message.

[0250] The response message contains at least one set of information, each set of information including the address information of at least one candidate training network element and the model information of the candidate training network element. The model information corresponds to the identification information of the analysis type in the request message of step 505 above. The content contained in the model information can be referred to the description of step 501 above.

[0251] It should be noted that the address information of the first network element in the model information of different candidate training network elements can be the same or different, and the address information of the second network element in the model information of different candidate training network elements can be the same or different.

[0252] As one implementation method, the data management network element can respond to the network element discovery request of the inference network element by calling the Nnrf_NFDiscovery_RequestResponse service operation. That is, the response message in step 506 can be the Nnrf_NFDiscovery_Request Response service operation.

[0253] Step 507: The inference network element selects either the training network element or the second network element.

[0254] If the response message in step 506 above contains multiple sets of information, the inference network element selects either the training network element or the second network element in the following order.

[0255] If one or more candidate training network elements corresponding to at least one set of information are of the same type as the inference network element vendor and the same type as the model deployment platform, then the inference network element selects one of the one or more candidate training network elements as a training network element, such as randomly selecting one or selecting one according to a predetermined rule.

[0256] If none of the candidate training network elements corresponding to the multiple sets of information are of the same type as the inference network element vendor and the same type as the model deployment platform, but there are one or more candidate training network elements corresponding to the multiple sets of information that are different from the inference network element vendor and the same type as the model deployment platform, then the inference network element selects one of the one or more candidate training network elements as the training network element, such as randomly selecting one or selecting one according to a predetermined rule.

[0257] If none of the candidate training network elements corresponding to the multiple sets of information are of the same type as the inference network element vendor and the same type as the model deployment platform, and none of the candidate training network elements corresponding to the multiple sets of information are of a different type than the inference network element vendor but the same type as the model deployment platform, then the inference network element selects a second network element based on the model information of the at least one candidate training network element corresponding to the multiple sets of information. For example, if the addresses of the second network elements in the model information of the at least one candidate training network element are all the same, then the address of a second network element is randomly selected. Or, if the addresses of the second network elements in the model information of the at least one candidate training network element are not completely identical, then one can be randomly selected or selected according to a predetermined rule.

[0258] If the inference network element is a training network element, then after step 507, steps 301 to 307 above can be executed, or steps 301 to 304 and steps 308 to 310 above can be executed.

[0259] If the inference network element selected is a second network element, then after step 507, steps 406 to 410 above can be executed, or steps 406 to 407 and steps 411 to 413 above can be executed.

[0260] The above solution enhances the functionality of the data management network element. The training network element first registers / updates the identification information of supported analysis types and the corresponding model information to the data management network element. Then, the inference network element discovers available training network elements or third-party network elements from the data management network element. Since the inference network element and training network element are deployed by different vendors, and the types of model deployment platforms they use may be the same or different, this solution provides a process for encrypted cross-vendor model distribution. This enhances the ability of the training network element to distribute models encryptedly, avoids the risk of the inference network element's deployment vendor stealing the model's framework and parameters, ensures the security of model information, and breaks the limitation of existing solutions that models can only be shared with one vendor.

[0261] It is understood that the data management network element in this embodiment of the invention is only an example. As a possible implementation method, the function played by the data management network element in this embodiment of the invention can be performed by other network elements (such as the model management network element).

[0262] It is understood that, in order to achieve the functions in the above embodiments, the inference network element, training network element, first network element, and second network element include hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, based on the units and method steps of the various examples described in conjunction with the embodiments disclosed in this application, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application scenario and design constraints of the technical solution.

[0263] Figure 6 and Figure 7 The diagram illustrates the possible communication devices provided in the embodiments of this application. These communication devices can be used to implement the functions of the inference network element, training network element, first network element, or second network element in the above method embodiments, and thus can also achieve the beneficial effects of the above method embodiments. In the embodiments of this application, the communication device can be an inference network element, training network element, first network element, or second network element, or it can be a module (such as a chip) applied to the inference network element, training network element, first network element, or second network element.

[0264] like Figure 6 As shown, the communication device 600 includes a processing unit 610 and a transceiver unit 620. The communication device 600 is used to implement the functions of the inference network element, training network element, first network element, or second network element in the above method embodiments.

[0265] In the first embodiment, when the communication device is an inference network element or a model (such as a chip) used for an inference network element, the transceiver unit 620 is configured to send a first request message to the training network element, the first request message including identification information of the analysis type, the first request message being used to request a model that supports the analysis type, the training network element and the inference network element having different vendor types, and the inference network element and the training network element having the same model deployment platform type; receive a first response message from the training network element, the first response message including an encrypted model or the address information of the encrypted model, the encrypted model supporting the analysis type; and the processing unit 610 is configured to obtain an encrypted analysis result based on the encrypted model; and obtain a decrypted analysis result based on the encrypted analysis result.

[0266] In one possible implementation, the transceiver unit 620 is used to send the encrypted analysis result to the training network element and receive the decrypted analysis result from the training network element.

[0267] In one possible implementation, the first response message further includes a first instruction message that instructs the training network element to decrypt the encrypted analysis result.

[0268] In one possible implementation, the transceiver unit 620 is used to send the encrypted analysis result and association identifier to the training network element. The association identifier is used by the training network element to determine the encryption algorithm corresponding to the encrypted model.

[0269] In one possible implementation, the first response message further includes the address information of the first network element; the processing unit 610 is configured to send the encrypted analysis result to the first network element through the transceiver unit 620 according to the address information of the first network element; and receive the decrypted analysis result from the first network element.

[0270] In one possible implementation, the processing unit 610 is used to send the encrypted analysis result and association identifier to the first network element through the transceiver unit 620 according to the address information of the first network element. The association identifier is used by the first network element to determine the encryption algorithm corresponding to the encryption model.

[0271] In one possible implementation, the first response message further includes second indication information, which indicates the data type of the input data corresponding to the encrypted model.

[0272] In one possible implementation, the first request message also includes the vendor type of the inference network element and the type of the model deployment platform of the inference network element.

[0273] In one possible implementation, the transceiver unit 620 is configured to send a second request message to the data management network element before sending the first request message to the training network element. The second request message includes identification information of the analysis type and is used to request a network element that supports the analysis type. The transceiver unit 620 is also configured to receive a second response message from the data management network element, the second response message including the address information of the training network element.

[0274] In the second embodiment, when the communication device is a training network element or a model (such as a chip) used for training the network element, the transceiver unit 620 is configured to receive a first request message from the inference network element, the first request message including identification information of the analysis type, the first request message being used to request a model that supports the analysis type, the training network element and the inference network element having different vendor types, and the inference network element and the training network element having the same model deployment platform type; send a first response message to the inference network element, the first response message including encrypted model or the address information of the encrypted model; receive encrypted analysis results from the inference network element, the encrypted analysis results being obtained based on the encrypted model; the processing unit 610 is configured to decrypt the encrypted analysis results to obtain decrypted analysis results; and the transceiver unit 620 is configured to send the decrypted analysis results to the inference network element.

[0275] In one possible implementation, the first request message further includes the vendor type of the inference network element and the type of the model deployment platform of the inference network element; the processing unit 610 is configured to determine, before the transceiver unit 620 sends the first response message to the inference network element, that the vendor type of the training network element is different from that of the inference network element, and that the type of the model deployment platform of the inference network element is the same as that of the training network element.

[0276] In one possible implementation, the first response message further includes a first instruction message that instructs the training network element to decrypt the encrypted analysis result.

[0277] In one possible implementation, the first response message further includes second indication information, which indicates the data type of the input data corresponding to the encrypted model.

[0278] In one possible implementation, the transceiver unit 620 is configured to send a registration request message to the data management network element before receiving the first request message from the inference network element. The registration request message includes identification information of the analysis type and model information of the training network element. The model information includes the vendor type of the training network element and the type of model deployment platform of the training network element.

[0279] In one possible implementation, the transceiver unit 620 is used to receive the encrypted analysis result and association identifier from the inference network element; the processing unit 610 is used to determine the encryption algorithm corresponding to the encrypted model based on the association identifier; determine the decryption algorithm based on the encryption algorithm; and decrypt the encrypted analysis result based on the decryption algorithm to obtain the decrypted analysis result.

[0280] In the third embodiment, when the communication device is an inference network element or a model (such as a chip) used for an inference network element, the transceiver unit 620 is configured to send a request message to the training network element. This request message includes identification information for the analysis type and is used to request a model that supports that analysis type. The training network element and the inference network element are from different vendors, and the model deployment platforms for the inference network element and the training network element are of different types. The transceiver unit 620 also receives a response message from the training network element. This response message includes first indication information and the address information of a second network element. The first indication information indicates a rejection of the request to support the analysis. The second network element supports a model deployment platform of the type of training network element; the processing unit 610 is used to send data to be analyzed to the second network element through the transceiver unit 620 according to the address information of the second network element. The data to be analyzed is used by the second network element to generate encrypted analysis results according to the encrypted model corresponding to the analysis type; the transceiver unit 620 is used to receive decrypted analysis results from the training network element or the first network element. The decrypted analysis results are obtained by the training network element or the first network element based on the encrypted analysis results.

[0281] In one possible implementation, the response message also includes a rejection reason value, which is that the training network element and the inference network element are of different vendor types and that the inference network element and the training network element are of different model deployment platform types.

[0282] In one possible implementation, the request message also includes the vendor type of the inference network element and the type of the model deployment platform for the inference network element.

[0283] In one possible implementation, the response message also includes second indication information, which indicates the data type of the input data corresponding to the encrypted model.

[0284] In one possible implementation, the processing unit 610 is used to send the data to be analyzed and the association identifier to the second network element through the transceiver unit 620 according to the address information of the second network element. The association identifier is used by the first network element or the training network element to determine the encryption algorithm corresponding to the encryption model.

[0285] In the fourth embodiment, when the communication device is a training network element or a model (such as a chip) used for training the network element, the transceiver unit 620 is used to receive a request message from the inference network element. The request message includes identification information of the analysis type. The request message is used to request a model that supports the analysis type. The training network element and the inference network element have different vendor types and different model deployment platform types. The transceiver unit 620 sends a response message to the inference network element. The response message includes first indication information and address information of the second network element. The first indication information indicates that the request to support the model that supports the analysis type is rejected. The type of model deployment platform supported by the second network element includes the type of model deployment platform of the training network element. The transceiver unit 620 receives an encrypted analysis result from the second network element. The encrypted analysis result is obtained by the second network element based on the data to be analyzed by the inference network element and the encrypted model corresponding to the analysis type. The processing unit 610 is used to decrypt the encrypted analysis result to obtain a decrypted analysis result. The transceiver unit 620 is used to send the decrypted analysis result to the inference network element.

[0286] In one possible implementation, the request message further includes the vendor type of the inference network element and the type of the model deployment platform of the inference network element; the processing unit 610 is configured to determine, before the transceiver unit 620 sends a response message to the inference network element, that the vendor type of the training network element is different from that of the inference network element, and that the type of the model deployment platform of the inference network element is different from that of the training network element.

[0287] In one possible implementation, the response message also includes a rejection reason value, which is that the training network element and the inference network element are of different vendor types and that the inference network element and the training network element are of different model deployment platform types.

[0288] In one possible implementation, the transceiver unit 620 is configured to send the identification information of the analysis type and the encrypted model corresponding to the analysis type to the second network element before receiving the request message from the inference network element.

[0289] In one possible implementation, the response message also includes second indication information, which indicates the data type of the input data corresponding to the encrypted model.

[0290] In one possible implementation, the transceiver unit 620 is used to receive the encrypted analysis result and association identifier from the second network element; the processing unit 610 is used to determine the encryption algorithm corresponding to the encryption model according to the association identifier; determine the decryption algorithm according to the encryption algorithm; and decrypt the encrypted analysis result according to the decryption algorithm to obtain the decrypted analysis result.

[0291] In the fifth embodiment, when the communication device is a first network element or a model (such as a chip) for the first network element, the transceiver unit 620 is used to receive the encrypted analysis result; the processing unit 610 is used to decrypt the encrypted analysis result to obtain the decrypted analysis result; and the transceiver unit 620 is used to send the decrypted analysis result to the inference network element.

[0292] In one possible implementation, the transceiver unit 620 is used to receive the encrypted analysis result from the inference network element.

[0293] In one possible implementation, the transceiver unit 620 is used to receive the encrypted analysis result and the address information of the inference network element from the second network element; the processing unit 610 is used to send the decrypted analysis result to the inference network element through the transceiver unit 620 according to the address information of the inference network element.

[0294] In one possible implementation, the transceiver unit 620 is configured to receive an association identifier from a training network element and an identifier of a decryption algorithm corresponding to the association identifier before receiving the encrypted analysis result; receive the encrypted analysis result and the association identifier; and the processing unit 610 is configured to determine the decryption algorithm based on the association identifier; and decrypt the encrypted analysis result according to the decryption algorithm to obtain the decrypted analysis result.

[0295] In the sixth embodiment, when the communication device is a second network element or a model (such as a chip) for the second network element, the transceiver unit 620 is used to receive the identification information of the analysis type and the encrypted model supporting the analysis type from the training network element, wherein the type of model deployment platform supported by the second network element includes the type of model deployment platform of the training network element; receive the data to be analyzed from the inference network element; the processing unit 610 is used to obtain the encrypted analysis result based on the encrypted model and the data to be analyzed; the transceiver unit 620 is used to send the encrypted analysis result to the training network element or the first network element and the address information of the inference network element for receiving the decrypted analysis result, wherein the decrypted analysis result is obtained by the training network element or the first network element based on the encrypted analysis result.

[0296] In the seventh embodiment, when the communication device is an inference network element or a model (such as a chip) used for an inference network element, the transceiver unit 620 is configured to send a request message to the data management network element, the request message including identification information of the analysis type, the request message being used to request a network element that supports the analysis type; receive a response message from the data management network element, the response message including at least one set of information, each set of information including address information of a candidate training network element and model information of the candidate training network element, the candidate training network element supporting the analysis type, the model information of the candidate training network element including the vendor type of the candidate training network element and the type of the model deployment platform of the candidate training network element; the processing unit 610 is configured to select one candidate training network element from the one or more candidate training network elements as the training network element when there are one or more candidate training network elements among the at least one set of information that have a different vendor type than the inference network element but the same model deployment platform type.

[0297] In one possible implementation, the processing unit 610 is configured to determine the address information of the second network element based on the at least one set of information when there is no candidate training network element among the at least one candidate training network element corresponding to the at least one set of information that has a different vendor type than the inference network element but the same model deployment platform type.

[0298] In one possible implementation, the model information of the candidate training network element includes the address information of the second network element; the processing unit 610 is used to obtain the address information of the second network element from the model information of the candidate training network element.

[0299] A more detailed description of the processing unit 610 and the transceiver unit 620 can be obtained directly from the relevant descriptions in the above method embodiments, and will not be repeated here.

[0300] like Figure 7 As shown, the communication device 700 includes a processor 710. As one possible implementation, the communication device 700 also includes an interface circuit 720. The processor 710 and the interface circuit 720 are coupled to each other. It is understood that the interface circuit 720 can be a transceiver or an input / output interface. As another possible implementation, the communication device 700 may also include a memory 730 for storing instructions executed by the processor 710, or storing input data required by the processor 710 to execute instructions, or storing data generated after the processor 710 executes instructions.

[0301] When the communication device 700 is used to implement the above method embodiment, the processor 710 is used to implement the function of the processing unit 610, and the interface circuit 720 is used to implement the function of the transceiver unit 620.

[0302] It is understood that the processor in the embodiments of this application may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. A general-purpose processor may be a microprocessor or any conventional processor.

[0303] The method steps in the embodiments of this application can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in random access memory, flash memory, read-only memory, programmable read-only memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, registers, hard disks, portable hard disks, CD-ROMs, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and storage medium can reside in an ASIC. Furthermore, the ASIC can reside in an access network device or a terminal device. Alternatively, the processor and storage medium can exist as discrete components in the access network device or terminal device.

[0304] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer programs or instructions. When the computer program or instructions are loaded and executed on a computer, the processes or functions described in the embodiments of this application are performed entirely or partially. The computer can be a general-purpose computer, a special-purpose computer, a computer network, an access network device, a terminal device, or other programmable device. The computer program or instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer program or instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium, such as a floppy disk, hard disk, or magnetic tape; it can also be an optical medium, such as a digital video optical disc; or it can be a semiconductor medium, such as a solid-state drive. The computer-readable storage medium may be a volatile or non-volatile storage medium, or may include both types of storage media.

[0305] In the various embodiments of this application, unless otherwise specified or in case of logical conflict, the terminology and / or descriptions between different embodiments are consistent and can be referenced by each other. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0306] In this application, "at least one" means one or more, and "more than one" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. In the textual description of this application, the character " / " generally indicates an "or" relationship between the preceding and following related objects; in the formulas of this application, the character " / " indicates a "division" relationship between the preceding and following related objects.

[0307] It is understood that the various numerical designations used in the embodiments of this application are merely for descriptive convenience and are not intended to limit the scope of the embodiments of this application. The order of the process numbers described above does not imply the order of execution; the execution order of each process should be determined by its function and internal logic.

Claims

1. A communication method, characterized in that, include: The inference network element sends a first request message to the training network element. The first request message includes identification information of the analysis type. The first request message is used to request a model that supports the analysis type. The training network element and the inference network element are of different vendor types, but the inference network element and the training network element have the same model deployment platform type. The inference network element receives a first response message from the training network element. The first response message includes an encrypted model or the address information of the encrypted model, and the encrypted model supports the analysis type. The inference network element obtains encrypted analysis results based on the encrypted model; The inference network element obtains the decryption analysis result based on the encrypted analysis result.

2. The method as described in claim 1, characterized in that, The inference network element obtains the decryption analysis result based on the encryption analysis result, including: The inference network element sends the encrypted analysis result to the training network element; The inference network element receives the decryption analysis results from the training network element.

3. The method as described in claim 2, characterized in that, The first response message also includes a first instruction, which instructs the training network element to decrypt the encrypted analysis result.

4. The method as described in claim 2, characterized in that, The inference network element sends the encrypted analysis result to the training network element, including: The inference network element sends the encrypted analysis result and association identifier to the training network element. The association identifier is used by the training network element to determine the encryption algorithm corresponding to the encrypted model.

5. The method as described in claim 3, characterized in that, The inference network element sends the encrypted analysis result to the training network element, including: The inference network element sends the encrypted analysis result and association identifier to the training network element. The association identifier is used by the training network element to determine the encryption algorithm corresponding to the encrypted model.

6. The method as described in claim 1, characterized in that, The first response message also includes the address information of the first network element; The inference network element obtains the decryption analysis result based on the encryption analysis result, including: The inference network element sends the encrypted analysis result to the first network element based on the address information of the first network element; The inference network element receives the decryption analysis result from the first network element.

7. The method as described in claim 6, characterized in that, The inference network element sends the encrypted analysis result to the first network element based on the address information of the first network element, including: The inference network element sends the encrypted analysis result and association identifier to the first network element based on the address information of the first network element. The association identifier is used by the first network element to determine the encryption algorithm corresponding to the encrypted model.

8. The method according to any one of claims 1 to 7, characterized in that, The first response message also includes second indication information, which is used to indicate the data type of the input data corresponding to the encrypted model.

9. The method according to any one of claims 1 to 7, characterized in that, The first request message also includes the vendor type of the inference network element and the type of the model deployment platform of the inference network element.

10. The method as described in claim 8, characterized in that, The first request message also includes the vendor type of the inference network element and the type of the model deployment platform of the inference network element.

11. The method according to any one of claims 1 to 7 and 10, characterized in that, Before the inference network element sends the first request message to the training network element, it also includes: The inference network element sends a second request message to the data management network element. The second request message includes the identification information of the analysis type and is used to request a network element that supports the analysis type. The inference network element receives a second response message from the data management network element, the second response message including the address information of the training network element.

12. The method as described in claim 8, characterized in that, Before the inference network element sends the first request message to the training network element, it also includes: The inference network element sends a second request message to the data management network element. The second request message includes the identification information of the analysis type and is used to request a network element that supports the analysis type. The inference network element receives a second response message from the data management network element, the second response message including the address information of the training network element.

13. The method as described in claim 9, characterized in that, Before the inference network element sends the first request message to the training network element, it also includes: The inference network element sends a second request message to the data management network element. The second request message includes the identification information of the analysis type and is used to request a network element that supports the analysis type. The inference network element receives a second response message from the data management network element, the second response message including the address information of the training network element.

14. A communication method, characterized in that, include: The training network element receives a first request message from the inference network element. The first request message includes identification information of the analysis type. The first request message is used to request a model that supports the analysis type. The training network element and the inference network element are of different vendor types, but the inference network element and the training network element are of the same model deployment platform type. The training network element sends a first response message to the inference network element, the first response message including the encrypted model or the address information of the encrypted model; The training network element receives encrypted analysis results from the inference network element, and the encrypted analysis results are obtained based on the encrypted model; The training network element decrypts the encrypted analysis result to obtain the decrypted analysis result; The training network element sends the decryption analysis result to the inference network element.

15. The method as described in claim 14, characterized in that, The first request message also includes the vendor type of the inference network element and the type of the model deployment platform of the inference network element; Before the training network element sends the first response message to the inference network element, the method further includes: The training network element is determined to be from a different vendor type than the inference network element, and the inference network element and the training network element are from the same model deployment platform type.

16. The method as described in claim 14, characterized in that, The first response message also includes a first instruction, which instructs the training network element to decrypt the encrypted analysis result.

17. The method as described in claim 15, characterized in that, The first response message also includes a first instruction, which instructs the training network element to decrypt the encrypted analysis result.

18. The method according to any one of claims 14 to 17, characterized in that, The first response message also includes second indication information, which is used to indicate the data type of the input data corresponding to the encrypted model.

19. The method according to any one of claims 14 to 17, characterized in that, Before the training network element receives the first request message from the inference network element, it also includes: The training network element sends a registration request message to the data management network element. The registration request message includes the identification information of the analysis type and the model information of the training network element. The model information includes the vendor type of the training network element and the type of the model deployment platform of the training network element.

20. The method as described in claim 18, characterized in that, Before the training network element receives the first request message from the inference network element, it also includes: The training network element sends a registration request message to the data management network element. The registration request message includes the identification information of the analysis type and the model information of the training network element. The model information includes the vendor type of the training network element and the type of the model deployment platform of the training network element.

21. The method according to any one of claims 14 to 17, 20, characterized in that, The training network element receives encrypted analysis results from the inference network element, including: The training network element receives the encrypted analysis results and association identifiers from the inference network element; The training network element decrypts the encrypted analysis result to obtain the decrypted analysis result, including: The training network element determines the encryption algorithm corresponding to the encrypted model based on the association identifier; The training network element determines the decryption algorithm based on the encryption algorithm; The training network element decrypts the encrypted analysis result according to the decryption algorithm to obtain the decrypted analysis result.

22. The method as described in claim 18, characterized in that, The training network element receives encrypted analysis results from the inference network element, including: The training network element receives the encrypted analysis results and association identifiers from the inference network element; The training network element decrypts the encrypted analysis result to obtain the decrypted analysis result, including: The training network element determines the encryption algorithm corresponding to the encrypted model based on the association identifier; The training network element determines the decryption algorithm based on the encryption algorithm; The training network element decrypts the encrypted analysis result according to the decryption algorithm to obtain the decrypted analysis result.

23. The method as described in claim 19, characterized in that, The training network element receives encrypted analysis results from the inference network element, including: The training network element receives the encrypted analysis results and association identifiers from the inference network element; The training network element decrypts the encrypted analysis result to obtain the decrypted analysis result, including: The training network element determines the encryption algorithm corresponding to the encrypted model based on the association identifier; The training network element determines the decryption algorithm based on the encryption algorithm; The training network element decrypts the encrypted analysis result according to the decryption algorithm to obtain the decrypted analysis result.

24. A communication method, characterized in that, include: The inference network element sends a request message to the training network element. The request message includes identification information of the analysis type. The request message is used to request a model that supports the analysis type. The training network element and the inference network element are from different vendors and the model deployment platforms of the inference network element and the training network element are from different types. The inference network element receives a response message from the training network element. The response message includes first indication information and address information of the second network element. The first indication information indicates that the request to support the model of the analysis type is rejected. The type of model deployment platform supported by the second network element includes the type of model deployment platform of the training network element. The inference network element sends the data to be analyzed to the second network element according to the address information of the second network element. The data to be analyzed is used by the second network element to generate encrypted analysis results according to the encrypted model corresponding to the analysis type. The inference network element receives the decrypted analysis result from the training network element or the first network element, and the decrypted analysis result is obtained by the training network element or the first network element based on the encrypted analysis result.

25. The method as described in claim 24, characterized in that, The response message also includes a rejection reason value, which is that the training network element and the inference network element are of different vendor types and that the inference network element and the training network element are of different model deployment platform types.

26. The method as described in claim 24, characterized in that, The request message also includes the vendor type of the inference network element and the type of the model deployment platform of the inference network element.

27. The method as described in claim 25, characterized in that, The request message also includes the vendor type of the inference network element and the type of the model deployment platform of the inference network element.

28. The method according to any one of claims 24 to 27, characterized in that, The response message also includes second indication information, which is used to indicate the data type of the input data corresponding to the encrypted model.

29. The method according to any one of claims 24 to 27, characterized in that, The encryption model is encrypted using one or more of the following algorithms: fully homomorphic encryption, random secure averaging, or differential privacy.

30. The method as described in claim 28, characterized in that, The encryption model is encrypted using one or more of the following algorithms: fully homomorphic encryption, random secure averaging, or differential privacy.

31. The method according to any one of claims 24 to 27 and 30, characterized in that, The inference network element sends the data to be analyzed to the second network element based on the address information of the second network element, including: The inference network element sends the data to be analyzed and the association identifier to the second network element based on the address information of the second network element. The association identifier is used by the first network element or the training network element to determine the encryption algorithm corresponding to the encrypted model.

32. The method as described in claim 28, characterized in that, The inference network element sends the data to be analyzed to the second network element based on the address information of the second network element, including: The inference network element sends the data to be analyzed and the association identifier to the second network element based on the address information of the second network element. The association identifier is used by the first network element or the training network element to determine the encryption algorithm corresponding to the encrypted model.

33. The method as described in claim 29, characterized in that, The inference network element sends the data to be analyzed to the second network element based on the address information of the second network element, including: The inference network element sends the data to be analyzed and the association identifier to the second network element based on the address information of the second network element. The association identifier is used by the first network element or the training network element to determine the encryption algorithm corresponding to the encrypted model.

34. A communication method, characterized in that, include: The training network element receives a request message from the inference network element. The request message includes identification information of the analysis type. The request message is used to request a model that supports the analysis type. The training network element and the inference network element are from different vendors and have different model deployment platforms. The training network element sends a response message to the inference network element. The response message includes first indication information and address information of the second network element. The first indication information indicates that the request to support the model of the analysis type is rejected. The type of model deployment platform supported by the second network element includes the type of model deployment platform of the training network element. The training network element receives encrypted analysis results from the second network element, and the encrypted analysis results are obtained by the second network element based on the data to be analyzed by the inference network element and the encrypted model corresponding to the analysis type. The training network element decrypts the encrypted analysis result to obtain the decrypted analysis result; The training network element sends the decryption analysis result to the inference network element.

35. The method as described in claim 34, characterized in that, The request message also includes the vendor type of the inference network element and the type of the model deployment platform of the inference network element; Before the training network element sends a response message to the inference network element, the method further includes: The training network element is determined to be from a different vendor type than the inference network element, and the inference network element is from a different model deployment platform type than the training network element.

36. The method as described in claim 34, characterized in that, The response message also includes a rejection reason value, which is that the training network element and the inference network element are of different vendor types and that the inference network element and the training network element are of different model deployment platform types.

37. The method as described in claim 35, characterized in that, The response message also includes a rejection reason value, which is that the training network element and the inference network element are of different vendor types and that the inference network element and the training network element are of different model deployment platform types.

38. The method according to any one of claims 34 to 37, characterized in that, Before the training network element receives the request message from the inference network element, it also includes: The training network element sends the identification information of the analysis type and the encrypted model corresponding to the analysis type to the second network element.

39. The method according to any one of claims 34 to 37, characterized in that, The response message also includes second indication information, which is used to indicate the data type of the input data corresponding to the encrypted model.

40. The method as described in claim 38, characterized in that, The response message also includes second indication information, which is used to indicate the data type of the input data corresponding to the encrypted model.

41. The method according to any one of claims 34 to 37, 40, characterized in that, The training network element receives encrypted analysis results from the second network element, including: The training network element receives the encrypted analysis results and association identifier from the second network element; The training network element decrypts the encrypted analysis result to obtain the decrypted analysis result, including: The training network element determines the encryption algorithm corresponding to the encrypted model based on the association identifier; The training network element determines the decryption algorithm based on the encryption algorithm; The training network element decrypts the encrypted analysis result according to the decryption algorithm to obtain the decrypted analysis result.

42. The method as described in claim 38, characterized in that, The training network element receives encrypted analysis results from the second network element, including: The training network element receives the encrypted analysis results and association identifier from the second network element; The training network element decrypts the encrypted analysis result to obtain the decrypted analysis result, including: The training network element determines the encryption algorithm corresponding to the encrypted model based on the association identifier; The training network element determines the decryption algorithm based on the encryption algorithm; The training network element decrypts the encrypted analysis result according to the decryption algorithm to obtain the decrypted analysis result.

43. The method as described in claim 39, characterized in that, The training network element receives encrypted analysis results from the second network element, including: The training network element receives the encrypted analysis results and association identifier from the second network element; The training network element decrypts the encrypted analysis result to obtain the decrypted analysis result, including: The training network element determines the encryption algorithm corresponding to the encrypted model based on the association identifier; The training network element determines the decryption algorithm based on the encryption algorithm; The training network element decrypts the encrypted analysis result according to the decryption algorithm to obtain the decrypted analysis result.

44. A communication device, characterized in that, The device includes a processor and a memory; the memory is used to store computer instructions, and when the device is running, the processor executes the computer instructions stored in the memory to cause the device to perform the method of any one of claims 1 to 13, 24 to 33, or the method of any one of claims 14 to 23, 34 to 43.

45. A computer-readable storage medium, characterized in that, The storage medium stores a computer program or instructions, which, when executed by a communication device, implement the method as described in any one of claims 1 to 43.

46. ​​A communication system, characterized in that, include: An inference network element, used to perform the method as described in any one of claims 1 to 13; as well as The training network element is used to send an encrypted model or the address information of the encrypted model to the inference network element.

47. A communication system, characterized in that, include: An inference network element is used to send a first request message to a training network element. The first request message includes identification information for the analysis type and is used to request a model that supports the analysis type. The training network element is used to perform the method as described in any one of claims 14 to 23.

48. A communication system, characterized in that, include: Inference network element, used to perform the method as described in any one of claims 24 to 33; as well as The training network element is used to send the address information of the second network element to the inference network element, wherein the type of model deployment platform supported by the second network element includes the type of model deployment platform of the training network element.

49. A communication system, characterized in that, include: An inference network element is used to send a request message to the training network element. The request message includes identification information for the analysis type and is used to request a model that supports the analysis type. The training network element is used to perform the method as described in any one of claims 34 to 43.

Images