METHOD OF COMMUNICATION, COMMUNICATION DEVICE, AND COMMUNICATION SYSTEM

Abstract

The present application provides a communication method, a communication apparatus, and a communication system. The method comprises: a network device acquiring model subscription information of a terminal device and configuration restriction information of a first model, wherein the model subscription information is used for indicating models allowed to be configured for the terminal device, the configuration restriction information is used for indicating terminal devices allowing for the configuration of the first model, and the first model is an AI model; and according to the model subscription information and the configuration restriction information, determining whether the configuration of the first model for the terminal device is allowed. The solution introduces configuration restriction information of a model, such that the privacy requirements of model configuration or model use can be satisfied, and the model security is improved.

Description

1 / 47 METHOD OF COMMUNICATION, COMMUNICATION DEVICE, AND COMMUNICATION SYSTEM TECHNICAL FIELD

[001] This application relates to the field of wireless communication technologies, and in particular, to a communication method, a communication apparatus, and a communication system. BACKGROUND

[002] Artificial intelligence (AI) is a technology for performing complex calculations by simulating the human brain. With improvements in data storage and computing capabilities, AI has found increasingly diverse applications. The 3rd generation partnership project (3GPP) proposed applying AI to wireless communication systems, including terminal devices, network devices, and other devices. In the life cycle management (LCM) of AI functions, data can be intelligently collected and analyzed to train and / or optimize AI models, improve the accuracy of inference and the computational speed of AI models, and enhance network performance and user experience.

[003] Before deploying AI models, lifecycle management may additionally include a model registration procedure. After model training, a resulting model is registered with a model management network element for subsequent management of model operations. These model operations may include at least one of the following: model transmission, model deployment, model configuration, model activation / deactivation, or other model operations.

[004] Considering privacy issues related to the model, it is necessary to implement restriction mechanisms for model settings after model registration. However, there is currently no corresponding solution. Petition 870250085540, dated 09 / 22 / 2025, page 8 / 59 2 / 47 SUMMARY

[005] This application provides a method of communication, a communication device, and a communication system for protecting the privacy of models.

[006] According to a first aspect, an embodiment of this application provides a method of communication. The method may be implemented by a network device or an apparatus used in the network device, for example, a chip. The method includes: obtaining model signature information from a terminal device and configuration restriction information from a first model, wherein the model signature information indicates a model that is permitted to be configured for the terminal device, the configuration restriction information indicates a terminal device that is permitted to be configured with the first model, and the first model is an AI model; and determining, based on the model signature information and the configuration restriction information, whether the terminal device is permitted to be configured with the first model.

[007] In the previous solution, template configuration restriction information is introduced. This can satisfy the privacy requirements of template configuration or usage, and help increase template security.

[008] In one possible implementation method, determining, based on model signature information and configuration restriction information, whether the terminal device is allowed to be configured with the first model includes: when the terminal device indicated by the configuration restriction information includes the terminal device, and the model indicated by the model signature information includes the first model, determine that the terminal device is allowed to be configured with the first model.

[009] In one possible implementation method, the method additionally includes: receiving indication information from the Petition 870250085540, dated 09 / 22 / 2025, page 9 / 59 3 / 47 terminal device, where the indication information indicates that the terminal device was configured with the first model and / or a state of the first model, and the state is an active state or an inactive state; or the indication information indicates that the terminal device was not configured with the first model.

[010] In the previous solution, the terminal device notifies the network device about a terminal device deployment case on the first model. This helps the network device to correctly manage the terminal device deployment on the first model.

[011] In one possible implementation method, when the terminal device is allowed to be configured with the first model and the terminal device has not been configured with the first model, the method additionally includes: sending model information from the first model to the terminal device.

[012] In one possible implementation method, model information includes model identification information and model management information. Model identification information includes identification information of the first model and / or identification information of a model function of the first model; and model management information includes model execution information of the first model and / or model resource information of the first model.

[013] In one possible implementation method, when the terminal device is allowed to be configured with the first model, and the terminal device was configured with the first model, but did not activate the first model, the method additionally includes: sending an activation indication to the terminal device, where the activation indication indicates to activate the first model.

[014] In one possible implementation method, the configuration constraint information includes at least one of the following Petition 870250085540, dated 09 / 22 / 2025, page 10 / 59 4 / 47 information: information about at least one terminal device, information about at least one terminal device manufacturer, or information about at least one chip manufacturer.

[015] According to a second aspect, one embodiment of this request provides a method of communication. The method can be implemented by a model management network element or an apparatus used in the model management network element, for example, a chip. The method includes: receiving a request message from a terminal device, where the request message is to request the registration of a first model with the model management network element, and the first model is an AI model; and sending configuration restriction information of the first model to a network device, where the configuration restriction information indicates a terminal device that is allowed to be configured with the first model.

[016] In the previous solution, template configuration restriction information is introduced. This can satisfy the privacy requirements of template configuration or usage, and help increase template security.

[017] In one possible implementation method, the method additionally includes: determining the configuration constraint information.

[018] In one possible implementation method, the request message includes indication information, and the indication information indicates that the first model has a usage range; and determining the configuration restriction information includes: determining the configuration restriction information based on the indication information.

[019] In the previous solution, the terminal device sends, to the model management network element, the indication information indicating that the first model has the usage range, Petition 870250085540, dated 09 / 22 / 2025, page 11 / 59 5 / 47 to trigger the model management network element to determine the configuration restriction information of the first model. This helps protect the privacy of the first model.

[020] In one possible implementation method, the method additionally includes: generating, for the first model, model identification information for the first model, where the model identification information identifies the first model.

[021] In the previous solution, the first model is identified based on the model identification information, which helps to manage the first model accurately.

[022] In one possible implementation method, the method additionally includes: sending the model identification information to the network device.

[023] In one possible implementation method, model identification information includes identification information of the first model and / or identification information of a model function of the first model.

[024] In one possible implementation method, the request message includes model management information for the first model, and the model management information includes model execution information for the first model and / or model resource information for the first model.

[025] In one possible implementation method, configuration constraint information includes at least one of the following: information about at least one terminal device, information about at least one terminal device manufacturer, or information about at least one chip manufacturer.

[026] According to a third aspect, an embodiment of this application provides a method of communication. The method may be implemented by a terminal device or an apparatus used in the terminal device, for example, a chip. The method includes: Petition 870250085540, dated 09 / 22 / 2025, page 12 / 59 6 / 47 send a request message to a model management network element, where the request message is to request the registration of a first model with the model management network element, and the first model is an AI model; and receive model identification information of the first model from the model management network element, where the model identification information identifies the first model.

[027] In the previous solution, the terminal device can register the first model with the model management network element, and this helps the model management network element to better maintain information related to the first model.

[028] In one possible implementation method, the method additionally includes: establishing an association between the model identification information and the model management information of the first model.

[029] In one possible implementation method, the request message includes model management information.

[030] In one possible implementation method, model management information includes model execution information from the first model and / or model resource information from the first model.

[031] In one possible implementation method, model identification information includes identification information of the first model and / or identification information of a model function of the first model.

[032] In one possible implementation method, the request message includes indication information, and the indication information indicates that the first model has a usage range.

[033] In the previous solution, the terminal device sends, to the model management network element, the indication information indicating that the first model has the usage range, Petition 870250085540, dated 09 / 22 / 2025, page 13 / 59 7 / 47 to trigger the model management network element to determine the configuration restriction information of the first model. This helps protect the privacy of the first model.

[034] In one possible implementation method, the indication information indicates that a terminal device from the same manufacturer as the terminal device is allowed to be configured with the first model; or the indication information indicates that a corresponding terminal device from the same chip manufacturer as the terminal device is allowed to be configured with the first model.

[035] In one possible implementation method, configuration constraint information includes at least one of the following: information about at least one terminal device, information about at least one terminal device manufacturer, or information about at least one chip manufacturer.

[036] According to a fourth aspect, an embodiment of this application provides a communication apparatus. The apparatus may be a network device or it may be a module (for example, a chip) used in the network device. The apparatus has the function of implementing any implementation method in the first aspect. The functions may be implemented by hardware, or they may be implemented by hardware implementing the corresponding software. The hardware or software includes one or more modules corresponding to the function.

[037] According to a fifth aspect, an embodiment of this application provides a communication device. The device may be a model management network element or it may be a module (e.g., a chip) used in the model management network element. The device has the function of implementing any implementation method in the second aspect. The functions may be implemented by hardware, or they may be implemented Petition 870250085540, dated 09 / 22 / 2025, p. 14 / 59 8 / 47 by hardware implementing the corresponding software. Hardware or software includes one or more modules corresponding to the function.

[038] According to a sixth aspect, an embodiment of this application provides a communication device. The device may be a terminal device, or it may be a module (e.g., a chip) used in the terminal device. The device has the function of implementing any implementation method in the third aspect. The functions may be implemented by hardware, or they may be implemented by hardware implementing the corresponding software. Hardware or software includes one or more modules corresponding to the function.

[039] According to a seventh aspect, an embodiment of this application provides a communication apparatus, including a processor and a memory. The memory is configured to store computer instructions. When the apparatus is running, the processor executes the computer instructions stored in the memory, to enable the apparatus to perform any implementation method from the first aspect to the third aspect.

[040] According to an eighth aspect, an embodiment of this application provides a communication apparatus, including a unit or means configured to perform steps of any implementation method from the first aspect to the third aspect.

[041] According to a ninth aspect, an embodiment of this application provides a communication apparatus, including a processor and an interface circuit. The processor is configured to communicate with another apparatus through the interface circuit, and to perform any implementation method from the first aspect to the third aspect. There is one or more processors.

[042] According to a tenth aspect, one embodiment of this application provides a communication apparatus, including a processor coupled to a memory. The processor is configured to invoke a program stored in memory, to perform Petition 870250085540, dated 09 / 22 / 2025, p. 15 / 59 9 / 47 Any method of implementing aspects one through three. Memory may be located inside or outside the device. Additionally, there may be one or more processors.

[043] According to an eleventh aspect, an embodiment of this application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions. When the instructions are executed on a communication device, any method of implementation from the first aspect to the third aspect is realized.

[044] According to a twelfth aspect, an embodiment of this application additionally provides a computer program product. The computer program product includes a computer program or instructions. When the computer program or instructions are executed by a communication device, any method of implementation from the first aspect to the third aspect is realized.

[045] According to a thirteenth aspect, an embodiment of this application further provides a chip system, including a processor configured to perform any implementation method in the first aspect or in the third aspect.

[046] According to a fourteenth aspect, an embodiment of this application further provides a communication system, including a network device configured to perform any implementation method in the first aspect and a model management network element configured to perform any implementation method in the second aspect.

[047] According to a fifteenth aspect, one embodiment of this request additionally provides a method of communication, including: A model management network element receives a request message from a terminal device, where the request message is to request the registration of a Petition 870250085540, dated 09 / 22 / 2025, page 16 / 59 10 / 47 first model with the model management network element, and the first model is an AI model; the model management network element sends configuration constraint information from the first model to a network device, where the configuration constraint information indicates an endpoint device that is allowed to be configured with the first model; the network device receives the configuration constraint information and obtains model subscription information from the endpoint device, where the model subscription information indicates a model that is allowed to be configured for the endpoint device; and the network device determines, based on the model subscription information and the configuration constraint information, whether the endpoint device is allowed to be configured with the first model. BRIEF DESCRIPTION OF THE DRAWINGS

[048] FIG. 1(a) is a diagram of a communication system architecture to which a modality of this request is applied; FIG. 1(b) is a diagram of a network device; FIG. 1(c) is a diagram of a communication system according to one embodiment of this application; FIG. 2(a) is a diagram of a neuron structure; FIG. 2(b) is a diagram of a neural network layer relationship; FIG. 3 is a schematic flowchart of a communication method according to one modality of this request; FIG. 4 is a schematic flowchart of a communication method according to one modality of this request; FIG. 5 is a diagram of the structure of a communication device according to one embodiment of this application; and FIG. 6 is a diagram of the structure of a communication device according to one embodiment of this application. Petition 870250085540, dated 09 / 22 / 2025, p. 17 / 59 11 / 47 DESCRIPTION OF THE MODALITIES

[049] FIG. 1(a) is a diagram of an architecture of a communication system to which an embodiment of this application is applied. The communication system 1000 shown in FIG. 1(a) includes an access network 100 and a core network 200. For example, the communication system 1000 additionally includes an internet 300. The access network 100 may include at least one access network device (for example, 110a and 110b in FIG. 1(a)) and may additionally include at least one terminal device (for example, 120a to 120j in FIG. 1(a)). The terminal device is connected to the access network device wirelessly, and the access network device is connected to the core network wirelessly or wired.A core network device and an access network device may be independent and distinct physical devices, or functions of the core network device and logical functions of the access network device may be integrated into the same physical device, or a portion of the functions of the core network device and a portion of the functions of the access network device may be integrated into a single physical device. Mutual connections between terminal devices and between access network devices may be implemented in a wired or wireless manner. FIG. 1(a) is only a diagram. The communication system may additionally include other access network devices, for example, it may additionally include a wireless relay device and a wireless backhaul device, which are not shown in FIG. 1(a).

[050] The access network device is an access device accessed by a terminal device in a communication system in a wired or wireless manner. The access network device can be a base station and includes, for example, an evolved NodeB (eNodeB), a transmission reception point (TRP), a Petition 870250085540, dated 09 / 22 / 2025, p. 18 / 59 12 / 47 A next-generation NodeB (gNB) in a 5th-generation (5G) mobile communication system, a next-generation base station in a 6th-generation (6G) mobile communication system, a base station in a future mobile communication system, and an access node in a wireless fidelity (Wi-Fi) system; or it can be a module or unit that completes part of the base station's functions, for example, it can be a central unit (CU) or a distributed unit (DU). The CU here performs functions of a radio resource control protocol layer and a packet data convergence protocol (PDCP) layer of the base station, and may additionally perform functions of a service data adaptation protocol (SDAP) layer.The DU completes the functions of a radio link control layer and a medium access control (MAC) layer of the base station, and may additionally complete some or all of the functions of a physical layer. For specific descriptions of the preceding protocol layers, refer to the related technical specifications in 3GPP. The access network device may be a macro base station (e.g., 110a in FIG. 1(a)), or it may be a micro base station or an indoor base station (e.g., 110b in FIG. 1(a)), or it may be a relay node, a donor node, or similar. A specific technology and a specific device form factor used by the access network device are not limited in the embodiments of this application.

[051] The terminal device is a device with wireless transceiver function, and can send a signal to the access network device or receive a signal from the access network device. The terminal device includes, but is not limited to, a terminal device, a terminal, equipment of Petition 870250085540, dated 09 / 22 / 2025, p. 19 / 59 13 / 47 user (user equipment, UE), a mobile station, a mobile terminal, and the like. The terminal device can be widely used in various scenarios such as device-to-device (D2D) communication, vehicle-to-everything (V2X) communication, machine-type communication (MTC), Internet of Things (IoT), virtual reality, augmented reality, industrial control, autonomous driving, telemedicine, smart grid, smart furniture, smart office, smart wearables, smart transportation, and smart city. The terminal device can specifically be a mobile phone, a tablet, a computer with wireless transceiver function, a wearable device, a vehicle, an airplane, a ship, a robot, a robotic arm, a smart home device, or similar.A specific technology and a specific device form used by the terminal device are not limited to embodiments of this application.

[052] The access network device and the terminal device may be in fixed positions or may be mobile. The access network device and the terminal device may be deployed on land, including an indoor or outdoor device, a portable device, or a vehicle-mounted device; they may be deployed on a water surface; or they may be deployed on an airplane, a balloon, and an artificial satellite in the air. The application scenarios of the access network device and the terminal device are not limited to the embodiments of this application.

[053] The functions of the access network device and the terminal device may be relative. For example, a helicopter or an unmanned aerial vehicle 120i in FIG. 1(a) may be configured as a mobile access network device and, for a terminal device 120j, accessing the access network. Petition 870250085540, dated 09 / 22 / 2025, page 20 / 59 14 / 47 From 100 to 120i, terminal device 120i is an access network device. However, for a first access network device 110a, 120i is a terminal device. In other words, communication between 110a and 120i is performed via a radio air interface protocol. Certainly, communication between 110a and 120i can alternatively be performed via an interface protocol between access network devices. In this case, for 110a, 120i is also an access network device. Therefore, both the access network device and the terminal device can be collectively called communication devices, 110a and 110b each in FIG. 1(a) can be called communication devices with the function of an access network device, and 120a to 120j each in FIG. 1(a) can be called communication devices with the function of a terminal device.

[054] Communication between the access network device and the terminal device, between access network devices, or between terminal devices may be carried out in a licensed spectrum, or may be carried out in an unlicensed spectrum, or may be carried out in a licensed spectrum and in an unlicensed spectrum. Communication may be carried out in a spectrum below 6 gigahertz (gigahertz, GHz), or may be carried out in a spectrum above 6 GHz, or may be carried out in a spectrum below 6 GHz and a spectrum above 6 GHz. A spectrum resource used for wireless communication is not limited in embodiments of this application.

[055] In embodiments of this application, a function of the access network device may alternatively be performed by a module (e.g., a chip) in the access network device, or it may be performed by a control subsystem including the function of the access network device. The control subsystem including the function of the access network device herein may be a control center in the preceding application scenarios, Petition 870250085540, dated 09 / 22 / 2025, page 21 / 59 15 / 47 such as a smart grid, industrial control, smart transportation, or a smart city. A terminal device function may alternatively be performed by a module (e.g., a chip or a modem) in the terminal device, or it may be performed by an appliance that includes the terminal device function.

[056] In this request, the access network device sends a downlink signal or downlink information to the terminal device, where the downlink information is carried on a downlink channel. The terminal device sends an uplink signal or uplink information to the access network device, where the uplink information is carried on an uplink channel. To communicate with the access network device, the terminal device needs to establish a wireless connection with a cell controlled by the access network device. The cell that establishes the wireless connection with the terminal device is called the terminal device's service cell.

[057] The 200 core network includes one or more core network devices. The core network device includes, but is not limited to, one or more of the following network elements: an authentication server function (AUSF), unified data management (UDM), a unified data repository (UDR), a network repository function (NRF), a network exposure function (NEF), an application function (AF), a policy control function (PCF), an access and mobility management function (AMF), a session management function (SMF), a user plane function. Petition 870250085540, dated 09 / 22 / 2025, page 22 / 59 Function 16 / 47, UPF), a linking support function (linking support function, BSF) and a network data analysis function (network data analysis function, NWDAF).

[058] FIG. 1(b) is a diagram of an access network device. As shown in FIG. 1(b), the access network device includes one or more CUs, one or more DUs, and one or more radio units (RUs). For clarity, FIG. 1(b) shows only one CU, one DU, and one RU. The CU is configured to connect to a core network and to one or more DUs. For example, a CU may have some of the functions of the core network. The CU may include a control plane (CP) and a user plane (UP).

[059] The CU and DU can be configured based on the protocol layer functions of a wireless network implemented by the CU and DU. For example, the CU is configured to implement functions of a packet data convergence protocol layer (PDCP) and functions of a protocol layer above the packet data convergence protocol layer (e.g., a radio resource control (RRC) layer and / or a service data adaptation protocol (SDAP) layer); and the DU is configured to implement functions of a protocol layer below the PDCP layer (e.g., a radio link control (RLC) layer, a medium access control (MAC) layer, and / or a physical (PHY) layer).In another example, the CU is configured to implement functions of a protocol layer above a PDCP layer (for example, an RRC layer and / or an SDAP layer); and the DU is configured to implement functions of a PDCP layer and functions of a protocol layer below the PDCP layer (for example, an RLC layer, a MAC layer, and / or a PHY layer).

[060] The previous CU and DU configurations are merely examples, and the functions of the CU or DU may also be Petition 870250085540, dated 09 / 22 / 2025, page 23 / 59 17 / 47 configured based on a requirement. For example, the CU or DU can be configured to have functions from more protocol layers; or the CU or DU can be configured to have a portion of the processing functions of the protocol layers. For example, a portion of the RLC layer functions and functions from a protocol layer above the RLC layer are defined in the CU; and the remaining RLC layer functions and functions from a protocol layer below the RLC layer are defined in the DU. In another example, the functions of the CU or DU can be divided based on a type of service or other system requirement, for example, divided based on a delay. A function whose processing time needs to satisfy a small delay requirement is defined in the DU, and a function whose processing time does not need to satisfy the delay requirement is defined in the CU.

[061] DU and RU can work together to implement PHY layer functions. A DU can be connected to one or more RUs. The functions of the DU and RU can be configured in various ways based on a project. For example, the DU is configured to implement a baseband function, and the RU is configured to implement an intermediate radio frequency function. In another example, the DU is configured to implement an upper layer function of the PHY layer, and the RU is configured to implement a lower layer function of the PHY layer or to implement the lower layer function and a radio frequency function. The upper layer function of the physical layer may include a portion of the physical layer functions, and that portion of the functions is closer to the MAC layer. The lower layer function of the physical layer may include another portion of the physical layer functions, and that portion of the functions is closer to an intermediate radio frequency side.

[062] In different systems, the CU (or the CU-CP and CU-UP), the DU or the RU may also have different names, but a person skilled in the art can understand their meanings. For example, in a Petition 870250085540, dated 09 / 22 / 2025, page 24 / 59 In an open RAN (ORAN) system, the CU can also be called an open CU (O-CU), the DU can also be called an ODU, the CU-CP can also be called an O-CU-CP, the CU-UP can also be called an O-CU-UP, and the RU can also be called an O-RU. For ease of description, the CU, CU-CP, CU-UP, DU, and RU are used as examples for description in this application. Any of the CU (or CU-CP or CU-UP), DU, and RU in this application can be implemented using a software module, a hardware module, or a combination of a software module and a hardware module.

[063] FIG. 1(c) is a diagram of a communication system according to an embodiment of this application. The communication system includes a network device and a model management network element. The network device may be the access network device or the core network device described above. The model management network element is a network element that manages an AI model or a network element that has an AI model management function. The model management network element may be a terminal device, a network device, an operation, administration, and maintenance (OAM) device, or a third-party service device. The network device may be the access network device or the core network device described above.The third-party service device can be an application service device (over-the-top, OTT) or another service provider device that can deliver an application service. It can be understood that the model described in this application is an AI model.

[064] In some embodiments of this application, an AI model may be deployed only on a network device or on a terminal device, or it may be deployed on a network device and on a terminal device, or it may be deployed simultaneously on three or more devices. The three or more Petition 870250085540, dated 09 / 22 / 2025, p. 25 / 59 19 / 47 devices can be devices of the same type, where, for example, all devices are terminal devices or all devices are network devices; or they can be devices of different types, where, for example, an AI model is deployed simultaneously on one network device and two terminal devices. Therefore, the AI ​​model in the embodiments of this application is classified into the following types. 1. One-sided model

[065] The one-sided model is a model deployed only on a network device or on an endpoint device, and the one-sided model includes a network-sided model (NW-sided model) and a UE-sided model (UE-sided model). 2. Two-sided model

[066] The two-sided model means that a model is divided into two parts: one part is deployed on an end device and the other part is deployed on a network device. The two parts are used together to complete the model's functions. 3. Multi-sided model

[067] The multi-sided model means that a model is divided into three or more parts. Each part is deployed on a device, and different parts are deployed on different devices. A plurality of devices are used together to complete the functions of the model.

[068] The AI ​​model in implementations of this application may include several deep neural networks. A neural network is a specific form of machine learning implementation. A neural network can be used to perform a classification task and a prediction task, and can also be used to establish a conditional probability distribution between variables. Common neural networks include a deep neural network (DNN), a generative neural network (GNN), and similar networks. Based on the way Petition 870250085540, dated 09 / 22 / 2025, page 26 / 59 20 / 47 of network construction, the DNN can include a feedforward neural network (FNN), a convolutional neural network (CNN), a recurrent neural network (RNN), and the like. The GNN includes a generative adversarial network (GAN) and a variational autoencoder (VAE). The neural network is built based on a neuron. Next, we describe a neural network calculation and optimization mechanism using a DNN as an example. It can be understood that a specific implementation of the neural network is not limited in modalities of this order. In the DNN, each neuron performs a weighted sum operation on an input value of the neuron, and generates an output of a weighted sum result using an activation function. FIG. 2(a) is a diagram of the structure of a neuron. It is assumed that an input to the neuron is x = [x0,x1,...[xn], a weight corresponding to the input is w = [w0, w-^... wn], a weighted sum offset is b, and the activation function forms can be diversified. For example, the activation function is y = / (x) = max{ 0,x}, and a neuron execution output is y = f(^n=owiXi + b) = max{0,%n=0wi.Xi + b], where WiXi represents a product of Wi and xj. The DNN usually has a multilayer structure. Each layer of the DNN can include a plurality of neurons. An input layer processes a received value using the neuron and then transfers the processed value to a hidden intermediate layer. Similarly, the hidden layer transfers a calculation result to a final output layer to generate a final output of the DNN. FIG. 2(b) is a diagram of a layer relationship of a neural network. A DNN typically has one or more hidden layers, and the hidden layer usually directly affects the information extraction and function tuning capabilities.Increasing the number of hidden layers in the DNN or increasing the number of neurons in each layer can improve the ability to adjust a function to the desired result. Petition 870250085540, dated 09 / 22 / 2025, page 27 / 59 21 / 47 DNN. The parameters of each neuron include a weight, an offset, and an activation function. A set that includes neuron parameters in the DNN is called a DNN parameter (also called a neural network parameter). The neuron's weight and offset can be optimized in a training process, so that the DNN has the ability to extract a data feature and express a mapping relationship.

[069] An AI model parameter includes information related to the neural network. For example, the AI ​​model parameter may include one or more of the following contents: (1) model architecture information, which can also be referred to as information related to a neural network structure, such as a type of neural network (e.g., deep neural network or generative neural network), a number of layers in the neural network, or a number of neurons; (2) a model parameter, which can also be referred to as a form of neuron connection or network layers; and (3) a neuron parameter, which can be understood as parameters of each neuron in a neural network, such as a weight, an offset, and an activation function.

[070] To protect the privacy of the model, this application provides corresponding solutions. Privacy here means restricting the use of a model to a device within a specific range, i.e., ensuring that the model is not used by a device outside the range.

[071] It can be understood that the method provided in this application can be implemented by a network device, a model management network element, and a terminal device, or it can be implemented by a module (e.g., a chip) used in the network device, the model management network element, and the terminal device. A communication method provided in this application using the following is described: Petition 870250085540, dated 09 / 22 / 2025, page 28 / 59 22 / 47 network device, the model management network element, and the terminal device as executing entities.

[072] FIG. 3 is a schematic flowchart of a communication method according to one modality of this request. The method includes the following steps.

[073] Step 301: A terminal device sends a request message to a model management network element. Correspondingly, the model management network element receives the request message.

[074] It can be understood that the terminal device can access a network device, and the terminal device sends the request message to the model management network element through the network device. For example, the network device and the model management network element are applicable to the communication system described in FIG. 1 (c).

[075] The request message is to request the registration of a first model with the model management network element. The first model can be obtained by the terminal device through training, or it can be obtained by a model provider through training and provided to the terminal device. The first model is an AI model.

[076] The model provider can train the first model and obtain management information from the first model. The model provider can be a network device, an OAM device, another terminal device, or a third-party service device. The model provider and the model management network element are different devices. It should be noted that if a model is trained by the terminal device, the fact that the terminal device trains the model can be understood as model training performed by at least one of the terminal devices, a terminal device manufacturer, or a chip manufacturer. Petition 870250085540, dated 09 / 22 / 2025, page 29 / 59 23 / 47 terminal device.

[077] For example, the request message includes the model management information of the first model, and the model management information of the first model includes model execution information of the first model and / or model resource information of the first model. Model execution information includes at least one of the following: model architecture information (e.g., at least one neural network type, a number of neural network layers, or a number of neurons) of the first model, at least one model parameter (e.g., a way of connecting neurons or network layers) of the first model, a neuron parameter (e.g., at least one weight, an offset, or an activation function) of at least one model layer of the first model, model function information of the first model, manufacturer information (e.g.,Information about a manufacturer that supplies the first model) related to the first model, chip information (e.g., information about a chip that can run the first model) related to the first model, provider information for the first model, application scenario information for the first model, input parameter information for the first model, output parameter information for the first model, performance information for the first model, or other model execution information for the first model. Application scenarios for the first model include, but are not limited to, energy saving, load balancing, mobility optimization, channel state information-reference signal (CSI-RS) feedback enhancement, beam management enhancement (BME),Positioning improvement, and similar. Petition 870250085540, dated 09 / 22 / 2025, page 30 / 59 24 / 47 These application scenarios can also be called AI use cases. The model feature information of the first model is also called first model feature information or first model. In one example, the model feature information of the first model can be used as a container, and mainly includes a first model model architecture, at least one first model model parameter, and a neuron parameter from at least one first model model layer.

[078] For example, the request message additionally includes indication information, and the indication information indicates that the first model has a usage range. The usage range may be a terminal device that is permitted to be configured with the first model. The indication information is also called privacy requirement information. In a specific example, for instance, the indication information indicates that a terminal device from the same manufacturer as the terminal device is permitted to be configured with the first model; or the indication information indicates that a terminal device corresponding to the same chip manufacturer as the terminal device is permitted to be configured with the first model.

[079] Step 302: The model management network element determines configuration constraint information for the first model.

[080] In one implementation method, when the request message includes the indication information, the model management network element determines the configuration constraint information of the first model based on the indication information. In another implementation method, when the request message does not include the indication information, the model management network element determines the configuration constraint information of the Petition 870250085540, dated 09 / 22 / 2025, page 31 / 59 25 / 47 first model based on a protocol preset.

[081] Configuration restriction information is also called indication information, configuration information, configuration indication information, or similar terms. Configuration restriction information indicates a terminal device that is permitted to be configured with the first model. For example, configuration restriction information includes at least one of the following: information about at least one terminal device, information about at least one terminal device manufacturer, or information about at least one chip manufacturer. The at least one terminal device permitted to be configured with the first model, and the at least one terminal device includes the terminal device described in step 301.A terminal device produced by at least one terminal device manufacturer may constitute the first model, and the terminal device produced by at least one terminal device manufacturer includes the previous terminal device. A terminal device that uses a chip produced by at least one chip manufacturer may constitute the first model, and the terminal device that uses the chip produced by at least one chip manufacturer includes the previous terminal device.

[082] In the previous solution, based on a model registration mechanism, the model management network element can obtain model management information from the first model and maintain the model management information.

[083] For example, in step 302, the model management network element can additionally allocate model identification information to the first model based on the model management information of the first model, where the model identification information identifies the first model. The model identification information of the first model includes identification information (e.g., an ID). Petition 870250085540, dated 09 / 22 / 2025, page 32 / 59 26 / 47 of the model) of the first model and / or identification information (e.g., a functionality ID) of a model function of the first model. The model function includes at least one of the following functions: channel state information (CSI) feedback enhancement, beam management enhancement, positioning enhancement, or other AI function. In one example, the model identification information may additionally indicate information about a first model trainer and / or first model execution information. The first model trainer is a manufacturer or device that obtains the first model through training.Information about the first model trainer includes one or more of the following: information about a terminal device manufacturer, information about a chip manufacturer, information about a network device manufacturer, information about a carrier, or information about a third-party service device.

[084] In this embodiment of this application, the first model’s model identification information and the first model’s model management information are collectively referred to as the first model’s model information. In other words, the first model’s model information includes the model identification information and the model management information.

[085] Step 303: The model management network element sends the configuration constraint information from the first model to the network device. Correspondingly, the network device receives the configuration constraint information from the first model.

[086] For example, the model management network element can additionally send the model information from the first model to the network device. The model information includes the model identification information and the Petition 870250085540, dated 09 / 22 / 2025, page 33 / 59 27 / 47 model management information. After receiving the model information, the network device stores it.

[087] If the model management network element additionally determines the model identification information of the first model in step 302, the model management network element will additionally send the model identification information of the first model to the network device in step 303. Correspondingly, the network device can establish an association between the model identification information of the first model and the model management information of the first model, and store the model information. The model information includes the model identification information of the first model and the model management information of the first model.

[088] The network device is a network device accessed by the terminal device. It can be understood that the model management network element can alternatively send at least one of the model identification information, the model management information, and the configuration restriction information to one or more other network devices connected to the model management network element.

[089] Step 304: The network device sends the model identification information of the first model to the terminal device. Correspondingly, the terminal device receives the model identification information of the first model.

[090] For example, if the model management network element updates the model execution information and / or the model resource information in the model management information of the first model, the network device may additionally send updated model management information to the endpoint device in step 304.

[091] For example, in step 304, the network device additionally sends the configuration restriction information of Petition 870250085540, dated 09 / 22 / 2025, page 34 / 59 28 / 47 first model for the terminal device.

[092] Step 305: The terminal device establishes an association between the model identification information and the model management information.

[093] After the terminal device establishes the association between the first model's model identification information and the first model's model management information, the terminal device and the network device can subsequently manage the first model based on the first model's model identification information. For example, when the network device needs to instruct the terminal device to activate the first model, the network device can send the first model's model identification information to the terminal device, or send the first model's model identification information and an activation indication, and the terminal device determines the first model's model management information based on the first model's model identification information.The first model's model management information includes the first model's model execution information and the first model's model resource information. Then, the first model is activated based on the first model's model resource information, or the first model is activated based on both the first model's model execution information and the first model's model resource information. Similarly, the terminal device can also deactivate the first model using a similar method. Details are not described again.

[094] In one implementation, when another terminal device accesses the network device, for a single-sided model and a two-sided model, the network device can determine, based on the configuration restriction information of the first Petition 870250085540, dated 09 / 22 / 2025, page 35 / 59 29 / 47 model, the first model must be configured for the other terminal device. For example, for the EU-sided model and the two-sided model, if the other terminal device is the terminal device indicated by the configuration restriction information of the first model, the network device can configure the first model for the other terminal device, and the network device can send the model information of the first model to the other terminal device. The other terminal device deploys the first model based on the model information of the first model. For specific descriptions, refer to the descriptions in the mode corresponding to FIG. 4.

[095] In one implementation, for a NW-side model, the network device can determine, based on the configuration constraint information of the first model, whether to activate the first model when communicating with another terminal device. For example, if the other terminal device is the terminal device indicated by the configuration constraint information of the first model, the first model may be activated when the network device communicates with the other terminal device.

[096] In the previous solution, the introduction of model configuration restriction information allows model configuration or model use on the terminal device corresponding to the configuration restriction information, thus ensuring model security and privacy.

[097] FIG. 4 is a schematic flowchart of a communication method according to one modality of this request. The method includes the following steps.

[098] Step 401: A terminal device accesses a network device.

[099] Step 402: The network device obtains model signature information from the terminal device. Petition 870250085540, dated 09 / 22 / 2025, page 36 / 59 30 / 47

[100] Model signature information indicates a model that is allowed to be configured for the terminal device. In an implementation method, model signature information includes at least one of the following: information about a terminal device manufacturer, information about a chip manufacturer, information about a model available for the terminal device, or information about a model compatible with the terminal device. For example, assuming that the model signature information includes information about the terminal device manufacturer, all models trained by the manufacturer indicated by the terminal device manufacturer information can be configured for the terminal device.

[101] The information about the model available for the terminal device can be understood as an indication of a model that the terminal device is permitted to use or a model that can be used by the terminal device within a scope of permission of the terminal device. The information about the model available for the terminal device may include identification information for at least one model. Alternatively, the information about the model available for the terminal device may indicate information about a terminal device manufacturer and / or information about a chip manufacturer. The information about the model available for the terminal device indicates information about the terminal device manufacturer. This indicates that the terminal device is permitted to use a model trained by the terminal device manufacturer.The information about the available model for the terminal device indicates information about the chip manufacturer. This indicates that the terminal device is authorized to use a model trained by the chip manufacturer.

[102] Information about the model supported by Petition 870250085540, dated 09 / 22 / 2025, page 37 / 59 31 / 47 terminal device information can be understood as an indication of a model that can be used by the terminal device within a capability range of the terminal device. Information about the model supported by the terminal device may include identifying information for at least one model that can be used by the terminal device within the capability range of the terminal device. Alternatively, information about the model supported by the terminal device may indicate information about a terminal device manufacturer and / or information about a chip manufacturer. Information about the model supported by the terminal device indicates information about the terminal device manufacturer. This indicates that the terminal device can use, within the capability range of the terminal device, a model trained by the terminal device manufacturer.The information about the model supported by the terminal device indicates information about the chip manufacturer. This indicates that the terminal device can use, within the terminal device's capability range, a model trained by the chip manufacturer.

[103] With reference to a specific example, the following describes a relationship and a difference between information about the model available for the terminal device and information about the model that the terminal device supports. Assume that terminal device manufacturers include a manufacturer 1, a manufacturer 2, and a manufacturer 3. Models supplied by manufacturer 1 include a model 1-1, a model 1-2, and a model 1-3. Models supplied by manufacturer 2 may include a model 2-1 and a model 2-2. A model supplied by manufacturer 3 may include a model 3-1. Model 1-1, model 1-2, model 2-1, and model 3-1 can be called low-capacity models, and the capacity of a terminal device required to configure these models is low. Model 1-3 and model 2-2 can be called high-capacity models, and the Petition 870250085540, dated 09 / 22 / 2025, pp. 38 / 59 32 / 47 The capability of a terminal device required to configure these models is high. In one example, the model information available for the terminal device includes information about manufacturer 1 and information about manufacturer 2. This indicates that the terminal device is permitted to use models trained by manufacturer 1 and manufacturer 2. For example, the terminal device is permitted to use at least one of the models 1-1, 1-2, 1-3, 2-1, and 2-2. It can be understood that the possibility of at least one of the previous models being configured for the terminal device is further determined with reference to the capability information of the terminal device and / or the model information that the terminal device supports.Assuming the terminal device's capacity is low, the terminal device can configure at least one of the following models: 1-1 trained by manufacturer 1, model 1-2 trained by manufacturer 1, and model 2-1 trained by manufacturer 2. In another example, still assuming the terminal device's capacity is low, it can be determined, based on a terminal's capacity, that the models that can be configured within the terminal device's capacity range include model 1-1, model 1-2, model 2-1, and model 3-1.Since the information about the model supported by the terminal device includes information about manufacturer 1 and information about manufacturer 2, indicating that the terminal device supports the use of models trained by manufacturer 1 and manufacturer 2 and that they have a low requirement on the terminal device's capacity, the terminal device can use, within the terminal device's capacity range, at least one of the models 1-1 trained by manufacturer 1, 1-2 trained by manufacturer 1, and 2-1 trained by manufacturer 2.

[104] In one implementation method, the network device can receive the template signature information sent Petition 870250085540, dated 09 / 22 / 2025, pp. 39 / 59 33 / 47 by the terminal device, or it can receive template signature information that is from the terminal device and that is sent by another network device. For example, when the network device is an access network device, the access network device can receive the terminal device template signature information from a core network device or from another access network device.

[105] Step 403: A model management network element sends model information from a first model and configuration constraint information from the first model to the network device. Correspondingly, the network device receives the model information from the first model and the configuration constraint information from the first model.

[106] For specific meanings of model information and configuration constraint information, see embodiment descriptions in FIG. 3.

[107] A specific implementation method for obtaining model identification information from the first model and configuration constraint information from the first model by the model management network element may be the same as the method in the embodiment of FIG. 3. Details are not described again.

[108] Model information and configuration constraint information can be sent in the same message or they can be sent separately in different messages.

[109] It should be noted that a sequence for executing steps 403, 401 and 402 is not limited.

[110] Step 404: The terminal device sends indication information to the network device. Correspondingly, the network device receives the indication information.

[111] The indication information indicates that the terminal device was configured with the first model and / or a first model state, and the state is an active state or a state Petition 870250085540, dated 09 / 22 / 2025, pages 40 / 59 34 / 47 inactive; or the indication information indicates that the terminal device has not been configured with the first template. The active state means that the first template has been deployed and activated, and a template deployment device (the network device and / or the terminal device) can use the first template to perform data processing or analysis. The inactive state means that the first template has been deployed but not activated, and the template deployment device cannot use the first template to perform data processing or analysis.

[112] Step 404 is an optional step.

[113] It can be understood that in step 402, if the network device receives the template signing information sent by the terminal device, in one implementation, steps 404 and 402 can be combined. For example, the terminal device can send the template signing information and the pointing information to the network device in the same message.

[114] Step 405: The network device determines, based on the template signature information and configuration restriction information, whether the terminal device is allowed to be configured with the first template.

[115] When a terminal device indicated by the configuration restriction information includes the terminal device, and the model indicated by the model signature information includes the first model, it is determined that the terminal device is allowed to be configured with the first model. When the terminal device indicated by the configuration restriction information does not include the terminal device, or the model indicated by the model signature information does not include the first model, it is determined that the terminal device is not allowed to configure the first model.

[116] In an implementation, when it is determined that the terminal device is allowed to be configured with the Petition 870250085540, dated 09 / 22 / 2025, pp. 41 / 59 35 / 47 first model, if the indication information in step 404 indicates that the terminal device was not configured with the first model, the following step 406 will be performed, or the following steps 406 and 407 will be performed; or if the indication information indicates that the terminal device was configured with the first model and the first model is in the inactive state, the following step 407 will be performed.

[117] Step 406: The network device sends the model information of the first model to the terminal device. Correspondingly, the terminal device receives the model information of the first model.

[118] The first model is configured for the terminal device based on the model information of the first model.

[119] Step 407: The network device sends a wake-up indication to the terminal device. Correspondingly, the terminal device receives the wake-up indication.

[120] The activation indicator indicates the activation of the first model.

[121] In one implementation, when the indication information in step 404 indicates that the terminal device has not been configured with the first model, and it is determined in step 405 that the terminal device is allowed to be configured with the first model, steps 406 and 407 can be combined to execute as step 406', to be specific, the network device sends the model information of the first model and the activation indication to the terminal device. In this implementation, the activation indication can indicate whether the first model should be activated.

[122] In another implementation, when the indication information in step 404 indicates that the terminal device has not been configured with the first model, and it is determined in step 405 that the terminal device is allowed to be configured with the first model, if step 406 is performed after step Petition 870250085540, dated 09 / 22 / 2025, pp. 42 / 59 36 / 47 405, but if step 407 is not performed, the first model is configured for the terminal device after the terminal device receives the model information from the first model, and the terminal device activates or deactivates the first model based on a protocol preset.

[123] It should be noted that, in one implementation, step 404 may alternatively be step 404'. Specifically, the network device may learn, in another way, whether the endpoint device deployed the first model and a state of the first model when the first model was deployed. For example, the network device learns the previous information from another network device.

[124] In the previous solution, the introduction of model configuration restriction information allows model configuration or model use on a terminal device matching the configuration restriction information, thus ensuring model security and privacy.

[125] It should be noted that, in the embodiments of FIG. 3 and FIG. 4, an example in which the model management network element and the network device are different network elements is used for description, i.e., a deployment position of a network element with a model management function is not on the network device. In another implementation method, if the model management network element and the network device are the same device, i.e., the model management function is deployed on the network device, the operations performed by the model management network element in the embodiments in FIG. 3 and FIG. 4 can be performed by the network device. For example, for the embodiment in FIG.3. The request message in step 301 can be sent to the network device, step 302 is performed by the network device, and step 303 is an operation between different functional modules on the network device, or step 303 may not be performed. Petition 870250085540, dated 09 / 22 / 2025, pp. 43 / 59 37 / 47 performed. For the mode in FIG. 4, step 403 is an operation between different functional modules in the network device, or step 403 may not be performed.

[126] In an implementation method, the embodiment in FIG. 3 and the embodiment in FIG. 4 can be combined for implementation. For example, the network device in the embodiment of FIG. 4 and the network device in the embodiment of FIG. 3 are the same network device, the model management network element in the embodiment of FIG. 4 and the model management network element in the embodiment of FIG. 3 are the same model management network element, and the terminal device in the embodiment of FIG. 4 and the terminal device in the embodiment of FIG. 3 are different terminal devices. In this case, step 403 in the embodiment of FIG. 4 and step 303 in the embodiment of FIG. 3 can be the same step.

[127] It may be understood that, in embodiments of this application, the terminal device, network device, or model management network element may perform some or all of the steps in embodiments of this application. These steps or operations are merely examples. In embodiments of this application, other operations or variations of various operations may be performed. Furthermore, the steps may be performed in a sequence different from a sequence presented in embodiments of this application, and it is possible that not all operations in embodiments of this application will be performed.

[128] It may be understood that, to implement functions in the preceding embodiments, the network device, the model management network element, or the terminal device includes corresponding hardware structures and / or software modules to perform the functions. A person skilled in the art should be aware that, in this application, units and steps of the method described in examples with reference to embodiments disclosed in this application may be implemented in a hardware form or a Petition 870250085540, dated 09 / 22 / 2025, pp. 44 / 59 38 / 47 combination of computer hardware and software. Whether a specific function is performed by hardware or hardware driven by computer software depends on specific application scenarios and design constraints of the technical solutions.

[129] FIGS. 5 and 6 are diagrams of the structures of possible communication devices according to the embodiments of this application. These communication devices can be configured to implement a function of the network device, the model management network element, or the terminal device in the previous method embodiments and, therefore, can also obtain beneficial effects from the previous method embodiments. In embodiments of this application, the communication device can be a network device, a model management network element, or a terminal device, or it can be a module (e.g., a chip) used in the network device, the model management network element, or the terminal device.

[130] The communication apparatus 500 shown in FIG. 5 includes a processing unit 510 and a transceiver unit 520. The communication apparatus 500 is configured to implement a function of the network device, the model management network element, or the terminal device in the previous method embodiments.

[131] When the communication device 500 is configured to implement the network device function in the previous modes, the processing unit 510 is configured to: obtain model signature information from a terminal device and configuration restriction information from a first model, where the model signature information indicates a model that is allowed to be configured for the terminal device, the configuration restriction information indicates a terminal device that is allowed to be configured with the first model, and the first model Petition 870250085540, dated 09 / 22 / 2025, pages 45 / 59 39 / 47 is an AI model; and determine, based on the model signature information and configuration restriction information, whether the terminal device is permitted to be configured with the first model.

[132] In one possible implementation method, processing unit 510 is configured to determine, based on model signature information and configuration restriction information, whether the terminal device is allowed to be configured with the first model, specifically including: when the terminal device indicated by the configuration restriction information includes the terminal device, and the model indicated by the model signature information includes the first model, determine that the terminal device is allowed to be configured with the first model.

[133] In one possible implementation method, the 520 transceiver unit is configured to receive indication information from the terminal device, where the indication information indicates that the terminal device has been configured with the first model and / or a state of the first model, and the state is an active state or an inactive state; or the indication information indicates that the terminal device has not been configured with the first model.

[134] In one possible implementation method, the 520 transceiver unit is additionally configured to: when the terminal device is allowed to be configured with the first model and the terminal device has not been configured with the first model, send model information of the first model to the terminal device.

[135] In one possible implementation method, model information includes model identification information and model management information. Model identification information includes identification information of the first model and / or identification information of a function. Petition 870250085540, dated 09 / 22 / 2025, pp. 46 / 59 40 / 47 of the first model; and model management information includes model execution information for the first model and / or model resource information for the first model.

[136] In one possible implementation method, the 520 transceiver unit is further configured to: when the terminal device is allowed to be configured with the first model, and the terminal device has been configured with the first model, but has not activated the first model, send an activation indication to the terminal device, where the activation indication indicates to activate the first model.

[137] In one possible implementation method, configuration constraint information includes at least one of the following: information about at least one terminal device, information about at least one terminal device manufacturer, or information about at least one chip manufacturer.

[138] When communication device 500 is configured to implement the model management network element function in the previous embodiments, transceiver unit 520 is configured to: receive a request message from a terminal device, where the request message is to request the registration of a first model with the model management network element, and the first model is an AI model; and send configuration restriction information of the first model to a network device, where the configuration restriction information indicates a terminal device that is allowed to be configured with the first model.

[139] In one possible implementation method, processing unit 510 is configured to determine the configuration constraint information.

[140] In one possible implementation method, the request message includes indication information, and the information of Petition 870250085540, dated 09 / 22 / 2025, pp. 47 / 59 41 / 47 indications indicate that the first model has a usage range; and that processing unit 510 is configured to determine configuration restriction information, including specifically: determining configuration restriction information based on indication information.

[141] In one possible implementation method, processing unit 510 is additionally configured to generate, for the first model, model identification information for the first model, where the model identification information identifies the first model.

[142] In one possible implementation method, the 520 transceiver unit is additionally configured to send model identification information to the network device.

[143] In one possible implementation method, model identification information includes identification information of the first model and / or identification information of a model function of the first model.

[144] In one possible implementation method, the request message includes model management information for the first model, and the model management information includes model execution information for the first model and / or model resource information for the first model.

[145] In one possible implementation method, configuration constraint information includes at least one of the following: information about at least one terminal device, information about at least one terminal device manufacturer, or information about at least one chip manufacturer.

[146] When the communication device 500 is configured to implement the function of the terminal device in the previous modes, the transceiver unit 520 is configured to: send a request message to a network element of Petition 870250085540, dated 09 / 22 / 2025, pp. 48 / 59 42 / 47 model management, where the request message is to request the registration of a first model with the model management network element, and the first model is an AI model; and receive model identification information of the first model from the model management network element, where the model identification information identifies the first model.

[147] In one possible implementation method, processing unit 510 is configured to establish an association between model identification information and model management information of the first model.

[148] In one possible implementation method, the request message includes model management information.

[149] In one possible implementation method, model management information includes model execution information from the first model and / or model resource information from the first model.

[150] In one possible implementation method, model identification information includes identification information of the first model and / or identification information of a model function of the first model.

[151] In one possible implementation method, the request message includes indication information, and the indication information indicates that the first model has a usage range.

[152] In one possible implementation method, the indication information indicates that a terminal device from the same manufacturer as the terminal device is allowed to be configured with the first model; or the indication information indicates that a corresponding terminal device from the same chip manufacturer as the terminal device is allowed to be configured with the first model.

[153] In one possible implementation method, the configuration constraint information includes at least one of the following Petition 870250085540, dated 09 / 22 / 2025, pp. 49 / 59 43 / 47 information: information about at least one terminal device, information about at least one terminal device manufacturer, or information about at least one chip manufacturer.

[154] For more detailed descriptions of the 510 processing unit and the 520 transceiver unit, refer directly to the related descriptions in the previous method embodiments. Details are not described again.

[155] The communication apparatus 600 shown in FIG. 6 includes a processor 610 and an interface circuit 620. The processor 610 and the interface circuit 620 are coupled to each other. It can be understood that the interface circuit 620 can be a transceiver or an input / output interface. For example, the communication apparatus 600 may additionally include a memory 630 configured to store instructions executed by the processor 610, store input data required by the processor 610 to execute the instructions, or store data generated after the processor 610 executes the instructions.

[156] When the communication device 600 is configured to implement the modes of the previous method, the processor 610 is configured to implement a function of the previous processing unit 510, and the interface circuit 620 is configured to implement a function of the previous transceiver unit 520.

[157] When the communication device is a chip used in a terminal device, the chip in the terminal device implements a function of the terminal device in the embodiments of the previous method. The chip in the terminal device receives information from another module (for example, a radio frequency module or an antenna) in the terminal device, and the information is sent by a network device to the terminal device. Alternatively, the chip in the terminal device sends information to another module (for example, a module of Petition 870250085540, dated 09 / 22 / 2025, pages 50 / 59 44 / 47 radio frequency or an antenna) in the terminal device, and the information is sent by the terminal device to a network device.

[158] When the communication device is a module used in a network device, the module in the network device implements a function of the network device in the previous method embodiments. The module in the network device receives information from another module (e.g., a radio frequency module or an antenna) in the network device, and the information is sent by a terminal device to the network device. Alternatively, the module in the network device sends information to another module (e.g., a radio frequency module or an antenna) in the network device, and the information is sent by the network device to a terminal device. The module in the network device here may be a baseband chip in the network device, or it may be a DU or other module. The DU here may be a DU in an open radio access network (O-RAN) architecture.

[159] It may be understood that the processor in the embodiments of this application may be a Central Processing Unit (CPU), or it may be another general-purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), or other programmable logic device, a transistor logic device, a hardware component, or any combination thereof. The general-purpose processor may be a microprocessor or any conventional processor.

[160] The steps of the method in modalities of this application can be implemented in a hardware manner or can be implemented by a processor executing software instructions. The Petition 870250085540, dated 09 / 22 / 2025, pp. 51 / 59 45 / 47 Software instructions may consist of corresponding software modules. Software modules may be stored in random access memory, flash memory, read-only memory, programmable read-only memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, a register, a hard disk, a removable hard disk, and compact disc read-only memory (CD-ROM), or any other form of storage medium well known in the art. For example, a storage medium is coupled to a processor so that the processor can read information from the storage medium and write information to the storage medium. Certainly, the storage medium may be a component of the processor. The processor and the storage medium may be located in an ASIC.Furthermore, the ASIC can be located in a network device, a model management network element, or a terminal device. Certainly, the processor and storage medium can alternatively exist in a network device, a model management network element, or a terminal device as discrete components.

[161] All or part of the foregoing modalities may be implemented by software, hardware, firmware, or any combination thereof. When software is used to implement modalities, all or part of the modalities may be implemented in the form of a computer program product. The computer program product includes one or more computer programs or instructions. A computer program is a group of instructions that instruct each step of an electronic computer or other message-processing device, is usually written in a programming language, and executes on a target architecture. When computer programs or instructions are loaded Petition 870250085540, dated 09 / 22 / 2025, pages 52 / 59 46 / 47 and executed on a computer, all or part of the procedures or functions in the modalities of this application are performed. The computer may be a general-purpose computer, a dedicated computer, a computer network, a network device, a terminal device, or other programmable apparatus. Computer programs or instructions may be stored on a computer-readable storage medium or may be transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, computer programs or instructions may be transmitted from one site, computer, server, or data center to another site, computer, server, or data center in a wired or wireless manner.Computer-readable storage media can be any usable media accessible by a computer or a data storage device, such as a server or data center, integrating one or more usable media. The usable media can be magnetic media, such as a floppy disk, a hard disk, or magnetic tape; or it can be optical media, for example, a digital video disc; or it can be semiconductor media, for example, a solid-state drive. Computer-readable storage media can be volatile or non-volatile storage media, or it can include two types of storage media: volatile and non-volatile storage media.

[162] The protocol described in this application may be a communication protocol or a specification, for example, a 3GPP communication protocol.

[163] In some embodiments of this application, unless otherwise specified or there is a logical conflict, the terms and / or descriptions in different embodiments are consistent and may be mutually referenced. Technical features in different embodiments may be combined based on a Petition 870250085540, dated 09 / 22 / 2025, pp. 53 / 59 47 / 47 Internal logical relationship, to form a new modality. In this request, at least one means one or more, and a plurality of means two or more. And / or describes an association relationship between associated objects and indicates that three relationships can exist. For example, A and / or B can indicate the following cases: Only A exists, A and B exist, and only B exists, where A and B can be singular or plural. In the text descriptions of this request, the / character generally indicates an OR relationship between associated objects. In a formula in this request, the / character indicates a division relationship between associated objects.

[164] It may be understood that various numbers in the modalities of this application are merely used for differentiation for ease of description, but are not used to limit the scope of the modalities of this application. Sequence numbers of previous processes do not signify execution sequences. The execution sequences of the processes must be determined based on the functions and internal logic of the processes. Petition 870250085540, dated 09 / 22 / 2025, pages 54 / 59

Claims

1 / 5 CLAIMS 1. A communication method, applied to a network device, characterized in that the method comprises: obtaining model signature information from a terminal device and configuration restriction information from a first model, wherein the model signature information indicates a model that is allowed to be configured for the terminal device, the configuration restriction information indicates a terminal device that is allowed to be configured with the first model, and the first model is an artificial intelligence model (AI); and determining, based on the model signature information and the configuration restriction information, whether the terminal device is allowed to be configured with the first model.

2. A method according to claim 1, characterized in that the determination, based on model signature information and configuration restriction information, of whether the terminal device is permitted to be configured with the first model comprises: when the terminal device indicated by the configuration restriction information comprises the terminal device, and the model indicated by the model signature information comprises the first model, determining that the terminal device is permitted to be configured with the first model.

3. Method, according to claim 1 or 2, characterized in that the method further comprises: receiving indication information from the terminal device, wherein the indication information indicates at least one of the following: that the terminal device has been configured with the first model or a state of the first model, and the state is an active state or an inactive state; or the indication information indicates that the terminal device has not been configured with the first model.

4. Method according to claim 3, characterized in that the method further comprises: when the terminal device is allowed to be configured with the first model and the terminal device has not been configured with the first model, sending model information of the first model to the terminal device.

5. A method according to claim 4, characterized in that the model information comprises model identification information and model management information; and the model identification information comprises at least one piece of first model identification information or first model function identification information; and the model management information comprises at least one piece of first model execution information or first model feature information.

6. Method according to claim 3, characterized in that the method further comprises: when the terminal device is allowed to be configured with the first model, and the terminal device has been configured with the first model but has not activated the first model, sending an activation indication to the terminal device, wherein the activation indication indicates to activate the first model.

7. A method according to any one of claims 1 to 6, characterized in that the configuration restriction information comprises at least one of the following: information about at least one terminal device, information about at least one terminal device manufacturer, or information about at least one chip manufacturer.

8. Communication method, applied to a model management network element, characterized in that the method comprises: receiving a request message from a terminal device, wherein the request message is to request the registration of a first model with the model management network element, and the first model is an AI artificial intelligence model; and sending configuration restriction information of the first model to a network device, wherein the configuration restriction information indicates a terminal device that is allowed to be configured with the first model.

9. Method according to claim 8, characterized in that the method further comprises: determining the configuration constraint information.

10. Method according to claim 9, characterized in that the request message comprises indication information, and the indication information indicates that the first model has a usage range; and determining the configuration restriction information comprises: determining the configuration restriction information based on the indication information.

11. A method according to any one of claims 8 to 10, characterized in that the method further comprises: generating, for the first model, model identification information for the first model, wherein the model identification information identifies the first model.

12. Method according to claim 11, characterized in that the method further comprises: sending the model identification information to the network device.

13. Method according to claim 11 or 12, Petition 870250068077, dated 04 / 08 / 2025, page 20 / 27 4 / 5 characterized in that the model identification information comprises at least one piece of first-model identification information or first-model function identification information.

14. A method according to any one of claims 8 to 13, characterized in that the request message comprises model management information of the first model, and the model management information comprises at least one piece of model execution information of the first model or model resource information of the first model.

15. A method according to any one of claims 8 to 14, characterized in that the configuration restriction information comprises at least one of the following: information about at least one terminal device, information about at least one terminal device manufacturer, or information about at least one chip manufacturer.

16. A communication apparatus, characterized in that it comprises a processor and an interface circuit, wherein the interface circuit is configured to receive a signal from another communication apparatus other than the communication apparatus and transmit the signal to the processor, or to send a signal from the processor to another communication apparatus other than the communication apparatus; and the processor is configured to implement the method as defined in any one of claims 1 to 7 or configured to implement the method as defined in any one of claims 8 to 15 using a logic circuit or by executing code instructions.

17. Communication apparatus, characterized in that it comprises a unit configured to implement the method as defined in any of claims 1 to 7 or configured for the method as defined in any of claims 8 to 15.

18. Computer-readable storage medium, characterized in that the storage medium stores a computer program or instructions; and when the computer program or instructions are executed by a communication device, the method as defined in any one of claims 1 to 7 is implemented, or the method as defined in any one of claims 8 to 15 is implemented.

19. Communication system, characterized in that it comprises a network device configured to perform the method as defined in any one of claims 1 to 7 and a model management network element configured to perform the method as defined in any one of claims 8 to 15. Petition 870250068077, dated 04 / 08 / 2025, pp. 22 / 27

Images