Communication method and communication apparatus

Abstract

The present application relates to the technical field of communications, and provides a communication method and a communication apparatus. The method comprises: receiving first information, the first information being used for requesting a first service; and sending second information to a second network element, the second information being used for instructing the second network element to perform an operation corresponding to the first service; wherein the first network element is configured to control model training, control model reasoning, and / or control data processing, and the second network element is configured for model training, model reasoning, and / or data processing. The method in embodiments of the present application is conducive to enhancing the scalability of a communication system.

Description

Communication methods and communication devices

[0001] This application claims priority to Chinese Patent Application No. 202510363764.5, filed on March 24, 2025, entitled "Communication Method and Communication Device", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of communication technology, specifically to a communication method and a communication device. Background Technology

[0003] With the development of communication technology, users have increasingly higher demands for communication systems. For example, future communication systems may simultaneously possess basic capabilities such as communication, computing, data processing, artificial intelligence (AI), and sensing. However, the existing communication system architecture may not be able to meet the growing user demands. Summary of the Invention

[0004] This application provides a communication method and a communication device, which helps to enhance the scalability of communication systems.

[0005] In a first aspect, a communication method is provided, the method being applied to a first network element or a component within the first network element (e.g., a processor, chip, chip system, circuit, or a functional module, etc.), the method comprising:

[0006] The system receives first information, which is used to request a first service; it sends second information to a second network element, which is used to instruct the second network element to perform the operation corresponding to the first service; wherein the first network element is used to control model training, control model inference and / or control data processing, and the second network element is used to train the model, infer the model and / or process the data.

[0007] In the embodiments of this application, the first network element is used to control model training, control model inference, and / or control data processing, and the second network element is used to perform model training, model inference, and / or data processing. The first network element receives first information for requesting a first service and instructs the second network element to perform the operation corresponding to the first service. This helps to decouple control functions (such as controlling model training, controlling model inference, and / or controlling data processing) from service processing functions (such as model training, model inference, and / or data processing), thereby helping to enhance the scalability of the communication system.

[0008] In some possible implementations, the method further includes: receiving third information from the second network element, the third information being used by the second network element for registration.

[0009] In this embodiment of the application, receiving third information helps the first network element to know the capabilities of the second network element, thereby helping the first network element to control the second network element to perform corresponding operations.

[0010] In some possible implementations, the third information includes one or more of the following: the identifier of the second network element, the federated learning capability information of the second network element, the analysis capability information of the second network element, and the computing capability information of the second network element.

[0011] In some possible implementations, the method further includes: determining, based on the third information, whether the second network element is a server-side network element or a client-side network element for federated learning.

[0012] In this embodiment of the application, the third information can be used to determine whether the second network element is more suitable as a server-side network element or a client-side network element in federated learning, thereby improving the efficiency of service execution.

[0013] In some possible implementations, the method further includes: receiving fourth information from a third network element, the fourth information being used to indicate model information stored by the third network element, the third network element being used to store data; and determining a first model based on the fourth information.

[0014] In this embodiment of the application, receiving the fourth information helps the first network element to know the model information stored by the third network element, helps the first network element to determine a suitable model, and thus helps to improve the execution efficiency of the service.

[0015] In some possible implementations, the second network element is a server-side network element of federated learning, and the second information is used to instruct the second network element to publish the first model to the client network element of federated learning, and / or the second network element to subscribe to the second model to the client network element of federated learning, wherein the second model is obtained based on the first model after training.

[0016] In some possible implementations, the method further includes: sending a fifth message to the second network element, the fifth message being used to instruct the second network element to subscribe to the first model from a third network element, the third network element being used to store data.

[0017] In this embodiment of the application, the fifth information is used to instruct the second network element to subscribe to the first model from the third network element. Sending the fifth information to the second network element helps the second network element to execute services based on a suitable model, thereby helping to improve the efficiency of service execution.

[0018] In some possible implementations, the method further includes: sending a sixth message to a third network element, the sixth message being used to instruct the third network element to publish the first model to the second network element, the third network element being used to store data.

[0019] In this embodiment of the application, the sixth information is used to instruct the third network element to publish the first model to the second network element. Sending the sixth information to the third network element helps the second network element to execute services based on a suitable model, thereby helping to improve the efficiency of service execution.

[0020] In some possible implementations, the second network element is a client network element of federated learning, and the second information is used to instruct the second network element to subscribe to the first model from the server network element of federated learning, and / or the second network element to publish the second model to the server network element of federated learning, the second model being obtained based on the first model after training.

[0021] In some possible implementations, the first network element is equipped with a first communication proxy, and the second network element is equipped with a second communication proxy, wherein the first communication proxy is used to transmit control signaling, and the second communication proxy is used to transmit service data.

[0022] In this embodiment, the first network element is attached to a first communication agent, and the second network element is attached to a second communication agent. In this way, control signaling can be transmitted through the first communication agent, and service data can be transmitted through the second communication agent, thereby improving the transmission efficiency of the communication system.

[0023] In some possible implementations, the first network element and the second network element are used for horizontal federated learning, or the first network element and the second network element are used for vertical federated learning.

[0024] Secondly, a communication method is provided, the method being applied to a second network element or a component within the second network element (e.g., a processor, chip, chip system, circuit, or a functional module, etc.), the method comprising:

[0025] The system receives second information from a first network element, the second information being used to instruct the second network element to perform the operation corresponding to the first service; and performs the corresponding operation according to the second information; wherein the first network element is used to control model training, control model inference, and / or control data processing, and the second network element is used for model training, model inference, and / or data processing.

[0026] In the embodiments of this application, the first network element is used to control model training, control model inference, and / or control data processing, and the second network element is used to perform model training, model inference, and / or data processing. The second network element receives second information from the first network element and performs corresponding operations based on the instructions of the first network element. This helps to decouple control functions (such as controlling model training, controlling model inference, and / or controlling data processing) from service processing functions (such as model training, model inference, and / or data processing), thereby helping to enhance the scalability of the communication system.

[0027] In some possible implementations, the method further includes: sending third information to the first network element, the third information being used by the second network element for registration.

[0028] In this embodiment of the application, sending third information to the first network element helps the first network element to know the capabilities of the second network element, thereby helping the first network element to control the second network element to perform corresponding operations.

[0029] In some possible implementations, the third information includes one or more of the following: the identifier of the second network element, the federated learning capability information of the second network element, the analysis capability information of the second network element, and the computing capability information of the second network element.

[0030] In some possible implementations, the second network element is a server-side network element of federated learning, and the second information is used to instruct the second network element to publish the first model to the client network element of federated learning, and / or the second network element to subscribe to the second model to the client network element of federated learning, wherein the second model is obtained based on the first model after training.

[0031] In some possible implementations, the method further includes: publishing the first model to the client network element of the federated learning through a second communication agent; and / or subscribing to the second model to the client network element of the federated learning through the second communication agent.

[0032] In some possible implementations, the method further includes: receiving fifth information from the first network element, the fifth information being used to instruct the second network element to subscribe to the first model from the third network element, the third network element being used to store data; and subscribing to the first model from the third network element through a second communication proxy according to the fifth information.

[0033] In this embodiment of the application, the fifth information is used to instruct the second network element to subscribe to the first model from the third network element. The second network element receives the fifth information from the first network element and subscribes to the first model according to the fifth information, which helps the second network element to execute services based on a suitable model, thereby helping to improve the execution efficiency of services.

[0034] In some possible implementations, the second network element is a client network element of federated learning, and the second information is used to instruct the second network element to subscribe to the first model from the server network element of federated learning, and / or the second network element to publish the second model to the server network element of federated learning, the second model being obtained based on the first model after training.

[0035] In some possible implementations, the method further includes: subscribing to the first model from the server-side network element of the federated learning via a second communication proxy; and / or publishing the second model to the server-side network element of the federated learning via the second communication proxy.

[0036] In some possible implementations, the first network element is equipped with a first communication proxy, and the second network element is equipped with a second communication proxy, wherein the first communication proxy is used to transmit control signaling, and the second communication proxy is used to transmit service data.

[0037] In this embodiment, the first network element is attached to a first communication agent, and the second network element is attached to a second communication agent. In this way, control signaling can be transmitted through the first communication agent, and service data can be transmitted through the second communication agent, thereby improving the transmission efficiency of the communication system.

[0038] In some possible implementations, the first network element and the second network element are used for horizontal federated learning, or the first network element and the second network element are used for vertical federated learning.

[0039] Thirdly, a communication method is provided, the method being applied to a third network element or a component within the third network element (e.g., a processor, chip, chip system, circuit, or a functional module, etc.), the method comprising:

[0040] The system receives a sixth message from a first network element, the sixth message being used to instruct the third network element to publish a first model to a second network element; the system publishes the first model to the second network element according to the sixth message; wherein the first network element is used to control model training, control model inference and / or control data processing, the second network element is used to train the model, infer the model and / or process the data, and the third network element is used to store the data.

[0041] In this embodiment, the sixth information is used to instruct the third network element to publish the first model to the second network element. Receiving the sixth information from the first network element and publishing the first model to the second network element according to the sixth information helps the second network element to execute services based on a suitable model, thereby helping to improve the efficiency of service execution.

[0042] In some possible implementations, the step of publishing the first model to the second network element according to the sixth information includes: publishing the first model to the second network element through a second communication proxy according to the sixth information, wherein the second communication proxy is used to transmit service data.

[0043] In some possible implementations, the method further includes: sending fourth information to the first network element, the fourth information being used to indicate the model information stored by the third network element.

[0044] In this embodiment of the application, sending the fourth information to the first network element helps the first network element to know the model information stored by the third network element, helps the first network element to determine a suitable model, and thus helps to improve the execution efficiency of the service.

[0045] Fourthly, a communication device is provided, comprising: the device can be used in a first network element of the first aspect; the device can be the first network element, or a device in the first network element (e.g., a chip, a chip system, a circuit, or a processor), or a device that can be matched with the first network element, or a logic module or software that can implement all or part of the first network element.

[0046] The device includes modules that perform the methods / operations / steps / actions described in the first aspect or any possible implementation of the first aspect. These modules can be hardware circuits, software, or a combination of hardware circuits and software.

[0047] Fifthly, a communication device is provided, comprising: the device can be used for a second network element in the second aspect; the device can be the second network element, or a device within the second network element (e.g., a chip, a chip system, a circuit, or a processor), or a device that can be used in conjunction with the second network element, or a logic module or software that can implement all or part of the second network element.

[0048] The device includes modules that perform the methods / operations / steps / actions described in the second aspect or any possible implementation of the second aspect. These modules can be hardware circuits, software, or a combination of hardware circuits and software.

[0049] In a sixth aspect, a communication device is provided, comprising: the device can be used for a third network element in the third aspect; the device can be a third network element, or a device within a third network element (e.g., a chip, a chip system, a circuit, or a processor), or a device that can be used in conjunction with a third network element, or a logic module or software that can implement all or part of a third network element.

[0050] The device includes modules that perform the methods / operations / steps / actions described in the first aspect or any possible implementation of the first aspect. These modules can be hardware circuits, software, or a combination of hardware circuits and software.

[0051] A seventh aspect provides a communication device comprising: a processor and a memory, the processor being coupled to the memory, the memory being used to store a computer program (also referred to as code or instructions), the computer program being executed by the processor causing the device to perform the method of the first aspect or any possible implementation thereof.

[0052] In some possible implementations, the device also includes a memory coupled to the processor.

[0053] In some possible implementations, there are one or more processors, and / or one or more memories.

[0054] In some possible implementations, the memory can be integrated with the processor, or the memory can be set up separately from the processor.

[0055] Eighthly, a communication device is provided, comprising: a processor and a memory, the processor being coupled to the memory, the memory being used to store a computer program (also referred to as code or instructions), the computer program being executed by the processor causing the device to perform the method of the second aspect or any possible implementation thereof.

[0056] In some possible implementations, the device also includes a memory coupled to the processor.

[0057] In some possible implementations, there are one or more processors, and / or one or more memories.

[0058] In some possible implementations, the memory can be integrated with the processor, or the memory can be set up separately from the processor.

[0059] A ninth aspect provides a communication device comprising: a processor and a memory, the processor being coupled to the memory, the memory being used to store a computer program (also referred to as code or instructions), the computer program being executed by the processor causing the device to perform the method of the third aspect or any possible implementation thereof.

[0060] In some possible implementations, the device also includes a memory coupled to the processor.

[0061] In some possible implementations, there are one or more processors, and / or one or more memories.

[0062] In some possible implementations, the memory can be integrated with the processor, or the memory can be set up separately from the processor.

[0063] In a tenth aspect, a computer-readable storage medium is provided, on which a computer program (also referred to as code or instructions) is stored, which, when run on a computer, causes the computer to perform the method of any of the above aspects or any possible implementation thereof.

[0064] Eleventhly, a computer program product is provided, comprising: a computer program (also referred to as code or instructions) that, when run on a computer, causes the computer to perform the method of any of the above aspects or any possible implementation thereof.

[0065] In a twelfth aspect, a chip is provided, comprising: a processor and a memory, the memory for storing a computer program (also referred to as code or instructions), the processor for calling and running the computer program stored in the memory, such that an apparatus or device on which the chip is mounted performs the method of any of the above aspects or any possible implementation thereof. Attached Figure Description

[0066] Figure 1 is a schematic diagram of a centralized service architecture for NWDAF in an embodiment of this application.

[0067] Figure 2 is a schematic diagram of a training-inference separated NWDAF service architecture in an embodiment of this application.

[0068] Figure 3 is a schematic diagram of a multi-NWDAF collaborative hierarchical intelligent architecture in an embodiment of this application.

[0069] Figure 4 is a schematic diagram of a data management architecture for NWDAF in an embodiment of this application.

[0070] Figure 5 is a schematic diagram of a horizontal federation architecture based on NWDAF in an embodiment of this application.

[0071] Figure 6 is a schematic diagram of a vertical federation architecture based on NWDAF in an embodiment of this application.

[0072] Figure 7 is a schematic diagram of a communication system architecture in an embodiment of this application.

[0073] Figure 8 is a schematic flowchart of a communication method provided in one embodiment of this application.

[0074] Figure 9 is a schematic diagram of a communication system architecture provided in one embodiment of this application.

[0075] Figure 10 is a schematic flowchart of a communication method provided in another embodiment of this application.

[0076] Figure 11 is a schematic diagram of a horizontal federation architecture based on NWDAF control and business separation in an embodiment of this application.

[0077] Figure 12 is a schematic diagram of a vertical federation architecture based on NWDAF control and business separation in an embodiment of this application.

[0078] Figure 13 is a schematic structural diagram of a communication device provided in one embodiment of this application.

[0079] Figure 14 is a schematic structural diagram of a communication device provided in another embodiment of this application.

[0080] Figure 15 is a schematic structural diagram of a communication device provided in another embodiment of this application.

[0081] Figure 16 is a schematic structural diagram of an apparatus provided in one embodiment of this application. Detailed Implementation

[0082] The technical solutions in the embodiments of this application will now be described with reference to the accompanying drawings.

[0083] In the description of this application, unless otherwise stated, " / " indicates that the objects before and after are in an "or" relationship. For example, A / B can represent A or B. "And / or" in this application merely 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, and B alone, where A and B can be singular or plural. Furthermore, in the description of this application, unless otherwise stated, "multiple" refers to two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple. Additionally, to facilitate a clear description of the technical solutions of the embodiments of this application, the terms "first" and "second" are used in the embodiments of this application to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first," "second," etc., do not limit the quantity or order of execution, and that "first," "second," etc., do not necessarily imply that they are different. It should be understood that in this application, descriptions such as "in the case of," "if," "when," "if," etc., can be used interchangeably.

[0084] The technical solutions of this application can be applied to various communication systems, such as 5th generation (5G) systems or new radio (NR), long term evolution (LTE) systems, LTE frequency division duplex (FDD) systems, LTE time division duplex (TDD) systems, satellite and other non-terrestrial communication systems, and communication systems that integrate terrestrial and non-terrestrial communication. The technical solutions provided in this application can also be applied to future communication systems.

[0085] Network data analytics function (NWDAF) is an important network function in communication systems (such as 5G core network (5GC)), providing specific intelligent analysis services to the network. The evolution of NWDAF network elements is explained in detail below.

[0086] The evolution of NWDAF can be divided into the following stages:

[0087] Phase 1: The 3rd generation partnership project (3GPP) Release 16 phase, which first proposed the NWDAF function.

[0088] As shown in Figure 1, NWDAF in version R16 is a centralized service architecture and can have the following functions:

[0089] Collect network data from data sources (such as 5GC, application function (AF) network elements, or operation, administration and management (OAM) network elements);

[0090] Request-feedback service model: Consumers can send subscription messages to NWDAF to request analysis results or data; NWDAF can notify consumers of the corresponding results (after analysis and processing).

[0091] Statistical results and predictive analysis: NWDAF can perform corresponding analysis and processing based on consumer requests.

[0092] The centralized service architecture of NWDAF has problems such as poor deployment flexibility and insufficient scalability.

[0093] Phase 2: R17, which proposed a training-inference separation NWDAF service architecture.

[0094] NWDAF can separate model training and analysis / inference functions to make the training platform more accessible. For example, as shown in Figure 2, it is divided into an analysis logical function (AnLF) and a model training logical function (MTLF). The AnLF implements the analysis / inference function of NWDAF, while the MTLF implements the model training function of NWDAF.

[0095] Multiple NWDAFs can also collaborate on processing. As shown in Figure 3, multiple child NWDAFs can receive sub-analysis tasks dispatched by the aggregation point NWDAF, perform analysis and processing, and return the sub-analysis results to the aggregation point NWDAF. The aggregation point NWDAF then analyzes and aggregates the results from the multiple sub-analysis tasks.

[0096] A data management framework can also be introduced, as shown in Figure 4. The data collection coordination function (DCCF) can be used to coordinate data collection and avoid duplication; the message framework adaptor function (MFAF) can be used for protocol conversion and information exchange between the message framework and the 3GPP system, reducing the communication overhead of performing data collection tasks; and the analytics data repository function (ADRF) can be used for data / analysis report storage management, alleviating the data storage burden on the NWDAF.

[0097] The third stage, R18, proposes a horizontal federated architecture based on NWDAF to achieve distributed federated learning.

[0098] As shown in Figure 5, the aggregation server is used for participating server discovery, model training task dispatch (such as initial model distribution), and model aggregation; participating servers (such as participant A, participant B, participant X, etc. (X is a positive integer)) can train based on local data and upload the trained model to the aggregation server.

[0099] Among them, the aggregation server can be a roaming exchange NWDAF (RE-NWDAF) to support roaming exchange; and enable the opening of data resources and analysis capabilities across public land mobile networks (PLMNs);

[0100] Phase 4: R19, proposes a vertical federated learning (VFL) architecture based on NWDAF to achieve cross-domain federated learning.

[0101] As shown in Figure 6, the VLF server-side NWDAF can dispatch analysis requests to the VLF client-side NWDAF. After completing the analysis and inference, the VLF client-side NWDAF can return sub-analysis results to the VLF server-side NWDAF, which then performs analysis and aggregation on the multiple sub-analysis results. Each node's training set contains the same sample data located in different feature spaces.

[0102] In the vertical federated architecture shown in Figure 6, the AF can also participate in cross-domain model construction and analysis reasoning. For example, the network exposure function (NEF) element can forward analysis requests and collect sub-analysis results to the VLF client AF.

[0103] As can be seen from the evolution process from R16 to R19 described above, NWDAF gradually developed the following functions:

[0104] Data capabilities:

[0105] NWDAF collects the required data from the data source based on the control parameters (such as analysis ID, analysis report target, report threshold, analysis period, analysis output strategy, etc.) carried in the data analysis request, performs the analysis, and returns the results.

[0106] The data collection / distribution coordination framework established by DCCF and MFAF decouples data consumers and providers, avoiding duplicate data subscriptions.

[0107] By leveraging ADRF's data storage and retrieval capabilities, the network further improves the utilization of historical data and reduces redundant data collection.

[0108] Regarding privacy protection, when the required data is related to an individual user, NWDAF ensures that data collection complies with the user's permission by retrieving the contracted data; the data does not leave the local area.

[0109] Model capabilities: Full lifecycle management of models.

[0110] The subscription / request process centered on NWDAF is open to any potential consumer, such as terminals, network management, and 5GC network functions, on demand, and supports specified model interoperability information.

[0111] The model can be stored and retrieved in ADRF, enabling flexible cross-domain co-governance and sharing;

[0112] Centralized and distributed federated learning;

[0113] Model performance monitoring mechanism;

[0114] Model retraining and resubscription.

[0115] Collaboration capabilities:

[0116] Analytical aggregation and analytical transfer techniques:

[0117] Analysis aggregation technology integrates multiple NWDAF data resources and analysis capabilities by breaking down and distributing complex analysis tasks; analysis migration technology supports the direct transfer of ongoing analysis subscriptions and analysis contexts from the original NWDAF to the new NWDAF, avoiding additional signaling overhead and improving the continuity of analysis services.

[0118] Horizontal federation and vertical federation enable collaborative computing;

[0119] Cross-domain collaboration and intelligent interaction with the management data analytics function (MDAF).

[0120] Application capabilities:

[0121] Network layer: Based on intelligent analysis of network data, it provides real-time perception of network operation status, thereby improving the efficiency of resource scheduling, management, and orchestration.

[0122] User layer: Analyze and mine historical data to build accurate user profiles and achieve personalized services and refined management at the user level;

[0123] Business layer: Integrates massive user terminal perception data to achieve fusion of sensing, computing, and communication, supporting the automatic intelligent adaptation and dynamic adjustment of network resources to meet business needs.

[0124] With the development of communication technology, users have increasingly higher demands for communication systems. For example, future communication systems may simultaneously possess basic capabilities such as communication, computing, data processing, artificial intelligence (AI), and sensing. However, the existing communication system architecture may not be able to meet the growing user demands.

[0125] The following section uses NWDAF as an example to illustrate the problems existing in current communication system architectures. As can be seen from the evolution process from R16 to R19 described above, the evolution of NWDAF from R16 to R19 gradually exhibits the following trend:

[0126] Architecture design is evolving from centralized to decentralized.

[0127] Model building has evolved from single-node training to distributed training.

[0128] Application scenarios have evolved from a single focus to a diversified range.

[0129] However, in the current NWDAF, control functions and business processing functions (such as intelligent analysis) are still integrated, which poses a challenge to the scalability of NWDAF.

[0130] For example, Figure 7 below shows a schematic diagram of a communication system architecture, including: a unified data repository (UDR), a network exposure function (NEF) network element, a network data analytics function (NWDAF) network element, an application function (AF) network element, a policy control function (PCF) network element, an operation, administration and management (OAM) network element, an access and mobility management function (AMF) network element, a session management function (SMF) network element, user equipment (UE), a radio access network (RAN) (which can be RAN or AN, represented by (R)AN in Figure 1), and a user plane function (UPF) network element, etc.

[0131] In this application, UE stands for Terminal Equipment, which can refer to a station, access terminal, user unit, user station, mobile station, mobile station (MS), remote station, remote terminal, mobile terminal (MT), user terminal, terminal, wireless communication equipment, user agent, or user device, etc. This application does not limit the specific type of terminal equipment. The terminal equipment in this application can also be a mobile phone, cellular phone, cordless phone, session initiation protocol (SIP) phone, wireless local loop (WLL) station, personal digital assistant (PDA), handheld device with wireless communication capabilities, computing device or other processing device connected to a wireless modem, large screen, in-vehicle device, wearable device, terminal equipment in a 5G network, or terminal equipment in a future public land mobile network (PLMN), etc. This application does not limit the specific type of terminal equipment. The terminal device in the embodiments of this application may also be a tablet computer, a laptop computer, a handheld computer, a mobile internet device (MID), a wearable device, a virtual reality (VR) device, an augmented reality (AR) device, a wireless terminal in industrial control, a wireless terminal in self-driving, a wireless terminal in remote medical surgery, a wireless terminal in a smart grid, a wireless terminal in transportation safety, a wireless terminal in a smart city, a wireless terminal in a smart home, etc., and this application does not limit it.

[0132] The NWDAF service processing functions can include model training, inference, data acquisition, and data storage. As shown in Figure 7, the control and service functions of NWDAF are integrated, which affects the scalability of NWDAF and is not conducive to the evolution of NWDAF. Multiple network elements are connected to the service-based interface (SBI), and control signaling and service data are transmitted through the SBI. However, the SBI is not suitable for transmitting large amounts of data and models, which may affect the transmission efficiency of the communication system.

[0133] To address one or more of the aforementioned technical problems, this application proposes a communication method and device that helps decouple control functions (such as control model training, control model inference, and / or control data processing) from service processing functions (such as model training, model inference, and / or data processing), thereby enhancing the scalability of the communication system. The communication method in the embodiments of this application is illustrated in detail below with reference to Figure 8.

[0134] Figure 8 is a schematic flowchart of a communication method provided in an embodiment of this application. The method 800 shown in Figure 8 may include steps S810, S820 and S830, as follows:

[0135] S810, the first network element receives the first information.

[0136] In some embodiments, the first information may be sent by a user, consumer, or service requester. For example, the first information may be sent by a task, operator, or third-party service requester.

[0137] The first network element can be used to implement the control function of NWDAF (e.g., the control function of NWDAF can be represented as NWDAF-C (NWDAF-control)). For example, the first network element can be used to control model training, control model inference, and / or control data processing.

[0138] In some embodiments, the first network element can be used for both horizontal federated learning (or horizontal federated inference) and vertical federated learning (or vertical federated inference). That is, the first network element in the embodiments of this application is applicable to both horizontal and vertical federated architectures.

[0139] In some embodiments, the first network element may be connected to a first communication agent. The first network element being connected to the first communication agent can be understood as the first network element being connected to the first communication agent.

[0140] The first communication agent can be used to transmit control signaling. For example, the first communication agent can be a service communication proxy (SCP) network element.

[0141] Figure 9 is a schematic diagram of a communication system architecture provided in an embodiment of this application. As shown in Figure 9, NWDAF-C can be connected to SCP. In addition, in order to transmit control signaling to network elements (or functions) connected to the New Service Communication Agent (xCP) (such as sending control signaling to a second or third network element), NWDAF-C can also be connected to xCP.

[0142] It should be noted that the first network element can be a newly defined network element (or function) in the communication system, or it can be an existing network element (or function) in the communication system. This application embodiment does not limit this.

[0143] In some embodiments, the first information may be used to request a first service. Optionally, the first service may be a model training service (such as federated learning), a model inference service (such as inference based on an AI model), or other services (such as data analysis or data processing).

[0144] S820: The first network element sends the second information to the second network element.

[0145] The second network element can be used to implement the NWDAF service processing function (e.g., the NWDAF service processing function can be represented as NWDAF-S (NWDAF-service)). Optionally, the second network element can be used for model training, model inference, and / or data processing.

[0146] For example, the second network element can be AnLF, used to perform model training; or it can be MTLF, used to perform model inference; or it can be a data analysis function, used to perform data analysis, etc.

[0147] In some embodiments, the second network element can be used for both horizontal federated learning (or horizontal federated inference) and vertical federated learning (or vertical federated inference). That is, the second network element in the embodiments of this application is applicable to both horizontal and vertical federated architectures.

[0148] In some embodiments, as shown in FIG9, the second network element can be attached to a second communication agent (as shown in FIG9, the new service communication agent (xCP)). The second network element attaching to the second communication agent can be understood as the second network element being connected to the second communication agent.

[0149] The second communication proxy can be used to transmit service data. For example, the second communication proxy can be a new service communication proxy (xCP), a data communication proxy (SCP) network element, or other newly defined network elements in the communication system, but this application embodiment is not limited to this.

[0150] As shown in Figure 9, NWDAF-S can be connected to xCP.

[0151] The MFAF described in the foregoing embodiments (as shown in Figure 4) is used to discover and transmit messages. Therefore, the MFAF can be merged into the second communication agent, that is, the second communication agent can implement some or all of the functions of the MFAF.

[0152] It should be noted that the second network element can be a newly defined network element (or function) in the communication system, or it can be an existing network element (or function) in the communication system. This application embodiment does not limit this.

[0153] In some embodiments, the second information can be used to instruct the second network element to perform the operation corresponding to the first service. For example, the second information can be used to instruct the second network element to perform model training (such as federated learning), to perform model inference (such as inference based on an AI model), or to perform other services (such as data analysis or data processing).

[0154] In some embodiments, the second network element may report capability information to the first network element. For example, before steps S810 and / or S820 above, method 800 may further include step S802, as follows:

[0155] S802, the second network element sends the third information to the first network element.

[0156] The third piece of information can be used for the registration of the second network element.

[0157] In some embodiments, the third information may include one or more of the following: the identifier (ID) of the second network element, the federated learning capability information of the second network element, the analysis capability information of the second network element, and the computing capability information of the second network element.

[0158] Among them, federated learning capability information can be used to indicate whether the second network element supports federated learning, whether the second network element supports horizontal federated learning, and / or whether the second network element supports vertical federated learning, etc.; analysis capability information can be used to indicate the processor of the second network element, the types of services that the second network element can analyze, and the amount of services that can be analyzed, etc.; computing capability information can be used to indicate the processor of the second network element, the types of services that the second network element can compute, and the amount of services that can be computed, etc.

[0159] In this embodiment of the application, the second network element sends third information to the first network element, which helps the first network element to know the capabilities of the second network element, thereby helping the first network element to control the second network element to perform corresponding operations.

[0160] In some embodiments, the first network element can determine whether the second network element is a server network element or a client network element in federated learning based on the third information. For example, before steps S810 and / or S820 above, method 800 may further include step S804, as follows:

[0161] S804, the first network element determines the second network element as either a server-side network element or a client-side network element for federated learning based on the third information.

[0162] In some embodiments, a third network element can report model information to a first network element. The third network element can be used to store data and models. For example, the third network element can be the ADRF described in the preceding embodiments (as shown in Figure 4). As shown in Figure 9, the third network element can be equipped with a second communication agent. Optionally, the ADRF and the data storage function (DSF) can be merged to achieve unified data / model storage management; that is, the merged network element (or function) can implement some or all of the functions of the ADRF and some or all of the functions of the DSF.

[0163] In this embodiment of the application, the first network element can determine, based on the third information, that the second network element is more suitable as a server-side network element or a client-side network element for federated learning, thereby improving the execution efficiency of the service.

[0164] For example, before steps S810 and / or S820 above, method 800 may also include steps S806 and S808, as follows:

[0165] S806, the third network element sends the fourth information to the first network element.

[0166] The fourth piece of information can be used to indicate model information stored in the third network element. For example, the fourth piece of information can indicate the model's identifier (ID), the model's service area, the model's data network access identifier (DNAI), etc.

[0167] In this embodiment of the application, the third network element sends fourth information to the first network element, which helps the first network element to know the model information stored by the third network element, helps the first network element to determine a suitable model, and thus helps to improve the execution efficiency of the service.

[0168] S808, the first network element determines the first model based on the fourth information.

[0169] The first model can refer to the foundational model used in federated learning.

[0170] For example, the first network element can determine the first model based on the model's ID, the model's service area, and the model's DNAI.

[0171] S830, the second network element performs the corresponding operation based on the second information.

[0172] In some embodiments, the second network element can be a server-side network element of the federated learning. Optionally, the second information can be used to instruct the second network element to publish the first model to the client network element of the federated learning, and / or the second network element to subscribe to the second model to the client network element of the federated learning. The second model can be obtained based on training the first model.

[0173] Optionally, the second network element may publish the first model to the client network element of the federated learning through the second communication agent based on the second information; and / or, the second network element may subscribe to the second model to the client network element of the federated learning through the second communication agent based on the second information.

[0174] In some embodiments, the first network element may instruct the second network element to subscribe to the base model used for federated learning from the third network element. For example, before step S820 above, method 800 may further include step S812, as follows:

[0175] S812, the first network element sends the fifth information to the second network element.

[0176] Among them, the fifth piece of information can be used to instruct the second network element to subscribe to the first model from the third network element.

[0177] In this embodiment of the application, the fifth information is used to instruct the second network element to subscribe to the first model from the third network element. The first network element sends the fifth information to the second network element, which helps the second network element to execute services based on a suitable model, thereby helping to improve the execution efficiency of services.

[0178] Optionally, the second network element can subscribe to the first model from the third network element through the second communication agent based on the fifth information.

[0179] In some embodiments, the first network element may instruct the third network element to publish the basic model to the second network element. For example, before step S820 above, method 800 may further include step S814, as follows:

[0180] S814, the first network element sends the sixth information to the third network element.

[0181] The sixth piece of information can be used to instruct the third network element to publish the first model to the second network element.

[0182] Optionally, the third network element can publish the first model to the second network element through the second communication agent based on the sixth information.

[0183] In some embodiments, the second network element can be a client network element of the federated learning. Optionally, the second information can be used to instruct the second network element to subscribe to the first model from the server-side network element of the federated learning, and / or the second network element to publish the second model to the server-side network element of the federated learning. The second model can be obtained based on training the first model.

[0184] In this embodiment of the application, the sixth information is used to instruct the third network element to publish the first model to the second network element. The first network element sends the sixth information to the third network element, which helps the second network element to execute services based on a suitable model, thereby helping to improve the execution efficiency of services.

[0185] Optionally, the second network element may subscribe to the first model from the server-side network element of the federated learning through the second communication agent based on the second information; and / or, the second network element may publish the second model to the server-side network element of the federated learning through the second communication agent based on the second information.

[0186] In the embodiments of this application, the first network element is used to control model training, control model inference, and / or control data processing, and the second network element is used to perform model training, model inference, and / or data processing. The first network element receives first information for requesting a first service and instructs the second network element to perform the operation corresponding to the first service. This helps to decouple control functions (such as controlling model training, controlling model inference, and / or controlling data processing) from service processing functions (such as model training, model inference, and / or data processing), thereby helping to enhance the scalability of the communication system.

[0187] The method in this application embodiment will be described in detail below with reference to Figure 10, taking the example of a business requester requesting federated learning services from NWDAF-C.

[0188] Figure 10 is a schematic flowchart of a communication method provided in another embodiment of this application. The method 1000 shown in Figure 10 may include steps S1001 to S1013, as follows:

[0189] S1001, NWDAF-S registers with NWDAF-C.

[0190] For example, an NWDAF-S can register its federated learning capabilities, analytical capabilities, and computational capabilities with an NWDAF-C. As shown in Figure 11, each NWDAF-S can register with an NWDAF-C separately.

[0191] It should be noted that method 1000 can be applied to the horizontal federated architecture based on NWDAF control and service separation shown in Figure 11, and can also be applied to the vertical federated architecture based on NWDAF control and service separation shown in Figure 12. This application embodiment does not limit this.

[0192] S1002, ADRF registers with NWDAF-C.

[0193] ADRF registers its model information with NWDAF-C.

[0194] S1003, NWDAF-C receives federated learning requests from service requesters.

[0195] NWDAF-C can receive federated learning requests from tasks, carriers, or third-party business requesters.

[0196] S1004, NWDAF-C determines the basic model.

[0197] NWDAF-C can select a base model based on the model information in ADRF. For example, NWDAF-C can determine the base model based on the model ID, service region, and DNAI stored in ADRF.

[0198] S1005, NWDAF-C determines the NWDAF-S on the server side and the NWDAF-S on the client side.

[0199] NWDAF-C can determine the NWDAF-S on the server side and the NWDAF-S on the client side based on federated learning capability information, analysis capability information, computing capability, etc.

[0200] S1006, NWDAF-C sends the first control signaling to the server.

[0201] The first control signaling can instruct the server to subscribe to topic 1, where topic 1 can correspond to the identifier of the basic model.

[0202] Accordingly, the server can subscribe to topic 1 via xCP.

[0203] S1007, NWDAF-C sends a second control signaling to ADRF.

[0204] The second control signaling can instruct ADRF to publish topic 1.

[0205] Accordingly, ADRF can publish topic 1 via xCP.

[0206] S1008, NWDAF-C sends a third control signaling to the server.

[0207] The third control signaling can instruct the server to publish topic 2 and subscribe to topic 3. Topic 2 can correspond to the identifier of the base model, and topic 3 can correspond to the identifier of the model obtained after the first round of training (i.e., the model obtained after training based on the base model).

[0208] Accordingly, the server can publish topic 2 via xCP and subscribe to topic 3 via xCP after the first round of training.

[0209] S1009, NWDAF-C sends the fourth control signaling to the client.

[0210] The fourth control signaling can instruct the client to subscribe to topic 2 and publish to topic 3.

[0211] Accordingly, the client can subscribe to topic 2 through xCP and perform local training based on the base model; after the first round of training, topic 3 can be published through xCP.

[0212] S1010, the server performs the first round of model aggregation.

[0213] The server can perform the first round of aggregation on the models obtained from the first round of training on multiple clients.

[0214] S1011, NWDAF-C sends the fifth control signaling to the server.

[0215] The fifth control signal can instruct the server to publish topic 4 and subscribe to topic 5. Topic 4 can correspond to the identifier of the model obtained in the first round of aggregation, and topic 5 can correspond to the identifier of the model obtained after the second round of training (i.e., the model obtained after training based on the model obtained in the first round of aggregation).

[0216] Accordingly, the server can publish topic 4 through xCP and perform local training based on the model obtained from the first round of aggregation; after the second round of training, it can subscribe to topic 5 through xCP.

[0217] S1012, NWDAF-C sends the sixth control signaling to the client.

[0218] The sixth control signaling can instruct the client to subscribe to topic 4 and publish to topic 5.

[0219] Accordingly, the client can subscribe to topic 4 via xCP and publish topic 5 via xCP after the first round of training.

[0220] S1013, the server performs the second round of model aggregation.

[0221] The server can perform a second round of aggregation on the models obtained from multiple clients after the second round of training.

[0222] One or more training and aggregation processes can be performed subsequently. The subsequent training and aggregation processes are similar to the steps in the above embodiments, and will not be repeated here.

[0223] The number of training and aggregation rounds N can be pre-configured or set by the user (such as a business requester), and is not limited in this embodiment. After N rounds of training and aggregation, a model generated by federated learning can be obtained.

[0224] It should be noted that the steps or the order of execution of the steps included in the above embodiments are merely examples and not limitations. The embodiments of this application may include more or fewer steps, or may include other steps. At the same time, the above steps may be executed in other orders, and the embodiments of this application are not limited in this regard.

[0225] The method embodiments of this application have been described in detail above with reference to Figures 1 to 12. The apparatus embodiments of this application will be described in detail below with reference to Figures 13 and 16. It should be understood that the descriptions of the method embodiments correspond to the descriptions of the apparatus embodiments; therefore, any parts not described in detail can be referred to the preceding method embodiments.

[0226] Figure 13 is a schematic structural diagram of a communication device provided in an embodiment of this application. The communication device 1300 shown in Figure 13 can be used in the first network element in the foregoing embodiments. The communication device 1300 can be the first network element, or a device in the first network element (e.g., a processor, chip, chip system, circuit, or a functional module, etc.), or a device that can be matched and used with the first network element, or a logic module or software that can implement all or part of the first network element.

[0227] As shown in Figure 13, the communication device 1300 includes a receiving unit 1310 and a transmitting unit 1320, as detailed below:

[0228] The receiving unit 1310 is used to receive first information, which is used to request a first service;

[0229] The sending unit 1320 is used to send second information to the second network element, the second information being used to instruct the second network element to perform the operation corresponding to the first service;

[0230] The first network element is used to control model training, model inference, and / or data processing, while the second network element is used for model training, model inference, and / or data processing.

[0231] Optionally, the receiving unit 1310 is further configured to: receive third information from the second network element, the third information being used by the second network element for registration.

[0232] Optionally, the third information includes one or more of the following: the identifier of the second network element, the federated learning capability information of the second network element, the analysis capability information of the second network element, and the computing capability information of the second network element.

[0233] Optionally, the device 1300 further includes a first determining unit 1330, configured to: determine, based on the third information, whether the second network element is a server-side network element or a client-side network element of federated learning.

[0234] Optionally, the receiving unit 1310 is further configured to: receive fourth information from a third network element, the fourth information being used to indicate model information stored by the third network element, the third network element being used to store data; the device 1300 further includes a second determining unit 1340, configured to: determine a first model based on the fourth information.

[0235] Optionally, the second network element is a server-side network element of federated learning, and the second information is used to instruct the second network element to publish the first model to the client network element of federated learning, and / or the second network element to subscribe to the second model to the client network element of federated learning, wherein the second model is obtained based on the first model after training.

[0236] Optionally, the sending unit 1320 is further configured to: send fifth information to the second network element, the fifth information being used to instruct the second network element to subscribe to the first model from the third network element, the third network element being used to store data.

[0237] Optionally, the sending unit 1320 is further configured to: send sixth information to a third network element, the sixth information being used to instruct the third network element to publish the first model to the second network element, the third network element being used to store data.

[0238] Optionally, the second network element is a client network element of federated learning, and the second information is used to instruct the second network element to subscribe to the first model from the server network element of federated learning, and / or the second network element to publish the second model to the server network element of federated learning, wherein the second model is obtained after training based on the first model.

[0239] Optionally, the first network element is equipped with a first communication agent, and the second network element is equipped with a second communication agent, wherein the first communication agent is used to transmit control signaling, and the second communication agent is used to transmit service data.

[0240] Optionally, the first network element and the second network element are used for horizontal federated learning, or the first network element and the second network element are used for vertical federated learning.

[0241] Figure 14 is a schematic structural diagram of a communication device provided in an embodiment of this application. The communication device 1400 shown in Figure 14 can be used in the second network element in the foregoing embodiments. The communication device 1400 can be the second network element, or a device in the second network element (e.g., a processor, chip, chip system, circuit, or a functional module, etc.), or a device that can be used in conjunction with the second network element, or a logic module or software that can implement all or part of the second network element.

[0242] As shown in Figure 14, the communication device 1400 includes a receiving unit 1410 and an execution unit 1420, as detailed below:

[0243] The receiving unit 1410 is configured to receive second information from the first network element, the second information being used to instruct the second network element to perform the operation corresponding to the first service;

[0244] Execution unit 1420 is used to perform corresponding operations based on the second information;

[0245] The first network element is used to control model training, model inference, and / or data processing, while the second network element is used for model training, model inference, and / or data processing.

[0246] Optionally, the device 1400 further includes a sending unit 1430, configured to: send third information to the first network element, the third information being used by the second network element for registration.

[0247] Optionally, the third information includes one or more of the following: the identifier of the second network element, the federated learning capability information of the second network element, the analysis capability information of the second network element, and the computing capability information of the second network element.

[0248] Optionally, the second network element is a server-side network element of federated learning, and the second information is used to instruct the second network element to publish the first model to the client network element of federated learning, and / or the second network element to subscribe to the second model to the client network element of federated learning, wherein the second model is obtained based on the first model after training.

[0249] Optionally, the apparatus 1400 further includes a sending unit 1430, configured to: publish the first model to the client network element of the federated learning through a second communication proxy; and / or, the apparatus 1400 further includes a subscription unit 1440, configured to: subscribe to the second model to the client network element of the federated learning through the second communication proxy.

[0250] Optionally, the receiving unit 1410 is further configured to: receive fifth information from the first network element, the fifth information being used to instruct the second network element to subscribe to the first model from the third network element, the third network element being used to store data; the device 1400 further includes a subscription unit 1440, configured to: subscribe to the first model from the third network element through a second communication proxy according to the fifth information.

[0251] Optionally, the second network element is a client network element of federated learning, and the second information is used to instruct the second network element to subscribe to the first model from the server network element of federated learning, and / or the second network element to publish the second model to the server network element of federated learning, wherein the second model is obtained after training based on the first model.

[0252] Optionally, the apparatus 1400 further includes a subscription unit 1440, configured to: subscribe to the first model from the server-side network element of the federated learning through a second communication proxy; and / or, the apparatus 1400 further includes a sending unit 1430, configured to: publish the second model to the server-side network element of the federated learning through the second communication proxy.

[0253] Optionally, the first network element is equipped with a first communication agent, and the second network element is equipped with a second communication agent, wherein the first communication agent is used to transmit control signaling, and the second communication agent is used to transmit service data.

[0254] Optionally, the first network element and the second network element are used for horizontal federated learning, or the first network element and the second network element are used for vertical federated learning.

[0255] Figure 15 is a schematic structural diagram of a communication device provided in an embodiment of this application. The communication device 1500 shown in Figure 15 can be used in the third network element in the foregoing embodiments. The communication device 1500 can be the third network element, or a device in the third network element (e.g., a processor, chip, chip system, circuit, or a functional module, etc.), or a device that can be matched with the third network element, or a logic module or software that can implement all or part of the third network element.

[0256] As shown in Figure 15, the communication device 1500 includes a receiving unit 1510 and a transmitting unit 1520, as detailed below:

[0257] The receiving unit 1510 is used to receive sixth information from the first network element, the sixth information being used to instruct the third network element to publish the first model to the second network element;

[0258] Sending unit 1520 is used to publish the first model to the second network element according to the sixth information;

[0259] The first network element is used to control model training, model inference, and / or data processing; the second network element is used to control model training, model inference, and / or data processing; and the third network element is used to store data.

[0260] Optionally, the sending unit 1520 is specifically used to: publish the first model to the second network element through the second communication proxy according to the sixth information, wherein the second communication proxy is used to transmit service data.

[0261] Optionally, the sending unit 1520 is further configured to: send fourth information to the first network element, the fourth information being used to indicate the model information stored by the third network element.

[0262] Figure 16 is a schematic structural diagram of an apparatus provided in an embodiment of this application. The dashed lines in Figure 16 indicate that the unit or module is optional. This apparatus 1600 can be used to implement the methods described in the above method embodiments. Apparatus 1600 can be a chip or a communication device.

[0263] Apparatus 1600 may include one or more processors 1610. The processor 1610 may support apparatus 1600 in implementing the methods described in the preceding method embodiments. The processor 1610 may be a general-purpose processor or a special-purpose processor. For example, the processor may be a central processing unit (CPU). Alternatively, the processor may be other general-purpose processors, microprocessor units (MPUs), microcontroller units (MCUs), graphics processing units (GPUs), artificial intelligence processors (AI processors) or neural processing units (NPUs), digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0264] The device 1600 may further include one or more memories 1620. The memories 1620 store a program that can be executed by the processor 1610, causing the processor 1610 to perform the methods described in the preceding method embodiments. The memories 1620 may be independent of the processor 1610 or integrated within the processor 1610. In this embodiment, the memories 1620 may include, but are not limited to, cache, read-only memory (ROM), random access memory (RAM), synchronous dynamic random access memory (SDRAM), hard disk drive (HDD) or solid-state drive (SSD), erasable programmable read-only memory (EPROM), or compact disc read-only memory (CD-ROM), etc.

[0265] The device 1600 may also include a transceiver 1630. The processor 1610 can communicate with other devices or chips via the transceiver 1630. For example, the processor 1610 can send and receive data with other devices or chips via the transceiver 1630.

[0266] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0267] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0268] This application also provides a computer-readable storage medium storing a computer program that, when run on a computer, causes the computer to perform the steps described in the various method embodiments above.

[0269] This application also provides a computer program product, which includes a computer program that, when run on a computer, causes the computer to perform the steps described in the various method embodiments above.

[0270] This application also provides a chip, which includes a processor and a memory. The memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory, so that a device or equipment (such as a communication device) with the chip installed performs the steps in the above-described method embodiments.

[0271] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate form. The computer-readable storage medium can include at least: any entity or device capable of carrying computer program code to a device / app, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some possible implementations, the computer-readable storage medium may not be an electrical carrier signal or a telecommunication signal.

[0272] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0273] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0274] In the embodiments provided in this application, it should be understood that the disclosed apparatus / devices and methods can be implemented in other ways. For example, the apparatus / device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0275] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0276] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A communication method, characterized in that, The method is applied to a first network element, and the method includes: Receive first information, which is used to request a first service; Send a second message to the second network element, the second message being used to instruct the second network element to perform the operation corresponding to the first service; The first network element is used to control model training, model inference, and / or data processing, while the second network element is used for model training, model inference, and / or data processing.

2. The method according to claim 1, characterized in that, The method further includes: The third information is received from the second network element, and the third information is used by the second network element to register.

3. The method according to claim 2, characterized in that, The third information includes one or more of the following: The identifier of the second network element, the federated learning capability information of the second network element, the analysis capability information of the second network element, and the computing capability information of the second network element.

4. The method according to claim 2 or 3, characterized in that, The method further includes: Based on the third information, the second network element is determined to be either a server-side network element or a client-side network element in federated learning.

5. The method according to any one of claims 1 to 4, characterized in that, The method further includes: Receive fourth information from a third network element, the fourth information being used to indicate model information stored by the third network element, the third network element being used to store data; The first model is determined based on the fourth information.

6. The method according to any one of claims 1 to 5, characterized in that, The second network element is a server-side network element of federated learning. The second information is used to instruct the second network element to publish the first model to the client network element of federated learning, and / or the second network element to subscribe to the second model to the client network element of federated learning. The second model is obtained based on the first model after training.

7. The method according to claim 6, characterized in that, The method further includes: A fifth message is sent to the second network element, the fifth message being used to instruct the second network element to subscribe to the first model from the third network element, the third network element being used to store data.

8. The method according to claim 6 or 7, characterized in that, The method further includes: A sixth message is sent to a third network element, the sixth message being used to instruct the third network element to publish the first model to the second network element, the third network element being used to store data.

9. The method according to any one of claims 1 to 5, characterized in that, The second network element is a client network element of federated learning. The second information is used to instruct the second network element to subscribe to the first model from the server network element of federated learning, and / or the second network element to publish the second model to the server network element of federated learning. The second model is obtained based on the first model after training.

10. The method according to any one of claims 1 to 9, characterized in that, The first network element is equipped with a first communication agent, and the second network element is equipped with a second communication agent. The first communication agent is used to transmit control signaling, and the second communication agent is used to transmit service data.

11. The method according to any one of claims 1 to 10, characterized in that, The first network element and the second network element are used for horizontal federated learning, or the first network element and the second network element are used for vertical federated learning.

12. A communication method, characterized in that, The method is applied to a second network element, and the method includes: Receive second information from the first network element, the second information being used to instruct the second network element to perform the operation corresponding to the first service; Perform the corresponding operation based on the second information; The first network element is used to control model training, model inference, and / or data processing, while the second network element is used for model training, model inference, and / or data processing.

13. The method according to claim 12, characterized in that, The method further includes: Send a third message to the first network element, the third message being used by the second network element for registration.

14. The method according to claim 13, characterized in that, The third information includes one or more of the following: The identifier of the second network element, the federated learning capability information of the second network element, the analysis capability information of the second network element, and the computing capability information of the second network element.

15. The method according to any one of claims 12 to 14, characterized in that, The second network element is a server-side network element of federated learning. The second information is used to instruct the second network element to publish the first model to the client network element of federated learning, and / or the second network element to subscribe to the second model to the client network element of federated learning. The second model is obtained based on the first model after training.

16. The method according to claim 15, characterized in that, The method further includes: The first model is published to the client network element of the federated learning through the second communication agent; and / or, The second model is subscribed to by the client network element of the federated learning through the second communication agent.

17. The method according to claim 15 or 16, characterized in that, The method further includes: The fifth information is received from the first network element, which is used to instruct the second network element to subscribe to the first model from the third network element, and the third network element is used to store data; Based on the fifth information, the first model is subscribed to by the third network element through the second communication agent.

18. The method according to any one of claims 12 to 14, characterized in that, The second network element is a client network element of federated learning. The second information is used to instruct the second network element to subscribe to the first model from the server network element of federated learning, and / or the second network element to publish the second model to the server network element of federated learning. The second model is obtained based on the first model after training.

19. The method according to claim 18, characterized in that, The method further includes: Subscribe to the first model from the server-side network element of the federated learning via a second communication agent; and / or, The second model is published to the server-side network element of the federated learning through the second communication agent.

20. The method according to any one of claims 12 to 19, characterized in that, The first network element is equipped with a first communication agent, and the second network element is equipped with a second communication agent. The first communication agent is used to transmit control signaling, and the second communication agent is used to transmit service data.

21. The method according to any one of claims 12 to 20, characterized in that, The first network element and the second network element are used for horizontal federated learning, or the first network element and the second network element are used for vertical federated learning.

22. A communication method, characterized in that, The method is applied to a third network element, and the method includes: Receive a sixth message from the first network element, the sixth message being used to instruct the third network element to publish the first model to the second network element; The first model is published to the second network element according to the sixth information; The first network element is used to control model training, model inference, and / or data processing; the second network element is used to control model training, model inference, and / or data processing; and the third network element is used to store data.

23. The method according to claim 22, characterized in that, The step of publishing the first model to the second network element based on the sixth information includes: According to the sixth information, the first model is published to the second network element through the second communication agent, and the second communication agent is used to transmit service data.

24. The method according to claim 22 or 23, characterized in that, The method further includes: Send a fourth message to the first network element, the fourth message being used to indicate the model information stored by the third network element.

25. A communication device, characterized in that, include: A module or unit for performing the method as described in any one of claims 1 to 24.

26. A communication device, characterized in that, include: A processor and a memory, the processor being coupled to the memory, the memory being used to store a computer program, which, when executed by the processor, causes the apparatus to perform the method as described in any one of claims 1 to 24.

27. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when run on a computer, causes the computer to perform the method as described in any one of claims 1 to 24.

28. A computer program product, characterized in that, include: A computer program that, when run on a computer, causes the computer to perform the method as described in any one of claims 1 to 24.

29. A chip, characterized in that, include: A processor and a memory, the memory for storing a computer program, the processor for calling and running the computer program stored in the memory, causing a device or apparatus on which the chip is mounted to perform the method as described in any one of claims 1 to 24.

Images