Wireless communication method, terminal device and network device

Abstract

Provided are a wireless communication method, a terminal device and a network device. The method comprises: a terminal device receiving or sending first information, wherein the first information is used for requesting that model information supported by the terminal device is indicated to a first network element, or the first information is used by the first network element to configure a model training task for the terminal device, the first network element being used for managing a model. In the embodiments of the present application, a first network element used for model management is introduced. In this way, a terminal device can interact with the first network element by means of first information, so as to manage a model in a communication system, thereby laying a solid foundation for a model-based communication process.

Description

Wireless communication methods, terminal devices, and network devices Technical Field

[0001] This application relates to the field of communication technology, and more specifically, to a wireless communication method, terminal device, and network device. Background Technology

[0002] Model-based communication processes are being introduced into future communication systems. For example, in intelligently-inherent 6G networks, each node can possess artificial intelligence (AI) / machine learning (ML) capabilities. Each node will support model operations locally or collaboratively support distributed model operations, achieving an intelligently-inherent network architecture to meet diverse AI / ML-based model-based needs. Therefore, future communication networks will contain a large number of complex model operations, posing significant demands for model lifecycle management. How to manage the large number of models in future networks is a pressing issue that needs to be addressed. Summary of the Invention

[0003] This application provides a wireless communication method, terminal device, and network device. The various aspects covered by this application are described below.

[0004] In a first aspect, a wireless communication method is provided, comprising: a terminal device receiving or sending first information, the first information being used to request a first network element to indicate model information supported by the terminal device, or the first information being used by the first network element to configure a model training task for the terminal device, the first network element being used to manage the model.

[0005] In a second aspect, a wireless communication method is provided, comprising: a first network element receiving or transmitting first information, the first network element being used to manage a model, wherein the first information is used to request the first network element to indicate model information supported by a terminal device, or the first information is used to configure a model training task for a model to be trained.

[0006] Thirdly, a wireless communication method is provided, comprising: a second network element receiving and / or sending first information, the first information being used to request the first network element to indicate model information supported by a terminal device, or the first information being used by the first network element to configure a model training task for the terminal device, the first network element being used to manage the model.

[0007] Fourthly, a terminal device is provided, comprising: a communication unit for receiving or sending first information, the first information being used to request a first network element to indicate model information supported by the terminal device, or the first information being used by the first network element to configure a model training task for the terminal device, the first network element being used to manage the model.

[0008] Fifthly, a network device is provided, the network device being a first network element, comprising: a communication unit for receiving or sending first information, the first network element for managing the model, wherein the first information is used to request the first network element to indicate model information supported by the terminal device, or the first information is used to configure a model training task for the model to be trained.

[0009] In a sixth aspect, a network device is provided, the network device being a second network element, comprising: a communication unit for receiving and / or sending first information, the first information being used to request the first network element to indicate model information supported by the terminal device, or the first information being used by the first network element to configure a model training task for the terminal device, the first network element being used to manage the model.

[0010] In a seventh aspect, a terminal device is provided, including a processor, a memory, and a communication interface, wherein the memory is used to store one or more computer programs, and the processor is used to invoke the computer programs in the memory to cause the terminal device to perform some or all of the steps in the methods described above.

[0011] Eighthly, a network device is provided, including a processor, a memory, and a transceiver, wherein the memory is used to store one or more computer programs, and the processor is used to invoke the computer programs in the memory to cause the network device to perform some or all of the steps of the methods described in the preceding aspects.

[0012] Ninthly, embodiments of this application provide a communication system including the aforementioned terminal device and / or network device. In another possible design, the system may further include other devices that interact with the terminal device or network device as described in the embodiments of this application.

[0013] In a tenth aspect, embodiments of this application provide a computer-readable storage medium storing a computer program that causes a communication device (e.g., a terminal device or a network device) to perform some or all of the steps in the methods described above.

[0014] Eleventhly, embodiments of this application provide a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program operable to cause a communication device (e.g., a terminal device or a network device) to perform some or all of the steps of the methods described in the foregoing aspects. In some implementations, the computer program product may be a software installation package.

[0015] In a twelfth aspect, embodiments of this application provide a chip including a memory and a processor, the processor being able to call and run a computer program from the memory to implement some or all of the steps described in the methods of the foregoing aspects.

[0016] In this embodiment, a first network element for managing the model is introduced. In this way, the terminal device can interact with the first network element through the first information to manage the model in the communication system, which helps to provide a solid foundation for the model-based communication process. Attached Figure Description

[0017] Figure 1 shows the wireless communication system 100 used in an embodiment of this application.

[0018] Figures 2A and 2B are schematic diagrams of the interfaces between other network elements of the network data analysis function (NWDAF) applicable to the embodiments of this application.

[0019] Figure 3 is a schematic diagram of the neural network applicable to the embodiments of this application.

[0020] Figure 4 is a schematic diagram of a convolutional neural network (CNN) to which the embodiments of this application are applicable. CNN is a deep neural network with a convolutional structure.

[0021] Figures 5 and 6 are schematic diagrams of model segmentation scenarios applicable to the embodiments of this application.

[0022] Figure 7 is a schematic diagram of a communication system incorporating a first network element according to an embodiment of this application.

[0023] Figure 8 is a schematic flowchart of a wireless communication method according to an embodiment of this application.

[0024] Figure 9 is a schematic flowchart of a terminal device requesting instruction model information from a first network element in an embodiment of this application.

[0025] Figure 10 is a schematic flowchart of other network elements requesting instruction model information from the first network element in an embodiment of this application.

[0026] Figure 11 is a schematic flowchart of a wireless communication method according to another embodiment of this application.

[0027] Figure 12 is a schematic flowchart of the scheme for the training task of the first network element configuration model in the embodiment of this application.

[0028] Figure 13 is a schematic diagram of a terminal device according to an embodiment of this application.

[0029] Figure 14 is a schematic diagram of the network device according to an embodiment of the application.

[0030] Figure 15 is a schematic diagram of a network device according to an embodiment of this application.

[0031] Figure 16 is a schematic structural diagram of a communication device according to an embodiment of this application. Detailed Implementation

[0032] The technical solutions of this application will now be described with reference to the accompanying drawings. For ease of understanding, the following description will first introduce a schematic diagram of the communication system architecture of an embodiment of this application with reference to Figure 1. Figure 1 is a schematic diagram of a communication system architecture 100 applicable to an embodiment of this application. This network architecture may include terminal devices, access network (AN) nodes, and core network nodes.

[0033] It should be understood that the technical solutions of the embodiments 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, etc. The technical solutions provided in this application can also be applied to future communication systems, such as 6th generation mobile communication systems, satellite communication systems, and so on.

[0034] The terminal device in this application embodiment can also be referred to as user equipment (UE), access terminal, user unit, user station, mobile station, mobile station (MS), mobile terminal, remote station, remote terminal, mobile device, user terminal, terminal, wireless core network node, user agent, or user device. The terminal device in this application embodiment can be a device that provides voice and / or data connectivity to a user, and can be used to connect people, objects, and machines, such as a handheld device with wireless connectivity, vehicle-mounted device, etc. The terminal devices in the embodiments of this application can be Internet of Things (IoT) devices, mobile phones, tablets, laptops, PDAs, mobile internet devices (MIDs), wearable devices, virtual reality (VR) devices, augmented reality (AR) devices, wireless terminals in industrial control, wireless terminals in self-driving, wireless terminals in remote medical surgery, wireless terminals in smart grids, wireless terminals in transportation safety, wireless terminals in smart cities, wireless terminals in smart homes, etc. Optionally, the terminal device can be used to act as a base station. For example, the terminal device can act as a dispatching entity, providing sidelink signals between terminal devices in vehicle-to-everything (V2X) or device-to-device (D2D) communications. For example, cellular phones and cars communicate with each other using sidelink signals. Cellular phones and smart home devices can communicate without relaying communication signals through base stations.

[0035] Access network nodes can be access network devices. Access network devices are devices that terminals use to wirelessly access the network architecture. They are primarily responsible for air interface-side radio resource management, Quality of Service (QoS) management, data compression, and encryption. Access network devices can also be called radio access network (RAN) devices, such as base stations. A base station can broadly encompass, or be replaced by, various names including: NodeB, evolved NodeB (eNB), next-generation NodeB (gNB), relay station, access point, transmitting and receiving point (TRP), transmitting point (TP), master eNB (MeNB), secondary eNB (SeNB), multi-standard radio (MSR) node, home base station, network controller, access node, wireless node, access point (AP), transmission node, transceiver node, baseband unit (BBU), remote radio unit (RRU), active antenna unit (AAU), remote radio head (RRH), central unit (CU), distributed unit (DU), positioning node, etc. A base station can be a macro base station, micro base station, relay node, donor node, or similar entities, or combinations thereof. A base station can also refer to a communication module, modem, or chip installed within the aforementioned equipment or apparatus. A base station can also be a mobile switching center, a device that performs base station functions in D2D, V2X, and machine-to-machine (M2M) communications, a network-side device in a 6G network, or a device that performs base station functions in future communication systems. A base station can support networks using the same or different access technologies. The embodiments of this application do not limit the specific technologies or device forms used in the access network equipment.

[0036] Base stations can be fixed or mobile. For example, a helicopter or drone can be configured to act as a mobile base station, and one or more cells can move depending on the location of the mobile base station. In other examples, a helicopter or drone can be configured as a device to communicate with another base station.

[0037] In some deployments, the access network device in this application embodiment may refer to a CU or a DU, or the access network device may include both a CU and a DU. The gNB may also include an AAU.

[0038] Core network nodes can be categorized into several types, including User Plane Function (UPF) nodes, Access and Mobility Management Function (AMF) nodes, Session Management Function (SMF) nodes, Policy Control Function (PCF) nodes, Application Function (AF) nodes, Data Network (DN) nodes, Network Slice Selection Function (NSSF) nodes, Authentication Server Function (AUSF) nodes, Unified Data Management (UDM) nodes, Network Exposure Function (NEF) nodes, Network Repository Function (NRF) nodes, and Network Slice-Specific Authentication and Authorization Function (NSSAAF) nodes. UPF nodes are primarily responsible for user data transmission, while the other nodes, often referred to as Control Plane Function nodes, are mainly responsible for authentication, authorization, registration management, session management, mobility management, and policy control to ensure reliable and stable user data transmission.

[0039] UPF nodes can be used to forward and receive data from terminals. For example, a UPF node can receive service data from the data network and transmit it to the terminal through access network equipment; a UPF node can also receive user data from the terminal through access network equipment and forward it to the data network. The transmission resources allocated and scheduled by the UPF node for the terminal are managed and controlled by the SMF node. The bearer between the terminal and the UPF node can include: the user plane connection between the UPF node and the access network equipment, and the establishment of a channel between the access network equipment and the terminal. The user plane connection is a QoS flow that can be established between the UPF node and the access network equipment for transmitting data.

[0040] AMF nodes can be used to manage terminal access to the core network, such as terminal location updates, network registration, access control, terminal mobility management, and terminal attachment and detachment. While providing services for a terminal's session, the AMF node can also provide control plane storage resources for that session to store the session identifier and the SMF node identifier associated with the session identifier.

[0041] SMF nodes can be used to select user plane nodes for terminals, redirect user plane nodes for terminals, assign Internet Protocol (IP) addresses to terminals, establish bearers (also known as sessions) between terminals and UPF nodes, modify and release sessions, and perform QoS control.

[0042] PCF nodes are used to provide policies to AMF and SMF nodes, such as QoS policies and slice selection policies.

[0043] AF nodes are used to interact with 3GPP core network nodes to support the routing of application-impacted data, access network exposure functions, and interact with PCF nodes for policy control, etc.

[0044] A Data Network (DN) can provide data services to users for networks such as IP Multimedia Service (IMS) and the Internet. A DN can contain various application servers (AS) that provide different application services, such as carrier services, Internet access, or third-party services. The AS can implement the functions of an Application Server (AF).

[0045] NSSF is used for network slice selection and supports the following functions: selecting a set of network slice instance examples to serve the end device; determining allowed network slice selection assistance information (NSSAI), and, when necessary, determining the mapping to the subscribed single-network slice selection assistance information (S-NSSAI); determining the configured NSSAI, and, when necessary, determining the mapping to the subscribed S-NSSAI; determining the set of AMFs that may be used to query the end device, or determining a list of candidate AMFs based on the configuration.

[0046] AUSF is used to receive AMF requests for terminal authentication. It requests a key from UDM and then forwards the issued key to AMF for authentication processing.

[0047] UDM includes functions such as generating and storing user subscription information and managing authentication data, and supports interaction with external third-party servers.

[0048] NEF is used for capability exposure, meaning that based on NEF, network capabilities can be exported to external networks. Untrusted external applications can access core network data through NEF to ensure network security. NEF can provide functions such as QoS capability exposure for external applications, event subscription, and AF request distribution.

[0049] The NRF (Network Request Framework) is used for core network node registration, management, and status monitoring, thereby enabling automated management of core network nodes. When a core network node starts up, it must register with the NRF to provide services. Registration information may include, for example, the core network node's type, address, and service list.

[0050] In some communication systems (such as 5G systems), core network nodes can also be called network functions (NFs).

[0051] The nodes in Figure 1 can be network elements in hardware devices, software functions running on dedicated hardware, or virtualization functions implemented on a platform (e.g., a cloud platform). It should be noted that the network architecture shown in the above figures is merely an illustrative representation of the nodes included in the overall network architecture. In this application embodiment, the number of nodes included in the entire network architecture is not limited.

[0052] Those skilled in the art will understand that the network architecture shown in Figure 1 does not constitute a limitation on the network architecture. In specific implementations, the network architecture may include more or fewer nodes than shown, or combine certain nodes, etc. It should be understood that AN or RAN is represented in Figure 1 as (R)AN.

[0053] In some scenarios, network devices and terminal devices can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; they can also be deployed on water; and they can also be deployed in the air on airplanes, balloons, and satellites. This application does not limit the scenarios in which the network devices and terminal devices are located.

[0054] By way of example and not limitation, in the embodiments of this application, the network device may have mobility characteristics; for example, the network device may be a mobile device. In some embodiments of this application, the network device may be a satellite or a balloon station. For example, the satellite may be a low Earth orbit (LEO) satellite, a medium Earth orbit (MEO) satellite, a geostationary Earth orbit (GEO) satellite, a high elliptical orbit (HEO) satellite, etc. In some embodiments of this application, the network device may also be a base station located on land, water, or other similar locations.

[0055] In this embodiment, the network device can provide services to a cell. The terminal device communicates with the network device through the transmission resources (e.g., frequency domain resources, or spectrum resources) used by the cell. The cell can be the cell corresponding to the network device (e.g., a base station). The cell can belong to a macro base station or to a base station corresponding to a small cell. The small cell can include: metro cell, micro cell, pico cell, femto cell, etc. These small cells have the characteristics of small coverage area and low transmission power, and are suitable for providing high-speed data transmission services.

[0056] NWDAF

[0057] In some scenarios, NWDAF has been introduced into the aforementioned wireless communication systems to provide network data analysis capabilities for operators. For example, NWDAF can collect data from various network elements in the core network and network management systems to perform big data statistics, analysis, or intelligent data analysis, deriving network-side analysis or predictive data, thereby assisting various network elements in more effectively controlling terminal device access based on the data analysis results.

[0058] As described above, NWDAF can collect data from other network elements for big data analysis. To this end, interfaces are defined between NWDAF and other network elements (e.g., NFs). Referring to Figure 2A, NWDAF can send the analysis results of its network data analysis function to other NFs through the Nnf interface. Referring to Figure 2B, other NFs can send requests to NWDAF through the Nnwdaf interface to request the analysis results of the network data analysis service (also known as the "analysis results of the network data analysis function").

[0059] In some implementations, the analysis results of the aforementioned network data analysis function can be identified by an analytics ID. In other embodiments, the aforementioned other NFs can be core network elements, such as AMF, SMF, NEF, PCF, etc. Of course, the aforementioned other NFs can also be trusted AFs or untrusted AFs.

[0060] In some implementations, the analysis results provided by NWDAF are shown in Table 1. The analysis results (also known as analytics information) provided by NWDAF may include: network performance information, UE mobility information, user data congestion information, and QoS sustainability, etc. Specifically, for network performance information, the corresponding request analysis ID is "network performance," and the corresponding response includes statistics and predictions of network load information for the area of ​​interest. Additionally, it can also count and predict the number of UEs in that area. For UE mobility information, the corresponding request analysis ID is "UE mobility," and the corresponding response includes statistics or predictions of UE mobility. For user data congestion information, the corresponding request analysis ID is "user data congestion," and the corresponding response includes statistics or predictions about user data congestion, for example, statistics or predictions of user plane transmission user data congestion, and / or statistics or predictions of control plane transmission user data congestion. For QoS sustainability information, the corresponding request analysis ID is QoS sustainability, and the statistical information in the corresponding response includes the location, time, and threshold exceeded when the QoS changed. The prediction information in the corresponding response includes information about the location and time when potential QoS changes may occur and the threshold that may be exceeded.

[0061] Table 1

[0062] In some implementations, the analysis results provided by NWDAF are shown in Table 1. The analysis results (also known as analytics information) provided by NWDAF may include: network performance information, UE mobility information, user data congestion information, and QoS sustainability, etc. Specifically, for network performance information, the corresponding request analysis ID is "network performance," and the corresponding response includes statistics and predictions of network load information for the area of ​​interest. Additionally, it can also count and predict the number of UEs in that area. For UE mobility information, the corresponding request analysis ID is "UE mobility," and the corresponding response includes statistics or predictions of UE mobility. For user data congestion information, the corresponding request analysis ID is "user data congestion," and the corresponding response includes statistics or predictions about user data congestion, for example, statistics or predictions of user plane transmission user data congestion, and / or statistics or predictions of control plane transmission user data congestion. For QoS sustainability information, the corresponding request analysis ID is QoS sustainability, and the statistical information in the corresponding response includes the location, time, and threshold exceeded when the QoS changed. The prediction information in the corresponding response includes information about the location and time when potential QoS changes may occur and the threshold that may be exceeded.

[0063] Neural Networks

[0064] In recent years, AI research, represented by neural networks, has achieved remarkable results in many fields and will continue to play an important role in people's production and daily lives for a long time to come. Common neural networks include CNNs, recurrent neural networks (RNNs), and deep neural networks (DNNs).

[0065] The neural network applicable to the embodiments of this application is described below with reference to Figure 3. The neural network shown in Figure 3 can be divided into three categories according to the position of different layers: input layer 310, hidden layer 320, and output layer 330. Generally speaking, the first layer is the input layer 310, the last layer is the output layer 330, and the intermediate layers between the first and last layers are all hidden layers 330.

[0066] The input layer 310 is used to input data, which may be, for example, a received signal received by a receiver. The hidden layer 330 is used to process the input data, for example, to decompress the received signal. The output layer 220 is used to output the processed output data, for example, to output the decompressed signal.

[0067] As shown in Figure 3, the neural network consists of multiple layers, each containing multiple neurons. The neurons between layers can be fully connected or partially connected. For connected neurons, the output of a neuron in the previous layer can serve as the input of a neuron in the next layer.

[0068] With the continuous development of neural network research, deep learning algorithms have been proposed in recent years. These algorithms introduce more hidden layers into neural networks, forming DNNs (Deep Neural Networks). More hidden layers allow DNNs to better depict complex situations in the real world. Theoretically, the more parameters a model has, the higher its complexity and the greater its "capacity," meaning it can accomplish more complex learning tasks. This type of neural network model is widely used in pattern recognition, signal processing, optimization, and anomaly detection.

[0069] CNN is a deep neural network with convolutional structures, as shown in Figure 4. It can include an input layer 410, a convolutional layer 420, a pooling layer 430, a fully connected layer 430, and an output layer 450.

[0070] Each convolutional layer 420 can include many convolution operators, also known as kernels. Their function can be seen as a filter that extracts specific information from the input signal. A convolution operator can essentially be a weight matrix, which is usually predefined.

[0071] The weight values ​​in these weight matrices need to be obtained through extensive training in practical applications. The weight matrices formed by the weight values ​​obtained through training can extract information from the input signal, thereby helping the CNN to make correct predictions.

[0072] When a CNN has multiple convolutional layers, the initial convolutional layers tend to extract more general features, which can also be called low-level features. As the depth of the CNN increases, the features extracted by later convolutional layers become more and more complex.

[0073] Pooling layer 430: Because it is often necessary to reduce the number of training parameters, pooling layers are often introduced periodically after convolutional layers. For example, it can be a pooling layer followed by a convolutional layer as shown in Figure 4, or it can be one or more pooling layers followed by multiple convolutional layers. In signal processing, the sole purpose of pooling layers is to reduce the spatial size of the extracted information.

[0074] After processing by convolutional layers 420 and pooling layers 430, the CNN is still insufficient to output the required information. As mentioned earlier, convolutional layers 420 and pooling layers 430 only extract features and reduce parameters introduced by the input data. However, to generate the final output information (e.g., the bitstream of the original information transmitted by the transmitter), the CNN still needs to utilize fully connected layers 430. Typically, fully connected layers 430 can include multiple hidden layers. The parameters contained in these hidden layers can be pre-trained based on training data relevant to a specific task type. For example, this task type could include decoding data signals received by a receiver, or it could include channel estimation based on pilot signals received by the receiver.

[0075] After the multiple hidden layers in the fully connected layer 430, which is the final layer of the entire CNN, is the output layer 450, used to output the result. Typically, this output layer 450 is equipped with a loss function (e.g., a loss function similar to the cross-entropy loss function for classification) to calculate the prediction error, or to evaluate the degree of difference between the output of the CNN model (also known as the predicted value) and the ideal result (also known as the true value).

[0076] Model lifecycle

[0077] In some implementations, the model lifecycle includes the following stages: requirements analysis, data preparation, model design, model training, model deployment, model maintenance, and model retirement.

[0078] For requirements analysis, it is necessary to clarify the model's objectives, application scenarios, and performance requirements. For example, known protocols provide three use cases for researching AI / ML-enabled wireless air interface technologies: AI + CSI feedback, AI + beam management, and AI + positioning technology.

[0079] AI+CSI feedback use cases can be understood as using AI / ML technology to compress and decompress CSI, reducing air interface transmission overhead and improving the accuracy of CSI feedback information.

[0080] AI+beam management use cases can be understood as using AI / ML technology to predict beam information in the time / spatial domain, reducing measurement overhead and latency, and improving the accuracy of beam selection.

[0081] AI+location technology use cases can be understood as using AI / ML technology to predict the location information of terminal devices, thereby improving the accuracy of the location information of terminal devices in NLOS scenarios.

[0082] For data preparation, data is the cornerstone of AI model training. The data preparation phase requires collecting and cleaning a large amount of training data. The quality and quantity of data directly affect the model's performance.

[0083] For model design, the model design phase involves selecting an appropriate model architecture and parameters. With the development of deep learning technology, many excellent model architectures have emerged. At this stage, it is necessary to select a suitable model architecture and set appropriate hyperparameters based on task requirements and data characteristics.

[0084] Model training is the core step in AI model development. This stage requires the use of high-performance computing resources to train the model. During training, the model's loss function and performance metrics need to be monitored and optimized as needed. After training, the model needs to be evaluated to ensure it meets the requirements of the use cases.

[0085] Model deployment involves integrating a trained model into an application system and providing services to the outside world. Before deployment, the model needs to be converted to a deployment-appropriate format and system testing should be conducted to ensure its stability and performance. After deployment, the model's running status and performance metrics need to be monitored, and any potential problems should be addressed promptly.

[0086] Model maintenance is a crucial part of the AI ​​model lifecycle. As data is continuously updated and business needs change, models require regular updates and fine-tuning. Furthermore, performance optimization and problem diagnosis are necessary to ensure the model continues to provide high-quality service.

[0087] Model retirement is necessary when a model can no longer meet business needs or its performance has significantly degraded. Before retirement, a detailed retirement plan must be developed, including data migration, resource release, and risk assessment. After retirement, the model needs to be archived and documented for future reference and reuse.

[0088] Model segmentation

[0089] In some scenarios, models may require multiple devices to collaborate, which involves model segmentation. Figure 5 illustrates a scheme for segmenting AI / ML model inference. As shown in Figure 5, based on the current AI / ML task and working environment, the AI / ML operation (or AI / ML model) is divided into multiple parts. The purpose is to transfer computationally intensive tasks to network-side nodes, while computations that are latency-sensitive or require to remain on the terminal under certain privacy protection rules remain on the terminal device. At this point, the terminal device executes the AI / ML operation to a specific part, or executes the AI / ML model (such as a neural network) to a specific layer, and sends the generated intermediate data to the network side. Correspondingly, the network-side nodes are responsible for executing the remaining parts of the AI / ML operation or the remaining layers of the AI / ML model, and feeding back the inference results to the terminal device. It should be noted that in the example shown in Figure 5, the final inference result is output by network-side AI / ML node 2. Depending on the actual use case, the inference result can also be output by other nodes, such as network-side AI / ML node 1.

[0090] Typically, to limit the required uplink data rate, potential split points can be set after pooling layers with relatively small output data volumes, as shown in Figure 6. Generally, the earlier the split point, the less computation the terminal device bears (e.g., potential split point 1). Conversely, the later the split point (e.g., potential split point 3), the less data needs to be transmitted. This is because the later the potential split point, the more pooling operations the device performs, and the greater the reduction in data volume.

[0091] Currently, model-based communication systems are a crucial characteristic of future communication system development. For example, AI / ML model-based communication is a key feature of 6G communication networks, applied to various communication processes. First, AI / ML models play a vital role in communication processes such as resource allocation, spectrum management, and mobility management. By adopting AI / ML technologies, 6G networks can allocate resources more efficiently, optimize spectrum usage, and provide more stable connections when users are mobile. Second, AI models also play a crucial role in the security and privacy protection of 6G networks. By introducing advanced encryption technologies and security protocols based on AI / ML models, user data security can be better protected, and privacy can be ensured during data transmission. Furthermore, AI models have demonstrated significant potential in 6G network application scenarios. For example, in ultra-high-definition communication and virtual reality applications, AI models can optimize data transmission speed and stability, providing a seamless user experience. In the Internet of Things (IoT) and smart cities, AI / ML models can handle the data transmission needs of a large number of devices, ensuring the reliability and efficiency of communication. Finally, AI / ML models also play an important role in fields such as autonomous driving and industrial automation. Through real-time data processing and edge computing technologies, AI models can meet the requirements of low latency and high reliability, ensuring the safe and stable operation of the system.

[0092] As discussed above, in an intelligently-intelligent 6G network, each node (e.g., one or more of terminal devices, access network devices, network nodes (NFs), and application centers (AFs)) can possess AI / ML capabilities. Each node will support model operations locally or collaboratively support distributed model operations to achieve an intelligently-intelligent network architecture, meeting diverse AI / ML-based model-based needs. Therefore, future communication networks will contain a large number of complex model operations, posing significant demands for model lifecycle management. How to manage the large number of models in future networks is a problem that urgently needs to be solved.

[0093] Therefore, to address the aforementioned issues, this application introduces a first network element for managing models within the communication system, thereby providing a solid foundation for model-based communication processes. In this application, the model may include the AI / ML model mentioned above.

[0094] In some implementations, the first network element used for model management may include managing processes such as model training, model deployment, model performance monitoring, and model updates. In some scenarios, the first network element is also called a model management function (MMF). This application does not limit the name of the first network element, and the first network element can be any network element with similar functions introduced in future communication systems.

[0095] In some implementations, the first network element can be a network element in the core network.

[0096] In some implementations, the first network element can interact with other network elements based on a service-based architecture (SBA). For example, as shown in Figure 7, taking the first network element as MMF, other network elements can communicate with MMF by calling its service interface Nmmf.

[0097] In some implementations, the first network element is used for one or more of the following: managing model information indicated (or registered) by other network elements; authorizing other network elements to use the model; configuring model tasks for other network elements; and determining model strategies.

[0098] Taking the first network element as an example of managing model information indicated by other network elements, in some implementations, the first network element is used to store model information indicated by other network elements to the first network element.

[0099] In some implementations, model information can refer to the configuration information of models supported by other network elements; therefore, model information is also called model profile. Alternatively, model information can refer to the model information supported by other network elements; therefore, model information is also called model capability information.

[0100] In some implementations, model information can indicate one or more of the following: models supported by other network elements; models allocated by other network elements; model functions supported by other network elements; model structures supported by other network elements; regions where other network elements run models; model interoperability supported by other network elements; model size supported by other network elements; and model complexity supported by other network elements.

[0101] Taking the model information indicating models supported for deployment by other network elements as an example, the model information supported for deployment by other network elements may include, for example, the model identifier of the model. Here, models supported for deployment by other network elements can be understood as models already deployed by other network elements and / or models supported by other network elements but not yet deployed.

[0102] Taking the model information indicating models already allocated to other network elements as an example, the model information already allocated to other network elements may include, for example, the model identifier of the model. Here, the models already allocated to other network elements can be understood as models allocated (or configured) for other network elements.

[0103] Taking the model information indicating the model functions supported by other network elements as an example, the model functions supported by other network elements are used to indicate which functions other network elements can support using the model. The supported model functions may include one or more of the following: mobility management functions, location-related functions, and perception-related functions. Of course, in the embodiments of this application, the supported model functions may also include other functions.

[0104] In some implementations, the model functions supported by other network elements can be indicated by model function identifiers.

[0105] Taking the model information indicating the model structure supported by other network elements as an example, the model structure supported by other network elements can be used to indicate that the model structure supported by other network elements includes linear model structures and / or neural network model structures. Of course, in the embodiments of this application, the model structure supported by other network elements may also include nonlinear structures, etc.

[0106] Taking the example of model information indicating the region where other network elements run the model, the region where other network elements run the model can be understood as the region where other network elements use the model. In this embodiment, the region is not limited. For example, the region could be a cell. Or, for example, the region could be a national geographical region. Of course, in this embodiment, the aforementioned region could also be a region specifically designated for model use.

[0107] In some implementations, the region where the model is used can be matched with the region corresponding to the training data used to train the model, which helps improve the accuracy of using the model. Here, the region corresponding to the training data can be the region where the training data was acquired.

[0108] In some implementations, the region where the model is running can be indicated by a region identifier or by the region's longitude and / or latitude information; however, this application does not limit this.

[0109] Taking model information indicating the model interoperability of models supported by other network elements as an example, model interoperability is used to indicate the model's ability to achieve seamless data and service interaction between different systems, platforms, or frameworks. For example, model interoperability can be used to indicate whether a model is a proprietary model or an open-source model.

[0110] In some implementations, if the model is a proprietary model, the model information may also include the vendor information of the model, which is used to indicate which vendors' models other network elements support / run.

[0111] Taking the model information indicating the model size supported by other network elements as an example, in some implementations, model size is an important indicator for measuring model complexity. It is closely related to the number of model parameters, computational cost, and storage requirements, and has a direct impact on model performance and training efficiency.

[0112] Taking the model information indicating the model complexity supported by other network elements as an example, in some implementations, the model complexity supported by other network elements can be the maximum model complexity supported by other network elements.

[0113] In some implementations, model complexity is an important metric for measuring the complexity and computational requirements of a deep learning model. It is typically measured from the following aspects: the number of model parameters, the computational cost (floating-point operations per second, FLOPs), the number of layers, the number of filters, the number of connections, activation functions, the amount of data, and task complexity.

[0114] Taking the example of a first network element authorizing other network elements to use a model, that is, the first network element can provide other network elements with the required model, for example, by assigning a model ID. It should be noted that the first network element itself does not need to generate the model, but it can tell other network elements which network element it can obtain the model from. Of course, in this embodiment, the first network element can also generate the model and directly provide it to other network elements.

[0115] Taking the first network element as an example of configuring model tasks for other network elements, that is to say, the first network element can support various model tasks according to the needs of the network inside or outside (for example, configuring model tasks for other network elements).

[0116] In some implementations, model tasks can include any stage of the model lifecycle. These stages can include one or more of the following: requirements analysis, data preparation, model design, model training, model deployment, model maintenance, and model retirement. Correspondingly, model tasks can target one or more of these stages. For example, model training tasks can include model training tasks. Further details about these stages can be found above.

[0117] Taking the first network element as an example of determining the model strategy, that is to say, the first network element can determine the appropriate model strategy for a certain requirement (such as positioning requirement, CSI requirement, or perception requirement).

[0118] In some implementations, a model strategy (also known as a model training strategy) is used to configure the model training task. The model strategy indicates one or more of the following: the model to be trained corresponding to the model training task; the model functionality of the model to be trained; the model structure of the model to be trained; the model operating region of the model to be trained; the model interoperability of the model to be trained; the model size of the model to be trained; the model complexity of the model to be trained; the input of the model to be trained; the output of the model to be trained; the training data information of the model to be trained; the parameter configuration information of the model to be trained; the model training time of the model to be trained; whether a model training task needs to be performed based on the initial model; and the nodes participating in the model training task.

[0119] Taking the model strategy as an example of indicating the model to be trained for the model training task, the model to be trained for the model training task can be understood as the model that the network element (e.g., terminal device, access network device, NF) needs to train.

[0120] In some implementations, the model strategy can indicate the model to be trained for the training task by carrying a model identifier of the model to be trained.

[0121] Taking a model strategy used to indicate the model functionality of a model to be trained as an example, the model functionality of the model to be trained indicates the functions that the model to be trained can support. The model functionality may include one or more of the following: mobility management functions, location-related functions, and perception-related functions. Of course, in the embodiments of this application, the model functionality may also include other functions.

[0122] In some implementations, the model functionality of the model to be trained can be indicated by a model functionality identifier.

[0123] Taking the model strategy used to indicate the model structure of the model to be trained as an example, the model structure includes a linear model structure and / or a neural network model structure. Of course, in the embodiments of this application, the model structure of the model to be trained may also include nonlinear structures, etc.

[0124] Taking the model strategy used to indicate the model operating region of the model to be trained as an example, the model operating region can be understood as the region where the model is used. In this embodiment, the region is not limited. For example, the region could be a neighborhood. Or, for example, the region could be a national geographic region. Of course, in this embodiment, the aforementioned region could also be a region specifically designated for model use.

[0125] In some implementations, the model's operating region can be matched with the region corresponding to the training data used to train the model, which helps improve the accuracy of using the model. Here, the region corresponding to the training data can be the region where the training data was acquired.

[0126] In some implementations, the model operating area can be indicated by a region identifier or by the region's longitude and / or latitude information; however, this application does not limit this.

[0127] Taking model policies as an example to indicate the interoperability of a model being trained, model interoperability indicates the model's ability to seamlessly interact with data and services across different systems, platforms, or frameworks. For instance, model interoperability can indicate whether a model is proprietary or open-source.

[0128] In some implementations, if the model is a proprietary model, the model information may also include the vendor information of the model, which is used to indicate which vendors' models other network elements support / run.

[0129] Taking the model strategy used to indicate the model size of the model to be trained as an example, in some implementations, model size is an important indicator of model complexity. It is closely related to the number of model parameters, computational cost and storage requirements, and has a direct impact on model performance and training efficiency.

[0130] Taking the model strategy as an example of indicating the model complexity of the model to be trained, in some implementations, model complexity can be used to indicate the computational cost of the model to be trained.

[0131] In some implementations, model complexity is an important metric for measuring the complexity and computational requirements of a deep learning model. It is typically measured from the following aspects: the number of model parameters, the computational cost (FLOPs), the number of layers, the number of filters, the number of connections, activation functions, the amount of data, and task complexity.

[0132] Taking the model strategy used to indicate the input of the model to be trained as an example, in some implementations, the input of the model to be trained may include the input format and / or input type.

[0133] Taking the model strategy used to indicate the output of the model to be trained as an example, in some implementations, the output of the model to be trained may include the output format and / or output type.

[0134] Taking the model strategy as an example of using training data information to indicate the model to be trained, in some implementations, the training data information includes the size and / or format of the training data.

[0135] Taking the model strategy as an example of using model parameter configuration information to indicate the model to be trained, in some implementations, the model parameter configuration information is used to indicate the model training accuracy and / or the learning rate of the model to be trained.

[0136] Taking the model training time as an example, where the model strategy is used to indicate the training time of the model to be trained, in some implementations, the model training time is used to indicate the maximum time to train the model. Of course, in the embodiments of this application, the model training time can also be used to indicate the average time to train the model.

[0137] Taking the model strategy as an example of indicating whether a model training task needs to be performed based on an initial model, in some implementations, the initial model can be understood as the original model used for model training.

[0138] Taking the model strategy as an example of indicating the nodes (or network elements) participating in the model training task, in some implementations, the node participating in the model training task is used to indicate the node receiving the intermediate data generated by the current node. For example, in a distributed training scenario, the node participating in the model training task is used to indicate the node receiving the intermediate data generated by the current node. Therefore, the node participating in the model training task is also called the "target node".

[0139] In some implementations, nodes participating in the model training task can be indicated by their identifiers or addresses, where the address may include, for example, an IP address or a MAC address.

[0140] In some implementations, the functions of the first network element can be provided to other network elements as services. For example, the first network element can provide one or more of the following services: model registration service (also known as "model information acquisition service"), model deployment service, model performance monitoring service, model update service, and model release service.

[0141] In some implementations, the model indication service can be represented as "Nmmf_ModelManagement_ModelRegister". Correspondingly, other network elements (also known as consumers) can indicate (or register with) the model information they support to the first network element by calling this service. For an introduction to model information, please refer to the above.

[0142] In some implementations, the model training service can be represented as "Nmmf_ModelManagement_ModelTraining", and other network elements (also known as consumers) can call this service to fulfill their model training needs.

[0143] In some implementations, the model deployment service can be represented as "Nmmf_ModelManagement_ModelDeployment". Correspondingly, other network elements (also known as consumers) can fulfill their model deployment needs by calling this service. For example, other network elements can request the first network element to allocate the corresponding model and determine the model ID by calling the first network element's model deployment service.

[0144] In some implementations, the model monitoring service can be represented as "Nmmf_ModelManaement_ModelMonitoring". Accordingly, other network elements (also known as consumers) can call this service to fulfill their model performance monitoring needs.

[0145] In some implementations, the model update service can be represented as "Nmmf_ModelManagement_ModelUpdate". Accordingly, other network elements (also known as consumers) can call this service to update the requested model and / or the deployed model.

[0146] In some implementations, the model release service can be represented as "Nmmf_ModelManagement_ModelRelease", and correspondingly, other network elements (also known as consumers) can call this service to release requested models and / or deployed models.

[0147] It should be noted that the embodiments of this application do not limit the other network elements. In some implementations, other network elements may include one or more of the following: terminal equipment, access network equipment, NF, and AF. Of course, in the embodiments of this application, other network elements also include other network elements introduced in future communication systems.

[0148] The first network element in the embodiments of this application has been introduced above. The following section describes the interaction between the terminal device and the first network element. In some implementations, the terminal device and the first network element can exchange first information used for model management. In the embodiments of this application, the interaction method between the first network element and the terminal device is not limited. In some implementations, the terminal device can communicate with the first network element by calling its service interface. For example, the terminal device can directly interact with the first network element by calling SBA N1. In other implementations, the terminal device can interact with the first network element through a second network element, which may be, for example, a mobility management network element. In still other implementations, the terminal device can be a remote terminal in a relay scenario; in this case, the terminal device can communicate with the first network element through a relay terminal.

[0149] In this application embodiment, the interaction between the terminal device and the first network element is different based on the different meanings of the first information. Therefore, the wireless communication method in this application embodiment is described below based on the different meanings of the first information, in conjunction with Embodiment 1 and Embodiment 2 respectively.

[0150] Example 1: The first information is used to request the first network element to indicate the model information supported by the terminal device.

[0151] Figure 8 is a schematic flowchart of a wireless communication method according to an embodiment of this application. The method shown in Figure 8 includes step S810. In step S810, the terminal device sends first information to a first network element, wherein the first information is used to request the first network element to indicate model information supported by the terminal device, or in other words, the first information is used to request the first network element to register the model information supported by the terminal device, or in other words, the first information is used to request the registration of model information supported by the terminal device.

[0152] In some implementations, the first information carries one or more of the following: the identifier of the terminal device; the identifier of the first network element; and model information supported by the terminal device, wherein the identifier of the first network element may include the fully qualified domain name (FQDN) of the first network element and / or the address of the first network element (e.g., IP address or media access control address (MAC) address).

[0153] In some implementations, the model information supported by the terminal device is used to indicate one or more of the following: models supported for deployment by the terminal device; models already allocated to the terminal device; model functionality supported by the terminal device; model structure supported by the terminal device; the region where the terminal device runs the model; model interoperability supported by the terminal device; model size supported by the terminal device; and model complexity supported by the terminal device. For a more detailed explanation of the model information supported by the terminal device, please refer to the above text.

[0154] In some implementations, the first information can be sent from the terminal device to the first network element through a service provided by the first network element. For example, the terminal device can send the first information by calling the model registration service "Nmmf_ModelManagement_ModelRegister" of the first network element, for which a description can be found above. In some scenarios, the first information sent by calling the above service is also called a model indication request, denoted as "Nmmf_ModelManagement_ModelRegister_request".

[0155] In other implementations, the first information can be sent from the terminal device to the first network element via the second network element. Of course, in this embodiment, the terminal device can be a remote terminal in a relay scenario. In this case, the terminal device sending the first information to the first network element can include the remote terminal sending the first information to the first network element via a relay terminal.

[0156] In some implementations, the model information supported by the terminal device is carried in a container, which helps improve the security of transmitting model information. For example, when the terminal device sends model information to the first network element through the second network element, carrying the model information in a container helps avoid the second network element parsing the model information in the container. In other words, the second network element can pass the container through to the first network element, which helps improve the security of transmitting model information.

[0157] In some implementations, the container that carries model information can be called a "ModelInfo Container".

[0158] In some implementations, if the first information is sent from the terminal device to the first network element through the second network element, the second network element, after receiving the first information, can determine which network element to send the first information to based on the identifier of the first network element in the first information.

[0159] In some implementations, the second network element can determine the service interface to be called from the first network element based on the container type (or container name) of the container. That is to say, the above method also includes: the second network element determining the first service interface to be called from the first network element based on the container type of the container, and the first service interface is used to register the model information supported by the terminal device with the first network element.

[0160] For example, when the container type is "model information container", the second network element can determine to call the model registration service "Nmmf_ModelManagement_ModelRegister", and related information can be found above.

[0161] In some implementations, the first information sent by the second network element to the first network element by calling the aforementioned service can be carried in a model registration request, denoted as "Nmmf_ModelManagement_ModelRegister_request". Of course, in this embodiment, the model registration request can also be called a "model information registration request", and this embodiment does not limit it in this way.

[0162] In some implementations, after receiving the first information, the first network element can determine whether to authorize the terminal device's model registration request based on the terminal device's identifier and / or the first network element's identifier. For example, if the first network element determines, based on the terminal device's identifier, that the terminal device has subscribed to model-related services, then the first network element can authorize the model registration request. As another example, if the first network element determines, based on its own identifier, that the first information is intended for itself, then the first network element can authorize the model registration request. Yet another example, if the first network element determines, based on its own identifier, that the first information is intended for itself, and based on the terminal device's identifier, determines that the terminal device has subscribed to model-related services, then the first network element can authorize the model registration request.

[0163] In some implementations, the above method further includes: the first network element sending second information to the terminal device, the second information being used to indicate whether the instruction (or registration) to the first network element for the model information supported by the terminal device has been completed.

[0164] In some implementations, the second information carries one or more of the following: the identifier of the terminal device; the identifier of the first network element; and information used to indicate to the first network element whether the model information supported by the terminal device is complete (or information indicating whether the model information supported by the terminal device is complete).

[0165] In some implementations, the second information can be sent directly from the first network element to the terminal device through its service. For example, the first network element sends the second information to the terminal device by calling its model registration service "Nmmf_ModelManagement_ModelRegister," for which a description can be found above. In some scenarios, the second information sent by calling the above service is also called the model registration response, denoted as "Nmmf_ModelManagement_ModelRegister_response."

[0166] In some implementations, the second information can be sent from the first network element to the terminal device via the second network element. Of course, in this embodiment, the terminal device can be a remote terminal in a relay scenario; in this case, the second information can be sent from the first network element to the terminal device via the relay network element.

[0167] In some implementations, the information used to indicate to the first network element whether the model information supported by the terminal device is complete can be carried in a container, which helps improve the security of transmitting the information. For example, when the first network element sends the information to the terminal device through the second network element, carrying the information in a container helps prevent the second network element from parsing the information in the container. In other words, the second network element can pass the container through to the terminal device, which helps improve the security of transmitting the information.

[0168] For ease of understanding, the process by which a terminal device requests indication model information from a first network element through a second network element in an embodiment of this application is described below with reference to Figure 9. Referring to Figure 9, it is assumed that the second network element is an AMF and the first network element is an MMF. The method shown in Figure 9 includes steps S910 to S960. It should be understood that Figure 9 mainly introduces the method flow in an embodiment of this application; the relevant terminology can be found above. Furthermore, in an embodiment of this application, indication model information and registration model information can be interchanged. For ease of description, the following description uses registration model information as an example.

[0169] In step S910, the terminal device sends the first information to the AMF.

[0170] In some implementations, the first information can be carried in a non-access stratum (NAS) message. This first information includes the terminal device ID, the MMF ID, and the model information supported by the terminal device. The MMF ID indicates the MMF the terminal device wishes to register. Furthermore, a description of the model information supported by the terminal device can be found in Table 2. This information may include one or more of the following: model ID, model function, model structure, model operating region, model interoperability information, model size, and model complexity.

[0171] Table 2

[0172] In some implementations, the model information supported by the terminal device is carried in the ModelInfo Container 1, which helps to avoid the AMF parsing or understanding the model information supported by the terminal device. This allows the AMF to directly forward the model information supported by the terminal device to the MMF, thereby improving the security of transmitting the model information supported by the terminal device.

[0173] In step S920, the AMF determines the MMF of the receiving model information container based on the MMF ID in the NAS message, and the AMF determines the model registration service of the MMF that needs to be called based on the container type of the container.

[0174] In step S930, the AMF sends a model registration request to the MMF.

[0175] In some implementations, the model registration request includes the terminal device ID, MMF ID, and model information container 1.

[0176] In step S940, the MMF authorizes the model registration request of the terminal device based on the terminal device ID and the MMF ID.

[0177] In some implementations, the MMF can store model information registered by the terminal device. Furthermore, this application does not limit the implementation method of the model registration request for the MMF-authorized terminal device; relevant details can be found above.

[0178] In step S950, the MMF sends a model registration response to the AMF.

[0179] In some implementations, the model registration response includes a terminal device ID, an MMF ID, and a model information container 2. The model information container 2 carries information indicating whether the registered terminal device has completed supporting model information. In this embodiment, by carrying the information indicating whether the registered terminal device has completed supporting model information in the model information container 2, it helps avoid the AMF parsing or understanding the information within the model information container, allowing the AMF to directly forward the model information container to the terminal device, thereby improving the security of transmitting the information within the model information container.

[0180] In step S960, the AMF sends a second message to the terminal device to notify the terminal device whether the model information registration is complete.

[0181] In some implementations, the second information is carried in the NAS message, and the second information includes the terminal device ID, MMF ID, and model information container 2.

[0182] The above example illustrates how a terminal device registers model information with a first network element. As mentioned earlier, other network elements (e.g., access network devices, NFs, or AFs) can also indicate (or register) their supported model information with the first network element. The process of other network elements indicating model information to the first network element is similar to the process of a terminal device calling the first network element's service interface to indicate model information; only the terminal device is replaced with another network element.

[0183] In some implementations, other network elements send first information to the first network element. The first information may carry one or more of the following: the identifier of the other network element (e.g., the address of the other network element); the identifier of the first network element (see above for related information); and model information supported by the other network element.

[0184] In some implementations, model information supported by other network elements is used to indicate one or more of the following: models supported for deployment by other network elements; models allocated by other network elements; model functions supported by other network elements; model structures supported by other network elements; regions where other network elements run models; model interoperability of models supported by other network elements; model size supported by other network elements; and model complexity supported by other network elements. For a more detailed explanation of model information, please refer to the above text.

[0185] In some implementations, the first information can be sent from other network elements to the first network element through its services. For example, other network elements can send the first information by calling the first network element's model registration service "Nmmf_ModelManagement_ModelRegister," for which a description can be found above. In some scenarios, the first information sent by calling the above service is also called a model registration request, denoted as "Nmmf_ModelManagement_ModelRegister_request."

[0186] In some implementations, after receiving the first information, the first network element can determine whether to authorize the model registration request of other network elements based on the identifiers of other network elements and / or its own identifier. For example, if the first network element determines, based on the identifiers of other network elements, that other network elements have subscribed to model-related services, then the first network element can authorize the model registration request. As another example, if the first network element determines, based on its own identifier, that the first information is intended for itself, then the first network element can authorize the model registration request. Yet another example, if the first network element determines, based on its own identifier, that the first information is intended for itself, and based on the identifiers of other network elements, determines that other network elements have subscribed to model-related services, then the first network element can authorize the model registration request.

[0187] In some implementations, the above method further includes: the first network element sending second information to other network elements, the second information being used to indicate whether the instruction (or registration) to the first network element for the model information supported by other network elements has been completed.

[0188] In some implementations, the second information carries one or more of the following: the identifier of other network elements; the identifier of the first network element; and information indicating whether the registration of model information supported by other network elements with the first network element has been completed.

[0189] In some implementations, the second information can be sent directly from the first network element to other network elements through its services. For example, the first network element can send the second information to other network elements by calling its model registration service "Nmmf_ModelManagement_ModelRegister," for which a description can be found above. In some scenarios, the second information sent by calling the above service is also called the model registration response, denoted as "Nmmf_ModelManagement_ModelRegister_response."

[0190] For ease of understanding, the process of other network elements requesting registration model information from the first network element in this embodiment of the application is described below with reference to Figure 10. Referring to Figure 10, it is assumed that the second network element is NF and the first network element is MMF. The method shown in Figure 10 includes steps S1010 to S1030. It should be understood that Figure 10 mainly introduces the method flow in this embodiment of the application; the relevant terminology can be referred to above. Furthermore, in this embodiment of the application, the indication model information and the registration model information can be interchanged. For ease of description, the following description uses the registration model information as an example.

[0191] In step S1010, NF sends a model registration request to MMF.

[0192] In some implementations, NF can determine that it needs to call the MMF's model registration service. Accordingly, NF sends a model registration request to the MMF, which includes the NF ID, the MMF ID, and information about the models supported by NF. The MMF ID indicates the MMF that NF wishes to register. For more information on the models supported by NF, please refer to the section above.

[0193] In step S1020, MMF authorizes the NF's model registration request based on the NF ID and MMF ID.

[0194] In some implementations, the MMF can store model information registered with the NF. Furthermore, this application does not limit the implementation method of the MMF authorizing the NF model registration request; relevant details can be found above.

[0195] In step S1030, MMF sends a model registration response to NF.

[0196] In some implementations, the model registration response includes the NF ID, the MMF ID, and information indicating whether the registration of NF-supported model information is complete.

[0197] Example 2: The first information is used to configure the model training task for the terminal device.

[0198] Figure 11 is a schematic flowchart of a wireless communication method according to another embodiment of this application. The method shown in Figure 11 includes step S1110. In step S1110, a first network element sends first information to a terminal device.

[0199] In some implementations, the first information is used to configure the model training task for the terminal device, or in other words, the first information is used by the first network element to configure the model training task for the terminal device.

[0200] In some implementations, the first piece of information (also known as model training task configuration) indicates one or more of the following: the model to be trained corresponding to the model training task; the model functionality of the model to be trained; the model structure of the model to be trained; the model operating region of the model to be trained; the model interoperability of the model to be trained; the model size of the model to be trained; the model complexity of the model to be trained; the input of the model to be trained; the output of the model to be trained; the training data information of the model to be trained; the model parameter configuration information of the model to be trained; the model training time of the model to be trained; whether a model training task based on the initial model is required; and the nodes participating in the model training task. See the above for related information.

[0201] In some implementations, the first information can be sent from the first network element to the terminal device by calling its service. For example, the first network element can send the first information by calling its model training service "Nmmf_ModelManagement_ModelTraining," for which a description can be found above. In some scenarios, the first information sent by calling the above service is also called the model training response, denoted as "Nmmf_ModelManagement_ModelTraining_response."

[0202] In some implementations, the first information can be sent from the first network element to the terminal device via the second network element. Of course, in this embodiment, the terminal device can be a remote terminal in a relay scenario; in this case, the first information can be sent from the first network element to the terminal device via the relay terminal.

[0203] In some implementations, the first information can be carried in a container, which helps improve the security of transmitting the first information. For example, when a terminal device sends the first information to a first network element through a second network element, carrying the first information in a container helps prevent the second network element from parsing the first information in the container. In other words, the second network element can pass the container through to the first network element, which helps improve the security of transmitting the first information.

[0204] In some implementations, the aforementioned first information may be requested by the terminal device from the first network element via third information. That is to say, the above method further includes: the terminal device sending third information to the first network element, the third information being used to request the execution of a model training task from the first network element, or in other words, the third information being used to request the execution of a model training task.

[0205] In some implementations, the third information carries one or more of the following: the identifier of the terminal device; the identifier of the first network element (see above for related information); and the model information of the model to be trained.

[0206] In some implementations, the model information of the model to be trained includes one or more of the following: the model functionality; the training completion time; the training accuracy; the input format; and the output format. See the above for further details.

[0207] In some implementations, the third information can be sent from the terminal device to the first network element by invoking its service. For example, the terminal device can send the third information by invoking the model training service "Nmmf_ModelManagement_ModelTraining" of the first network element, the details of which can be found above. In some scenarios, the third information sent by invoking the above service is also called a model training request, denoted as "Nmmf_ModelManagement_ModelTraining_request".

[0208] In some implementations, the third information can be sent from the terminal device to the first network element via the second network element. Of course, in this embodiment, the third information can be sent from the terminal device to the first network element via a relay terminal.

[0209] In some implementations, the model information of the model to be trained can be carried in a container, which helps improve the security of transmitting the model information. For example, when a terminal device sends the model information of the model to be trained to a first network element through a second network element, carrying the model information in a container helps avoid the second network element parsing the model information in the container. In other words, the second network element can pass the container through to the first network element, which helps improve the security of transmitting the model information.

[0210] In some implementations, the container that holds the model information of the model to be trained can be called a "model training container".

[0211] In some implementations, if the third information is sent from the terminal device to the first network element through the second network element, the second network element, after receiving the first information, can determine which network element to send the third information to based on the identifier of the first network element in the third information.

[0212] In some implementations, the second network element can determine the service interface to be called from the first network element based on the container type (or container name). That is to say, the above method also includes: the second network element determines the second service interface to be called from the first network element based on the container type, and the second service interface is used to call the services provided by the first network element related to the model training task.

[0213] For example, when the container type is "model training container", the second network element can determine to call the model training service "Nmmf_ModelManagement_ModelTraining", and related information can be found above.

[0214] In some implementations, the terminal device needs to train the model based on the initial model. In this case, the terminal device can request the initial model from the first network element. That is to say, the above method also includes: the terminal device sending fourth information to the first network element, the fourth information being used to request the initial model corresponding to the model training task from the first network element; and / or the terminal device receiving fifth information sent by the first network element, the fifth information being used to indicate the initial model corresponding to the model training task.

[0215] In some implementations, the fourth piece of information can carry the model identifier of the initial model.

[0216] In some implementations, the fifth information may carry a model identifier and / or the initial model. Of course, in the embodiments of this application, the fifth information may indicate the identifier of a network element capable of obtaining the initial model, so that the terminal device can obtain the initial model from that network element.

[0217] It should be noted that when configuring the model strategy for the terminal device, the first network element can determine whether the terminal device needs to train based on the initial model. Therefore, when the terminal device needs to train based on the initial model, the terminal device can choose not to send the fourth information to the first network element, and instead, the first network element can directly send the fifth information to the terminal device. Of course, in this embodiment, the first network element can configure the initial model for the terminal device based on the fourth information sent by the terminal device.

[0218] In some implementations, the fourth piece of information can be sent from the terminal device to the first network element through the service interface of the first network element. For example, the terminal device can send the fourth piece of information by calling the model deployment service "Nmmf_ModelManagement_ModelDeployment" of the first network element, where a description of this service can be found above. In some scenarios, the fourth piece of information sent by calling the above service is also called a model deployment request, denoted as "Nmmf_ModelManagement_ModelDeployment_request".

[0219] In some implementations, the fourth information can be sent by the terminal device to the first network element through the second network element. Of course, in the embodiments of this application, the fourth information can be sent by the terminal device to the first network element through a relay terminal.

[0220] In some implementations, the model identifier of the initial model can be carried in a container, which helps improve the security of transmitting the model identifier of the initial model. For example, when a terminal device sends the model identifier of the initial model to a first network element through a second network element, carrying the model identifier of the initial model in a container helps prevent the second network element from parsing the model identifier of the initial model in the container. In other words, the second network element can pass the container through to the first network element, which helps improve the security of transmitting the model identifier of the initial model.

[0221] In some implementations, the container that carries the model identifier of the initial model can be called a "ModelDeployment Container".

[0222] In some implementations, if the fourth information is sent from the terminal device to the first network element through the second network element, the second network element, after receiving the first information, can determine which network element to send the fourth information to based on the identifier of the first network element in the fourth information.

[0223] In some implementations, the second network element can determine the service interface to be called from the first network element based on the container type (or container name). That is to say, the above method also includes: the second network element determining the third service interface to be called from the first network element based on the container type, and the third service interface is used to call the first network element to deploy the initial model.

[0224] For example, when the container type is "model deployment container", the second network element can determine to call the model deployment service "Nmmf_ModelManagement_ModelDeployment", and related information can be found above.

[0225] The transmission method of the fourth information in the embodiments of this application has been described above. The transmission method of the fifth information in the embodiments of this application is described below. In some implementations, the fifth information can be sent by the first network element to the terminal device through the service interface of the first network element. For example, the first network element can send the fifth information by calling the model deployment service "Nmmf_ModelManagement_ModelDeployment" of the first network element. The relevant introduction of this service can be found above. In some scenarios, the fifth information sent by calling the above service is also called the model deployment response, represented as "Nmmf_ModelManagement_ModelDeployment_response".

[0226] In some implementations, the fifth information can be sent from the first network element to the terminal device via the second network element. Of course, in this embodiment, the fifth information can be sent from the first network element to the terminal device via a relay terminal.

[0227] In some implementations, the model identifier and / or initial model of the initial model can be carried in a container, which helps improve the security of transmitting the model identifier and / or initial model. For example, when the first network element sends the model identifier and / or initial model of the initial model to the terminal device through the second network element, carrying the model identifier and / or initial model of the initial model in a container helps prevent the second network element from parsing the model identifier and / or initial model of the initial model in the container. In other words, the second network element can pass the container through to the terminal device, which helps improve the security of transmitting the model identifier and / or initial model of the initial model.

[0228] In some implementations, the container that carries the model identifier and / or the initial model can be called a "ModelDeployment Container".

[0229] In some scenarios, the model training process may be distributed training. Assuming the network elements participating in distributed training include terminal devices and network elements (NFs), the terminal device can send intermediate results of the trained model to the NF so that the NF can continue model training based on these intermediate results. Similarly, after completing model training, the NF needs to notify the terminal device that model training is complete (i.e., the model training task is complete) through the first network element. Therefore, the above method also includes: the first network element sending a first indication message to the terminal device, which indicates that the model training task is complete.

[0230] In some implementations, the first instruction information can be sent from the first network element to the terminal device through its service interface. For example, the first network element can send the first instruction information by calling its model training service "Nmmf_ModelManagement_ModelTraining," for which a description can be found above. In some scenarios, the first instruction information sent by calling the above service can be carried in the model training response, represented as "Nmmf_ModelManagement_ModelTraining_response."

[0231] In some implementations, the first instruction information can be sent from the first network element to the terminal device via the second network element. Of course, in this embodiment, the first instruction information can be sent from the first network element to the terminal device via a relay terminal.

[0232] In some implementations, the first indication information can be carried in a container, which helps improve the security of transmitting the first indication information. For example, when the first network element sends the first indication information to the terminal device through the second network element, carrying the first indication information in a container helps prevent the second network element from parsing the first indication information in the container. In other words, the second network element can pass the container through to the terminal device, which helps improve the security of transmitting the first indication information.

[0233] The above example illustrates how the first network element configures model training tasks for a terminal device. In some scenarios, other network elements (e.g., access network devices, NFs, or AFs) can also have their model training tasks configured by the first network element. The process of the first network element configuring model training tasks for other network elements is similar to the process of the first network element calling its service interface to configure model training tasks for a terminal device; only the terminal device described above needs to be replaced with another network element.

[0234] In some implementations, the first network element sends the first information to other network elements.

[0235] In some implementations, the first information is used to configure the model training task. This first information indicates one or more of the following: the model to be trained corresponding to the model training task; the model functionality of the model to be trained; the model structure of the model to be trained; the model operating region of the model to be trained; the model interoperability of the model to be trained; the model size of the model to be trained; the model complexity of the model to be trained; the input of the model to be trained; the output of the model to be trained; the training data information of the model to be trained; the model parameter configuration of the model to be trained; the model training time of the model to be trained; whether a model training task based on the initial model is required; and the nodes participating in the model training task. See the above for related details.

[0236] In some implementations, the first information can be sent from the first network element to other network elements through its service interface. For example, the first network element can send the first information by calling its model training service "Nmmf_ModelManagement_ModelTraining," for which a description can be found above. In some scenarios, the first information sent by calling the above service is also called the model training response, denoted as "Nmmf_ModelManagement_ModelTraining_response."

[0237] In some implementations, the aforementioned first information may be requested by other network elements from the first network element via third information. That is to say, the above method also includes: other network elements sending third information to the first network element, the third information being used to request the first network element to perform a model training task.

[0238] In some implementations, the third information carries one or more of the following: the identifier of other network elements; the identifier of the first network element (see above for related information); and the model information of the model to be trained.

[0239] In some implementations, the model information of the model to be trained includes one or more of the following: the model functionality; the training completion time; the training accuracy; the input format; and the output format. See the above for further details.

[0240] In some implementations, the third information can be sent from other network elements to the first network element through its service interface. For example, other network elements can send the third information by calling the first network element's model training service "Nmmf_ModelManagement_ModelTraining," for which a description can be found above. In some scenarios, the third information sent by calling the above service is also called a model training request, denoted as "Nmmf_ModelManagement_ModelTraining_request."

[0241] In some implementations, other network elements need to train the model based on the initial model. In this case, the other network elements can request the initial model from the first network element. That is to say, the above method also includes: other network elements sending a fourth message to the first network element, the fourth message being used to request the initial model corresponding to the model training task from the first network element; and / or other network elements receiving a fifth message sent by the first network element, the fifth message being used to indicate the initial model corresponding to the model training task.

[0242] In some implementations, the fourth piece of information can carry the model identifier of the initial model.

[0243] In some implementations, the fifth information may carry the model identifier of the initial model and / or the initial model itself. Of course, in the embodiments of this application, the fifth information may indicate the identifier of the network element capable of obtaining the initial model, so that other network elements can obtain the initial model from that network element.

[0244] It should be noted that when the first network element configures model strategies for other network elements, it can know whether those other network elements need to be trained based on the initial model. Therefore, when other network elements need to train their models based on the initial model, they can choose not to send the fourth information to the first network element, but instead, the first network element can directly send the fifth information to them. Of course, in this embodiment, the first network element can configure the initial model for other network elements based on the fourth information sent by those other network elements.

[0245] In some implementations, the fourth piece of information can be sent from other network elements to the first network element through its service interface. For example, other network elements can send the fourth piece of information by calling the first network element's model deployment service "Nmmf_ModelManagement_ModelDeployment," for which a description can be found above. In some scenarios, the fourth piece of information sent by calling the above service is also called a model deployment request, denoted as "Nmmf_ModelManagement_ModelDeployment_request."

[0246] The transmission method of the fourth information in the embodiments of this application has been described above. The transmission method of the fifth information in the embodiments of this application is described below. In some implementations, the fifth information can be sent by the first network element to other network elements through the service interface of the first network element. For example, the first network element can send the fifth information by calling the model deployment service "Nmmf_ModelManagement_ModelDeployment" of the first network element. The relevant introduction of this service can be found above. In some scenarios, the fifth information sent by calling the above service is also called the model deployment response, represented as "Nmmf_ModelManagement_ModelDeployment_response".

[0247] In some scenarios, model training may be distributed. Assuming the network elements participating in distributed training include other network elements and the network function (NF), other network elements can send intermediate results generated during model training to the NF so that the NF can continue training the model based on these intermediate results. Similarly, after completing model training, the NF needs to notify other network elements that training is complete (i.e., the model training task is complete) through the first network element. Therefore, the above method also includes: the first network element sending a first indication message to other network elements, which indicates that the model training task is complete.

[0248] In some implementations, the first instruction information can be sent by the first network element to other network elements through its service interface. For example, the first network element can send the first instruction information by calling its model training service "Nmmf_ModelManagement_ModelTraining," for which a description can be found above. In some scenarios, the first instruction information sent by calling the above service can be carried in the model training response, represented as "Nmmf_ModelManagement_ModelTraining_response."

[0249] It should be noted that in some implementations, Embodiment 1 and Embodiment 2 can be used independently. In other implementations, Embodiment 1 and Embodiment 2 can be used in combination. For example, when the first network element configures the model training task for the terminal device based on the scheme of Embodiment 2, it can use the model information supported by each network element obtained through the scheme of Embodiment 1 to determine the network elements participating in the model training task. For a related description, please refer to step S1216 shown in Figure 12.

[0250] For ease of understanding, the scheme for the first network element configuration model training task in this application embodiment is described below with reference to Figure 12. Assume the first network element is MMF and the second network element is AMF. The method described in Figure 12 includes steps S1210 to S1246.

[0251] In step S1210, the terminal device sends third information to the AMF, which is used to request the execution of the model training task.

[0252] In some implementations, when a terminal device wants to perform a model training task, it can send a NAS message to the AMF. This NAS message includes third information, which includes the identifier of the terminal device; the identifier of the first network element; and the model information of the model to be trained. For a description of this information, please refer to the above text. For example, if the terminal device wants to train a perception model, the model information of the model to be trained can indicate the training completion time, the model's accuracy, input / output format, etc.

[0253] In some implementations, the model information of the model to be trained can be carried in a model training container, which helps to avoid the AMF parsing the model information of the model to be trained in the container. In other words, the AMF can pass the model training container through to the MMF, which helps to improve the security of transmitting the model information of the model to be trained.

[0254] In step S1212, the AMF determines the MMF that receives the model training container based on the MMF ID in the NAS message, and the AMF determines the model training service that needs to be called from the MMF based on the container type of the container.

[0255] In step S1214, the AMF sends a model training request to the MMF.

[0256] In some implementations, the model training request includes the terminal device ID, MMF ID, and model training container; see the previous text for related information.

[0257] In step S1216, the MMF determines the network elements (or nodes) participating in the model training task based on the model requirements corresponding to the model training task and the model information supported by each stored network element. For example, the MMF determines that the network elements participating in the model training task include terminal devices and NFs, where the NF can be, for example, an NWDAF, or the NF can be an MTLF within an NWDAF.

[0258] It should be noted that the process by which the MMF obtains the model information supported by each network element can be based on the scheme shown in Figure 10 and / or Figure 9. Of course, in the embodiments of this application, the MMF can also obtain the model information supported by each network element through other means.

[0259] In step S1218, if the training method corresponding to the model training task is distributed model training, then MMF sends model training response 1 to the terminal device and NF respectively to configure the model training task.

[0260] In some implementations, the model training response 1 carries a model training strategy, which can be found in Table 3. The model training strategy is used to indicate one or more of the following information about the model to be trained: model ID, model function, model structure, model operating region, model format, model size, model complexity, model input / output, required dataset, model parameter configuration, training time, initial model, and target network element ID or address.

[0261] Table 3

[0262] In step S1220, the terminal device and NF respectively send response information to MMF to indicate whether the model training response has been successfully received.

[0263] In step S1222, the terminal device sends fourth information to the AMF, which is used to request the initial model corresponding to the model training task.

[0264] In some implementations, the terminal device determines whether to train the model based on the initial model based on the model training information of the model to be trained. If the model training information of the model to be trained indicates that model training should be based on the initial model, the terminal device sends the fourth information to the AMF.

[0265] In some implementations, the fourth information is carried in the NAS message. The fourth information can carry the model identifier of the initial model. The model identifier of the initial model is carried in the model deployment container, which helps to avoid AMF parsing the model identifier of the initial model in the model deployment container. In other words, AMF can pass the model deployment container through MMF to improve the security of transmitting the model identifier of the initial model.

[0266] In step S1224, the AMF sends a model deployment request to the MMF to request an initial model.

[0267] In some implementations, AMF can determine the service-oriented interface related to calling the MMF model deployment service based on the container type of the model deployment container carried in the NAS message.

[0268] In some implementations, the model deployment request carries the terminal device ID, MMF ID, and model deployment container; see the above for details.

[0269] In step S1226, the MMF sends a model deployment response to the AMF to indicate the initial model corresponding to the model training task.

[0270] In some implementations, the model deployment response can carry the terminal device ID, MMF ID, model identifier of the initial model, and the initial model. The model identifier of the initial model and the initial model are carried in the model deployment container, which helps to avoid AMF parsing the information in the model deployment container. In other words, AMF can pass the model deployment container through to MMF to improve the security of transmitting the model identifier of the initial model and the initial model.

[0271] In step S1228, the AMF sends the fifth message to the terminal device to indicate the initial model corresponding to the model training task.

[0272] In some implementations, the fifth information can be carried in the NAS message. The fifth information carries the terminal device ID, MMF ID, model identifier of the initial model, and the initial model. The model identifier of the initial model and the initial model are carried in the model deployment container, which helps to avoid AMF parsing the information in the model deployment container. That is to say, AMF can pass the model deployment container through to the terminal device to improve the security of transmitting the model identifier of the initial model and the initial model.

[0273] In step S1230, NF sends a model deployment request to MMF to request the initial model corresponding to the model training task.

[0274] In some implementations, the NF determines whether to train the model based on the model training information of the model to be trained. If the model training information of the model to be trained indicates that model training should be based on the initial model, the NF sends a model deployment request to the MMF.

[0275] In some implementations, the model deployment request carries the NF ID, MMF ID, and the model identifier of the initial model.

[0276] In step S1232, MMF sends a model deployment response to NF to indicate the initial model corresponding to the model training task.

[0277] In some implementations, the model deployment response may carry the NF ID, MMF ID, model identifier of the initial model, and the initial model.

[0278] In step S1234, the terminal device performs local data collection and local model training based on the model information of the model to be trained.

[0279] In step S1236, the terminal device sends the intermediate results of local training to NF.

[0280] In step S1238, NF performs model training based on the intermediate results.

[0281] In some implementations, NF collects local data and aggregates intermediate results sent by terminal devices according to the model training strategy configuration in order to continue local model training.

[0282] In step S1240, NF sends the trained results to the terminal device.

[0283] It should be noted that in the embodiments of this application, steps S1234 to S1240 can be iterated multiple times according to the model training requirements.

[0284] In step S1242, NF sends a model training response to MMF to indicate that model training is complete.

[0285] In some implementations, NF determines that model training is complete based on the model training completion time and / or model training accuracy indicated by the model information of the model to be trained.

[0286] In step S1244, the MMF sends a model training response to the AMF to indicate that model training is complete.

[0287] In some implementations, the model training response includes the terminal device ID, MMF ID, model identifier (i.e., the model identifier of the model to be trained), and information indicating that model training is complete. The model identifier and the information indicating that model training is complete can be carried in a container, which helps to improve the security of transmitting the model identifier and the information indicating that model training is complete.

[0288] In step S1246, the AMF sends a first instruction message to the terminal device, indicating that the terminal device has completed model training.

[0289] In some implementations, the first indication information is carried in a NAS message. The first indication information includes the terminal device ID, MMF ID, model identifier, and information indicating that the model training is complete. The model identifier and the information indicating that the model training is complete can be carried in a container, which helps to improve the security of transmitting the model identifier and the information indicating that the model training is complete.

[0290] 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 to 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.

[0291] Figure 13 is a schematic diagram of a terminal device according to an embodiment of this application. The terminal device 1300 shown in Figure 13 includes a communication unit 1310.

[0292] The communication unit 1310 is used to receive or send first information, the first information being used to request the first network element to indicate the model information supported by the terminal device, or the first information being used by the first network element to configure a model training task for the terminal device, the first network element being used to manage the model.

[0293] In some implementations, the first information is used to request the first network element to indicate the model information supported by the terminal device, and the first information carries one or more of the following: the identifier of the terminal device; the identifier of the first network element; and the model information supported by the terminal device.

[0294] In some implementations, the model information supported by the terminal device is used to indicate one or more of the following: models supported for deployment by the terminal device; models allocated by the terminal device; model functions supported by the terminal device; model structures supported by the terminal device; the region where the terminal device runs models; model interoperability of the models supported by the terminal device; model size supported by the terminal device; and model complexity supported by the terminal device.

[0295] In some implementations, the model information supported by the terminal device is carried in a container.

[0296] In some implementations, the communication unit is further configured to send the first information to the first network element.

[0297] In some implementations, the first information is sent from the terminal device to the first network element via the second network element.

[0298] In some implementations, the communication unit is configured to receive second information sent by the first network element, the second information being used to indicate to the first network element whether the model information supported by the terminal device has been completed.

[0299] In some implementations, the second information carries one or more of the following: the identifier of the terminal device; the identifier of the first network element; and information used to indicate to the first network element whether the model information supported by the terminal device is complete.

[0300] In some implementations, the information used to indicate to the first network element whether the model information supported by the terminal device is complete is carried in a container.

[0301] In some implementations, the second information is received by the terminal device from the first network element through the second network element.

[0302] In some implementations, the first information is used to configure the model training task, and the first information is used to indicate one or more of the following: the model to be trained corresponding to the model training task; the model functionality of the model to be trained; the model structure of the model to be trained; the model operating region of the model to be trained; the model interoperability of the model to be trained; the model size of the model to be trained; the model complexity of the model to be trained; the input of the model to be trained; the output of the model to be trained; the training data information of the model to be trained; the model parameter configuration information of the model to be trained; the model training time of the model to be trained; whether the model training task needs to be performed based on an initial model; and the nodes participating in the model training task.

[0303] In some implementations, the communication unit is used to receive the first information sent by the first network element.

[0304] In some implementations, the first information is received by the terminal device from the first network element through the second network element.

[0305] In some implementations, the communication unit is used to send third information to the first network element, the third information being used to request the first network element to execute the model training task.

[0306] In some implementations, the third information carries one or more of the following: the identifier of the terminal device; the identifier of the first network element; and the model information of the model to be trained.

[0307] In some implementations, the model information of the model to be trained includes one or more of the following: the model functionality of the model to be trained; the training completion time of the model to be trained; the model training accuracy of the model to be trained; the model input format of the model to be trained; and the model output format of the model to be trained.

[0308] In some implementations, the model information of the model to be trained is carried in a container.

[0309] In some implementations, the third information is received by the terminal device from the first network element through the second network element.

[0310] In some implementations, the communication unit is used to send fourth information to the first network element, the fourth information being used to request an initial model corresponding to the model training task from the first network element; and / or the communication unit is used to receive fifth information sent by the first network element, the fifth information being used to indicate the initial model corresponding to the model training task.

[0311] In some implementations, the fourth information is sent by the terminal device to the first network element through the second network element, and / or the fifth information is received by the terminal device from the first network element through the second network element.

[0312] In some implementations, the communication unit is used to receive first indication information sent by the first network element, the first indication information being used to indicate that the model training task is completed.

[0313] In some implementations, the first indication information is received by the terminal device from the first network element through the second network element.

[0314] In some implementations, the first network element is used for one or more of the following: managing model information indicated to the first network element by other network elements; authorizing other network elements to use the model; configuring model training tasks for other network elements; and determining model strategies.

[0315] Figure 14 is a schematic diagram of a network device according to an embodiment of the application. The network device 1400 shown in Figure 14 is a first network element, and the network device 1400 includes: a communication unit 1410.

[0316] The communication unit 1410 is used to receive or send first information, the first network element is used to manage the model, wherein the first information is used to request the first network element to indicate the model information supported by the terminal device, or the first information is used to configure the model training task of the model to be trained.

[0317] In some implementations, the first information is used to request the first network element to indicate the model information supported by the terminal device, and the first information carries one or more of the following: the identifier of the terminal device; the identifier of the first network element; and the model information supported by the terminal device.

[0318] In some implementations, the model information supported by the terminal device is used to indicate one or more of the following: models supported for deployment by the terminal device; models allocated by the terminal device; model functions supported by the terminal device; model structures supported by the terminal device; the region where the terminal device runs models; model interoperability of the models supported by the terminal device; model size supported by the terminal device; and model complexity supported by the terminal device.

[0319] In some implementations, the model information supported by the terminal device is carried in a container.

[0320] In some implementations, the communication unit is used to receive the first information sent by the terminal device.

[0321] In some implementations, the first information is received by the first network element from the terminal device through the second network element.

[0322] In some implementations, the communication unit is used to send second information to the terminal device, the second information being used to indicate to the first network element whether the model information supported by the terminal device has been completed.

[0323] In some implementations, the second information carries one or more of the following: the identifier of the terminal device; the identifier of the first network element; and information used to indicate to the first network element whether the model information supported by the terminal device is complete.

[0324] In some implementations, the information used to indicate to the first network element whether the model information supported by the terminal device is complete is carried in a container.

[0325] In some implementations, the second information is sent from the first network element to the terminal device via the second network element.

[0326] In some implementations, the first information is used to configure the model training task of the model to be trained. The first information is used to indicate one or more of the following: the model to be trained; the model functionality of the model to be trained; the model structure of the model to be trained; the model operating region of the model to be trained; the model interoperability of the model to be trained; the model size of the model to be trained; the model complexity of the model to be trained; the input of the model to be trained; the output of the model to be trained; the training data information of the model to be trained; the model parameter configuration information of the model to be trained; the model training time of the model to be trained; whether the model training task needs to be performed based on the initial model; and the nodes participating in the model training task.

[0327] In some implementations, the communication unit is used to send the first information to the terminal device or NF.

[0328] In some implementations, the first information is sent from the first network element to the terminal device via a second network element.

[0329] In some implementations, the communication unit is used to receive third information sent by the terminal device, the third information being used to instruct the terminal device to request the execution of the model training task.

[0330] In some implementations, the third information carries one or more of the following: the identifier of the terminal device; the identifier of the first network element; and the model information of the model to be trained.

[0331] In some implementations, the model information of the model to be trained includes one or more of the following: the model functionality of the model to be trained; the training completion time of the model to be trained; the model training accuracy of the model to be trained; the model input format of the model to be trained; and the model output format of the model to be trained.

[0332] In some implementations, the model information of the model to be trained is carried in a container.

[0333] In some implementations, the third information is received by the first network element from the terminal device through the second network element.

[0334] In some implementations, the communication unit is configured to receive fourth information sent by the terminal device, the fourth information being used to request an initial model corresponding to the model training task; and / or the communication unit is configured to send fifth information to the terminal device, the fifth information being used to indicate the initial model corresponding to the model training task.

[0335] In some implementations, the communication unit is configured to receive first indication information sent by NF, the first indication information being used to indicate that the model training task is completed; and / or the communication unit is configured to send the first indication information to the terminal device.

[0336] In some implementations, the first instruction information is sent by the first network element to the terminal device through the second network element.

[0337] In some implementations, the first network element is used for one or more of the following: managing model information indicated to the first network element by other network elements; authorizing other network elements to use the model; configuring model training tasks for other network elements; and determining model strategies.

[0338] Figure 15 is a schematic diagram of a network device according to an embodiment of this application. The network device 1500 shown in Figure 15 is a second network element, and the network device 1500 includes: a communication unit 1510.

[0339] The communication unit 1510 is used to receive and / or send first information, the first information being used to request the first network element to indicate the model information supported by the terminal device, or the first information being used by the first network element to configure a model training task for the terminal device, the first network element being used to manage the model.

[0340] In some implementations, the first information is used to request the first network element to indicate the model information supported by the terminal device, and the first information carries one or more of the following: the identifier of the terminal device; the identifier of the first network element; and the model information supported by the terminal device.

[0341] In some implementations, the model information supported by the terminal device is used to indicate one or more of the following: models supported for deployment by the terminal device; models allocated by the terminal device; model functions supported by the terminal device; model structures supported by the terminal device; the region where the terminal device runs models; model interoperability of the models supported by the terminal device; model size supported by the terminal device; and model complexity supported by the terminal device.

[0342] In some implementations, the model information supported by the terminal device is carried in a container.

[0343] In some implementations, the communication device further includes: a first processing unit, configured to determine, based on the container type of the container, a first service interface to be invoked on the first network element, wherein the first service interface is configured to indicate to the first network element the model information supported by the terminal device.

[0344] In some implementations, the communication unit is configured to: receive the first information sent by the terminal device; and send the first information to the first network element.

[0345] In some implementations, the communication unit is configured to: receive second information sent by the first network element, the second information being used to indicate to the first network element whether the model information supported by the terminal device has been completed; and send the second information to the terminal device.

[0346] In some implementations, the second information carries one or more of the following: the identifier of the terminal device; the identifier of the first network element; and information used to indicate to the first network element whether the model information supported by the terminal device is complete.

[0347] In some implementations, the information used to indicate to the first network element whether the model information supported by the terminal device is complete is carried in a container.

[0348] In some implementations, the first information is used to configure the model training task, and the first information is used to indicate one or more of the following: the model to be trained corresponding to the model training task; the model functionality of the model to be trained; the model structure of the model to be trained; the model operating region of the model to be trained; the model interoperability of the model to be trained; the model size of the model to be trained; the model complexity of the model to be trained; the input of the model to be trained; the output of the model to be trained; the training data information of the model to be trained; the model parameter configuration information of the model to be trained; the model training time of the model to be trained; whether the model training task needs to be performed based on an initial model; and the nodes participating in the model training task.

[0349] In some implementations, the communication unit is configured to: receive the first information sent by the first network element; and send the first information to the terminal device.

[0350] In some implementations, the communication unit is configured to: receive third information sent by the terminal device, the third information being used to request the first network element to execute the model training task; and send the third information to the terminal device.

[0351] In some implementations, the third information carries one or more of the following: the identifier of the terminal device; the identifier of the first network element; and the model information of the model to be trained.

[0352] In some implementations, the model information of the model to be trained includes one or more of the following: the model functionality of the model to be trained; the training completion time of the model to be trained; the model training accuracy of the model to be trained; the model input format of the model to be trained; and the model output format of the model to be trained.

[0353] In some implementations, the model information of the model to be trained is carried in a container.

[0354] In some implementations, the communication device further includes: a second processing unit, configured to determine, based on the container type of the container, a second service interface for invoking the first network element, the second service interface being used to invoke the first network element to provide services related to the model training task.

[0355] In some implementations, the communication unit is configured to: receive fourth information sent by the terminal device, the fourth information being used to request an initial model corresponding to the model training task from the first network element; and send the fourth information to the first network element.

[0356] In some implementations, the communication unit is configured to: receive fifth information sent by the first network element, the fifth information being used to indicate the initial model corresponding to the model training task; and send the fifth information to the terminal device.

[0357] In some implementations, the communication unit is configured to: receive first indication information sent by the first network element, the first indication information being used to indicate that the model training task is completed; and send the first indication information to the terminal device.

[0358] In some implementations, the first network element is used for one or more of the following: managing model information indicated to the first network element by other network elements; authorizing other network elements to use the model; configuring model training tasks for other network elements; and determining model strategies.

[0359] In an optional embodiment, the communication unit 1310 may be a transceiver 1630. The terminal device 1300 may also include a processor 1610 and a memory 1620, as shown in FIG16.

[0360] In an optional embodiment, the communication unit 1410 may be a transceiver 1630. The network device 1400 may also include a transceiver 1630 and a memory 1620, as shown in FIG16.

[0361] In an optional embodiment, the communication unit 1510 may be a transceiver 1630. The network device 1500 may also include a transceiver 1630 and a memory 1620, as shown in FIG16.

[0362] Figure 16 is a schematic structural diagram of a communication device according to an embodiment of this application. The dashed lines in Figure 16 indicate that the unit or module is optional. This device 1600 can be used to implement the methods described in the above method embodiments. Device 1600 can be a chip, a terminal device, or a network device.

[0363] 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, 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.

[0364] The apparatus 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.

[0365] 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.

[0366] This application also provides a computer-readable storage medium for storing a program. This computer-readable storage medium can be applied to a terminal or network device provided in this application, and the program causes a computer to execute the methods performed by the terminal or network device in various embodiments of this application.

[0367] This application also provides a computer program product. The computer program product includes a program. The computer program product can be applied to a terminal or network device provided in this application embodiment, and the program causes a computer to execute the methods performed by the terminal or network device in various embodiments of this application.

[0368] This application also provides a computer program. This computer program can be applied to the terminal or network device provided in this application, and the computer program causes the computer to execute the methods performed by the terminal or network device in various embodiments of this application.

[0369] It should be understood that the terms "system" and "network" in this application can be used interchangeably. Furthermore, the terminology used in this application is only for explaining specific embodiments of the application and is not intended to limit the application. The terms "first," "second," "third," and "fourth," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. In addition, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion.

[0370] In the embodiments of this application, the term "instruction" can be a direct instruction, an indirect instruction, or an indication of a relationship. For example, A instructing B can mean that A directly instructs B, such as B being able to obtain information through A; it can also mean that A indirectly instructs B, such as A instructing C, so B can obtain information through C; or it can mean that there is a relationship between A and B.

[0371] In the embodiments of this application, "B corresponding to A" means that B is associated with A, and B can be determined based on A. However, it should also be understood that determining B based on A does not mean that B is determined solely based on A; B can also be determined based on A and / or other information.

[0372] In the embodiments of this application, the term "correspondence" can indicate a direct or indirect correspondence between two things, or an association between two things, or a relationship such as instruction and being instructed, configuration and being configured.

[0373] In this application embodiment, "predefined" or "preconfigured" can be implemented by pre-storing corresponding codes, tables, or other means that can be used to indicate relevant information in the device (e.g., including terminal devices and network devices). This application does not limit the specific implementation method. For example, predefined can refer to what is defined in the protocol.

[0374] In this application embodiment, the "protocol" may refer to a standard protocol in the field of communication, such as the LTE protocol, the NR protocol, and related protocols applied to future communication systems. This application does not limit this.

[0375] In the embodiments of this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0376] In the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0377] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of 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.

[0378] 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.

[0379] In addition, the functional units in the various embodiments of this application 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.

[0380] 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 instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can read or a data storage device such as a server or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., digital video discs, DVDs) or semiconductor media (e.g., solid-state disks, SSDs), etc.

[0381] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for wireless communication, characterized in that, include: The terminal device receives or sends first information, which is used to request the first network element to indicate the model information supported by the terminal device, or the first information is used by the first network element to configure a model training task for the terminal device, and the first network element is used to manage the model.

2. The method as described in claim 1, characterized in that, The first information is used to request the first network element to indicate the model information supported by the terminal device, and the first information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; The model information supported by the terminal device.

3. The method as described in claim 2, characterized in that, The model information supported by the terminal device is used to indicate one or more of the following: The terminal device supports the deployment of the following models; The terminal device has been assigned a model; The terminal device supports the following model functions; The model structure supported by the terminal device; The region of the terminal device operating model; Model interoperability of the models supported by the terminal device; The terminal device supports the following model sizes; The model complexity supported by the terminal device.

4. The method as described in claim 2 or 3, characterized in that, The model information supported by the terminal device is carried in a container.

5. The method according to any one of claims 2-4, characterized in that, The terminal device receives or sends first information, including: The terminal device sends the first information to the first network element.

6. The method as described in claim 5, characterized in that, The first information is sent from the terminal device to the first network element via the second network element.

7. The method according to any one of claims 2-6, characterized in that, The method further includes: The terminal device receives second information sent by the first network element, the second information being used to indicate to the first network element whether the model information supported by the terminal device has been completed.

8. The method as described in claim 7, characterized in that, The second information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; Information used to indicate to the first network element whether the model information supported by the terminal device is complete.

9. The method as described in claim 7 or 8, characterized in that, The information used to indicate to the first network element whether the model information supported by the terminal device is complete is carried in a container.

10. The method according to any one of claims 7-9, characterized in that, The second information is received by the terminal device from the first network element through the second network element.

11. The method as described in claim 1, characterized in that, The first information is used to configure the model training task, and the first information is used to indicate one or more of the following: The model to be trained corresponding to the model training task; The model functionality of the model to be trained; The model structure of the model to be trained; The model operating region of the model to be trained; Model interoperability of the models to be trained; The size of the model to be trained; The model complexity of the model to be trained; The input to the model to be trained; The output of the model to be trained; The training data information of the model to be trained; The model parameter configuration information of the model to be trained; The training time of the model to be trained; Is it necessary to perform the model training task based on the initial model? The nodes that participate in the model training task.

12. The method as described in claim 11, characterized in that, The terminal device receives or sends first information, including: The terminal device receives the first information sent by the first network element.

13. The method as described in claim 12, characterized in that, The first information is received by the terminal device from the first network element through the second network element.

14. The method as described in claim 12 or 13, characterized in that, The method further includes: The terminal device sends third information to the first network element, the third information being used to request the first network element to execute the model training task.

15. The method as described in claim 14, characterized in that, The third information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; The model information of the model to be trained.

16. The method as described in claim 15, characterized in that, The model information of the model to be trained includes one or more of the following: The model functionality of the model to be trained; The training completion time of the model to be trained; The training accuracy of the model to be trained; The model input format of the model to be trained; The output format of the model to be trained.

17. The method as described in claim 15 or 16, characterized in that, The model information of the model to be trained is carried in a container.

18. The method according to any one of claims 14-17, characterized in that, The third piece of information is received by the terminal device from the first network element through the second network element.

19. The method according to any one of claims 11-18, characterized in that, The method further includes: The terminal device sends fourth information to the first network element, the fourth information being used to request the initial model corresponding to the model training task from the first network element; and / or The terminal device receives the fifth information sent by the first network element, the fifth information being used to indicate the initial model corresponding to the model training task.

20. The method as described in claim 19, characterized in that, The fourth information is sent by the terminal device to the first network element through the second network element, and / or The fifth piece of information is received by the terminal device from the first network element through the second network element.

21. The method according to any one of claims 11-20, characterized in that, The method further includes: The terminal device receives a first indication message sent by the first network element, the first indication message being used to indicate that the model training task is completed.

22. The method as described in claim 21, characterized in that, The first indication information is received by the terminal device from the first network element through the second network element.

23. The method according to any one of claims 1-22, characterized in that, The first network element is used for one or more of the following: Manage the model information indicated by other network elements to the first network element; Authorize other network elements to use the model; Configure model training tasks for other network elements; Determine the model strategy.

24. A method for wireless communication, characterized in that, include: The first network element receives or sends first information, and the first network element is used to manage the model. The first information is used to request the first network element to indicate the model information supported by the terminal device, or the first information is used to configure the model training task of the model to be trained.

25. The method as described in claim 24, characterized in that, The first information is used to request the first network element to indicate the model information supported by the terminal device, and the first information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; The model information supported by the terminal device.

26. The method as described in claim 25, characterized in that, The model information supported by the terminal device is used to indicate one or more of the following: The terminal device supports the deployment of the following models; The terminal device has been assigned a model; The terminal device supports the following model functions; The model structure supported by the terminal device; The region of the terminal device operating model; Model interoperability of the models supported by the terminal device; The terminal device supports the following model sizes; The model complexity supported by the terminal device.

27. The method as described in claim 25 or 26, characterized in that, The model information supported by the terminal device is carried in a container.

28. The method according to any one of claims 24-27, characterized in that, The first network element receives or sends first information, including: The first network element receives the first information sent by the terminal device.

29. The method as described in claim 28, characterized in that, The first information is received by the first network element from the terminal device through the second network element.

30. The method as described in claim 28 or 29, characterized in that, The method further includes: The first network element sends second information to the terminal device, the second information being used to indicate to the first network element whether the model information supported by the terminal device has been completed.

31. The method as described in claim 30, characterized in that, The second information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; Information used to indicate to the first network element whether the model information supported by the terminal device is complete.

32. The method as described in claim 30 or 31, characterized in that, The information used to indicate to the first network element whether the model information supported by the terminal device is complete is carried in a container.

33. The method according to any one of claims 30-32, characterized in that, The second information is sent from the first network element to the terminal device through the second network element.

34. The method as described in claim 24, characterized in that, The first information is used to configure the model training task of the model to be trained, and the first information is used to indicate one or more of the following: The model to be trained; The model functionality of the model to be trained; The model structure of the model to be trained; The model operating region of the model to be trained; Model interoperability of the models to be trained; The size of the model to be trained; The model complexity of the model to be trained; The input to the model to be trained; The output of the model to be trained; The training data information of the model to be trained; The model parameter configuration information of the model to be trained; The training time of the model to be trained; Is it necessary to perform the model training task based on the initial model? The nodes that participate in the model training task.

35. The method as described in claim 34, characterized in that, The first network element receives or sends first information, including: The first network element sends the first information to the terminal device or NF.

36. The method as described in claim 35, characterized in that, The first information is sent from the first network element to the terminal device through the second network element.

37. The method according to any one of claims 34-36, characterized in that, The method further includes: The first network element receives third information sent by the terminal device, the third information being used to instruct the terminal device to request the execution of the model training task.

38. The method as described in claim 37, characterized in that, The third information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; The model information of the model to be trained.

39. The method as described in claim 38, characterized in that, The model information of the model to be trained includes one or more of the following: The model functionality of the model to be trained; The training completion time of the model to be trained; The training accuracy of the model to be trained; The model input format of the model to be trained; The output format of the model to be trained.

40. The method as described in claim 38 or 39, characterized in that, The model information of the model to be trained is carried in a container.

41. The method according to any one of claims 37-40, characterized in that, The third information is received by the first network element from the terminal device through the second network element.

42. The method according to any one of claims 34-41, characterized in that, The method further includes: The first network element receives fourth information sent by the terminal device, the fourth information being used to request the initial model corresponding to the model training task; and / or The first network element sends a fifth piece of information to the terminal device, the fifth piece of information being used to indicate the initial model corresponding to the model training task.

43. The method according to any one of claims 34-42, characterized in that, The method further includes: The first network element receives a first indication message sent by the NF, the first indication message being used to indicate that the model training task is complete; and / or The first network element sends the first instruction information to the terminal device.

44. The method as described in claim 43, characterized in that, The first instruction information is sent by the first network element to the terminal device through the second network element.

45. The method according to any one of claims 24-44, characterized in that, The first network element is used for one or more of the following: Manage the model information indicated by other network elements to the first network element; Authorize other network elements to use the model; Configure model training tasks for other network elements; Determine the model strategy.

46. ​​A method for wireless communication, characterized in that, include: The second network element receives and / or sends first information, which is used to request the first network element to indicate the model information supported by the terminal device, or the first information is used by the first network element to configure model training tasks for the terminal device, and the first network element is used to manage the model.

47. The method as described in claim 46, characterized in that, The first information is used to request the first network element to indicate the model information supported by the terminal device, and the first information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; The model information supported by the terminal device.

48. The method as described in claim 47, characterized in that, The model information supported by the terminal device is used to indicate one or more of the following: The terminal device supports the deployment of the following models; The terminal device has been assigned a model; The terminal device supports the following model functions; The model structure supported by the terminal device; The region of the terminal device operating model; Model interoperability of the models supported by the terminal device; The terminal device supports the following model sizes; The model complexity supported by the terminal device.

49. The method as described in claim 47 or 48, characterized in that, The model information supported by the terminal device is carried in a container.

50. The method as described in claim 49, characterized in that, The method further includes: The second network element determines to invoke the first service interface of the first network element based on the container type of the container. The first service interface is used to indicate to the first network element the model information supported by the terminal device.

51. The method according to any one of claims 46-50, characterized in that, The second network element receives and / or sends the first information, including: The second network element receives the first information sent by the terminal device; The second network element sends the first information to the first network element.

52. The method according to any one of claims 47-51, characterized in that, The method further includes: The second network element receives the second information sent by the first network element, the second information being used to indicate to the first network element whether the model information supported by the terminal device has been completed; The second network element sends the second information to the terminal device.

53. The method as described in claim 52, characterized in that, The second information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; Information used to indicate to the first network element whether the model information supported by the terminal device is complete.

54. The method as described in claim 53, characterized in that, The information used to indicate to the first network element whether the model information supported by the terminal device is complete is carried in a container.

55. The method as described in claim 46, characterized in that, The first information is used to configure the model training task, and the first information is used to indicate one or more of the following: The model to be trained corresponding to the model training task; The model functionality of the model to be trained; The model structure of the model to be trained; The model operating region of the model to be trained; Model interoperability of the models to be trained; The size of the model to be trained; The model complexity of the model to be trained; The input to the model to be trained; The output of the model to be trained; The training data information of the model to be trained; The model parameter configuration information of the model to be trained; The training time of the model to be trained; Is it necessary to perform the model training task based on the initial model? The nodes that participate in the model training task.

56. The method as described in claim 55, characterized in that, The second network element receives and / or sends the first information, including: The second network element receives the first information sent by the first network element; The second network element sends the first information to the terminal device.

57. The method as described in claim 55 or 56, characterized in that, The method further includes: The second network element receives third information sent by the terminal device, the third information being used to request the first network element to execute the model training task; The second network element sends the third information to the terminal device.

58. The method as described in claim 57, characterized in that, The third information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; The model information of the model to be trained.

59. The method as described in claim 58, characterized in that, The model information of the model to be trained includes one or more of the following: The model functionality of the model to be trained; The training completion time of the model to be trained; The training accuracy of the model to be trained; The model input format of the model to be trained; The output format of the model to be trained.

60. The method as described in claim 58 or 59, characterized in that, The model information of the model to be trained is carried in a container.

61. The method as described in claim 60, characterized in that, The method further includes: The second network element determines the second service interface of the first network element to call based on the container type of the container. The second service interface is used to call the first network element to provide services related to the model training task.

62. The method according to any one of claims 53-61, characterized in that, The method further includes: The second network element receives the fourth information sent by the terminal device, the fourth information being used to request the initial model corresponding to the model training task from the first network element; The second network element sends the fourth information to the first network element.

63. The method according to any one of claims 55-62, characterized in that, The method further includes: The second network element receives the fifth information sent by the first network element, the fifth information being used to indicate the initial model corresponding to the model training task; The second network element sends the fifth information to the terminal device.

64. The method according to any one of claims 55-63, characterized in that, The method further includes: The second network element receives the first indication information sent by the first network element, the first indication information being used to indicate that the model training task is completed; The second network element sends the first instruction information to the terminal device.

65. The method according to any one of claims 46-64, characterized in that, The first network element is used for one or more of the following: Manage the model information indicated by other network elements to the first network element; Authorize other network elements to use the model; Configure model training tasks for other network elements; Determine the model strategy.

66. A terminal device, characterized in that, include: A communication unit is used to receive or send first information, the first information being used to request the first network element to indicate the model information supported by the terminal device, or the first information being used by the first network element to configure a model training task for the terminal device, the first network element being used to manage the model.

67. The terminal device as described in claim 66, characterized in that, The first information is used to request the first network element to indicate the model information supported by the terminal device, and the first information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; The model information supported by the terminal device.

68. The terminal device as described in claim 67, characterized in that, The model information supported by the terminal device is used to indicate one or more of the following: The terminal device supports the deployment of the following models; The terminal device has been assigned a model; The terminal device supports the following model functions; The model structure supported by the terminal device; The region of the terminal device operating model; Model interoperability of the models supported by the terminal device; The terminal device supports the following model sizes; The model complexity supported by the terminal device.

69. The terminal device as described in claim 67 or 68, characterized in that, The model information supported by the terminal device is carried in a container.

70. The terminal device as described in any one of claims 67-69, characterized in that, The communication unit is also used to send the first information to the first network element.

71. The terminal device as described in claim 70, characterized in that, The first information is sent from the terminal device to the first network element via the second network element.

72. The terminal device as described in any one of claims 67-71, characterized in that, The communication unit is configured to receive second information sent by the first network element, the second information being used to indicate to the first network element whether the model information supported by the terminal device has been completed.

73. The terminal device as described in claim 72, characterized in that, The second information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; Information used to indicate to the first network element whether the model information supported by the terminal device is complete.

74. The terminal device as described in claim 72 or 73, characterized in that, The information used to indicate to the first network element whether the model information supported by the terminal device is complete is carried in a container.

75. The terminal device as described in any one of claims 72-74, characterized in that, The second information is received by the terminal device from the first network element through the second network element.

76. The terminal device as described in claim 66, characterized in that, The first information is used to configure the model training task, and the first information is used to indicate one or more of the following: The model to be trained corresponding to the model training task; The model functionality of the model to be trained; The model structure of the model to be trained; The model operating region of the model to be trained; Model interoperability of the models to be trained; The size of the model to be trained; The model complexity of the model to be trained; The input to the model to be trained; The output of the model to be trained; The training data information of the model to be trained; The model parameter configuration information of the model to be trained; The training time of the model to be trained; Is it necessary to perform the model training task based on the initial model? The nodes that participate in the model training task.

77. The terminal device as described in claim 76, characterized in that, The communication unit is used to receive the first information sent by the first network element.

78. The terminal device as described in claim 77, characterized in that, The first information is received by the terminal device from the first network element through the second network element.

79. The terminal device as described in claim 77 or 78, characterized in that, The communication unit is used to send third information to the first network element, the third information being used to request the first network element to execute the model training task.

80. The terminal device as described in claim 79, characterized in that, The third information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; The model information of the model to be trained.

81. The terminal device as described in claim 80, characterized in that, The model information of the model to be trained includes one or more of the following: The model functionality of the model to be trained; The training completion time of the model to be trained; The training accuracy of the model to be trained; The model input format of the model to be trained; The output format of the model to be trained.

82. The terminal device as described in claim 80 or 81, characterized in that, The model information of the model to be trained is carried in a container.

83. The terminal device as described in any one of claims 79-82, characterized in that, The third piece of information is received by the terminal device from the first network element through the second network element.

84. The terminal device as described in any one of claims 76-83, characterized in that, The communication unit is used to send fourth information to the first network element, and the fourth information is used to request the initial model corresponding to the model training task from the first network element. and / or The communication unit is used to receive the fifth information sent by the first network element, the fifth information being used to indicate the initial model corresponding to the model training task.

85. The terminal device as described in claim 84, characterized in that, The fourth information is sent by the terminal device to the first network element through the second network element, and / or The fifth piece of information is received by the terminal device from the first network element through the second network element.

86. The terminal device as described in any one of claims 76-85, characterized in that, The communication unit is used to receive first indication information sent by the first network element, the first indication information being used to indicate that the model training task is completed.

87. The terminal device as described in claim 86, characterized in that, The first indication information is received by the terminal device from the first network element through the second network element.

88. The terminal device as described in any one of claims 66-87, characterized in that, The first network element is used for one or more of the following: Manage the model information indicated by other network elements to the first network element; Authorize other network elements to use the model; Configure model training tasks for other network elements; Determine the model strategy.

89. A network device, characterized in that, The network device is the first network element, including: A communication unit is used to receive or send first information, and the first network element is used to manage the model. The first information is used to request the first network element to indicate the model information supported by the terminal device, or the first information is used to configure the model training task of the model to be trained.

90. The network device as described in claim 89, characterized in that, The first information is used to request the first network element to indicate the model information supported by the terminal device, and the first information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; The model information supported by the terminal device.

91. The network device as described in claim 90, characterized in that, The model information supported by the terminal device is used to indicate one or more of the following: The terminal device supports the deployment of the following models; The terminal device has been assigned a model; The terminal device supports the following model functions; The model structure supported by the terminal device; The region of the terminal device operating model; Model interoperability of the models supported by the terminal device; The terminal device supports the following model sizes; The model complexity supported by the terminal device.

92. The network device as described in claim 90 or 91, characterized in that, The model information supported by the terminal device is carried in a container.

93. The network device as described in any one of claims 89-92, characterized in that, The communication unit is used to receive the first information sent by the terminal device.

94. The network device as described in claim 93, characterized in that, The first information is received by the first network element from the terminal device through the second network element.

95. The network device as described in claim 93 or 94, characterized in that, The communication unit is used to send second information to the terminal device, the second information being used to indicate to the first network element whether the model information supported by the terminal device has been completed.

96. The network device as described in claim 95, characterized in that, The second information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; Information used to indicate to the first network element whether the model information supported by the terminal device is complete.

97. The network device as described in claim 95 or 96, characterized in that, The information used to indicate to the first network element whether the model information supported by the terminal device is complete is carried in a container.

98. The network device as described in any one of claims 95-97, characterized in that, The second information is sent from the first network element to the terminal device through the second network element.

99. The network device as described in claim 89, characterized in that, The first information is used to configure the model training task of the model to be trained, and the first information is used to indicate one or more of the following: The model to be trained; The model functionality of the model to be trained; The model structure of the model to be trained; The model operating region of the model to be trained; Model interoperability of the models to be trained; The size of the model to be trained; The model complexity of the model to be trained; The input to the model to be trained; The output of the model to be trained; The training data information of the model to be trained; The model parameter configuration information of the model to be trained; The training time of the model to be trained; Is it necessary to perform the model training task based on the initial model? The nodes that participate in the model training task.

100. The network device as described in claim 99, characterized in that, The communication unit is used to send the first information to the terminal device or NF.

101. The network device as described in claim 100, characterized in that, The first information is sent from the first network element to the terminal device through the second network element.

102. The network device as described in any one of claims 99-101, characterized in that, The communication unit is used to receive third information sent by the terminal device, the third information being used to instruct the terminal device to request the execution of the model training task.

103. The network device as described in claim 102, characterized in that, The third information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; The model information of the model to be trained.

104. The network device as described in claim 103, characterized in that, The model information of the model to be trained includes one or more of the following: The model functionality of the model to be trained; The training completion time of the model to be trained; The training accuracy of the model to be trained; The model input format of the model to be trained; The output format of the model to be trained.

105. The network device as described in claim 103 or 104, characterized in that, The model information of the model to be trained is carried in a container.

106. The network device as described in any one of claims 102-105, characterized in that, The third information is received by the first network element from the terminal device through the second network element.

107. The network device as described in any one of claims 99-106, characterized in that, The communication unit is configured to receive fourth information sent by the terminal device, the fourth information being used to request an initial model corresponding to the model training task; and / or The communication unit is used to send fifth information to the terminal device, the fifth information being used to indicate the initial model corresponding to the model training task.

108. The network device as described in any one of claims 99-107, characterized in that, The communication unit is configured to receive first indication information sent by NF, the first indication information being used to indicate that the model training task is complete; and / or The communication unit is used to send the first indication information to the terminal device.

109. The network device as described in claim 108, characterized in that, The first instruction information is sent by the first network element to the terminal device through the second network element.

110. The network device as described in any one of claims 89-109, characterized in that, The first network element is used for one or more of the following: Manage the model information indicated by other network elements to the first network element; Authorize other network elements to use the model; Configure model training tasks for other network elements; Determine the model strategy.

111. A network device, characterized in that, The network device is a second network element, including: A communication unit is used to receive and / or send first information, the first information being used to request the first network element to indicate the model information supported by the terminal device, or the first information being used by the first network element to configure a model training task for the terminal device, the first network element being used to manage the model.

112. The network device as described in claim 111, characterized in that, The first information is used to request the first network element to indicate the model information supported by the terminal device, and the first information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; The model information supported by the terminal device.

113. The network device as described in claim 112, characterized in that, The model information supported by the terminal device is used to indicate one or more of the following: The terminal device supports the deployment of the following models; The terminal device has been assigned a model; The terminal device supports the following model functions; The model structure supported by the terminal device; The region of the terminal device operating model; Model interoperability of the models supported by the terminal device; The terminal device supports the following model sizes; The model complexity supported by the terminal device.

114. The network device as described in claim 112 or 113, characterized in that, The model information supported by the terminal device is carried in a container.

115. The network device as described in claim 114, characterized in that, The network device also includes: The first processing unit is configured to determine, based on the container type of the container, to invoke a first service interface of the first network element, wherein the first service interface is configured to indicate to the first network element the model information supported by the terminal device.

116. The network device as described in any one of claims 111-115, characterized in that, The communication unit is used for: Receive the first information sent by the terminal device; Send the first information to the first network element.

117. The network device as described in any one of claims 112-116, characterized in that, The communication unit is used for: The system receives second information sent by the first network element, wherein the second information is used to indicate to the first network element whether the model information supported by the terminal device has been completed. The second information is sent to the terminal device.

118. The network device as described in claim 117, characterized in that, The second information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; Information used to indicate to the first network element whether the model information supported by the terminal device is complete.

119. The network device as described in claim 118, characterized in that, The information used to indicate to the first network element whether the model information supported by the terminal device is complete is carried in a container.

120. The network device as claimed in claim 111, characterized in that, The first information is used to configure the model training task, and the first information is used to indicate one or more of the following: The model to be trained corresponding to the model training task; The model functionality of the model to be trained; The model structure of the model to be trained; The model operating region of the model to be trained; Model interoperability of the models to be trained; The size of the model to be trained; The model complexity of the model to be trained; The input to the model to be trained; The output of the model to be trained; The training data information of the model to be trained; The model parameter configuration information of the model to be trained; The training time of the model to be trained; Is it necessary to perform the model training task based on the initial model? The nodes that participate in the model training task.

121. The network device as claimed in claim 120, characterized in that, The communication unit is used for: Receive the first information sent by the first network element; The first information is sent to the terminal device.

122. The network device as described in claim 120 or 121, characterized in that, The communication unit is used for: The third information sent by the terminal device is received, and the third information is used to request the first network element to execute the model training task; The third information is sent to the terminal device.

123. The network device as described in claim 122, characterized in that, The third information carries one or more of the following: The identifier of the terminal device; The identifier of the first network element; The model information of the model to be trained.

124. The network device as described in claim 123, characterized in that, The model information of the model to be trained includes one or more of the following: The model functionality of the model to be trained; The training completion time of the model to be trained; The training accuracy of the model to be trained; The model input format of the model to be trained; The output format of the model to be trained.

125. The network device as described in claim 123 or 124, characterized in that, The model information of the model to be trained is carried in a container.

126. The network device as described in claim 125, characterized in that, The network device also includes: The second processing unit is used to determine, based on the container type of the container, a second service interface to be invoked from the first network element. The second service interface is used to invoke the first network element to provide services related to the model training task.

127. The network device as described in any one of claims 118-126, characterized in that, The communication unit is used for: The terminal device sends a fourth message, which is used to request the initial model corresponding to the model training task from the first network element. The fourth information is sent to the first network element.

128. The network device as described in any one of claims 118-127, characterized in that, The communication unit is used for: Receive the fifth information sent by the first network element, wherein the fifth information is used to indicate the initial model corresponding to the model training task; The fifth message is sent to the terminal device.

129. The network device as described in any one of claims 118-128, characterized in that, The communication unit is used for: Receive first indication information sent by the first network element, the first indication information being used to indicate that the model training task is completed; The first instruction information is sent to the terminal device.

130. The network device as described in any one of claims 111-129, characterized in that, The first network element is used for one or more of the following: Manage the model information indicated by other network elements to the first network element; Authorize other network elements to use the model; Configure model training tasks for other network elements; Determine the model strategy.

131. A terminal device, characterized in that, The device includes a transceiver, a memory, and a processor. The memory stores a program, and the processor invokes the program in the memory and controls the transceiver to receive or send signals so that the terminal device performs the method as described in any one of claims 1-23.

132. A network device, characterized in that, The device includes a transceiver, a memory, and a processor. The memory stores a program, and the processor invokes the program in the memory and controls the transceiver to receive or transmit signals so that the network device performs the method as described in any one of claims 24-65.

133. An apparatus, characterized in that, Includes a processor for calling a program from memory to cause the device to perform the method as described in any one of claims 1-65.

134. A chip, characterized in that, Includes a processor for calling a program from memory, causing a device on which the chip is mounted to perform the method as described in any one of claims 1-65.

135. A computer-readable storage medium, characterized in that, It contains a program that causes a computer to perform the method as described in any one of claims 1-65.

136. A computer program product, characterized in that, Includes a program that causes a computer to perform the method as described in any one of claims 1-65.

137. A computer program, characterized in that, The computer program causes the computer to perform the method as described in any one of claims 1-65.

Images