WO2024093503A1 - 一种处理模型的方法和装置 - Google Patents

一种处理模型的方法和装置 Download PDF

Info

Publication number
WO2024093503A1
WO2024093503A1 PCT/CN2023/116867 CN2023116867W WO2024093503A1 WO 2024093503 A1 WO2024093503 A1 WO 2024093503A1 CN 2023116867 W CN2023116867 W CN 2023116867W WO 2024093503 A1 WO2024093503 A1 WO 2024093503A1
Authority
WO
WIPO (PCT)
Prior art keywords
model
terminal device
information
network element
management network
Prior art date
Application number
PCT/CN2023/116867
Other languages
English (en)
French (fr)
Inventor
叶宽
曾宇
耿婷婷
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Publication of WO2024093503A1 publication Critical patent/WO2024093503A1/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/10—Interfaces, programming languages or software development kits, e.g. for simulating neural networks
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/147—Network analysis or design for predicting network behaviour
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/16—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using machine learning or artificial intelligence
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements

Definitions

  • the embodiments of the present application relate to the field of artificial intelligence, and more specifically, to a method and device for processing a model.
  • Artificial intelligence is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.
  • artificial intelligence is a branch of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence.
  • Artificial intelligence is to study the design principles and implementation methods of various intelligent machines so that machines have the functions of perception, reasoning and decision-making.
  • Research in the field of artificial intelligence includes robotics, natural language processing, computer vision, decision-making and reasoning, human-computer interaction, recommendation and search, basic AI theory, etc.
  • the present application provides a method and device for processing models, which can ensure the unification of models between terminal devices and model management network elements, so that the model management network element can achieve efficient management of the models of terminal devices, thereby ensuring the performance of the system.
  • an embodiment of the present application provides a method for processing a model, which can be executed by a model management network element or by a component of the model management network element (such as a chip or a chip system, etc.), and the present application does not limit this.
  • the method includes: the model management network element receives a first request message, the first request message includes model information of a first model, and the first request message is used to request registration of the first model.
  • the model management network element determines the first model based on the model information of the first model.
  • the model management network element sends a first response message, and the first response message is used to notify whether the registration of the first model is successful.
  • the successful registration of the model can be understood as the model management network element determining that the model supports online registration based on the model information or model of the received model.
  • the model management network element determines that the terminal device that sends the model information or model of the model has the authority to register the model (or, it is understood that the terminal device's request to register the model is in compliance with the agreement, etc.).
  • the method for processing models provided in the embodiment of the present application sends the model information of the model to be registered to the model management network element through the terminal device, so that the terminal device can request the model management network element to register the model that the terminal device does not have, thereby realizing the unification of the model between the terminal device and the model management network element.
  • the first request message also includes identification information of the terminal device, and the identification information of the terminal device is used by the management network element to identify the terminal device.
  • the identification information of the terminal device is the default credential of the terminal device.
  • the model management network element identifies the first model registered by the terminal device through the identification information of the terminal device, and determines whether the terminal device supports online registration of the first model, which can improve the reliability of the registration process.
  • the method also includes: when the first model is successfully registered, the model management network element sends model configuration information of the first model, the model configuration information of the first model includes model identification information of the first model, model information of the first model and at least one of the first models, wherein the model identification information of the first model is used to associate the first model.
  • the model identification information of the successfully registered model is sent to the terminal device through the model management network element to achieve efficient management of the model.
  • the model configuration information of the first model is carried in the first response message.
  • the method further includes: when the first model is successfully registered, the model management network element assigns model identification information of the first model to the first model.
  • the model identification information is allocated through the model management network element to achieve unified management of the model.
  • the method when the first model is successfully registered, the method further includes: the model management network element receives a second request message, the second request message includes second indication information, and the second indication information is used to indicate that the first model is updated to a second model.
  • the model management network element receives model update information of the first model.
  • the model management network element determines an update result based on the second indication information and the model update information of the first model.
  • the model management network element sends a second response message, the second response message includes the update result, and the second response message is used to notify the update result.
  • the successful update of the model can be understood as the model management network element determines that the model supports the update based on the received model update information or model of the model.
  • the model management network element determines that the terminal device that sends the model update information or model of the model has the authority to update the model (or, it is understood that the terminal device requests to update the model in accordance with the agreement of the protocol, etc.).
  • the method for processing models provided in the embodiment of the present application sends the model update information of the model to be updated to the model management network element through the terminal device, so that the terminal device can request the model management network element to update the locally updated model to the model management network element, thereby realizing the unification of the model between the terminal device and the model management network element.
  • the second request message includes identification information of the terminal device, and the identification information of the terminal device is used by the management network element to identify the terminal device.
  • the model management network element identifies the terminal device through the identification information of the terminal device and determines whether the terminal device supports online updating of the first model, which can improve the reliability of the update process.
  • the model update information of the first model includes gradient information and at least one of the second model, and model identification information of the first model.
  • the model update information of the first model is carried in the second request message.
  • the method also includes: when the update result is successful, the model management network element sends model update information of the second model, the model update information of the second model includes model identification information of the second model, and the model identification information of the second model is used to associate the second model.
  • the model update information of the successfully updated model is sent to the terminal device through the model management network element, thereby achieving efficient management of the model.
  • the method further includes: when the update result is successful, the model management network element assigns model identification information of the second model to the second model.
  • the updated model identification information is distributed through the model management network element to achieve unified management of the updated model.
  • the model update information of the second model is carried in the second response message.
  • the model management network element receives the first request message, including: the model management network element receives the first request message from a base station. Alternatively, the management network element receives the first request message from a terminal device.
  • the model management network element receives the second request message, including: the model management network element receives the second request message from the base station. Alternatively, the model management network element receives the second request message from the terminal device.
  • the model management network element receives request messages from different devices, so that the model management network element can be deployed in Different locations in the system to meet different application scenarios.
  • the method before the model management network element receives the first request message, the method further includes:
  • the model management network element sends at least one of model registration support information and model information, wherein the model registration support information is used to indicate that a terminal device receiving the model registration support information supports model registration, and the model information is used to indicate a model that the terminal device receiving the model information supports registration.
  • an embodiment of the present application provides a method for processing a model, which can be executed by a model management network element or by a component of the model management network element (such as a chip or a chip system, etc.), and the present application does not limit this.
  • the method includes: the model management network element receives a first request message, and the first request message is used to request registration of the first model.
  • the model management network element receives the first model.
  • the model management network element determines that the first model supports registration based on the first model.
  • the model management network element sends a first response message, and the first response message is used to notify whether the registration of the first model is successful.
  • the method for processing models provided in the embodiment of the present application sends the model to be registered to the model management network element through the terminal device, and can register the model that the model management network element does not have into the model management network element for unified management, thereby realizing the unification of the model between the terminal device and the model management network element.
  • the first request message also includes identification information of the terminal device, and the identification information of the terminal device is used by the management network element to identify the terminal device.
  • the first request message also includes the first model.
  • the first request message also includes first indication information, and the first indication information is used to instruct the model management network element to receive the first model.
  • the method also includes: when the first model is successfully registered, the model management network element sends model configuration information of the first model, the model configuration information of the first model includes model identification information of the first model, model information of the first model and at least one of the first models, wherein the model identification information of the first model is used to associate the first model.
  • the model configuration information of the first model is carried in the first response message.
  • the method further includes: when the first model is successfully registered, the model management network element assigns model identification information of the first model to the first model.
  • the method when the first model is successfully registered, the method further includes: the model management network element receives a second request message, the second request message includes second indication information, and the second indication information is used to indicate that the first model is updated to a second model.
  • the model management network element receives model update information of the first model.
  • the model management network element determines an update result based on the second indication information and the model update information of the first model.
  • the model management network element sends a second response message, the second response message includes the update result, and the second response message is used to notify the update result.
  • the second request message includes identification information of the terminal device, and the identification information of the terminal device is used by the management network element to identify the terminal device.
  • the model update information of the first model includes gradient information and at least one of the second model, and model identification information of the first model.
  • the model update information of the first model is carried in the second request message.
  • the method also includes: when the update result is successful, the model management network element sends model update information of the second model, the model update information of the second model includes model identification information of the second model, and the model identification information of the second model is used to associate the second model.
  • the method further includes: when the update result is successful, the model management network element assigns model identification information of the second model to the second model.
  • the model update information of the second model is carried in the second response message.
  • the model management network element receiving a first request message includes: the model management network element receiving the first request message from a base station. Or, the management network element receiving the first request message from a terminal device. A request message.
  • the model management network element receives the second request message, including: the model management network element receives the second request message from the base station. Alternatively, the model management network element receives the second request message from the terminal device.
  • the method before the model management network element receives the first request message, the method further includes:
  • the model management network element sends at least one of model registration support information and model information, wherein the model registration support information is used to indicate that a terminal device receiving the model registration support information supports model registration, and the model information is used to indicate a model that the terminal device receiving the model information supports registration.
  • an embodiment of the present application provides a method for processing a model, which can be executed by a model management network element or by a component of a model management network element (such as a chip or a chip system, etc.), and the present application does not limit this.
  • the method includes: the model management network element receives a first request message, the first request message includes model information of a first model, and the first request message is used to request registration of the first model.
  • the model management network element receives the first model.
  • the model management network element determines that the first model and the first model support registration based on the model information of the first model and the first model.
  • the model management network element sends a first response message, and the first response message is used to notify whether the registration of the first model is successful.
  • the method for processing models provided in the embodiment of the present application sends the model information and model of the model to be registered to the model management network element through the terminal device, and can register the model that the model management network element does not have into the model management network element for unified management, thereby realizing the unification of the model between the terminal device and the model management network element.
  • the first request message also includes identification information of the terminal device, and the identification information of the terminal device is used by the management network element to identify the terminal device.
  • the first request message also includes the first model.
  • the first request message also includes first indication information, and the first indication information is used to instruct the model management network element to receive the first model.
  • the method also includes: when the first model is registered successfully, the model management network element sends model configuration information of the first model, the model configuration information of the first model includes model identification information of the first model, model information of the first model and at least one of the first models, wherein the model identification information of the first model is used to associate the first model.
  • the model configuration information of the first model is carried in the first response message.
  • the method further includes: when the first model is successfully registered, the model management network element assigns model identification information of the first model to the first model.
  • the method when the first model is successfully registered, the method further includes: the model management network element receives a second request message, the second request message includes second indication information, and the second indication information is used to indicate that the first model is updated to the second model.
  • the model management network element receives model update information of the first model.
  • the model management network element determines an update result based on the second indication information and the model update information of the first model.
  • the model management network element sends a second response message, the second response message includes the update result, and the second response message is used to notify the update result.
  • the second request message includes identification information of the terminal device, and the identification information of the terminal device is used by the management network element to identify the terminal device.
  • the model update information of the first model includes gradient information and at least one of the second model, and model identification information of the first model.
  • the model update information of the first model is carried in the second request message.
  • the method also includes: when the update result is successful, the model management network element sends model update information of the second model, the model update information of the second model includes model identification information of the second model, and the model identification information of the second model is used to associate the second model.
  • the method further includes: when the update result is successful, the model management network element assigns model identification information of the second model to the second model.
  • the model update information of the second model is carried in the second response message.
  • the model management network element receives the first request message, including: the model management network element receives the first request message from a base station. Alternatively, the management network element receives the first request message from a terminal device.
  • the model management network element receives the second request message, including: the model management network element receives the second request message from the base station. Alternatively, the model management network element receives the second request message from the terminal device.
  • the method before the model management network element receives the first request message, the method further includes:
  • the model management network element sends at least one of model registration support information and model information, wherein the model registration support information is used to indicate that a terminal device receiving the model registration support information supports model registration, and the model information is used to indicate a model that the terminal device receiving the model information supports registration.
  • an embodiment of the present application provides a method for processing a model, which can be executed by a model management network element or by a component of a model management network element (such as a chip or a chip system, etc.), and the present application does not limit this.
  • the method includes: the model management network element receives a second request message, the second request message includes second indication information, and the second indication information is used to indicate that the first model is updated to a second model.
  • the model management network element receives model update information of the first model.
  • the model management network element determines an update result based on the second indication information and the model update information of the first model.
  • the model management network element sends a second response message, the second response message includes the update result, and the second response message is used to notify the update result.
  • the second request message includes identification information of the terminal device, and the identification information of the terminal device is used by the management network element to identify the terminal device.
  • the model update information of the first model includes gradient information and at least one of the second model, and model identification information of the first model.
  • the model update information of the first model is carried in the second request message.
  • the method also includes: when the update result is successful, the model management network element sends model update information of the second model, the model update information of the second model includes model identification information of the second model, and the model identification information of the second model is used to associate the second model.
  • the method further includes: when the update result is successful, the model management network element assigns model identification information of the second model to the second model.
  • the model update information of the second model is carried in the second response message.
  • the model management network element receives the second request message, including: the model management network element receives the second request message from the base station. Alternatively, the model management network element receives the second request message from the terminal device.
  • an embodiment of the present application provides a method for processing a model, which can be executed by a terminal device or by a component of a terminal device (such as a chip or a chip system, etc.), and the present application does not limit this.
  • the method includes: the terminal device sends a first request message, the first request message includes model information of a first model, and the first request message is used to request registration of the first model.
  • the terminal device receives a first response message, and the first response message is used to notify whether the registration of the first model is successful.
  • the first request message also includes identification information of the terminal device, and the identification information of the terminal device is used by the management network element to identify the terminal device.
  • the method further includes:
  • the terminal device receives model configuration information of the first model, the model configuration information of the first model includes model identification information of the first model, model information of the first model and at least one of the first models, wherein the model identification information of the first model is used to associate the first model.
  • the model configuration information of the first model is carried in the first response message.
  • the method when the first model is successfully registered, the method further includes: the terminal device sends a second request message, the second request message includes second indication information, and the second indication information is used to indicate that the first model is updated to a second model.
  • the terminal device sends model update information of the first model.
  • the terminal device receives a second response message, the second response message includes the update result, and the second response message is used to notify the update result.
  • the second request message includes identification information of the terminal device, and the identification information of the terminal device is used by the management network element to identify the terminal device.
  • the model update information of the first model includes gradient information and at least one of the second model, and model identification information of the first model.
  • the model update information of the first model is carried in the second request message.
  • the method also includes: when the update result is successful, the terminal device receives model update information of the second model, the model update information of the second model includes model identification information of the second model, and the model identification information of the second model is used to associate the second model.
  • the model update information of the second model is carried in the second response message.
  • the terminal device receiving the first response message includes: the terminal device receiving the first response message from a base station. Alternatively, the terminal device receiving the first response message from a model management network element.
  • the terminal device receiving the second response message includes: the terminal device receiving the second response message from a base station.
  • the terminal device receiving the second response message from a model management network element.
  • an embodiment of the present application provides a method for processing a model, which can be executed by a terminal device or by a component of a terminal device (such as a chip or a chip system, etc.), and the present application does not limit this.
  • the method includes: the terminal device sends a first request message, and the first request message is used to request registration of the first model.
  • the terminal device sends the first model.
  • the terminal device receives a first response message, and the first response message is used to notify whether the registration of the first model is successful.
  • the first request message also includes identification information of the terminal device, and the identification information of the terminal device is used by the management network element to identify the terminal device.
  • the first request message also includes the first model.
  • the first request message also includes first indication information, and the first indication information is used to instruct the model management network element to receive the first model.
  • the method also includes: when the first model is successfully registered, the terminal device receives model configuration information of the first model, the model configuration information of the first model includes model identification information of the first model, model information of the first model and at least one of the first models, wherein the model identification information of the first model is used to associate the first model.
  • the model configuration information of the first model is carried in the first response message.
  • the method when the first model is successfully registered, the method further includes: the terminal device sends a second request message, the second request message includes second indication information, and the second indication information is used to indicate that the first model is updated to a second model.
  • the terminal device sends model update information of the first model.
  • the terminal device receives a second response message, the second response message includes the update result, and the second response message is used to notify the update result.
  • the second request message includes identification information of the terminal device, and the identification information of the terminal device is used by the management network element to identify the terminal device.
  • the model update information of the first model includes gradient information and at least one of the second model, and model identification information of the first model.
  • the model update information of the first model is carried in the second request message.
  • the method further includes: when the update result is successful, The terminal device receives model update information of the second model, where the model update information of the second model includes model identification information of the second model, and the model identification information of the second model is used to associate with the second model.
  • the model update information of the second model is carried in the second response message.
  • the terminal device receiving the first response message includes: the terminal device receiving the first response message from a base station. Alternatively, the terminal device receiving the first response message from a model management network element.
  • the terminal device receiving the second response message includes: the terminal device receiving the second response message from a base station.
  • the terminal device receiving the second response message from a model management network element.
  • an embodiment of the present application provides a method for processing a model, which can be executed by a terminal device or by a component of a terminal device (such as a chip or a chip system, etc.), and the present application does not limit this.
  • the method includes: the terminal device sends a first request message, the first request message includes model information of a first model, and the first request message is used to request registration of the first model.
  • the terminal device sends the first model.
  • the terminal device receives a first response message, and the first response message is used to notify whether the registration of the first model is successful.
  • the first request message also includes identification information of the terminal device, and the identification information of the terminal device is used by the management network element to identify the terminal device.
  • the first request message also includes the first model.
  • the first request message also includes first indication information, and the first indication information is used to instruct the model management network element to receive the first model.
  • the method further includes:
  • the terminal device receives model configuration information of the first model, the model configuration information of the first model includes model identification information of the first model, model information of the first model and at least one of the first models, wherein the model identification information of the first model is used to associate the first model.
  • the model configuration information of the first model is carried in the first response message.
  • the method when the first model is successfully registered, the method further includes: the terminal device sends a second request message, the second request message includes second indication information, and the second indication information is used to indicate that the first model is updated to a second model.
  • the terminal device sends model update information of the first model.
  • the terminal device receives a second response message, the second response message includes the update result, and the second response message is used to notify the update result.
  • the second request message includes identification information of the terminal device, and the identification information of the terminal device is used by the management network element to identify the terminal device.
  • the model update information of the first model includes gradient information and at least one of the second model, and model identification information of the first model.
  • the model update information of the first model is carried in the second request message.
  • the method also includes: when the update result is successful, the terminal device receives model update information of the second model, the model update information of the second model includes model identification information of the second model, and the model identification information of the second model is used to associate the second model.
  • the model update information of the second model is carried in the second response message.
  • the terminal device receiving the first response message includes: the terminal device receiving the first response message from a base station. Alternatively, the terminal device receiving the first response message from a model management network element.
  • the terminal device receiving the second response message includes: the terminal device receiving the second response message from a base station.
  • the terminal device receiving the second response message from a model management network element.
  • an embodiment of the present application provides a method for processing a model, which can be executed by a terminal device or by a component of a terminal device (such as a chip or a chip system, etc.), and the present application does not limit this.
  • the method includes: the terminal device sends a second request message, the second request message includes second indication information, and the second indication information is used to indicate that the first model is updated to a second model.
  • the terminal device sends model update information of the first model.
  • the terminal device receives a second response message, the second response message includes the update result, and the second response message is used to notify the update result.
  • the second request message includes identification information of the terminal device, and the identification information of the terminal device is used by the management network element to identify the terminal device.
  • the model update information of the first model includes gradient information and at least one of the second model, and model identification information of the first model.
  • the model update information of the first model is carried in the second request message.
  • the method also includes: when the update result is successful, the terminal device receives model update information of the second model, the model update information of the second model includes model identification information of the second model, and the model identification information of the second model is used to associate the second model.
  • the model update information of the second model is carried in the second response message.
  • the terminal device receiving the second response message includes: the terminal device receiving the second response message from a base station.
  • the terminal device receiving the second response message from a model management network element.
  • an embodiment of the present application provides a device for processing a model.
  • the device is used to execute the method provided in the first aspect or the second aspect or the third aspect or the fourth aspect.
  • the device for processing the model may include units and/or modules for executing the method provided in the first aspect or any one of the above-mentioned implementations of the first aspect, such as a processing unit and an acquisition unit.
  • the device for processing the model may include units and/or modules for executing the method provided in the second aspect or any one of the above-mentioned implementations of the second aspect, such as a processing unit and an acquisition unit.
  • the device for processing the model may include units and/or modules for executing the method provided in the third aspect or any one of the above-mentioned implementations of the third aspect, such as a processing unit and an acquisition unit.
  • the device for processing the model may include units and/or modules for executing the method provided in the fourth aspect or any one of the above-mentioned implementations of the fourth aspect, such as a processing unit and an acquisition unit.
  • the device for processing the model is a model management network element.
  • the acquisition unit may be a transceiver, or an input/output interface; the processing unit may be at least one processor.
  • the transceiver may be a transceiver circuit.
  • the input/output interface may be an input/output circuit.
  • the device for processing the model is a chip, a chip system or a circuit in a model management network element.
  • the acquisition unit may be an input/output interface, an interface circuit, an output circuit, an input circuit, a pin or a related circuit on the chip, the chip system or the circuit;
  • the processing unit may be at least one processor, a processing circuit or a logic circuit.
  • an embodiment of the present application provides a device for processing a model, which is used to execute the method provided in the fifth aspect or the sixth aspect or the seventh aspect or the eighth aspect.
  • the device for processing the model may include units and/or modules for executing the method provided in the fifth aspect or any one of the above-mentioned implementations of the fifth aspect, such as a processing unit and an acquisition unit.
  • the device for processing the model may include units and/or modules for executing the method provided in the sixth aspect or any one of the above-mentioned implementations of the sixth aspect, such as a processing unit and an acquisition unit.
  • the device for processing the model may include units and/or modules for executing the method provided in the seventh aspect or any one of the above-mentioned implementations of the seventh aspect, such as a processing unit and an acquisition unit.
  • the device for processing the model may include units and/or modules for executing the method provided in the eighth aspect or any one of the above-mentioned implementations of the eighth aspect, such as a processing unit and an acquisition unit.
  • the device for processing the model is a model management network element.
  • the acquisition unit may be a transceiver, or an input/output interface; the processing unit may be at least one processor.
  • the transceiver may be a transceiver circuit.
  • the input/output interface may be an input/output circuit.
  • the device of the processing model is a chip, a chip system or a circuit in a terminal device.
  • the acquisition unit may be an input/output interface, an interface circuit, an output circuit, an input circuit, a pin or a related circuit on the chip, the chip system or the circuit;
  • the processing unit may be at least one processor, a processing circuit or a logic circuit.
  • an embodiment of the present application provides a processor for executing the methods provided in the above aspects.
  • an embodiment of the present application provides a computer-readable storage medium.
  • the computer-readable storage medium stores a program code for execution by a device, and the program code includes a method for executing any one of the first to eighth aspects above, and any one of the implementations of any one of the first to eighth aspects.
  • an embodiment of the present application provides a computer program product comprising instructions.
  • the computer program product When the computer program product is run on a computer, the computer executes any one of the first to eighth aspects above, and a method provided by any one of the implementations of any one of the first to eighth aspects.
  • an embodiment of the present application provides a chip, which includes a processor and a communication interface.
  • the processor reads instructions stored in a memory through the communication interface to execute any one of the first to eighth aspects above, and any one of the implementation methods of any one of the first to eighth aspects.
  • the chip also includes a memory, in which a computer program or instructions are stored, and the processor is used to execute the computer program or instructions stored in the memory.
  • the processor is used to execute any one of the first to eighth aspects above, and the method provided by any one of the implementation methods of any one of the first to eighth aspects.
  • an embodiment of the present application provides a system for processing models, including the device for processing models described in the ninth aspect and the device for processing models described in the tenth aspect.
  • FIG1 is a schematic diagram of an AI module applicable to an embodiment of the present application.
  • FIG. 2 is a schematic flowchart of a first model processing method 200 provided in an embodiment of the present application.
  • FIG3 is a schematic flowchart of a second model processing method 300 provided in an embodiment of the present application.
  • FIG. 4 is a schematic flowchart of a third model processing method 400 provided in an embodiment of the present application.
  • FIG. 5 is a schematic flowchart of a fourth model processing method 500 provided in an embodiment of the present application.
  • FIG. 6 is a schematic flowchart of a fifth model processing method 600 provided in an embodiment of the present application.
  • FIG. 7 is a schematic flowchart of a sixth model processing method 700 provided in an embodiment of the present application.
  • FIG8 is a schematic block diagram of a seventh model processing method 800 provided in an embodiment of the present application.
  • FIG. 9 is a schematic block diagram of a first model processing device 900 provided in an embodiment of the present application.
  • FIG. 10 is a schematic block diagram of a device 1000 for processing a second model provided in an embodiment of the present application.
  • FIG11 is a schematic diagram of the structure of a model management network element provided in an embodiment of the present application.
  • FIG. 12 is a schematic diagram of the structure of a terminal device provided in an embodiment of the present application.
  • a and/or B can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural.
  • the character “/” generally indicates that the objects associated with each other are in an "or” relationship.
  • the size of the serial number of each process below does not mean the order of execution.
  • the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application. For example, in the embodiments of the present application, "301", “401", The words “501” and the like are merely identifiers for the convenience of description and do not limit the order of execution of the steps.
  • references to "one embodiment” or “some embodiments” described in this specification mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment.
  • words such as “exemplary” or “for example” are used to indicate examples, illustrations or explanations. Any embodiment or design described as “exemplary” or “for example” in the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs.
  • the use of words such as “exemplary” or “for example” is intended to present related concepts in a specific way.
  • the terms “including”, “comprising”, “having” and their variations all mean “including but not limited to”, unless otherwise specifically emphasized.
  • "used to indicate” may include being used for direct indication and being used for indirect indication.
  • the indication information When describing that a certain indication information is used to indicate A, it may include that the indication information directly indicates A or indirectly indicates A, but it does not mean that the indication information must carry A.
  • the technical solution of the embodiment of the present application can be applied to various communication systems, such as: long term evolution (LTE) system, LTE frequency division duplex (FDD) system, LTE time division duplex (TDD), universal mobile telecommunication system (UMTS), worldwide interoperability for microwave access (Wi-MAX) communication system, fifth generation (5G) system or new radio (NR), future sixth generation (6G) system, inter-satellite communication and satellite communication and other non-terrestrial communication network (NTN) systems.
  • LTE long term evolution
  • FDD frequency division duplex
  • TDD LTE time division duplex
  • UMTS universal mobile telecommunication system
  • Wi-MAX worldwide interoperability for microwave access
  • 5G fifth generation
  • NR new radio
  • future sixth generation (6G) system inter-satellite communication and satellite communication and other non-terrestrial communication network (NTN) systems.
  • the satellite communication system includes a satellite base station and a terminal device.
  • the satellite base station provides communication services for the terminal device.
  • the satellite can refer to a drone, a hot air balloon, a low-orbit satellite, a medium-orbit satellite, a high-orbit satellite, etc. Satellite may also refer to a non-ground base station or non-ground equipment, etc.
  • the embodiments of the present application can be applied to terminal devices.
  • the terminal device can be a wireless terminal or a wired terminal.
  • the wireless terminal can be a device that provides voice and/or data connectivity to a user, a handheld device with a wireless connection function, or other processing devices connected to a wireless modem.
  • the wireless terminal can communicate with one or more core networks via a radio access network (abbreviation: RAN).
  • RAN radio access network
  • the wireless terminal can be a mobile terminal, such as a mobile phone (or "cellular" phone) and a computer with a mobile terminal, for example, a portable, pocket-sized, handheld, computer-built-in or vehicle-mounted mobile device, which exchanges language and/or data with the wireless access network.
  • a wireless terminal may also be called a system, a subscriber unit (SU), a subscriber station (SS), a mobile station (MB), a mobile, a remote station (RS), an access point (AP), a remote terminal (RT), an access terminal (AT), a user terminal (UT), a user agent (UA), a terminal device (UD), or a user equipment (UE).
  • SU subscriber unit
  • SS subscriber station
  • MB mobile station
  • RS remote station
  • AP access point
  • RT remote terminal
  • AT access terminal
  • U user agent
  • U terminal device
  • UE user equipment
  • the device for realizing the function of the terminal device may be the terminal device; or it may be a device capable of supporting the terminal device to realize the function, such as a chip system.
  • the device may be installed in the terminal device or used in combination with the terminal device.
  • the chip system may be composed of a chip, or may include a chip and other discrete devices.
  • the technical solution in the embodiment of the present application can also be applied to access network equipment.
  • the access network equipment can be a device that can access a terminal device to a wireless network.
  • the access network equipment can also be called a radio access network (RAN) node, a radio access network device, or a network device.
  • RAN radio access network
  • the access network equipment can be a base station.
  • the base station in the embodiment of the present application can broadly cover the following various names, or be replaced with the following names, such as: NodeB, evolved NodeB (eNB), base station gNB in 5G network, relay station, access point, transmitting and receiving point (TRP), transmitting point (TP), master eNodeB (MeNB), secondary eNodeB (SeNB), multi-standard radio (MSR) node, home base station, network controller, access node, wireless node, access point (AP), transmission node, transceiver node, baseband unit (BBU), remote radio unit (RRU), active antenna unit (active antenna unit A base station may be a macro base station, a micro base station, a relay node, a donor node, or the like, or a combination thereof.
  • a base station may also refer to a communication module, a modem, or a chip provided in the aforementioned device or apparatus.
  • a base station may also be a network-side device in a 6G network, a device that assumes the function of a base station in a future communication system, or the like.
  • a base station may support networks with the same or different access technologies.
  • the base station can be a centralized unit (CU) and distributed unit (DU) separated architecture.
  • RAN can be connected to the core network (for example, it can be the core network of long-term evolution (LTE) or the core network of 5G, etc.).
  • CU and DU can be understood as the division of base stations from the perspective of logical functions.
  • CU and DU can be physically separated or deployed together.
  • Multiple DUs can share one CU.
  • One DU can also be connected to multiple CUs.
  • CU and DU can be connected through an interface, such as an F1 interface.
  • CU and DU can be divided according to the protocol layer of the wireless network.
  • CU is used to perform the functions of the radio resource control (RRC) layer, the service data adaptation protocol (SDAP) layer, and the packet data convergence protocol (PDCP) layer
  • DU is used to perform the functions of the radio link control (RLC) layer, the media access control (MAC) layer, the physical layer, etc.
  • RRC radio resource control
  • SDAP service data adaptation protocol
  • PDCP packet data convergence protocol
  • DU is used to perform the functions of the radio link control
  • RLC radio link control
  • MAC media access control
  • the physical layer etc.
  • CU or DU can be divided into functions with more protocol layers.
  • CU or DU can also be divided into partial processing functions with protocol layers.
  • the functions of CU or DU can also be divided according to the service type or other system requirements. For example, according to the latency, the functions that need to meet the latency requirements are set in the DU, and the functions that do not need to meet the latency requirements are set in the CU.
  • the CU can also have one or more functions of the core network.
  • One or more CUs can be set centrally or separately.
  • the CU can be set on the network side for centralized management.
  • the DU can have multiple RF functions, or the RF function can be set remotely.
  • the functions of CU can be implemented by one entity or by different entities.
  • the functions of CU can be further divided, for example, the control plane (CP) and the user plane (UP) are separated, that is, the control plane (CU-CP) of CU and the user plane (CU-UP) of CU.
  • CU-CP and CU-UP can be implemented by different functional entities and connected through the E1 interface.
  • the CU-CP and CU-UP can be coupled with DU to jointly complete the functions of the base station.
  • the control plane CU-CP of CU also includes a further divided architecture, that is, the existing CU-CP is further divided into CU-CP1 and CU-CP2.
  • CU-CP1 includes various wireless resource management functions
  • CU-CP2 only includes RRC functions and PDCP-C functions (that is, the basic functions of control plane signaling at the PDCP layer).
  • Base stations can be fixed or mobile.
  • a helicopter or drone can be configured to act as a mobile base station, and one or more cells can move based on the location of the mobile base station.
  • a helicopter or drone can be configured to act as a device that communicates with another base station.
  • OAM operation administration and maintenance
  • OAM refers to the division of network management work into three categories: operation, administration, and maintenance, referred to as OAM, based on the actual needs of the operator's network operation.
  • OAM can also be called OAM entity or function.
  • Operation mainly completes the analysis, prediction, planning and configuration of daily networks and services; maintenance mainly involves daily operational activities such as testing and fault management of the network and its services.
  • OAM can detect network operation status, optimize network connections and performance, improve network operation stability, and reduce network maintenance costs.
  • OTT third-party application services
  • This application is different from the communication services currently provided by operators. It only uses the operator's network, and the service is provided by a third party outside the operator.
  • typical OTT services include Internet TV services, application stores, etc.
  • the technical solution in the embodiment of the present application can be applied to network equipment.
  • the network equipment can be a base station or a core network equipment.
  • the core network equipment may include a mobility management entity (MME), an access and mobility management function (AMF) network element, a network data analysis function (NWDAF) network element, a session management function (SMF) network element, or other core network equipment.
  • MME mobility management entity
  • AMF access and mobility management function
  • NWDAF network data analysis function
  • SMF session management function
  • AI model also known as AI algorithm (or AI operator)
  • AI algorithm or AI operator
  • AI model is a general term for mathematical algorithms built on the principles of artificial intelligence, and is also the basis for using AI to solve specific problems.
  • the type of AI model is not limited in the embodiments of the present application.
  • the AI model can be a machine learning model, a deep learning model, a reinforcement learning model, or a federated learning model.
  • Machine learning is a method to achieve artificial intelligence.
  • the goal of this method is to design and analyze some algorithms (also known as models) that allow computers to "learn" automatically.
  • the designed algorithms are called machine learning models.
  • Machine learning models are a type of algorithm that automatically analyzes data to obtain patterns and uses the patterns to predict unknown data. There are many types of machine learning models. Depending on whether the model training needs to rely on the labels corresponding to the training data, machine learning models can be divided into: 1. Supervised learning models; 2. Unsupervised learning models.
  • Deep learning is a new technical field that emerged in the process of machine learning research. Specifically, deep learning is a method in machine learning based on deep representation learning of data. Deep learning interprets data by establishing a neural network that simulates the human brain for analysis and learning. Since in machine learning methods, almost all features need to be determined by industry experts and then encoded. However, deep learning algorithms try to learn features from data by themselves. Algorithms designed based on deep learning ideas are called deep learning models.
  • Reinforcement learning is a special field in machine learning. It is a process of continuously learning the optimal strategy, making sequential decisions, and obtaining the maximum reward through the interaction between the agent and the environment. In layman's terms, reinforcement learning is learning "what to do (i.e. how to map the current situation into actions) to maximize the numerical benefit signal". The agent will not be told what action to take, but must try to find out which actions will produce the most lucrative benefits. Reinforcement learning is different from supervised learning and unsupervised learning in the field of machine learning. Supervised learning is the process of learning from labeled training data provided externally (task-driven), and unsupervised learning is the process of finding implicit structures in unlabeled data (data-driven). Reinforcement learning is the process of finding a better solution through "trials”. The agent must develop existing experience to obtain benefits, and also conduct trials so that it can obtain a better action selection space in the future (i.e. learn from mistakes). The algorithm designed based on reinforcement learning is called a reinforcement learning model.
  • Federated learning also known as collaborative learning
  • collaborative learning is a machine learning technique that trains algorithms on multiple decentralized edge devices or servers holding local data samples without exchanging them. This approach is in stark contrast to traditional centralized machine learning techniques, where all local datasets are uploaded to a single server for training.
  • Federated learning enables multiple participants to build a common, robust machine learning model without sharing data, allowing key issues such as data privacy, data security, data access rights, and access to heterogeneous data to be addressed.
  • AI model training refers to the process of using a specified initial model to calculate the training data, and adjusting the parameters in the initial model using a certain method based on the calculation results, so that the model gradually learns certain rules and has specific functions.
  • AI model reasoning is the process of using a trained AI model to calculate the input data and obtain the predicted reasoning results (also called output data).
  • the AI module is a module with AI learning and computing capabilities.
  • the AI module can be located in the OAM, in the gNB (the separation architecture is located in the CU), in some UEs, or as a separate network element entity.
  • the main function of the AI module in the wireless communication system is to perform a series of AI calculations such as model building, training approximation, and reinforcement learning based on input data (for example, in a wireless communication system, the input data can be network operation data provided by the RAN side or monitored by OAM, such as network load, channel quality, etc.).
  • the trained model provided by the AI module has the function of predicting network changes on the RAN side, and can usually be used for load prediction, UE trajectory prediction, etc.
  • the AI module can also perform policy reasoning from the perspectives of network energy saving and mobility optimization based on the prediction results of the trained model on the RAN network performance, so as to obtain reasonable and efficient energy saving strategies, mobility optimization strategies, etc.
  • the AI module When the AI module is located in OAM, its communication with the gNB on the RAN side can reuse the current northbound interface; when the AI module is located in the gNB or CU, the current F1, Xn, Uu and other interfaces can be reused; when the AI module becomes an independent network entity, it is necessary to re-establish the communication link to the OAM and RAN side, such as based on a wired link or a wireless link.
  • Fig. 1 is a schematic diagram of an AI module applicable to an embodiment of the present application.
  • the AI module 100 shown in Fig. 1 includes a database module 101, a training module 102, a model module 103 and an execution module 104.
  • the database module 101 can store training data.
  • the training data can also come from the terminal device.
  • the training data can come from the network device.
  • the training data can come from the base station (such as gNB) or the functional unit (such as CU or DU) constituting the base station.
  • the training data can come from other network devices other than the base station.
  • a gateway a management entity (such as mobile management (mobile management entity, MME), a core network device, etc.
  • the training module 102 analyzes the training data provided by the database module 101 to obtain an AI model.
  • the training module 102 can send the trained AI model to the model module 103.
  • the training module 102 can also update the trained model and send the update parameters used to update the model to the model module 103.
  • the model module 103 can also collect some model operation data during the operation of the AI model and send the operation data to the training module 102.
  • the training module 102 can update the AI model based on the operation data.
  • the model module 103 can determine the output data based on the AI model and the input data.
  • the output data may include the prediction results of the network operation obtained based on the input data and the AI model.
  • the output data may also include an adjustment strategy determined based on the input data and the AI model.
  • the network device and/or the terminal device may directly send the input data to the model module 103.
  • the database module 101 may also collect data from the network device and/or the terminal device, determine the input data and send the input data to the model module 103.
  • the execution module 104 can be used to execute the adjustment strategy determined by the model module 103.
  • the execution module 104 can also collect the specific performance of the network after the adjustment strategy is applied, such as performance parameters in the network, and feed this information back to the data module 101.
  • the database module 101 can store this feedback information. This feedback information can be used for subsequent model training or improving the AI model.
  • AI models can be used to intelligently collect and analyze data, improving network performance and user experience.
  • AI can be applied to channel state information (CSI) feedback enhancement.
  • CSI is the channel attribute of the communication link and is the channel quality information reported by the terminal device to the base station.
  • the terminal device reports the downlink channel quality information to the base station so that a more appropriate modulation and coding scheme (MCS) can be selected for the terminal device, so that it can better adapt to the changing wireless channel.
  • MCS modulation and coding scheme
  • AI can also be applied to beam management (BM).
  • BM is mainly used to find the strongest transmit/receive beam pair.
  • AI-based beam prediction can improve prediction accuracy.
  • AI can also be applied to positioning accuracy enhancements.
  • Positioning accuracy is related to the number of total radiated power (TRP) antennas.
  • TRP total radiated power
  • TDOA time difference of arrival
  • RTT round trip time
  • LOS line of sight
  • AI-based positioning can improve positioning accuracy in scenarios with fewer line-of-sight paths.
  • AI can also be applied to network energy saving.
  • Network energy saving can be achieved through cell activation/deactivation, load reduction, coverage improvement or other RAN setting adjustments.
  • the optimal energy saving decision depends on factors such as the load of different RAN nodes, RAN node capabilities, key performance indicators (KPI) requirements, quality of service (QoS) requirements, number of activated users and mobility of terminal devices, and cell utilization.
  • KPI key performance indicators
  • QoS quality of service
  • improving network energy efficiency is a complex process. Wrong cell closure and wrong traffic offloading operations will cause a decrease in network performance and even energy efficiency.
  • AI technology can be used to optimize energy saving decisions by utilizing data collected in the RAN network.
  • AI algorithms can predict the energy efficiency and load status of the next cycle, which can be used to make better decisions on cell activation/deactivation to save energy. Based on the predicted load, the system can dynamically configure energy saving strategies to maintain a balance between system performance and energy efficiency and reduce energy consumption.
  • AI can also be applied to load balancing.
  • load balancing is to evenly distribute the load between cells and between areas within a cell, or to transfer part of the traffic from congested cells, or to allow terminal devices to be diverted on a cell, carrier or access standard to improve network performance. This can be achieved by optimizing switching parameters and switching actions. This optimized automation can provide a high-quality user experience while improving system capacity and minimizing manual intervention in network management and optimization tasks. At present, load balancing decisions that rely on the current/past cell load status are not enough. In addition, the overall network and service performance is difficult to guarantee during load balancing.
  • solutions based on AI models can be introduced to improve load balancing performance, such as inputting various measurements and feedbacks from users and network nodes, historical data, etc. into AI models to improve load balancing performance, so as to provide a higher quality user experience and increase system capacity.
  • AI can also be applied to mobility optimization.
  • Mobility management is a solution to ensure service continuity during the mobility of terminal devices by minimizing dropped calls, radio link failures (RLF), unnecessary switching, and ping-pong effects.
  • RLF radio link failures
  • ping-pong effects For future high-frequency networks, as the coverage area of a single node decreases, the frequency of terminal devices switching between nodes will become very high, especially for high-mobility terminal devices.
  • QoE quality of experience
  • model management network element can be located in different network devices (such as base stations, core network devices, etc.) in the wireless communication system formed by the terminal device, or in OAM, or in the application service OTT provided by a third party, or the model management network element is used as an independent network element, etc., to realize the function of model management. It should be understood that any network element or entity or application service where the above-mentioned model management network element is located and the terminal device can exchange information with the terminal device based on the method for processing the model provided by the present application.
  • the model management network element is located in the base station (corresponding to the following method 200, method 500,), or the model management network element is located in AMF/MME/model management network element as a separate network element connected to the base station (corresponding to the following method 300, method 600) or the model management network element is located in OAM/OTT/SMF/NWDAF (corresponding to the following method 400, method 700)
  • the interaction process between the two is used as an example to describe in detail the method for processing the model provided in the embodiment of the present application.
  • the first model is the model requested to be registered by the terminal device.
  • the second model is the model after the third model is updated.
  • the third model may also be the model requested to be registered by the terminal before being updated. In other words, before the third model is updated to the second model, the third model may be the model requested to be registered by the terminal device.
  • model information, model information, and model update information are involved, among which the model information further involves model type information, model function information, model range identification information, model upload information, and model operation information.
  • model information further involves model type information, model function information, model range identification information, model upload information, and model operation information.
  • model information further involves model type information, model function information, model range identification information, model upload information, and model operation information.
  • the present invention provides a method for processing a model, which includes the following steps:
  • Step 1 The terminal device sends a first request message to the model management network element, where the first request message is used to request registration of a first model.
  • the model management network element may be a base station, a core network device, OAM, OTT, or a network element for model management.
  • the core network device may include MME, AMF, NWDAF, SMF, or at least one of other network devices.
  • the first request message is a radio resource control (RRC) message exchanged between the terminal device and the base station or a non-access stratum (NAS) message/information in the RRC message.
  • RRC radio resource control
  • NAS non-access stratum
  • Step 2 The model management network element determines the registration result according to the first request message.
  • Step 3 The model management network element sends a first response message to the terminal device, where the first response message is used to notify the terminal device whether the first model registration is successful.
  • FIG2 is a schematic flow chart of a processing model method 200 provided by an embodiment of the present application from the perspective of device interaction.
  • the model management network element may be a base station, or a network element with model management that is directly connected to a terminal device.
  • the method 200 may include S201 to S203. The steps in the method 200 are described in detail below.
  • a terminal device sends a first request message to a model management network element, where the first request message is used to request registration of a first model.
  • the terminal device After the terminal device is online, when the terminal device wants to register a model, the terminal device sends a first request message to the model management network element, and the first request message is used to request registration of the first model.
  • the model management network element receives the first request message from the terminal device, and learns from the first request message that the terminal device has initiated a model registration request for the first model.
  • the first request message carries model information of the first model.
  • the first request message carries the first model.
  • the first request message carries model information of the first model and the first model.
  • the first model is carried in other information or messages sent by the terminal device to the model management network element.
  • the first model is carried in an uplink data packet sent by the terminal device to the model management network element.
  • the first request message carries first indication information, and the first indication information is used to instruct the model management network element to receive the first model.
  • the terminal device instructs the model management network element through the first indication information in the first request message that the terminal device has sent the first model to the model management network element.
  • the first model when the data of the first model is relatively large (for example, the data corresponding to the first model is 1G bit, and the data corresponding to the first model occupies a large amount of memory of the terminal device), the first model cannot be carried in the first request message, and the first model is sent to the model management network element through the uplink data packet.
  • the first request message when the first model requested to be registered by the terminal device is a model that is not saved by the terminal device, the first request message carries the first model information regardless of whether the first model is stored in the model management network element. In other words, when the first request message is used to request registration of a model that is not saved by the terminal device, the terminal device sends the first model information to the model management network element through the first request message.
  • the first request message carries at least one of the first model information and the first model.
  • the first request message carries the first indication information, and the first model is carried in other information or messages.
  • the terminal device sends at least one of the first model information and the first model to the model management network element through the first request message.
  • the terminal device sends the first indication information to the model management network element through the first request message, and at the same time sends the first model to the model management network element through other messages or information.
  • the model information may include at least one of the following information: model type information, model function information, model range identification information, model upload information, and model operation information.
  • model type information is used to indicate whether the model is a unilateral model or a bilateral model.
  • model function information is used to indicate the function of the model, wherein the function of the model includes: measurement result prediction, measurement result compression, positioning, beam prediction, and beam management.
  • the model range identification information includes but is not limited to the manufacturer information of the terminal device that supports model registration, the chip manufacturer information, the application information (it may also be the application layer information, which is used to indicate the scenario in which the model corresponding to the model information can be applied, or which applications can be used to indicate which model corresponding to the model information can be called) or at least one of other information.
  • the model upload information is used to indicate whether it is necessary or allowed for the terminal device to send the model corresponding to the model information when performing model registration.
  • the model operation information is used to indicate the input information and/or output information of the model (the model that allows model registration), wherein the input information may be the input data set type or data set format of the model, and the output information may be the output data type or output format of the model.
  • model function information may include at least one of the model functions listed above.
  • model range identification information may include at least one of the model functions listed above.
  • model information of the first model refers to the model information corresponding to the first model.
  • model type information is used to indicate whether the first model is a unilateral model or a bilateral model.
  • model function information is used to indicate the function of the first model.
  • model range identification information, model upload information, and model operation information they are all used to indicate the model range identification information, model upload information, and model operation information corresponding to the first model, which will not be repeated here.
  • the model refers to an AI model.
  • the terminal device sends the first model to the model management network element, it may refer to the terminal device sending the first AI model or the first AI operator to the model management network element.
  • the first request message sent by the terminal device is a radio resource control (RRC) message exchanged between the terminal device and the base station or a non-access stratum (NAS) message/information in the RRC message.
  • RRC radio resource control
  • NAS non-access stratum
  • the model management network element determines a registration result according to the first request message.
  • the process by which the model management network element determines the registration result according to the first request message is as follows: the model management network element selects a model from the model pool based on the model information of the first model. If the model management network element successfully selects one or more models that meet the requirements, the model management network element selects any one of the multiple models that meet the requirements as the first model. At the same time, the model management network element confirms that the terminal device supports online registration of the first model. At this point, the model management network element confirms that the first model requested to be registered by the terminal device is successfully registered. Alternatively, if the model management network element does not select a model that meets the requirements, the model management network element confirms that the first model requested to be registered by the terminal device has failed to be registered.
  • the process by which the model management network element determines the registration result according to the first request message is as follows: the model management network element determines whether the first model is valid based on the first model, that is, the model management network element determines whether the first model can be registered online. If the first model is valid, that is, the first model can be registered online. At the same time, the model management network element confirms that the terminal device supports online registration of the first model. At this point, the model management network element confirms that the first model is successfully registered. Alternatively, if the model is invalid, that is, the first model cannot be registered online by the terminal device, the model management network element confirms that the first model registration has failed.
  • the model management network element confirms that the terminal device supports online registration of the first model, which can be understood as the model management network element verifying the terminal device or verifying the authority of the terminal device or verifying the legitimacy of the terminal device.
  • the model management network element can identify the terminal device based on the identification information of the terminal device and determine whether the terminal device supports online registration of the first model.
  • the model management network element can receive identification information of the terminal device from the terminal device or other network devices.
  • the terminal device can carry the identification information of the terminal device in the first request message of S201.
  • the identification information of the terminal device may include: The identifier assigned to identify the terminal device on the interface between the base station and the core network device, the default credentials of the terminal device, the System Architecture Evolution Temporary Mobile Subscriber Identity (S-TMSI) of the terminal device, the international mobile subscriber identity (IMSI) of the terminal device, or other identification information.
  • S-TMSI System Architecture Evolution Temporary Mobile Subscriber Identity
  • IMSI international mobile subscriber identity
  • the default credentials of the terminal device can be understood as information used to authenticate or verify the terminal device.
  • the model management network element confirms that the terminal device supports online registration of the first model, and the model management network element verifies the terminal device according to the identification information of the terminal device. At this time, the terminal device sends the identification information of the terminal device to the model management network element. Optionally, the terminal device sends the identification information of the terminal device to the model management network element in S201 through a first request message. That is, the model management network element verifies the authority of the terminal device or verifies the legitimacy of the terminal device or verifies whether the terminal device has the authority to register the first device online according to the identification information of the terminal device.
  • the model management network element confirms that the terminal device supports online registration of the first model.
  • the model management network element may determine the first model based on the terminal device information stored in the model management network element or carried by other messages, which is not limited here.
  • the model management network element sends a first response message to the terminal device, where the first response message is used to notify the terminal device whether the first model registration is successful.
  • the first response message is used to notify the terminal device that the first model registration is successful.
  • the model management network element determines that the first model requested by the terminal device is unsuccessful, the first response message is used to notify the terminal device that the first model registration is unsuccessful.
  • the model management network element sends the model configuration information of the first model that is successfully registered to the terminal device.
  • the model configuration information includes at least one of model identification information, model information, and model.
  • the model identification information is used to identify/recognize/associate the model, and the model identification information can be the index of the model in the model pool, or can be determined based on the identification information of the terminal device or part of the identification information of the terminal device, which is not limited in this application.
  • model configuration information of the first model may include at least one of model identification information of the first model (hereinafter also referred to as first model identification information), the first model information and the first model.
  • the model management network element determines that the terminal device stores the first model
  • the model configuration information of the first model sent by the model management network element to the terminal device carries the first model identification information.
  • the terminal device After the terminal device receives the model configuration information of the first model, the terminal device associates the first model with the first model identification information according to the first model identification information in the model configuration information of the first model.
  • the model management network element determines that the terminal device stores the first model
  • the model configuration information of the first model sent by the model management network element to the terminal device carries the model information of the first model.
  • the model management network element determines that the terminal device stores the first model
  • the model configuration information of the first model sent by the model management network element to the terminal device carries the first model identification information and the model information of the first model.
  • the terminal device After the terminal device receives the model configuration information of the first model, it associates the first model with the first model identification information according to the first model identification information in the model configuration information of the first model.
  • the terminal device associates the first model, the model information of the first model, and the first model identification information.
  • the model management network element determines that the terminal device does not store the first model, the model management network element sends the first model configuration information of the first model to the terminal device, and the model management network element carries the first model.
  • the configuration information of the first model may also carry at least one of the first model identification information and the model information of the first model.
  • the terminal device associates the first model identification information with the first model after receiving the model configuration information of the first model.
  • the terminal device associates the model information of the first model with the first model after receiving the model configuration information of the first model.
  • the terminal device associates the first model information, the first model and the first model identification information after receiving the model configuration information of the first model.
  • the method also includes the model management network element allocating the first model identification information to the first model.
  • model management network element determines that the first model is successfully registered, if the model management network element receives model information of the first model from the terminal device, the model management network element associates the model information of the first model with the first model identification information.
  • the model management network element when the model management network element receives the first model from the terminal device, the model management network element stores the first model and The first model is associated with first model identification information.
  • the model management network element when the model management network element receives the model information of the first model and the first model from the terminal device, the model management network element stores the first model and associates the model information of the first model, the first model, and the first model identification information.
  • the model configuration information may be partially or completely carried in the first response message.
  • the model management network element may send at least one of the model information of the first model and the first model identification information to the base station.
  • the model management network element may send the first model to the terminal device, and the first model may be sent to the terminal device through a NAS message/information or in a data packet or in other ways.
  • the base station may send the model information of the first model and the identification information of the first model to the terminal device through an RRC message.
  • the base station may also send the first model to the terminal device through a NAS message/information in an RRC message or in a data packet or in other ways.
  • the model management network element may send at least one of the model information of the first model and the first model identification information to the base station.
  • the model management network element may send model configuration information to the terminal device, and the model configuration information includes the first model.
  • the model configuration information also includes at least one of the model information of the first model and the first model identification information.
  • the model configuration information may be sent to the terminal device through a NAS message/information or in a data packet or in other ways.
  • the base station may send the model information of the first model and the identification information of the first model to the terminal device through an RRC message, or the base station may send the model configuration information to the terminal device through a NAS message/information in an RRC message or in a data packet or in other ways. It should be understood that the message carrying the above-mentioned model configuration information is only an example, and the embodiments of the present application are not limited to this.
  • the model management network element determines that the registration of the model requested by the terminal device has failed, the model management network element does not send a first response message to the terminal device. That is, the first response message can only be used to notify the terminal device that the model registration is successful.
  • the terminal device can set a waiting time T1 after sending the first request message. After T1 time, if the terminal device does not receive the first response message, the terminal device defaults to model registration failure.
  • the terminal device can set a countdown timer of T1 time after sending the first request message. When the time of the timer is 0, the terminal device has not received the first response message, and the terminal device defaults to model registration failure.
  • the method 200 further includes the following steps:
  • the model management network element sends a first message to the terminal device, where the first message is used to indicate whether the terminal device can register the model.
  • the model management network element may first send the first message to the terminal device to indicate whether the terminal device can initiate model registration.
  • the first message may be an RRC message, or a NAS signaling, or a system message.
  • the first message includes third indication information
  • the third indication information is used to indicate whether the terminal supports model registration of the terminal device.
  • the model management network element is a base station
  • the third indication information is used to indicate whether the service cell or the main service of the terminal device supports model registration.
  • the terminal device receives the first message, it can determine that it can initiate model registration to the model management network element based on the third indication information.
  • the number of bits included in the indication field occupied by the third indication information is equal to 1. If the value of the indication field is 0, it is used to indicate that the cell covered by the base station does not support model registration of the terminal device. If the value of the indication field is 1, it is used to indicate that the cell covered by the base station supports model registration of the terminal device. It should be understood that the number of bits occupied by the third indication information and the meaning corresponding to the values of different indication fields are only examples and not limitations.
  • the first message is used to indicate the model information of the model supported for registration by the terminal device (hereinafter also referred to as model registration support information).
  • the terminal device After the terminal device receives the first message, it can initiate model registration to the model management network element according to the model information of the model supported for registration.
  • the description of the model information can refer to the description in step S201, which will not be repeated here.
  • the first message includes the third indication information and the model information of the model supported by the terminal device for registration.
  • the terminal device can determine that the model registration can be initiated to the model management network element according to at least one of the third indication information and the model information of the model supported by the terminal device for registration.
  • the functions of the third indication information and the model information of the model supported by the terminal device for registration can be referred to the above description respectively, and will not be repeated here.
  • the model management network element sends fourth indication information to the terminal device, where the fourth indication information is used to indicate whether the first model supports online update.
  • the terminal device when the terminal device receives the fourth indication information sent by the model management network element, it can determine whether the first model supports online update according to the fourth indication information.
  • the number of bits included in the indication field occupied by the fourth indication information is equal to 1. If the value of the indication field is 0, it is used to indicate that the first model does not support online update. If the value of the indication field is 1, it is used to indicate that the first model supports online update.
  • the fourth indication information may be carried in an RRC message.
  • FIG 3 is a schematic flow chart of a processing model method 300 provided by an embodiment of the present application, shown from the perspective of device interaction.
  • the model management network element can be AMF, MME or an independent network element connected to the base station.
  • the model management network element is AMF or MME, which means that the AMF or MME can perform the operation of the model management network element.
  • the AMF or MME can implement the function of the model management network element.
  • the model management network element can be an independent network element connected to the base station, which means that the model management network element can be other network elements connected to the base station not listed here, or the model management network element can be a separately defined network element.
  • the method 300 may include S301 to S305. The steps in the method 300 are described in detail below.
  • the terminal device sends a third request message to the base station, where the third request message is used to request registration of the first model.
  • the terminal device when the terminal device wants to register a model, the terminal device sends a third request message to the base station, and the third request message is used to request registration of the first model.
  • the base station receives the third request message from the terminal device, and learns from the third request message that the terminal device has initiated a model registration request for the first model.
  • the third request message carries model information of the first model.
  • the third request message carries the first model.
  • the third request message carries model information of the first model and the first model.
  • the terminal device carries the model information of the first model in the third request message.
  • the third request message also includes first indication information.
  • the first indication information is used to indicate that the terminal device has sent the first model to the model management network element.
  • the description of the first indication information can refer to the relevant description in step S201, which is not repeated here.
  • the first model can be carried in other information or messages sent by the terminal device to the base station.
  • the model information of the first model is carried in the third request message.
  • the terminal device sends the model information of the first model to the base station through the third request message.
  • the third request message carries the model information of the first model and at least one of the first models.
  • the third request message carries the model information of the first model and the first indication information, and the first model is carried in other information or messages.
  • the terminal device sends the model information of the first model and at least one of the first models to the base station through the third request message.
  • the terminal device sends the first indication information to the base station through the third request message, and at the same time sends the first model to the base station through other messages or information.
  • the model information of the first model can refer to the relevant instructions in the above S201, which will not be repeated here.
  • the third request message carries the first model
  • the first model can refer to the relevant instructions in the above S201, which will not be repeated here.
  • the first request message sent by the terminal device is an RRC message exchanged between the terminal device and the base station or NAS information carried in the RRC message.
  • the base station sends a first request message to the model management network element.
  • the first request message is used to request the model management network element to perform online registration of the model.
  • the base station sends a first request message to the model management network element, and correspondingly, the model management network element receives the first request message from the base station, wherein the first request message is determined according to the third request message in S301.
  • the first request message may carry model information of the first model.
  • the base station may send a first model registration request to the model management network element.
  • the first request message may also carry identification information of the terminal device.
  • the description of the identification information of the terminal device can refer to the description in step S202, which will not be repeated here.
  • the model management network element determines a registration result according to the first request message.
  • the model management network element sends a first response message to the base station, where the first response message is used to notify the base station whether the first model registration is successful.
  • the first response message is used to notify the base station that the first model registration is successful.
  • the first response message is used to notify the base station that the first model registration fails.
  • the model management network element can send the first model successfully registered to the base station.
  • the model configuration information of the first model please refer to the above S203, which will not be repeated here.
  • the base station sends a third response message to the terminal device, where the third response message is used to notify the terminal device whether the first model registration is successful.
  • the base station sends a third response message to the terminal device, and correspondingly, the terminal device receives the third response message from the base station.
  • the third response message is determined according to the first response message in S304. Specifically, when the first response message is used to notify the base station that the first model is successfully registered, the third response message is used to notify the terminal device that the first model is successfully registered. When the first response message is used to notify the base station that the first model registration fails, the third response message is used to notify the terminal device that the first model registration fails.
  • the third response message may include model configuration information.
  • model configuration information For a specific description, please refer to step S203 and will not be repeated here.
  • the method may not include S306.
  • the third request message carries the identification information of the AMF, and the base station can send the information carried in the third request message directly to the model management network element through the first request according to the identification information of the AMF.
  • the third request message received by the base station from the terminal device is RRC signaling
  • the first request message sent by the base station to the model management network element is RRC signaling
  • the model management network element is AMF/MME
  • the base station determines the AMF/MME where the model management network element is located.
  • the base station selects AMF/MME according to the third request sent by the terminal device and the locally maintained AMF information/MME information.
  • the base station may make a selection based on whether the model function information (e.g., a function list) in the model information of the model supported by the AMF/MME for online registration includes the model function information requested by the terminal device.
  • the model function information e.g., a function list
  • the method 300 further includes the following steps:
  • the model management network element sends an indication message to the base station as to whether model registration is supported.
  • the description of the indication information can refer to the description of the third indication information and/or the model registration support information in step S2011, which will not be repeated here.
  • the indication information may include at least one of model registration support information and model information supporting model registration.
  • the base station receives at least one of model registration support information and model information supporting model registration from the model management network element, wherein the model registration support information refers to the model supporting terminal device registration, and the model information supporting model registration refers to the model information corresponding to the model that the base station can support terminal device registration.
  • the base station receives model registration support information and/or model information supporting model registration sent by a model management network element.
  • the base station sends a first message to the terminal device, where the first message is used to notify the terminal device to register a model.
  • the base station may first send the first message to the terminal device to notify the terminal device that the base station can provide terminal device initiated model registration.
  • the first message may be an RRC message, or an RRC message used for transparent transmission of NAS signaling, or a system message.
  • the first message may include at least one of the third indication information and model information of the model supported by the terminal device for registration.
  • the first message reference may be made to the above S2011, which will not be repeated here.
  • the model management network element sends fourth indication information to the base station, where the fourth indication information is used to indicate whether the first model supports online update.
  • the fourth indication information may be carried in an RRC message or NAS signaling.
  • the base station forwards the fourth indication information to the terminal device.
  • the model management network element sends fifth indication information to the base station, where the fifth indication information is used to indicate whether to send the first model and/or model information of the first model to the base station.
  • the model management network element when the model management network element determines that the base station does not store the first model and/or model information of the first model, the model management network element sends fifth indication information to indicate whether the terminal device sends the first model and/or model information of the first model to the base station.
  • the base station forwards the fifth indication information to the terminal device.
  • the base station forwards the fifth indication information to the terminal device, and accordingly, the terminal device receives the fifth indication information and determines whether to send the first model and/or the model information of the first model to the base station according to the fifth indication information.
  • the number of bits included in the indication field occupied by the fifth indication information is equal to 2
  • the value of the indication field is 00
  • the value of the indication field is 01
  • it is used to indicate that the model information of the first model is sent to the base station.
  • the value of the indication field is 10
  • the terminal device sends model configuration information to the base station.
  • the model configuration information of the terminal device may be obtained through the third response message in S305.
  • the model configuration information is sent to the base station by the model management network element through other messages, the base station forwards the model configuration information included in the other messages to the terminal device.
  • the terminal device may send the first model to the base station.
  • the terminal device may send the model information of the first model to the base station.
  • the terminal device may send the first model and the model information of the first model to the base station.
  • the terminal device determines that it is necessary to send the first model and/or model information of the first model to the base station based on the fifth indication information sent by the model management network element, the terminal device sends the first model and/or model information of the first model to the base station based on the fifth indication information.
  • the model management network element sends the first model and/or model information of the first model to the base station.
  • the model management network element determines that the base station does not store the first model, the model management network element sends the first model and/or model information of the first model to the base station.
  • the model management network element may send the model information of the first model to the base station.
  • the model management network element sends the first model to the base station according to the seventh indication information.
  • the model management network element sends the model information of the first model to the base station according to the eighth indication message.
  • the model management network element sends sixth indication information to the base station, and the sixth indication information is used to indicate whether to send the first model and/or the model information of the first model to the terminal device.
  • the number of bits included in the indication field occupied by the sixth indication information is 2. If the value of the indication field is 00, it is used to indicate that the base station sends the first model to the terminal device. If the value of the indication field is 01, it is used to indicate that the base station sends the model information of the first model to the terminal device. If the value of the indication field is 10, it is used to indicate that the base station sends the first model and the model information of the first model to the terminal device.
  • the value of the indication field is 11, it is used to indicate that the base station does not send any one of the first model and the model information of the first model to the terminal device. It should be understood that the number of bits occupied by the sixth indication information and the meanings corresponding to the values of different indication fields are only examples and not limitations.
  • the base station determines the information carried in the third response message according to the sixth indication information.
  • the base station determines that it is not necessary to send any one of the first model and the model information of the first model to the terminal device according to the sixth indication information, the base station believes that the first model and the model information of the first model are stored in the terminal device, and at this time, the base station only sends the first model identification information to the terminal device in the third response message.
  • FIG4 is a schematic flow chart of a processing model method 400 provided by an embodiment of the present application from the perspective of device interaction.
  • the model management network element may be SMF, OAM, OTT, NWDAF or an independent network element connected to AMF.
  • the model management network element is SMF, OAM, NWDAF or OTT, which means that SMF, OAM, NWDAF or OTT can perform the operation of the model management network element.
  • SMF, OAM, NWDAF or OTT can implement the function of the model management network element.
  • the model management network element may be an independent network element connected to AMF, which means that the model management network element may be other network elements connected to AMF not listed here or the model management network element may be a separately defined network element.
  • the method 400 may include S401 to S407. The steps in the method 400 are described below.
  • the difference between the method 400 and the method 300 is that in the method 400 shown in FIG. 4 , the message or information sent by the model management network element to the base station is forwarded to the base station via the AMF, while in the method 300 shown in FIG. 3 , the model management network element sends the message or information to the base station via the AMF. The message or information sent is directly sent to the base station. Therefore, compared with the method 300, the method 400 has the following different steps.
  • the base station sends a fourth request message to the AMF.
  • the base station After the base station receives the third request message sent by the terminal device, the base station sends a fourth request message to the AMF.
  • the AMF receives the fourth request message.
  • the fourth request message can be RRC signaling or NAS signaling, optionally.
  • the fourth request message is determined based on the third request message. In other words, after receiving the third request message, the base station may not process the received third request message, but directly forward the third request message to the AMF in the form of a fourth request message. Or after receiving the third request message, the base station determines the fourth request message based on the information carried in the third request message.
  • the base station when the fourth request message is NAS signaling, after receiving the third request message, the base station directly transparently transmits the third request message from the terminal device to the AMF in the form of the fourth request message.
  • the method when the fourth request message is RRC signaling, after the base station receives the third request message, since the third request message is RRC signaling, at this time, the method also includes S408.
  • the base station determines the AMF.
  • the base station determines the AMF based on the third request message of the terminal device and the locally maintained AMF information.
  • the base station can determine the AMF based on whether the model function information (for example, the function list) in the model information of the model supported by the AMF for online registration includes the model function information requested by the terminal device.
  • the base station determines that the function list of any AMF includes the function information corresponding to the first model, the base station can determine the AMF as the target AMF.
  • the base station After the base station determines the target AMF, the base station sends a fourth request message to the target AMF.
  • AMF sends a fourth response message to the base station.
  • the AMF forwards the received first response message to the base station, and the response message sent by the AMF to the base station is the fourth response message.
  • the AMF can send the fourth response message to the base station through the N2 interface with the base station.
  • the optional steps included in the method 400 shown in Figure 4, namely S4012, S4015, S4018, S4022, and S4024 are all information forwarded by the AMF to the base station. It should be understood that the forwarding process can be implemented through the N2 interface between the AMF and the base station. In addition, various information forwarded by the AMF to the base station. For example, model registration support information, model information supporting model registration, fourth indication information, fifth indication information, and sixth indication information, etc., can all refer to the relevant descriptions in Figure 3 above, which will not be repeated here.
  • the terminal device sends a third request message to the base station. This step can refer to S301, and will not be repeated here.
  • AMF sends a first request message to the model management network element.
  • This step can refer to S302, the base station sends a first request message to the model management network element, which will not be repeated here.
  • the model management network element determines the registration result according to the first request message. This step can refer to S303 and will not be described in detail here.
  • the model management network element sends a first response message to the AMF.
  • This step can refer to S304, the model management network element sends a first response message to the base station, which will not be repeated here.
  • the base station sends a third response message to the terminal device. This step can refer to S305 and will not be repeated here.
  • the model management network element sends model registration support information and/or model information supporting model registration to the AMF.
  • This step can refer to S3011, the model management network element sends model registration support information and/or model information supporting model registration to the base station, which will not be repeated here.
  • the base station sends a first message to the terminal device. This step can refer to S3012 and will not be repeated here.
  • the model management network element sends the fourth indication information to the AMF.
  • This step can refer to S3013, the model management network element sends the fourth indication information to the base station, which will not be repeated here.
  • the base station sends a fourth indication message to the terminal device. This step can refer to S3014 and will not be repeated here.
  • the model management network element sends the fifth indication information to the AMF. This step can refer to S3015, the model management network element sends the fourth indication information to the base station, which will not be repeated here.
  • the base station sends a fifth indication message to the terminal device. This step can refer to S3016 and will not be repeated here.
  • the terminal device sends model configuration information to the base station. This step can refer to S3017 and will not be repeated here.
  • the model management network element sends the first model and/or model information of the first model to the AMF.
  • This step can refer to S3018, The model management network element sends the first model and/or model information of the first model to the base station, which will not be repeated here.
  • the model management network element sends the sixth indication information to the AMF.
  • This step can refer to S3019, the model management network element sends the sixth indication information to the base station, which will not be repeated here.
  • Fig. 5 is a schematic flow chart of a processing model method 500 provided by an embodiment of the present application from the perspective of device interaction.
  • the model management network element is a base station.
  • the method 500 may include S501 to S504. The steps in the method 500 are described in detail below.
  • the terminal device sends a second request message to the model management network element (a base station in the embodiment of the present application), where the second request message is used to request to update the third model to the second model.
  • the model management network element a base station in the embodiment of the present application
  • the second model is generated.
  • the terminal device sends a second request message to the model management network element, and the second request message is used to request to update the third model to the second model.
  • the second request message carries second indication information, and the second indication information is used to indicate that the process requested by the terminal device is an update process. That is, the second indication information is used to indicate to update the third model to the second model.
  • the model update process performed by the terminal device may be a process of optimizing the local model rather than retraining.
  • the second request message may be RRC signaling or NAS signaling.
  • the second request message carries identification information (default credentials) of the terminal device, and the identification information of the terminal device is used by the model management network element to identify the terminal device and determine whether the terminal device supports online updating of the third model.
  • S502 The terminal device sends model update information of the third model to the model management network element.
  • the terminal device when the terminal device wants to update the third model to the second model, the terminal device sends model update information of the third model to the model management network element.
  • the model update information includes at least one of the gradient information and the model, and the model identification information.
  • the model identification information can refer to the relevant description above, which will not be repeated here.
  • the gradient information can be understood as a parameter information generated during the update of the third model to the second model.
  • the model management network element receives the gradient information, the third model can be updated to the second model according to the gradient information.
  • the model can refer to the relevant description above, which will not be repeated here.
  • model update information of the third model sent by the terminal device to the model management network element may include model identification information and gradient information of the third model. Or include model identification information of the third model and the second model. Or include model identification information of the third model, gradient information and the second model.
  • the model identification information of the third model is carried in the third model update information.
  • the terminal device sends the model identification information of the third model to the model management network element.
  • the third model update information carries the gradient information and at least one of the second model, and the model identification information of the third model.
  • the terminal device sends the gradient information and at least one of the second model, and the model identification information of the third model to the model management network element.
  • the parent model refers to the model sent by the network side device to the terminal device, or refers to the model stored in the network device, or may refer to the model managed by the network device.
  • the child model refers to the model updated locally by the terminal device, or refers to the model stored only in the terminal device, or may refer to the model managed only by the terminal device.
  • the model update information of the third model can be carried in the second request message, or can be carried by a message or information sent by other terminal devices to the base station.
  • the model update information of the third model includes the second model
  • the model update information of the third model is sent to the model management network element via an uplink data packet.
  • the model management network element determines an update result according to the second request message and the model update information of the third model.
  • the process in which the model management network element determines the update result according to the second request message is as follows: the model management network element updates the parameters of the third model based on the model identification information and gradient information of the third model. At the same time, the model management network element confirms that the terminal device supports online updating of the third model. At this point, the model management network element confirms that the third model requested to be updated by the terminal device is updated successfully. Otherwise, the model management network element confirms that the third model requested to be updated by the terminal device fails to be updated.
  • the process of the model management network element determining the update result according to the second request message is as follows: the model management network element confirms that the third model identification is valid, and receives the second model, and at the same time, the model management network element confirms that the terminal device supports online update of the third model. At this time, the model management network element confirms that the third model requested to be updated by the terminal device is updated successfully. Otherwise, the model management network element confirms that the third model requested to be updated by the terminal device fails to be updated.
  • the model management network element confirms that the terminal device supports online updating of the third model, which can be understood as the model management network element verifying the terminal device or verifying the authority of the terminal device or verifying the legitimacy of the terminal device.
  • This process can be identified by the model management network element according to the identification information of the terminal device.
  • the identification information of the terminal device can refer to the relevant instructions in S202, which will not be repeated here.
  • the model management network element when the model management network element confirms that the third model requested to be updated by the terminal device is successfully updated, the model management network element generates a second model, and the model management network element can add the second model as a parent model to the model pool of the parent model managed by the model management network element.
  • the second model is managed as an updated version of the third model. That is, the model management network element manages the model version based on the second model.
  • the model management network element updates the parent model version based on the second model and updates the model pool.
  • the model management network element adds the second model as a parent model to the model pool of the parent model managed by the model management network element, the model management network element allocates the model identification information of the second model (hereinafter also referred to as the second model identification information) to the second model.
  • the model management network element updates the function type or model identification that supports online update in its model pool, and generates a model version management list as shown in Table 1 below.
  • the model management network element manages the second model as a sub-model version, that is, manages the second model as an updated version of the third model, at this time, the model management network element updates the sub-model version corresponding to the third model in its model pool, and generates a model version management list as shown in Table 2 below.
  • the model management network element updates the child model version corresponding to the third model in its model pool, and generates a model version management list as shown in Table 3 below.
  • the model management network element sends a second response message to the terminal device, where the second response message is used to notify the update result.
  • the second response message is used to notify the terminal device that the third model is successfully updated.
  • the second response message is used to notify the terminal device that the third model is failed to be updated.
  • the model management network element sends the model update information of the second model to the terminal device.
  • the model management network element sends the model update information of the second model to the terminal device.
  • model update information of the second model may include model identification information of the second model.
  • the model identification information of the second model is used to associate with the second model.
  • model update information of the second model may be partially or completely carried in the second response message, or the model update information of the second model may be carried in other messages sent by the base station to the terminal device.
  • the model management network element determines that the model update requested by the terminal device has failed, the model management network element does not send a second response message to the terminal device. That is, the second response message can only be used to notify the terminal device that the model update is successful.
  • the terminal device can set a waiting time T2 after sending the second request message. After T2, if the terminal device does not receive the second response message, the terminal device defaults to the model update failure.
  • the terminal device can set a countdown timer for T2 time after sending the second request message. When the time of the timer is 0, the terminal device has not received the second response message, and the terminal device defaults to the model update failure.
  • the model management network element can first determine whether the third model supports updating according to the second request message. When the model management network element determines that the third model supports updating, it updates the third model to the second model according to the model update information of the third model received in S502.
  • the second response message may be RRC signaling or NAS signaling.
  • the method 500 further includes the following steps:
  • the terminal device updates the third model to the second model.
  • the terminal device updates the third model to the second model according to local data.
  • FIG. 6 is a schematic flow chart of a processing model method 600 provided by an embodiment of the present application, shown from the perspective of device interaction.
  • the model management network element may be an AMF, an MME, or an independent network element connected to the base station.
  • the model management network element being an AMF or an MME means that the AMF or the MME can perform the operations of the model management network element.
  • the AMF or the MME can implement the functions of the model management network element.
  • the model management network element may be an independent network element connected to the base station, which means that the model management network element may be other network elements connected to the base station not listed here, or the model management network element may be a separately defined network element.
  • the method 600 may include S601 to S607. The steps in the method 600 are described in detail below.
  • the terminal device sends a fifth request message to the base station, where the fifth request message is used to request to update the third model to the second model.
  • the terminal device sends a fifth request message to the base station, and the base station receives the fifth request message.
  • the terminal device After the terminal device completes the update of the third model locally, it generates a second model.
  • the terminal device sends a fifth request message to the base station, and the fifth request message is used to request to update the third model to the second model.
  • the second request message carries second indication information, and the second indication information is used to indicate that the process requested by the terminal device is an update process. That is, the second indication information is used to indicate to update the third model to the second model.
  • the model update process performed by the terminal device may be a process of optimizing the local model rather than retraining.
  • the fifth request message may be RRC signaling or NAS signaling.
  • the fifth request message carries identification information of the terminal device, and is used by the model management network element to identify the terminal device according to the identification information of the terminal device.
  • the fifth request message is NAS signaling
  • the fifth request signaling carries the identification information of the AMF. Therefore, the base station directly transmits the fifth request message from the terminal device in the form of a second request message to the model management network element according to the identification information of the AMF.
  • the method when the fifth request message is RRC signaling, after the base station receives the fifth request message, the method further includes S608.
  • the base station determines AMF/MME.
  • the base station determines the AMF based on the fifth request message of the terminal device and the locally maintained AMF information.
  • the base station can determine the AMF based on whether the model function information (for example, the function list) in the model information of the model supported by the AMF for online registration includes the model function information requested by the terminal device.
  • the base station determines that the function list of any AMF includes the function information corresponding to the first model, the base station can determine the AMF as the target AMF.
  • the base station After the base station determines the target AMF, the base station sends a second request message to the target AMF.
  • the method does not include S608.
  • the fifth request message received by the base station from the terminal device is RRC signaling
  • the second request message sent by the base station to the model management network element is RRC signaling
  • the model management network element is AMF
  • the base station is used to determine the target MME.
  • the process of determining the target MME is the same as the process of determining the target AMF, which will not be repeated here.
  • the base station sends a second request message to the model management network element, where the second request message is used to request updating the third model to the second model.
  • the base station determines the second request message according to the fifth request message.
  • the second request message carries model information of the second model.
  • S603 The terminal device sends model update information of the third model to the base station.
  • the terminal device when the terminal device wants to update the third model to the second model, the terminal device sends model update information of the third model to the base station.
  • the base station receives the model update information of the third model.
  • the model update information of the third model can refer to the description in S502 above, which will not be repeated here.
  • the model update information of the third model sent by the terminal device to the base station can be carried in the fifth request message, or can be carried by messages or information sent by other terminal devices to the base station, such as carried in uplink data packets, which is not repeated here.
  • the base station sends model update information of the third model to the model management network element.
  • the base station after receiving the model update information of the third model from the terminal device, the base station forwards the model update information of the third model to the model management network element.
  • the model update information of the third model sent by the base station to the model management network element can be carried in the second request message, or can be carried by a message or information sent to the base station by other terminal devices, such as carried in an uplink data packet, which is not repeated here.
  • the model management network element determines an update result according to the second request message and the model update information of the third model.
  • the process may refer to the above S503, which will not be described in detail here.
  • the model management network element sends a second response message to the base station, where the second response message is used to notify the update result.
  • the second response message is used to notify the terminal device that the third model is successfully updated.
  • the second response message is used to notify the terminal device that the third model fails to be updated.
  • the information carried in the second response message can refer to the above S504, which will not be repeated here.
  • model management network when the model management network generates a model version management list such as Table 1, Table 2 or Table 3 during the update process, Table 1, Table 2 or Table 3 can be sent to AMF through the second response message or other messages.
  • the base station sends a fifth response message to the terminal device, where the fifth response message is used to notify the update result.
  • the fifth response message is determined based on the second response message, wherein the information carried in the fifth response message can refer to the above S504 and will not be repeated here.
  • the fifth response message may be RRC signaling or NAS signaling.
  • the method 600 further includes the following steps:
  • the terminal device updates the third model to the second model.
  • the terminal device updates the third model to the second model according to local data.
  • FIG. 7 is a schematic flow chart of a processing model method 700 provided by an embodiment of the present application, shown from the perspective of device interaction.
  • the model management network element may be SMF, OAM, OTT, NWDAF or an independent network element connected to AMF.
  • the model management network element is SMF, OAM, NWDAF or OTT, which means that SMF, OAM, NWDAF or OTT can perform the operations of the model management network element.
  • SMF, OAM, NWDAF or OTT can implement the functions of the model management network element.
  • the model management network element may be an independent network element connected to AMF, which means that the model management network element may be other network elements not listed here.
  • the network element to which the AMF is connected or the model management network element may be a network element defined separately.
  • the method 700 may include S701 to S710. The steps in the method 700 are described below.
  • the difference between the method 700 and the method 600 is that in the method 700 shown in FIG7 , the message or information sent by the model management network element to the base station is forwarded to the base station via the AMF, while in the method 600 shown in FIG6 , the message or information sent by the model management network element to the base station is directly sent to the base station. Therefore, compared with the method 600, the method 700 has the following multiple different steps.
  • S702 The base station sends a sixth request message to the AMF.
  • the base station After the base station receives the fifth request message sent by the terminal device, the base station sends a sixth request message to the AMF.
  • the AMF receives the sixth request message.
  • the sixth request message can be RRC signaling or NAS signaling, optionally.
  • the sixth request message is determined based on the fifth request message. In other words, after receiving the fifth request message, the base station may not process the received fifth request message, but directly forward the fifth request message to the AMF in the form of the sixth request message. Or after receiving the fifth request message, the base station determines the sixth request message based on the information carried in the fifth request message.
  • the base station when the sixth request message is NAS signaling, after the base station receives the fifth request message, since the fifth request message is NAS signaling, at this time, the fifth request signaling carries the identification information of AMF. Therefore, the base station directly transmits the fifth request message from the terminal device to AMF in the form of the sixth request message according to the identification information of AMF.
  • the method when the sixth request message is RRC signaling, after the base station receives the fifth request message, since the fifth request message is RRC signaling, at this time, the method also includes S711.
  • the base station determines the AMF.
  • the base station determines the AMF based on the fifth request message of the terminal device and the locally maintained AMF information.
  • the base station can determine the AMF based on whether the model function information (for example, the function list) in the model information of the model supported by the AMF for online registration includes the model function information requested by the terminal device.
  • the base station determines that the function list of any AMF includes the function information corresponding to the first model, the base station can determine the AMF as the target AMF.
  • the base station After the base station determines the target AMF, the base station sends the sixth request message to the target AMF.
  • S705 The base station sends model update information of the third model to the AMF.
  • the terminal device when the terminal device wants to update the third model to the second model, the terminal device sends the model update information of the third model to the base station.
  • the base station forwards the received model update information of the third model to the AMF.
  • the model update information of the third model can refer to the description in S502 above, which will not be repeated here.
  • AMF sends a sixth response message to the base station.
  • the AMF forwards the received second response message to the base station, and the response message sent by the AMF to the base station is the sixth response message.
  • the AMF can send the sixth response message to the base station through the N2 interface with the base station.
  • Fig. 8 is a schematic flowchart of a model processing method 800 provided in an embodiment of the present application. Specifically, the method 800 includes the following steps.
  • the method 800 can be regarded as a scenario in which the above-mentioned methods 200 to 700 are used in a reasonable combination.
  • S801 can be one of the above-mentioned methods 200 to 400
  • S802 can be one of the above-mentioned methods 500 to 700.
  • the scenarios used by S801 and S802 are the same.
  • the model management network element is a base station
  • S801 executes method 200
  • S801 executes method 500.
  • the model management network element is, for example, AMF
  • S801 executes method 300
  • S801 executes method 600.
  • S801 executes method 400
  • S801 executes method 700.
  • Fig. 9 is a schematic block diagram of a device 900 for processing a model provided in an embodiment of the present application.
  • the device 900 includes a receiving module 901, which can be used to implement a corresponding receiving function.
  • the receiving module 901 can also be called a receiving unit.
  • the device 900 further includes a processing module 902 , which can be used to implement corresponding processing functions.
  • the device 900 also includes a sending module 903, which can be used to implement the corresponding sending function.
  • the sending module 903 can also It is called a sending unit.
  • the device 900 also includes a storage unit, which can be used to store instructions and/or data.
  • the processing unit 902 can read the instructions and/or data in the storage unit so that the device implements the actions of the relevant devices in the aforementioned method embodiments.
  • the device 900 can be used to execute the actions performed by the terminal device or model management network element in the above method embodiments.
  • the device 900 can be a component of the terminal device or model management network element
  • the receiving module 901 is used to execute the reception-related operations of the terminal device or model management network element in the above method embodiments
  • the processing module 902 is used to execute the processing-related operations of the terminal device or model management network element in the above method embodiments
  • the sending module 903 is used to execute the sending-related operations of the terminal device or model management network element in the above method embodiments.
  • the device 900 is used to perform the actions performed by any device in the above method embodiments (method 200 to method 800).
  • the device of the processing model can be used to perform the operations of the terminal device in Figures 2 to 8 above. For example:
  • the receiving module 901 is used to receive a first response message, where the first response message is used to notify whether the registration of the first model is successful.
  • the sending module 903 is used to send a first request message, where the first request message includes model information of the first model, and the first request message is used to request registration of the first model.
  • the first request message also includes identification information of the terminal device, and the identification information of the terminal device is used by the model management network element to identify the terminal device.
  • the receiving module 901, the processing module 902 and the sending module 903 in the device 900 of the processing model can also implement other operations or functions of the terminal device in the above method, which will not be repeated here.
  • the device 900 of the processing model may be a device including a terminal device.
  • the device 900 of the processing model may be a component configured in the terminal device, for example, a chip in the terminal device.
  • the receiving module 901 and the sending module 903 may be interface circuits, pins, etc.
  • the interface circuit may include an input circuit and an output circuit, wherein the receiving module 901 may include an input circuit, the sending module 903 may include an output circuit, and the processing module 902 may include a processing circuit.
  • the device for processing the model can be used to perform the operations of the model management network element in Figures 2 to 8 above.
  • the device for processing the model can be used to perform the operations of the model management network element in Figures 2 to 8 above.
  • the device for processing the model can be used to perform the operations of the model management network element in Figures 2 to 8 above. For example:
  • the receiving module 901 is used to receive a first request message, where the first request message includes model information of a first model, and the first request message is used to request registration of the first model.
  • the first request message also includes identification information of the terminal device, and the identification information of the terminal device is used by the model management network element to identify the terminal device.
  • Processing module 902 for determining the first model according to the model information of the first model
  • the sending module 903 is used to send a first response message, where the first response message is used to notify whether the registration of the first model is successful.
  • receiving module 901, processing module 902 and sending module 903 in the device 900 for processing the model can also implement other operations or functions of the model management network element in the above method, which will not be repeated here.
  • the device 900 for processing the model may be a device including a model management network element.
  • the device 900 for processing the model may be a component configured in the model management network element, for example, a chip in the model management network element.
  • the receiving module 901 and the sending module 903 may be interface circuits, pins, etc.
  • the interface circuit may include an input circuit and an output circuit, wherein the receiving module 901 may include an input circuit, the sending module 903 may include an output circuit, and the processing module 902 may include a processing circuit.
  • FIG 10 is a schematic structural diagram of another device for processing a model provided in an embodiment of the present application.
  • the device 1000 for processing the model includes a processor 1001.
  • the device for processing the model may also include at least one memory 1002 for storing computer programs or instructions or and/or data.
  • the memory 1002 is coupled to the processor 1001, and the processor 1001 is used to execute the computer program or instructions and/or data stored in the memory 1002, so that the method (method 200 to method 800) in the above method embodiment is executed.
  • the coupling in the embodiment of the present application is an indirect coupling or communication connection between devices, units or modules, which can be electrical, mechanical or other forms for information exchange between devices, units or modules.
  • the processor 1001 may operate in conjunction with the memory 1002. At least one of the at least one memory 1002 may be included in the processor 1001.
  • the device 1000 for processing the model includes one or more processors 1001.
  • the memory 1002 may be integrated with the processor 1001 or provided separately.
  • the device 1000 for processing a model may also include a transceiver 1003 for processing the model with other devices via a transmission medium, so that the device can process the model with other devices.
  • the transceiver 1003 may be an interface, a bus, a circuit or a device capable of implementing a transceiver function.
  • the device in the transceiver 1003 for implementing the receiving function may be regarded as a receiving module
  • the device in the transceiver 1003 for implementing the sending function may be regarded as a sending module, that is, the transceiver 1003 includes a receiver and a transmitter.
  • the specific connection medium between the processor 1001, the memory 1002 and the transceiver 1003 is not limited in the embodiment of the present application.
  • the processor 1001, the memory 1002 and the transceiver 1003 are connected via a bus 1004, which is represented by a thick line in FIG10 .
  • the connection between other components is only for schematic illustration and is not intended to be limiting.
  • the bus can be divided into an address bus, a data bus, a control bus, etc.
  • FIG. 10 only uses one thick line, but this does not mean that there is only one bus or one type of bus.
  • the device 1000 for processing the model may further include a transceiver 1003 and/or a communication interface, and the transceiver 1003 and/or the communication interface are used to receive and/or send signals.
  • the processor 1001 is used to control the transceiver 1003 and/or the communication interface to receive and/or send data.
  • a transceiver may also be sometimes referred to as a transceiver, a transceiver module, or a transceiver circuit, etc.
  • a receiver may also be sometimes referred to as a receiver, a receiving module, or a receiving circuit, etc.
  • a transmitter may also be sometimes referred to as a transmitter, a transmitter, a transmitting module, or a transmitting circuit, etc.
  • the processor 1001 is configured to perform other operations or functions of the terminal device.
  • the transceiver 1003 is used to implement the processing model between the device of the processing model and the model management network element.
  • the processor 1001 is configured to model other operations or functions of the management network element.
  • the transceiver 1003 is used to implement the processing model between the apparatus of the processing model and the terminal device.
  • One or more of the above modules or units can be implemented by software, hardware or a combination of the two.
  • the software exists in the form of computer program instructions and is stored in a memory, and the processor can be used to execute the program instructions and implement the above method flow.
  • the processor may include but is not limited to at least one of the following: a central processing unit (CPU), a microprocessor, a digital signal processor (DSP), a microcontroller (MCU), or an artificial intelligence processor and other types of computing devices that run software, each of which may include one or more cores for executing software instructions for calculation or processing.
  • the processor may be built into an SoC (system on chip) or an application specific integrated circuit (ASIC), or it may be an independent semiconductor chip.
  • processors In addition to the core used to execute software instructions for calculation or processing in the processor, it may further include necessary hardware accelerators, such as a field programmable gate array (FPGA), a PLD (programmable logic device), or a logic circuit that implements a dedicated logic operation.
  • FPGA field programmable gate array
  • PLD programmable logic device
  • logic circuit that implements a dedicated logic operation.
  • the hardware can be any one or any combination of a CPU, a microprocessor, a DSP, an MCU, an artificial intelligence processor, an ASIC, a SoC, an FPGA, a PLD, a dedicated digital circuit, a hardware accelerator or a non-integrated discrete device, which can run the necessary software or not rely on the software to execute the above method flow.
  • the above modules or units When the above modules or units are implemented using software, they can be implemented in whole or in part in the form of a computer program product.
  • the computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part.
  • the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.
  • the computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium.
  • the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means.
  • the computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated.
  • the available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid state drive (SSD)), etc.
  • the embodiment of the present application provides a device 1100 for processing a model, which may be a model management network element or a chip.
  • the device 1100 may be used to execute the operations performed by the model management network element in the above method embodiments (methods 200 to 800).
  • FIG. 11 shows a simplified schematic diagram of the structure of the model management network element.
  • the model management network element includes part 1110 and part 1120.
  • Part 1110 includes an antenna and a radio frequency circuit.
  • the antenna is mainly used for transmitting and receiving radio frequency signals
  • the radio frequency circuit is mainly used for converting radio frequency signals into baseband signals.
  • Part 1120 includes a memory and a processor, which are mainly used for baseband processing and controlling the model management network element.
  • Part 1110 can generally be referred to as a transceiver unit, a transceiver, a transceiver circuit, or a transceiver.
  • Part 1120 is generally the control center of the model management network element, which can generally be referred to as a processing unit, and is used to control the model management network element to perform the processing operations on the model management network element side in the above method embodiment.
  • the device for implementing the receiving function in part 1110 may be regarded as a receiving unit, and the device for implementing the sending function may be regarded as a sending unit, that is, part 1110 includes a receiving unit and a sending unit.
  • the receiving unit may also be referred to as a receiver, a receiver, or a receiving circuit, etc.
  • the sending unit may be referred to as a transmitter, a transmitter, or a transmitting circuit, etc.
  • the processor When data needs to be sent, the processor performs baseband processing on the data to be sent and outputs the baseband signal to the RF circuit.
  • the RF circuit performs RF processing on the baseband signal and then sends the RF signal outward in the form of electromagnetic waves through the antenna.
  • the RF circuit receives the RF signal through the antenna, converts the RF signal into a baseband signal, and outputs the baseband signal to the processor.
  • the processor converts the baseband signal into data and processes the data.
  • Part 1120 may include one or more boards, each of which may include one or more processors and one or more memories. For ease of explanation, only one memory and processor are shown in FIG11. The processor is used to read and execute programs in the memory to implement baseband processing functions and control the model management network element. If there are multiple boards, the boards can be interconnected to enhance processing capabilities. As an optional implementation, multiple boards may share one or more processors, or multiple boards may share one or more memories.
  • FIG11 is merely an example and not a limitation, and the above-mentioned model management network element including the transceiver unit and the processing unit may not rely on the structure shown in FIG11.
  • the chip When the device 1100 of the processing model is a chip, the chip includes a transceiver unit and a processing unit, wherein the transceiver unit may be an input/output circuit or a communication interface; and the processing unit may be a processor or a microprocessor or an integrated circuit integrated on the chip.
  • the embodiment of the present application also provides another device 1200 for processing a model, which can be a terminal device or a chip.
  • the device 1200 can be used to execute the operations executed by the terminal device in the above method embodiments (methods 200 to 800).
  • FIG12 shows a simplified schematic diagram of the structure of the terminal device.
  • the terminal device includes a processor, a memory, a radio frequency circuit, an antenna, and an input-output device.
  • the processor is mainly used to process communication protocols and communication data, and to control the terminal device, execute software programs, process software program data, etc.
  • the memory is mainly used to store software programs and data.
  • the radio frequency circuit is mainly used for conversion between baseband signals and radio frequency signals and processing of radio frequency signals.
  • the antenna is mainly used to send and receive radio frequency signals in the form of electromagnetic waves.
  • Input-output devices such as touch screens, display screens, keyboards, etc., are mainly used to receive data input by users and output data to users. It should be noted that some types of terminal devices may not have input-output devices.
  • the processor When data needs to be sent, the processor performs baseband processing on the data to be sent, and then outputs the baseband signal to the RF circuit.
  • the RF circuit performs RF processing on the baseband signal and then sends the RF signal outward in the form of electromagnetic waves through the antenna.
  • the RF circuit receives the RF signal through the antenna, converts the RF signal into a baseband signal, and outputs the baseband signal to the processor.
  • the processor converts the baseband signal into data and processes the data.
  • only one memory and processor are shown in Figure 12. In an actual terminal device product, there may be one or more processors and one or more memories.
  • the memory may also be referred to as a storage medium or a storage device, etc.
  • the memory may be set independently of the processor or integrated with the processor, and the embodiments of the present application do not limit this.
  • the antenna and the radio frequency circuit with transceiver functions can be regarded as the transceiver unit of the terminal device, and the processor with processing function can be regarded as the processing unit of the terminal device.
  • the terminal device includes a transceiver unit 10 and a processing unit 20.
  • the transceiver unit 10 may also be called a transceiver, a transceiver, a transceiver device or a transceiver circuit, etc.
  • the processing unit 20 may also be called a processor, a processing board, a processing module, a processing device, etc.
  • the device used to implement the receiving function in the transceiver unit 10 can be regarded as a receiving unit
  • the device used to implement the sending function in the transceiver unit 10 can be regarded as a sending unit
  • the transceiver unit 10 includes a receiving unit and a sending unit.
  • the receiving unit may also be sometimes referred to as a receiver, a receiver, a receiving device, or a receiving circuit, etc.
  • the sending unit may also be sometimes referred to as a transmitter, a transmitter, a transmitting device, or a transmitting circuit, etc.
  • FIG12 is merely an example and not a limitation, and the terminal device including the transceiver unit and the processing unit may not rely on the structure shown in FIG12.
  • the chip When the device 1200 of the processing model is a chip, the chip includes a transceiver unit and a processing unit, wherein the transceiver unit may be an input/output circuit or a communication interface; and the processing unit may be a processor or a microprocessor or an integrated circuit integrated on the chip.
  • the present application also provides a computer program product, which includes: computer program code, when the computer program code is run on a computer, the computer executes the method of the terminal device in the aforementioned method embodiment.
  • the present application also provides a computer program product, which includes: computer program code, when the computer program code runs on a computer, the computer executes the method of model management network element in the aforementioned method embodiment.
  • the present application also provides a computer-readable medium, which stores a program code.
  • the program code runs on a computer, the computer executes the method of the terminal device in the aforementioned method embodiment.
  • the present application also provides a computer-readable medium, which stores a program code.
  • the program code runs on a computer, the computer executes the method of model management network elements in the aforementioned method embodiment.
  • An embodiment of the present application also provides a processing device, including a processor and an interface; the processor is used to execute the method for processing the model in any of the above method embodiments.
  • An embodiment of the present application also provides a system for processing models, which includes multiple model management network elements and a terminal device in the above embodiment.
  • a component can be, but is not limited to, a process running on a processor, a processor, an object, an executable file, an execution thread, a program and/or a computer.
  • applications and computing devices running on a computing device can be components.
  • One or more components may reside in a process and/or an execution thread, and a component may be located on a computer and/or distributed between two or more computers.
  • these components may be executed from various computer-readable media having various data structures stored thereon.
  • Components may, for example, communicate through local and/or remote processes according to signals having one or more data packets (e.g., data from two components interacting with another component between a local system, a distributed system and/or a network, such as the Internet interacting with other systems through signals).
  • signals having one or more data packets (e.g., data from two components interacting with another component between a local system, a distributed system and/or a network, such as the Internet interacting with other systems through signals).
  • At least one in the embodiments of the present application refers to one or more, and “plurality” refers to two or more.
  • “And/or” describes the association relationship of associated objects, indicating that three relationships may exist.
  • a and/or B can represent: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural.
  • the character “/” generally indicates that the associated objects before and after are in an "or” relationship.
  • At least one of the following” or similar expressions refers to any combination of these items, including any combination of single items or plural items.
  • At least one of a, b and c can represent: a, or, b, or, c, or, a and b, or, a and c, or, b and c, or, a, b and c.
  • a, b and c can be single or multiple, respectively.
  • the disclosed systems, devices and methods can be implemented in other ways.
  • the device embodiments described above are only schematic.
  • the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.
  • Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
  • the units described as separate components may or may not be physically separate, and the components shown as units may be Or it may not be a physical unit, that is, it may be located in one place, or it may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
  • each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.
  • the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art.
  • the computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application.
  • the aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Computing Systems (AREA)
  • Physics & Mathematics (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Computational Linguistics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

本申请提供一种处理模型的方法和装置，能够保证终端设备与模型管理网元之间的模型的统一，使得模型管理网元能够实现对终端设备的模型的高效管理，进而保证系统的性能。该方法包括：模型管理网元接收第一请求消息，并接收第一模型的模型信息和第一模型中的至少一个，其中，该第一请求消息用于请求注册所述第一模型。模型管理网元根据第一模型的模型信息和第一模型中的至少一个，确定第一模型的注册结果，并发送第一响应消息来通知第一模型的注册是否成功。

Description

一种处理模型的方法和装置
本申请要求于2022年11月03日提交中国专利局、申请号为202211372686.8、申请名称为“一种处理模型的方法和装置”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请实施例涉及人工智能领域，更具体地，涉及一种处理模型的方法和装置。
背景技术
人工智能(artificial intelligence，AI)是利用数字计算机或者数字计算机控制的机器模拟、延伸和扩展人的智能，感知环境、获取知识并使用知识获得最佳结果的理论、方法、技术及应用系统。换句话说，人工智能是计算机科学的一个分支，它企图了解智能的实质，并生产出一种新的能以人类智能相似的方式作出反应的智能机器。人工智能也就是研究各种智能机器的设计原理与实现方法，使机器具有感知、推理与决策的功能。人工智能领域的研究包括机器人，自然语言处理，计算机视觉，决策与推理，人机交互，推荐与搜索，AI基础理论等。
发明内容
本申请提供一种处理模型的方法和装置，能够保证终端设备与模型管理网元之间的模型的统一，使得模型管理网元能够实现对终端设备的模型的高效管理，进而保证系统的性能。
第一方面，本申请实施例提供了一种处理模型的方法，该方法可以由模型管理网元或者由模型管理网元的部件(如芯片或芯片系统等)执行，本申请对此不作限定。该方法包括：模型管理网元接收第一请求消息，所述第一请求消息包括第一模型的模型信息，所述第一请求消息用于请求注册所述第一模型。所述模型管理网元根据所述第一模型的模型信息，确定所述第一模型。所述模型管理网元发送第一响应消息，所述第一响应消息用于通知所述第一模型的注册是否成功。
需要说明的是，在本申请实施例中模型的注册成功可以理解为，模型管理网元根据接收到的模型的模型信息或者模型，确定该模型支持在线注册。同时模型管理网元确定发送该模型的模型信息或者模型的终端设备具备注册该模型的权限(或者，理解为终端设备请求注册该模型是符合协议等约定的)。
基于该方案，本申请实施例提供的处理模型的方法，通过终端设备将要注册的模型的模型信息发送给模型管理网元，使得终端设备可以向模型管理网元请求注册终端设备没有的模型，从而实现终端设备与模型管理网元之间的模型的统一。
结合第一方面，在第一方面的某些实现方式中，所述第一请求消息还包括终端设备的标识信息，所述终端设备的标识信息用于所述管理网元识别所述终端设备。
其中，该终端设备的标识信息为终端设备的默认凭证。
基于该方案，模型管理网元通过终端设备的标识信息对终端设备注册第一模型的进行识别，确定终端设备是否支持在线注册该第一模型，能够提升注册过程的可靠性。
结合第一方面，在第一方面的某些实现方式中，所述方法还包括：当所述第一模型注册成功时，所述模型管理网元发送所述第一模型的模型配置信息，所述第一模型的模型配置信息包括第一模型的模型标识信息、所述第一模型的模型信息和所述第一模型中的至少一个，其中，所述第一模型的模型标识信息用于关联所述第一模型。
基于该方案，当模型注册成功时，通过模型管理网元向终端设备发送该注册成功的模型的模型标识信息，实现对该模型的高效管理。
结合第一方面，在第一方面的某些实现方式中，所述第一模型的模型配置信息携带在所述第一响应消息中。
基于该方案，通过将模型的模型配置信息通过响应消息发送，能够节约信息传递的资源，提升注 册的效率。
结合第一方面，在第一方面的某些实现方式中，所述方法还包括：当所述第一模型注册成功时，所述模型管理网元为所述第一模型分配所述第一模型的模型标识信息。
基于该方案，当模型注册成功时，通过模型管理网元分配该模型标识信息，实现对该模型的统一管理。
结合第一方面，在第一方面的某些实现方式中，当所述第一模型注册成功时，所述方法还包括：所述模型管理网元接收第二请求消息，所述第二请求消息包括第二指示信息，所述第二指示信息用于指示更新所述第一模型为第二模型。所述模型管理网元接收第一模型的模型更新信息。所述模型管理网元根据所述第二指示信息和所述第一模型的模型更新信息，确定更新结果。所述模型管理网元发送第二响应消息，所述第二响应消息包括所述更新结果，所述第二响应消息用于通知所述更新结果。
需要说明的是，在本申请实施例中模型的更新成功可以理解为，模型管理网元根据接收到的模型的模型更新信息或者模型，确定该模型支持更新。同时模型管理网元确定发送该模型的模型更新信息或者模型的终端设备具备更新该模型的权限(或者，理解为终端设备请求更新该模型是符合协议等约定的)。
基于该方案，本申请实施例提供的处理模型的方法，通过终端设备将要更新的模型的模型更新信息发送给模型管理网元，使得终端设备可以将本地更新的模型向模型管理网元请求更新到模型管理网元中，从而实现终端设备与模型管理网元之间的模型的统一。
结合第一方面，在第一方面的某些实现方式中，所述第二请求消息包括终端设备的标识信息，所述终端设备的标识信息用于所述管理网元识别所述终端设备。
基于该方案，模型管理网元通过终端设备的标识信息对终端设备进行识别，确定终端设备在是否支持在线更新该第一模型，能够提升更新过程的可靠性。
结合第一方面，在第一方面的某些实现方式中，所述第一模型的模型更新信息包括梯度信息和所述第二模型中的至少一个，以及第一模型的模型标识信息。
结合第一方面，在第一方面的某些实现方式中，所述第一模型的模型更新信息携带在所述第二请求消息中。
基于该方案，通过将模型的模型更新信息通过请求消息发送，能够节约信息传递的资源，提升更新的效率。
结合第一方面，在第一方面的某些实现方式中，所述方法还包括：当所述更新结果为成功时，所述模型管理网元发送第二模型的模型更新信息，所述第二模型的模型更新信息包括第二模型的模型标识信息，所述第二模型的模型标识信息用于关联所述第二模型。
基于该方案，当模型更新成功时，通过模型管理网元向终端设备发送该更新成功的模型的模型更新信息，实现对该模型的高效管理。
结合第一方面，在第一方面的某些实现方式中，所述方法还包括：当所述更新结果为成功时，所述模型管理网元为所述第二模型分配所述第二模型的模型标识信息。
基于该方案，当模型更新成功时，通过模型管理网元分配更新后的模型标识信息，实现对更新的模型的统一管理。
结合第一方面，在第一方面的某些实现方式中，所述第二模型的模型更新信息携带在所述第二响应消息中。
基于该方案，通过将更新后的模型的模型更新信息通过响应消息发送，能够节约信息传递的资源，提升更新的效率。
结合第一方面，在第一方面的某些实现方式中，所述模型管理网元接收第一请求消息，包括：所述模型管理网元接收来自基站的所述第一请求消息。或者，所述管理网元接收来自终端设备的所述第一请求消息。
结合第一方面，在第一方面的某些实现方式中，所述模型管理网元接收第二请求消息，包括：所述模型管理网元接收来自基站的所述第二请求消息。或者，所述模型管理网元接收来自终端设备的所述第二请求消息。
基于上述方案，通过模型管理网元接收来自不同设备的请求消息，使得模型管理网元可以部署在 系统中的不同位置，从而满足不同的应用场景。
结合第一方面，在第一方面的某些实现方式中，在所述模型管理网元接收第一请求消息之前，所述方法还包括：
所述模型管理网元发送模型注册支持信息和模型信息中的至少一个，所述模型注册支持信息用于指示接收所述模型注册支持信息的终端设备支持模型注册，所述模型信息用于指示接收所述模型信息的所述终端设备支持注册的模型。
第二方面，本申请实施例提供了一种处理模型的方法，该方法可以由模型管理网元或者由模型管理网元的部件(如芯片或芯片系统等)执行，本申请对此不作限定。该方法包括：模型管理网元接收第一请求消息，所述第一请求消息用于请求注册所述第一模型。所述模型管理网元接收所述第一模型。所述模型管理网元根据所述第一模型，确定所述第一模型支持注册。所述模型管理网元发送第一响应消息，所述第一响应消息用于通知所述第一模型的注册是否成功。
基于该方案，本申请实施例提供的处理模型的方法，通过终端设备将要注册的模型发送给模型管理网元，能够将模型管理网元没有的模型注册到模型管理网元中进行统一管理，从而实现终端设备与模型管理网元之间的模型的统一。
结合第二方面，在第二方面的某些实现方式中，所述第一请求消息还包括终端设备的标识信息，所述终端设备的标识信息用于所述管理网元识别所述终端设备。
结合第二方面，在第二方面的某些实现方式中，所述第一请求消息还包括所述第一模型。
结合第二方面，在第二方面的某些实现方式中，所述第一请求消息还包括第一指示信息，所述第一指示信息用于指示所述模型管理网元接收第一模型。
结合第二方面，在第二方面的某些实现方式中，所述方法还包括：当所述第一模型注册成功时，所述模型管理网元发送所述第一模型的模型配置信息，所述第一模型的模型配置信息包括第一模型的模型标识信息、所述第一模型的模型信息和所述第一模型中的至少一个，其中，所述第一模型的模型标识信息用于关联所述第一模型。
结合第二方面，在第二方面的某些实现方式中，所述第一模型的模型配置信息携带在所述第一响应消息中。
结合第二方面，在第二方面的某些实现方式中，所述方法还包括：当所述第一模型注册成功时，所述模型管理网元为所述第一模型分配所述第一模型的模型标识信息。
结合第二方面，在第二方面的某些实现方式中，当所述第一模型注册成功时，所述方法还包括：所述模型管理网元接收第二请求消息，所述第二请求消息包括第二指示信息，所述第二指示信息用于指示更新所述第一模型为第二模型。所述模型管理网元接收第一模型的模型更新信息。所述模型管理网元根据所述第二指示信息和所述第一模型的模型更新信息，确定更新结果。所述模型管理网元发送第二响应消息，所述第二响应消息包括所述更新结果，所述第二响应消息用于通知所述更新结果。
结合第二方面，在第二方面的某些实现方式中，所述第二请求消息包括终端设备的标识信息，所述终端设备的标识信息用于所述管理网元识别所述终端设备。
结合第二方面，在第二方面的某些实现方式中，所述第一模型的模型更新信息包括梯度信息和所述第二模型中的至少一个，以及第一模型的模型标识信息。
结合第二方面，在第二方面的某些实现方式中，所述第一模型的模型更新信息携带在所述第二请求消息中。
结合第二方面，在第二方面的某些实现方式中，所述方法还包括：当所述更新结果为成功时，所述模型管理网元发送第二模型的模型更新信息，所述第二模型的模型更新信息包括第二模型的模型标识信息，所述第二模型的模型标识信息用于关联所述第二模型。
结合第二方面，在第二方面的某些实现方式中，所述方法还包括：当所述更新结果为成功时，所述模型管理网元为所述第二模型分配所述第二模型的模型标识信息。
结合第二方面，在第二方面的某些实现方式中，所述第二模型的模型更新信息携带在所述第二响应消息中。
结合第二方面，在第二方面的某些实现方式中，所述模型管理网元接收第一请求消息，包括：所述模型管理网元接收来自基站的所述第一请求消息。或者，所述管理网元接收来自终端设备的所述第 一请求消息。
结合第二方面，在第二方面的某些实现方式中，所述模型管理网元接收第二请求消息，包括：所述模型管理网元接收来自基站的所述第二请求消息。或者，所述模型管理网元接收来自终端设备的所述第二请求消息。
结合第二方面，在第二方面的某些实现方式中，在所述模型管理网元接收第一请求消息之前，所述方法还包括：
所述模型管理网元发送模型注册支持信息和模型信息中的至少一个，所述模型注册支持信息用于指示接收所述模型注册支持信息的终端设备支持模型注册，所述模型信息用于指示接收所述模型信息的所述终端设备支持注册的模型。
第三方面，本申请实施例提供了一种处理模型的方法，该方法可以由模型管理网元或者由模型管理网元的部件(如芯片或芯片系统等)执行，本申请对此不作限定。该方法包括：模型管理网元接收第一请求消息，所述第一请求消息包括第一模型的模型信息，所述第一请求消息用于请求注册所述第一模型。所述模型管理网元接收所述第一模型。所述模型管理网元根据所述第一模型的模型信息和所述第一模型，确定所述第一模型和所述第一模型支持注册。所述模型管理网元发送第一响应消息，所述第一响应消息用于通知所述第一模型的注册是否成功。
基于该方案，本申请实施例提供的处理模型的方法，通过终端设备将要注册的模型的模型信息和模型发送给模型管理网元，能够将模型管理网元没有的模型注册到模型管理网元中进行统一管理，从而实现终端设备与模型管理网元之间的模型的统一。
结合第三方面，在第三方面的某些实现方式中，所述第一请求消息还包括终端设备的标识信息，所述终端设备的标识信息用于所述管理网元识别所述终端设备。
结合第三方面，在第三方面的某些实现方式中，所述第一请求消息还包括所述第一模型。
结合第三方面，在第三方面的某些实现方式中，所述第一请求消息还包括第一指示信息，所述第一指示信息用于指示所述模型管理网元接收第一模型。
结合第三方面，在第三方面的某些实现方式中，所述方法还包括：当所述第一模型注册成功时，所述模型管理网元发送所述第一模型的模型配置信息，所述第一模型的模型配置信息包括第一模型的模型标识信息、所述第一模型的模型信息和所述第一模型中的至少一个，其中，所述第一模型的模型标识信息用于关联所述第一模型。
结合第三方面，在第三方面的某些实现方式中，所述第一模型的模型配置信息携带在所述第一响应消息中。
结合第三方面，在第三方面的某些实现方式中，所述方法还包括：当所述第一模型注册成功时，所述模型管理网元为所述第一模型分配所述第一模型的模型标识信息。
结合第三方面，在第三方面的某些实现方式中，当所述第一模型注册成功时，所述方法还包括：所述模型管理网元接收第二请求消息，所述第二请求消息包括第二指示信息，所述第二指示信息用于指示更新所述第一模型为第二模型。所述模型管理网元接收第一模型的模型更新信息。所述模型管理网元根据所述第二指示信息和所述第一模型的模型更新信息，确定更新结果。所述模型管理网元发送第二响应消息，所述第二响应消息包括所述更新结果，所述第二响应消息用于通知所述更新结果。
结合第三方面，在第三方面的某些实现方式中，所述第二请求消息包括终端设备的标识信息，所述终端设备的标识信息用于所述管理网元识别所述终端设备。
结合第三方面，在第三方面的某些实现方式中，所述第一模型的模型更新信息包括梯度信息和所述第二模型中的至少一个，以及第一模型的模型标识信息。
结合第三方面，在第三方面的某些实现方式中，所述第一模型的模型更新信息携带在所述第二请求消息中。
结合第三方面，在第三方面的某些实现方式中，所述方法还包括：当所述更新结果为成功时，所述模型管理网元发送第二模型的模型更新信息，所述第二模型的模型更新信息包括第二模型的模型标识信息，所述第二模型的模型标识信息用于关联所述第二模型。
结合第三方面，在第三方面的某些实现方式中，所述方法还包括：当所述更新结果为成功时，所述模型管理网元为所述第二模型分配所述第二模型的模型标识信息。
结合第三方面，在第三方面的某些实现方式中，所述第二模型的模型更新信息携带在所述第二响应消息中。
结合第三方面，在第三方面的某些实现方式中，所述模型管理网元接收第一请求消息，包括：所述模型管理网元接收来自基站的所述第一请求消息。或者，所述管理网元接收来自终端设备的所述第一请求消息。
结合第三方面，在第三方面的某些实现方式中，所述模型管理网元接收第二请求消息，包括：所述模型管理网元接收来自基站的所述第二请求消息。或者，所述模型管理网元接收来自终端设备的所述第二请求消息。
结合第三方面，在第三方面的某些实现方式中，在所述模型管理网元接收第一请求消息之前，所述方法还包括：
所述模型管理网元发送模型注册支持信息和模型信息中的至少一个，所述模型注册支持信息用于指示接收所述模型注册支持信息的终端设备支持模型注册，所述模型信息用于指示接收所述模型信息的所述终端设备支持注册的模型。
第四方面，本申请实施例提供了一种处理模型的方法，该方法可以由模型管理网元或者由模型管理网元的部件(如芯片或芯片系统等)执行，本申请对此不作限定。该方法包括：模型管理网元接收第二请求消息，所述第二请求消息包括第二指示信息，所述第二指示信息用于指示更新所述第一模型为第二模型。所述模型管理网元接收第一模型的模型更新信息。所述模型管理网元根据所述第二指示信息和所述第一模型的模型更新信息，确定更新结果。所述模型管理网元发送第二响应消息，所述第二响应消息包括所述更新结果，所述第二响应消息用于通知所述更新结果。
结合第四方面，在第四方面的某些实现方式中，所述第二请求消息包括终端设备的标识信息，所述终端设备的标识信息用于所述管理网元识别所述终端设备。
结合第四方面，在第四方面的某些实现方式中，所述第一模型的模型更新信息包括梯度信息和所述第二模型中的至少一个，以及第一模型的模型标识信息。
结合第四方面，在第四方面的某些实现方式中，所述第一模型的模型更新信息携带在所述第二请求消息中。
结合第四方面，在第四方面的某些实现方式中，所述方法还包括：当所述更新结果为成功时，所述模型管理网元发送第二模型的模型更新信息，所述第二模型的模型更新信息包括第二模型的模型标识信息，所述第二模型的模型标识信息用于关联所述第二模型。
结合第四方面，在第四方面的某些实现方式中，所述方法还包括：当所述更新结果为成功时，所述模型管理网元为所述第二模型分配所述第二模型的模型标识信息。
结合第四方面，在第四方面的某些实现方式中，所述第二模型的模型更新信息携带在所述第二响应消息中。
结合第四方面，在第四方面的某些实现方式中，所述模型管理网元接收第二请求消息，包括：所述模型管理网元接收来自基站的所述第二请求消息。或者，所述模型管理网元接收来自终端设备的所述第二请求消息。
第五方面，本申请实施例提供了一种处理模型的方法，该方法可以由终端设备或者由终端设备的部件(如芯片或芯片系统等)执行，本申请对此不作限定。该方法包括：终端设备发送第一请求消息，所述第一请求消息包括第一模型的模型信息，所述第一请求消息用于请求注册所述第一模型。所述终端设备接收第一响应消息，所述第一响应消息用于通知所述第一模型的注册是否成功。
结合第五方面，在第五方面的某些实现方式中，所述第一请求消息还包括所述终端设备的标识信息，所述终端设备的标识信息用于管理网元识别所述终端设备。
结合第五方面，在第五方面的某些实现方式中，所述方法还包括：
当所述第一模型注册成功时，所述终端设备接收所述第一模型的模型配置信息，所述第一模型的模型配置信息包括第一模型的模型标识信息、所述第一模型的模型信息和所述第一模型中的至少一个，其中，所述第一模型的模型标识信息用于关联所述第一模型。
结合第五方面，在第五方面的某些实现方式中，所述第一模型的模型配置信息携带在所述第一响应消息中。
结合第五方面，在第五方面的某些实现方式中，当所述第一模型注册成功时，所述方法还包括：所述终端设备发送第二请求消息，所述第二请求消息包括第二指示信息，所述第二指示信息用于指示更新所述第一模型为第二模型。所述终端设备发送第一模型的模型更新信息。所述终端设备接收第二响应消息，所述第二响应消息包括所述更新结果，所述第二响应消息用于通知所述更新结果。
结合第五方面，在第五方面的某些实现方式中，所述第二请求消息包括所述终端设备的标识信息，所述终端设备的标识信息用于管理网元识别所述终端设备。
结合第五方面，在第五方面的某些实现方式中，所述第一模型的模型更新信息包括梯度信息和所述第二模型中的至少一个，以及第一模型的模型标识信息。
结合第五方面，在第五方面的某些实现方式中，所述第一模型的模型更新信息携带在所述第二请求消息中。
结合第五方面，在第五方面的某些实现方式中，所述方法还包括：当所述更新结果为成功时，所述终端设备接收第二模型的模型更新信息，所述第二模型的模型更新信息包括第二模型的模型标识信息，所述第二模型的模型标识信息用于关联所述第二模型。
结合第五方面，在第五方面的某些实现方式中，所述第二模型的模型更新信息携带在所述第二响应消息中。
结合第五方面，在第五方面的某些实现方式中，所述终端设备接收第一响应消息，包括：所述终端设备接收来自基站的所述第一响应消息。或者，所述终端设备接收来自模型管理网元的所述第一响应消息。
结合第五方面，在第五方面的某些实现方式中，所述终端设备接收所述第二响应消息，包括：所述终端设备接收来自基站的所述第二响应消息。或者，所述终端设备接收来自模型管理网元的所述第二响应消息。
第六方面，本申请实施例提供了一种处理模型的方法，该方法可以由终端设备或者由终端设备的部件(如芯片或芯片系统等)执行，本申请对此不作限定。该方法包括：终端设备发送第一请求消息，所述第一请求消息用于请求注册所述第一模型。所述终端设备发送所述第一模型。所述终端设备接收第一响应消息，所述第一响应消息用于通知所述第一模型的注册是否成功。
结合第六方面，在第六方面的某些实现方式中，所述第一请求消息还包括所述终端设备的标识信息，所述终端设备的标识信息用于管理网元识别所述终端设备。
结合第六方面，在第六方面的某些实现方式中，所述第一请求消息还包括所述第一模型。
结合第六方面，在第六方面的某些实现方式中，所述第一请求消息还包括第一指示信息，所述第一指示信息用于指示所述模型管理网元接收第一模型。
结合第六方面，在第六方面的某些实现方式中，所述方法还包括：当所述第一模型注册成功时，所述终端设备接收所述第一模型的模型配置信息，所述第一模型的模型配置信息包括第一模型的模型标识信息、所述第一模型的模型信息和所述第一模型中的至少一个，其中，所述第一模型的模型标识信息用于关联所述第一模型。
结合第六方面，在第六方面的某些实现方式中，所述第一模型的模型配置信息携带在所述第一响应消息中。
结合第六方面，在第六方面的某些实现方式中，当所述第一模型注册成功时，所述方法还包括：所述终端设备发送第二请求消息，所述第二请求消息包括第二指示信息，所述第二指示信息用于指示更新所述第一模型为第二模型。所述终端设备发送第一模型的模型更新信息。所述终端设备接收第二响应消息，所述第二响应消息包括所述更新结果，所述第二响应消息用于通知所述更新结果。
结合第六方面，在第六方面的某些实现方式中，所述第二请求消息包括所述终端设备的标识信息，所述终端设备的标识信息用于管理网元识别所述终端设备。
结合第六方面，在第六方面的某些实现方式中，所述第一模型的模型更新信息包括梯度信息和所述第二模型中的至少一个，以及第一模型的模型标识信息。
结合第六方面，在第六方面的某些实现方式中，所述第一模型的模型更新信息携带在所述第二请求消息中。
结合第六方面，在第六方面的某些实现方式中，所述方法还包括：当所述更新结果为成功时，所 述终端设备接收第二模型的模型更新信息，所述第二模型的模型更新信息包括第二模型的模型标识信息，所述第二模型的模型标识信息用于关联所述第二模型。
结合第六方面，在第六方面的某些实现方式中，所述第二模型的模型更新信息携带在所述第二响应消息中。
结合第六方面，在第六方面的某些实现方式中，所述终端设备接收第一响应消息，包括：所述终端设备接收来自基站的所述第一响应消息。或者，所述终端设备接收来自模型管理网元的所述第一响应消息。
结合第六方面，在第六方面的某些实现方式中，所述终端设备接收所述第二响应消息，包括：所述终端设备接收来自基站的所述第二响应消息。或者，所述终端设备接收来自模型管理网元的所述第二响应消息。
第七方面，本申请实施例提供了一种处理模型的方法，该方法可以由终端设备或者由终端设备的部件(如芯片或芯片系统等)执行，本申请对此不作限定。该方法包括：终端设备发送第一请求消息，所述第一请求消息包括第一模型的模型信息，所述第一请求消息用于请求注册所述第一模型。所述终端设备发送所述第一模型。所述终端设备接收第一响应消息，所述第一响应消息用于通知所述第一模型的注册是否成功。
结合第七方面，在第七方面的某些实现方式中，所述第一请求消息还包括所述终端设备的标识信息，所述终端设备的标识信息用于管理网元识别所述终端设备。
结合第七方面，在第七方面的某些实现方式中，所述第一请求消息还包括所述第一模型。
结合第七方面，在第七方面的某些实现方式中，所述第一请求消息还包括第一指示信息，所述第一指示信息用于指示所述模型管理网元接收第一模型。
结合第七方面，在第七方面的某些实现方式中，所述方法还包括：
当所述第一模型注册成功时，所述终端设备接收所述第一模型的模型配置信息，所述第一模型的模型配置信息包括第一模型的模型标识信息、所述第一模型的模型信息和所述第一模型中的至少一个，其中，所述第一模型的模型标识信息用于关联所述第一模型。
结合第七方面，在第七方面的某些实现方式中，所述第一模型的模型配置信息携带在所述第一响应消息中。
结合第七方面，在第七方面的某些实现方式中，当所述第一模型注册成功时，所述方法还包括：所述终端设备发送第二请求消息，所述第二请求消息包括第二指示信息，所述第二指示信息用于指示更新所述第一模型为第二模型。所述终端设备发送第一模型的模型更新信息。所述终端设备接收第二响应消息，所述第二响应消息包括所述更新结果，所述第二响应消息用于通知所述更新结果。
结合第七方面，在第七方面的某些实现方式中，所述第二请求消息包括所述终端设备的标识信息，所述终端设备的标识信息用于管理网元识别所述终端设备。
结合第七方面，在第七方面的某些实现方式中，所述第一模型的模型更新信息包括梯度信息和所述第二模型中的至少一个，以及第一模型的模型标识信息。
结合第七方面，在第七方面的某些实现方式中，所述第一模型的模型更新信息携带在所述第二请求消息中。
结合第七方面，在第七方面的某些实现方式中，所述方法还包括：当所述更新结果为成功时，所述终端设备接收第二模型的模型更新信息，所述第二模型的模型更新信息包括第二模型的模型标识信息，所述第二模型的模型标识信息用于关联所述第二模型。
结合第七方面，在第七方面的某些实现方式中，所述第二模型的模型更新信息携带在所述第二响应消息中。
结合第七方面，在第七方面的某些实现方式中，所述终端设备接收第一响应消息，包括：所述终端设备接收来自基站的所述第一响应消息。或者，所述终端设备接收来自模型管理网元的所述第一响应消息。
结合第七方面，在第七方面的某些实现方式中，所述终端设备接收所述第二响应消息，包括：所述终端设备接收来自基站的所述第二响应消息。或者，所述终端设备接收来自模型管理网元的所述第二响应消息。
第八方面，本申请实施例提供了一种处理模型的方法，该方法可以由终端设备或者由终端设备的部件(如芯片或芯片系统等)执行，本申请对此不作限定。该方法包括：终端设备发送第二请求消息，所述第二请求消息包括第二指示信息，所述第二指示信息用于指示更新所述第一模型为第二模型。所述终端设备发送第一模型的模型更新信息。所述终端设备接收第二响应消息，所述第二响应消息包括所述更新结果，所述第二响应消息用于通知所述更新结果。
结合第八方面，在第八方面的某些实现方式中，所述第二请求消息包括所述终端设备的标识信息，所述终端设备的标识信息用于管理网元识别所述终端设备。
结合第八方面，在第八方面的某些实现方式中，所述第一模型的模型更新信息包括梯度信息和所述第二模型中的至少一个，以及第一模型的模型标识信息。
结合第八方面，在第八方面的某些实现方式中，所述第一模型的模型更新信息携带在所述第二请求消息中。
结合第八方面，在第八方面的某些实现方式中，所述方法还包括：当所述更新结果为成功时，所述终端设备接收第二模型的模型更新信息，所述第二模型的模型更新信息包括第二模型的模型标识信息，所述第二模型的模型标识信息用于关联所述第二模型。
结合第八方面，在第八方面的某些实现方式中，所述第二模型的模型更新信息携带在所述第二响应消息中。
结合第八方面，在第八方面的某些实现方式中，所述终端设备接收所述第二响应消息，包括：所述终端设备接收来自基站的所述第二响应消息。或者，所述终端设备接收来自模型管理网元的所述第二响应消息。
第九方面，本申请实施例提供了一种处理模型的装置。该装置用于执行上述第一方面或第二方面或第三方面或第四方面提供的方法。具体地，该处理模型的装置可以包括用于执行第一方面或第一方面的上述任意一种实现方式提供的方法的单元和/或模块，如处理单元和获取单元。或者，该处理模型的装置可以包括用于执行第二方面或第二方面的上述任意一种实现方式提供的方法的单元和/或模块，如处理单元和获取单元。或者，该处理模型的装置可以包括用于执行第三方面或第三方面的上述任意一种实现方式提供的方法的单元和/或模块，如处理单元和获取单元。或者，该处理模型的装置可以包括用于执行第四方面或第四方面的上述任意一种实现方式提供的方法的单元和/或模块，如处理单元和获取单元。
在一种实现方式中，该处理模型的装置为模型管理网元。获取单元可以是收发器，或，输入/输出接口；处理单元可以是至少一个处理器。可选地，收发器可以为收发电路。可选地，输入/输出接口可以为输入/输出电路。
在另一种实现方式中，该处理模型的装置为模型管理网元中的芯片、芯片系统或电路。获取单元可以是该芯片、芯片系统或电路上的输入/输出接口、接口电路、输出电路、输入电路、管脚或相关电路等；处理单元可以是至少一个处理器、处理电路或逻辑电路等。
第十方面，本申请实施例提供了一种处理模型的装置，该装置用于执行上述第五方面或第六方面或第七方面或第八方面提供的方法。具体地，该处理模型的装置可以包括用于执行第五方面或第五方面的上述任意一种实现方式提供的方法的单元和/或模块，如处理单元和获取单元。或者，该处理模型的装置可以包括用于执行第六方面或第六方面的上述任意一种实现方式提供的方法的单元和/或模块，如处理单元和获取单元。或者，该处理模型的装置可以包括用于执行第七方面或第七方面的上述任意一种实现方式提供的方法的单元和/或模块，如处理单元和获取单元。或者，该处理模型的装置可以包括用于执行第八方面或第八方面的上述任意一种实现方式提供的方法的单元和/或模块，如处理单元和获取单元。
在一种实现方式中，该处理模型的装置为模型管理网元。获取单元可以是收发器，或，输入/输出接口；处理单元可以是至少一个处理器。可选地，收发器可以为收发电路。可选地，输入/输出接口可以为输入/输出电路。
在另一种实现方式中，该处理模型的装置为终端设备中的芯片、芯片系统或电路。获取单元可以是该芯片、芯片系统或电路上的输入/输出接口、接口电路、输出电路、输入电路、管脚或相关电路等；处理单元可以是至少一个处理器、处理电路或逻辑电路等。
第十一方面，本申请实施例提供了一种处理器，用于执行上述各方面提供的方法。
对于处理器所涉及的发送和获取/接收等操作，如果没有特殊说明，或者，如果未与其在相关描述中的实际作用或者内在逻辑相抵触，则可以理解为处理器输出和接收、输入等操作，也可以理解为由射频电路和天线所进行的发送和接收操作，本申请对此不做限定。
第十二方面，本申请实施例提供了一种计算机可读存储介质。该计算机可读存储介质存储用于设备执行的程序代码，该程序代码包括用于执行上述第一方面至第八方面中的任意一方面，以及第一方面至第八方面的任意一方面中的任意一种实现方式提供的方法。
第十三方面，本申请实施例提供了一种包含指令的计算机程序产品。当该计算机程序产品在计算机上运行时，使得计算机执行上述第一方面至第八方面中的任意一方面，以及第一方面至第八方面的任意一方面中的任意一种实现方式提供的方法。
第十四方面，本申请实施例提供了一种芯片，芯片包括处理器与通信接口，处理器通过通信接口读取存储器上存储的指令，执行上述第一方面至第八方面中的任意一方面，以及第一方面至第八方面的任意一方面中的任意一种实现方式提供的方法。
可选地，作为一种实现方式，芯片还包括存储器，存储器中存储有计算机程序或指令，处理器用于执行存储器上存储的计算机程序或指令，当计算机程序或指令被执行时，处理器用于执行上述第一方面至第八方面中的任意一方面，以及第一方面至第八方面的任意一方面中的任意一种实现方式提供的方法。
第十五方面，本申请实施例提供了一种处理模型的系统，包括第九方面所述的处理模型的装置和第十方面所述的处理模型的装置。
上述第二方面至第十五方面带来的有益效果具体可以参考第一方面中有益效果的描述，此处不再赘述。
附图说明
图1是本申请实施例适用于本申请实施例的一种AI模块的示意图。
图2是本申请实施例提供的第一种处理模型方法200的示意性流程图。
图3是本申请实施例提供的第二种处理模型方法300的示意性流程图。
图4是本申请实施例提供的第三种处理模型方法400的示意性流程图。
图5是本申请实施例提供的第四种处理模型方法500的示意性流程图。
图6是本申请实施例提供的第五种处理模型方法600的示意性流程图。
图7是本申请实施例提供的第六种处理模型方法700的示意性流程图。
图8是本申请实施例提供的第七种处理模型方法800的示意性框图。
图9是本申请实施例提供的第一种处理模型的装置900的示意性框图。
图10是本申请实施例提供的第二种处理模型的装置1000的示意性框图。
图11是本申请实施例提供的一例模型管理网元的结构示意图。
图12是本申请实施例提供的一例终端设备的结构示意图。
具体实施方式
下面将结合附图，对本申请实施例中的技术方案进行描述。
以下实施例中所使用的术语只是为了描述特定实施例的目的，而并非旨在作为对本申请的限制。如在本申请的说明书和所附权利要求书中所使用的那样，单数表达形式“一个”、“一种”和“该”旨在包括例如“一个或多个”这种表达形式，除非其上下文中明确地有相反指示。还应当理解，在本申请以下各实施例中，“至少一个”、“至少一项”、“一个或多个”是指一个、两个或两个以上。“第一”、“第二”以及各种数字编号只是为了描述方便进行的区分，并不用来限制本申请实施例的范围。“和/或”，用于描述对应对象的对应关系，表示可以存在三种关系，例如，“A和/或B”可以表示：只存在A，只存在B以及同时存在A和B三种情况，其中A，B可以是单数或复数。字符“/”一般表示前后关联对象是一种“或”的关系。下文各过程的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本申请实施例的实施过程构成任何限定。例如，本申请实施例中，“301”、“401”、 “501”等字样仅为了描述方便作出的标识，并不是对执行步骤的次序进行限定。
在本说明书中描述的参考“一个实施例”或“一些实施例”等意味着在本申请的一个或多个实施例中包括结合该实施例描述的特定特征、结构或特点。本申请中，“示例性的”或者“例如”等词用于表示作例子、例证或说明。本申请中被描述为“示例性的”或者“例如”的任何实施例或设计方案不应被解释为比其他实施例或设计方案更优选或更具优势。确切而言，使用“示例性的”或者“例如”等词旨在以具体方式呈现相关概念。术语“包括”、“包含”、“具有”及它们的变形都意味着“包括但不限于”，除非是以其他方式另外特别强调。在本申请实施例中，“当……时”、“在……的情况下”、“若”以及“如果”等描述均指在某种客观情况下设备会做出相应的处理，并非是限定时间，且也不要求设备在实现时一定要有判断的动作，也不意味着存在其它限定。
在本申请中，“用于指示”可以包括用于直接指示和用于间接指示。当描述某一指示信息用于指示A时，可以包括该指示信息直接指示A或间接指示A，而并不代表该指示信息中一定携带有A。在本申请实施例中，“当……时”、“在……的情况下”、“若”以及“如果”等描述均指在某种客观情况下设备会做出相应的处理，并非是限定时间，且也不要求设备在实现时一定要有判断的动作，也不意味着存在其它限定。
本申请实施例的技术方案可以应用于各种通信系统，例如：长期演进(long term evolution，LTE)系统、LTE频分双工(frequency division duplex，FDD)系统、LTE时分双工(time division duplex，TDD)、通用移动通信系统(universal mobile telecommunication system，UMTS)、全球互联微波接入(worldwide interoperability for microwave access，Wi-MAX)通信系统、第五代(5th generation，5G)系统或新空口(new radio，NR)、未来的第六代(6th generation，6G)系统、星间通信和卫星通信等非陆地通信网络(non-terrestrial network，NTN)系统。卫星通信系统包括卫星基站以及终端设备。该卫星基站为终端设备提供通信服务。卫星基站也可以与基站进行通信。卫星可作为基站，也可作为终端设备。其中，卫星可以是指无人机，热气球，低轨卫星，中轨卫星，高轨卫星等。卫星也可以是指非地面基站或非地面设备等。
本申请的实施例可以应用于终端设备。终端设备可以是无线终端也可以是有线终端，无线终端可以是指向用户提供语音和/或数据连通性的设备，具有无线连接功能的手持式设备、或连接到无线调制解调器的其他处理设备。无线终端可以经无线接入网(radio access cetwork；缩写：RAN)与一个或多个核心网进行通信，无线终端可以是移动终端，如移动电话(或称为“蜂窝”电话)和具有移动终端的计算机，例如，可以是便携式、袖珍式、手持式、计算机内置的或者车载的移动装置，它们与无线接入网交换语言和/或数据。例如，个人通信业务(personal communication service，PCS)电话、无绳电话、会话发起协议(session initiation protocol，SIP)话机、无线本地环路(wireless local loop，WLL)站、个人数字助理(personal digital assistant，PDA)等设备。无线终端也可以称为系统、订户单元(subscriber unit，SU)、订户站(subscriber station，SS)，移动站(mobile station，MB)、移动台(mobile)、远程站(remote station，RS)、接入点(access point，AP)、远程终端(remote terminal，RT)、接入终端(access terminal，AT)、用户终端(user terminal，UT)、用户代理(user agent，UA)、终端设备(user device，UD)、或用户装备(user equipment，UE)。
本申请实施例中用于实现终端设备的功能的装置可以是终端设备；也可以是能够支持终端设备实现该功能的装置，例如芯片系统。该装置可以被安装在终端设备中或者和终端设备匹配使用。本申请实施例中，芯片系统可以由芯片构成，也可以包括芯片和其他分立器件。
本申请实施例中的技术方案还可以应用于接入网设备。接入网设备可以是能够将终端设备接入到无线网络的设备。该接入网设备还可以称为无线接入网(radio access network，RAN)节点、无线接入网设备、网络设备。示例性的，该接入网设备可以是基站。
本申请实施例中的基站可以广义的覆盖如下中的各种名称，或与如下名称进行替换，比如：节点B(NodeB)、演进型基站(evolved NodeB，eNB)、5G网络中的基站gNB、中继站、接入点、收发点(transmitting and receiving point，TRP)、发射点(transmitting point，TP)、主站(master eNodeB，MeNB)、辅站(secondary eNodeB，SeNB)、多制式无线(multi standard radio，MSR)节点、家庭基站、网络控制器、接入节点、无线节点、接入点(access point，AP)、传输节点、收发节点、基带单元(base band unit，BBU)、射频拉远单元(remote radio unit，RRU)、有源天线单元(active antenna  unit，AAU)、射频头(remote radio head，RRH)、定位节点等。基站可以是宏基站、微基站、中继节点、施主节点或类似物，或其组合。基站还可以指用于设置于前述设备或装置内的通信模块、调制解调器或芯片。基站还可以是6G网络中的网络侧设备、未来的通信系统中承担基站功能的设备等。基站可以支持相同或不同接入技术的网络。
基站可以是集中式单元(centralized unit，CU)和分布式单元(distributed unit，DU)分离架构。RAN可以与核心网相连(例如可以是长期演进(long term evolution，LTE)的核心网，也可以是5G的核心网等)。CU和DU可以理解为是对基站从逻辑功能角度的划分。CU和DU在物理上可以是分离的也可以部署在一起。多个DU可以共用一个CU。一个DU也可以连接多个CU。CU和DU之间可以通过接口相连，例如可以是F1接口。CU和DU可以根据无线网络的协议层划分。例如其中一种可能的划分方式是：CU用于执行无线资源控制(radio resource control，RRC)层、业务数据适配协议(service data adaptation protocol，SDAP)层以及分组数据汇聚层协议(packet data convergence protocol，PDCP)层的功能，而DU用于执行无线链路控制(radio link control，RLC)层，媒体接入控制(media access control，MAC)层，物理(physical)层等的功能。可以理解对CU和DU处理功能按照这种协议层的划分仅仅是一种举例，也可以按照其他的方式进行划分。例如可以将CU或者DU划分为具有更多协议层的功能。例如，CU或DU还可以划分为具有协议层的部分处理功能。在一设计中，将RLC层的部分功能和RLC层以上的协议层的功能设置在CU，将RLC层的剩余功能和RLC层以下的协议层的功能设置在DU。在另一种设计中，还可以按照业务类型或者其他系统需求对CU或者DU的功能进行划分。例如按时延划分，将处理时间需要满足时延要求的功能设置在DU，不需要满足该时延要求的功能设置在CU。在另一种设计中，CU也可以具有核心网的一个或多个功能。一个或者多个CU可以集中设置，也分离设置。例如CU可以设置在网络侧方便集中管理。DU可以具有多个射频功能，也可以将射频功能拉远设置。
CU的功能可以由一个实体来实现也可以由不同的实体实现。例如，可以对CU的功能进行进一步切分，例如，将控制面(control plane，CP)和用户面(user plane，UP)分离，即CU的控制面(CU-CP)和CU用户面(CU-UP)。例如，CU-CP和CU-UP可以由不同的功能实体来实现，并通过E1接口相连，所述CU-CP和CU-UP可以与DU相耦合，共同完成基站的功能。CU的控制面CU-CP还包括一种进一步切分的架构，即把现有的CU-CP进一步切分为CU-CP1和CU-CP2。其中CU-CP1包括各种无线资源管理功能，CU-CP2仅包括RRC功能和PDCP-C功能(即控制面信令在PDCP层的基本功能)。
基站可以是固定的，也可以是移动的。例如，直升机或无人机可以被配置成充当移动基站，一个或多个小区可以根据该移动基站的位置移动。在其他示例中，直升机或无人机可以被配置成用作与另一基站通信的设备。
本申请实施例中的技术方案还可以应用于操作维护管理(operation administration and maintenance，OAM)。OAM是指根据运营商网络运营的实际需要，通常将网络的管理工作划分为3大类：操作(operation)、管理(administration)、维护(maintenance)，简称OAM，OAM也可以称为OAM实体或功能。操作主要完成日常网络和业务进行的分析、预测、规划和配置工作；维护主要是对网络及其业务的测试和故障管理等进行的日常操作活动，OAM可以检测网络运行状态、优化网络连接和性能，提升网络运行稳定性，降低网络维护成本。
本申请实施例中的技术方案还可以应用于第三方应用服务，例如，过顶传球(over the top，OTT)，OTT是指通过互联网向用户提供各种应用服务。这种应用和目前运营商所提供的通信业务不同，它仅利用运营商的网络，而服务由运营商之外的第三方提供。目前，典型的OTT业务有互联网电视业务，应用商店等。
本申请实施例中的技术方案可以应用于网络设备。网络设备可以为基站、核心网设备。其中，核心网设备可以包括移动管理模块(mobility management entity，MME)、接入和移动性管理功能(access and mobility management function，AMF)网元、网络数据分析功能(network data analytics function，NWDAF)网元、会话管理功能(session management function，SMF)网元，或者其他核心网网络设备。
AI模型也即AI算法(或者AI算子)，是基于人工智能的原理构建的数学算法的统称，也是利用AI解决特定问题的基础。本申请实施例中对AI模型的类型并不限定。例如，AI模型可以是机器学习模型，也可以是深度学习模型，也可以是强化学习模型，还可以是联邦学习等。
机器学习是实现人工智能的一种方法，该方法的目标是设计和分析一些让计算机可以自动“学习”的算法(也即模型)，所设计的算法称为机器学习模型。机器学习模型是一类从数据中自动分析获得规律，并利用规律对未知数据进行预测的算法。机器学习模型包括多种多样，根据模型训练时是否需要依赖训练数据对应的标签，机器学习模型可以分为：1、有监督学习模型；2、无监督学习模型。
深度学习是机器学习研究过程中产生的一个新的技术领域，具体地，深度学习是机器学习中一种基于对数据进行深层次表征学习的方法，深度学习通过建立模拟人脑进行分析学习的神经网络来解释数据。由于在机器学习方法中，几乎所有的特征都需要通过行业专家确定，然后对特征进行编码。然而深度学习算法试图自己从数据中学习特征，根据深度学习思想设计的算法称为深度学习模型。
强化学习是机器学习中的一个特殊领域，是通过智能体(agent)和环境(environment)的相互作用，不断学习最优策略，做出序列决策，并获得最大回报的过程。通俗言之，强化学习是学习“做什么(即如何把当前的情景映射成动作)才能使得数值化的收益信号最大化”。智能体不会被告知应该采取什么动作，而是必须自己通过尝试去发现哪些动作会产生最丰厚的收益。强化学习与机器学习领域中的有监督学习和无监督学习不同，有监督学习是从外部提供的带标签的训练数据中进行学习的过程(任务驱动型)，无监督学习是寻找未标注的数据中隐含结构的过程(数据驱动型)。强化学习是通过“试探”寻找更优解的过程。智能体必须开发已有的经验来获取收益，同时也要进行试探，使得未来可以获得更好的动作选择空间(即从错误中学习)。根据强化学习设计的算法称为强化学习模型。
联邦学习(也称为协作学习)是一种机器学习技术，它可以在多个分散的边缘设备或持有本地数据样本的服务器上训练算法，而不交换它们。这种方法与传统的集中式机器学习技术形成了鲜明对比，即集中式机器学习中所有本地数据集都上传到一台服务器从而完成训练。联邦学习使多个参与者能够在不共享数据的情况下构建一个共同的、健壮的机器学习模型，从而允许解决数据隐私、数据安全、数据访问权限和访问异构数据等关键问题。
任何一个AI模型在用于解决特定的技术问题之前，都需要经过训练。AI模型的训练是指利用指定初始模型对训练数据进行计算，根据计算的结果采用一定的方法对初始模型中的参数进行调整，使得该模型逐渐学习到一定的规律，具备特定的功能的过程。经过训练后具有稳定功能的AI模型即可用于推理。AI模型的推理是利用训练完成的AI模型对输入数据进行计算，获得预测的推理结果(也可以称为输出数据)的过程。
AI模块为具备AI学习计算能力的模块。在无线通信系统中，AI模块可位于OAM中，也可位于gNB中(分离架构位于CU中)，可以位于部分UE中，也可以单独成为一个网元实体。AI模块在无线通信系统中主要功能为根据输入数据(例如，在无线通信系统中，输入数据可以是RAN侧提供的、或OAM监测的网络运行数据，例如网络负载、信道质量等)，进行模型建立、训练逼近、强化学习等一系列AI计算。AI模块提供的已训练完成的模型，具备针对RAN侧网络变化的预测功能，通常可以用于负载预测，UE轨迹预测等。此外，AI模块还可以根据训练完成的模型对RAN网络性能的预测结果，从网络节能、移动性优化等角度进行策略推理，以得到合理高效的节能策略、移动性优化策略等。当AI模块位于OAM中时，其与RAN侧gNB的通信，可以复用当前的北向接口；当AI模块位于gNB或CU中时，可以复用当前的F1、Xn、Uu等接口；当AI模块独立成一个网络实体时，需要重新建立到OAM和RAN侧等的通信链路，例如基于有线链路，或无线链路。
图1是本申请实施例适用于本申请实施例的一种AI模块的示意图。如图1所示的AI模块100包括数据库模块101，训练模块102，模型模块103和执行模块104。
数据库模块101可以存储训练数据。训练数据还可以来自于终端设备。训练数据可以来自于网络设备。例如，训练数据可以来自于基站(例如gNB)或者组成基站的功能单元(例如CU或DU)。又如，训练数据可以来自于除基站以外的其他网络设备。例如，网关、管理实体(例如移动管理(mobile management entity，MME)、核心网设备等。
训练模块102对数据库模块101提供的训练数据进行分析，得到AI模型。训练模块102可以将训练好的AI模型发送给模型模块103。在完成AI模型训练后，训练模块102还可以更新已训练好的模型，并将用于更新模型的更新参数发送给模型模块103。模型模块103在运行该AI模型的过程中还可以收集一些模型的运行数据，并将这些运行数据发送给训练模块102。训练模块102可以根据这些运行数据更新该AI模型。
模型模块103可以根据该AI模型和输入数据，确定输出数据。输出数据可以包括基于输入数据和AI模型得到的网络运行的预测结果。输出数据还可以包括根据输入数据和AI模型确定的调整策略。在一些实施例中，网络设备和/或终端设备可以直接将输入数据发送给模型模块103。在另一些实施例中，数据库模块101也可以收集来自于网络设备和/或终端设备的数据，确定输入数据并将输入数据发送给模型模块103。
执行模块104可以用于执行模型模块103确定的调整策略。执行模块104还可以收集应用了该调整策略后的网络的具体表现，例如网络中的性能参数等，并将这些信息反馈给数据模块101。该数据库模块101可以存储这些反馈信息。这些反馈信息可以用于后续的模型训练或者改善该AI模型。
下面对于AI移动通信系统中的应用介绍。在移动通信系统中，利用AI模型可以实现智能收集和分析数据，提升网络性能和用户体验。
例如，AI可以应用于信道状态信息(channel state information，CSI)反馈增强(CSI feedback enhancement)。CSI是通信链路的信道属性，是终端设备上报给基站的信道质量信息。终端设备通过将下行信道质量信息上报给基站，以便为终端设备选择更加合适的调制编码方案(modulation and coding scheme，MCS)，这样可以更好地适应变换的无线信道。
又如，AI还可以应用于波束管理(beam management，BM)。BM主要是用于要发现最强的发射/接收波束对(beam pair)。基于AI的波束预测可以提升预测准确度。
又如，AI还可以应用于定位精度增强(positioning accuracy enhancements)。定位精度和全向辐射功率(total radiated power，TRP)天线数量相关。一般而言，TRP天线数量越多，定位精度越高。利用AI模型可以在较少TRP天线数量的情况下实现较高的定位精度。例如，传统的定位手段，比如到达时间差(time difference of arrival,TDOA)、往返时间(round trip time,RTT)等，依赖于终端设备和TRP天线之间视距(line of sight，LOS)路径的信息收集。对于室内场景，可能没有足够数量的视距路径，传统的定位手段无法很好地工作。因此，基于AI的定位可以改进视距路径较少的场景下的定位准确度。
又如，AI还可以应用于网络节能(network energy saving)。网络节能可以通过小区激活/去激活(cell activation/deactivation)、减少负载、改进覆盖或者其他RAN设置调整。最优的节能决策取决于不同RAN节点的负载情况、RAN节点能力、关键绩效指标(key performance indication，KPI)要求、服务质量(quality of service，QoS)要求、激活用户数和终端设备的移动性、小区利用率等因素。然而提升网络能效是很复杂的过程，错误的小区关闭以及错误的流量卸载操作等都会引起网络性能的下降甚至能效的下降。AI技术可用于通过利用在RAN网络中收集的数据来优化节能决策。AI算法可以预测下一个周期的能效和负载状态，这可以用于更好地决策小区激活/去激活，以节省能源。基于预测的负载，系统可以动态配置节能策略，以保持系统性能和能效之间的平衡，并降低能耗。
又如，AI还可以应用于负载均衡(load balancing)。负载均衡的目的是使得负载在小区之间和小区内各区域之间均匀分布，或将部分流量从拥塞小区转移，或让终端设备在一个小区、载波或接入制式上进行分流，以提高网络性能。这可以通过优化切换参数和切换动作来实现。这种优化的自动化可以提供高质量的用户体验，同时提高系统容量，并最大限度地减少对网络管理和优化任务的人工干预。目前，依赖当前/过去时刻的小区负载状态的负载均衡决策是不够的。另外，负载均衡时整体网络和业务性能难以保证。因此可以引入基于AI模型的解决方案来提高负载均衡性能，如将用户和网络节点的各种测量和反馈、历史数据等输入AI模型来提高负载均衡性能，以提供更高质量的用户体验，提高系统容量。
又如，AI还可以应用于移动性优化(mobility optimization)。移动性管理是通过最大限度地减少掉话、无线链路失败(radio link failure，RLF)、不必要的切换和乒乓效应来保证终端设备移动期间业务连续性的方案。对于未来的高频网络，随着单节点覆盖区域的减少，终端设备在节点之间切换的频率会变得很高，尤其是对于高移动性终端设备。此外，对于可靠性、时延等QoS要求严格的应用，体验质量(quality of experience，QoE)对切换的性能较为敏感，因此移动性管理应避免切换失败，并减少切换过程中的时延。然而，对于传统的方法来说，基于试错的方案要实现几乎零失败的切换是有挑战性的。因此可以利用基于AI的解决方案来在以下几方面增强移动性管理：1)降低意外事件发生的概率，2)终端设备位置/移动性/性能预测，3)流量引导。
以下结合附图详细说明本申请所提供的处理模型的方法。由于模型管理网元可以位于与终端设备构成的无线通信系统中的不同的网络设备之中(例如基站、核心网设备等)，或者位于OAM中，或者位于第三方提供的应用服务OTT中，或者模型管理网元作为一个独立的网元等，以实现模型管理的功能。应理解，上述提到的模型管理网元所在的任何一个网元或实体或应用服务与终端设备之间均可以基于本申请所提供的处理模型的方法与终端设备进行信息交互。以下，模型管理网元位于基站(对应下文方法200、方法500、)，或者模型管理网元位于AMF/MME/模型管理网元作为单独的与基站连接的网元(对应下文方法300、方法600)或者模型管理网元位于OAM/OTT/SMF/NWDAF(对应下文方法400、方法700)之间的交互过程为例详细说明本申请实施例提供的处理模型的方法。
需要说明的是，在本申请实施例中，第一模型为终端设备请求注册的模型。第二模型为将第三模型更新后的模型。应理解，第三模型在更新前，也可以是终端请求注册的模型。换句话说，当第三模型更新为第二模型之前，该第三模型可以是终端设备请求注册的模型。
此外，在本申请实施例中，涉及模型信息、模型信息以及模型更新信息等，其中，模型信息又涉及模型类型信息、模型功能信息、模型范围标识信息、模型上传信息、模型运行信息。应理解，上述信息的名称仅为方便说明本实施例时所采用的一种名称，并不对本申请的保护范围造成影响。在未来技术的发展过程中，这些信息可能出现其他名称，但若这些信息的作用与本申请实施例中说明的作用相同，也应在本申请的保护范围之内。
本申请实施例提供一种处理模型的方法，该方法包括如下多个步骤：
步骤1：终端设备向模型管理网元发送第一请求消息，该第一请求消息用于请求注册第一模型。
需要说明的是，在本申请实施例中，该模型管理网元可以为基站、核心网设备、OAM、OTT、或者为用于模型管理的网元。其中，核心网设备可以包括MME、AMF、NWDAF、SMF，或者其他网络设备的至少一种。
可选地，该第一请求消息为终端设备和基站之间交互的无线资源控制(radio resource control，RRC)消息或者RRC消息中的非接入层(non-access stratum，NAS)消息/信息。
步骤2：模型管理网元根据第一请求消息确定注册结果。
步骤3：模型管理网元向终端设备发送第一响应消息，该第一响应消息用于通知终端设备第一模型注册是否成功。
以下根据模型管理网元的不同场景，结合图2到图8示例的场景具体说明上述处理模型的方法。
图2是从设备交互的角度示出的本申请实施例提供的一种处理模型方法200的示意性流程图。其中，在图2中，模型管理网元可以为基站，或者直接与终端设备连接的具备模型管理的网元。如图所示，该方法200可以包括S201至S203。下面详细说明方法200中的各步骤。
S201，终端设备向模型管理网元发送第一请求消息，该第一请求消息用于请求注册第一模型。
具体地，终端设备上线之后，当终端设备要进行模型注册时，终端设备向模型管理网元发送第一请求消息，该第一请求消息用于请求注册第一模型。相应地，模型管理网元接收终端设备的第一请求消息，并根据第一请求消息获知该终端设备发起了第一模型的模型注册请求。
可选地，该第一请求消息中携带第一模型的模型信息。或者，该第一请求消息中携带第一模型。或者该第一请求消息中携带第一模型的模型信息和第一模型。
可选地，第一模型携带在终端设备向模型管理网元发送的其他信息或消息中，示例性的，第一模型携带在终端设备向模型管理网元发送的上行数据包中。此时，该第一请求消息中携带第一指示信息，该第一指示信息用于指示模型管理网元接收第一模型。换句话说，当终端设备不通过第一请求消息发送第一模型时，终端设备通过第一请求消息中的第一指示信息指示模型管理网元终端设备向模型管理网元发送了该第一模型。示例性地，当第一模型的数据较为庞大时(例如该第一模型对应的数据为1G bit，该第一模型对应的数据占据大量终端设备的内存)，该第一模型无法携带在第一请求消息中，则第一模型通过上行数据包发送给模型管理网元。
在一种可实现的方式中，当终端设备请求注册的第一模型为终端设备没有保存的模型时，无论模型管理网元中是否存储该第一模型，该第一请求消息中携带第一模型信息。换句话说，第一请求消息用于请求注册的第一模型为终端设备未保存的模型时，终端设备通过该第一请求消息向模型管理网元发送第一模型信息。
在另一种可实现的方式中，当终端设备请求注册的第一模型为保存在终端设备中的模型时，无论模型管理网元中是否存储该第一模型，该第一请求消息中携带第一模型信息和第一模型中的至少一种。或者，该第一请求消息中携带第一指示信息，第一模型携带在其他信息或者消息中。换句话说，第一请求消息用于请求注册的第一模型为终端设备的本地模型时，终端设备通过该第一请求消息向模型管理网元发送第一模型信息和第一模型中的至少一个。或者，终端设备通过该第一请求消息向模型管理网元发送第一指示信息，同时通过其他消息或信息向模型管理网元发送该第一模型。
需要说明的是，在本申请实施例中模型信息可以包括如下至少一种信息：模型类型信息、模型功能信息、模型范围标识信息、模型上传信息、模型运行信息。其中，模型类型信息用于表示模型为单边模型还是双边模型。模型功能信息用于表示模型的功能，其中，模型的功能包括：测量结果预测、测量结果压缩、定位、波束预测和波束管理。模型范围标识信息包括但不限于支持模型注册的终端设备的厂商信息、芯片厂商信息、应用信息(也可以应用层信息，用于表示模型信息对应的模型可以应用的场景，或者用于表示模型信息对应的模型可以被哪些应用所调用)或者其它信息中的至少一种。模型上传信息用于表示是否需要或允许终端设备进行模型注册时发送模型信息对应的模型。模型运行信息用于表示模型(允许模型注册的模型)的输入信息和/或输出信息，其中，输入信息可以是模型的输入数据集类型或数据集格式，输出信息可以是模型的输出数据类型或输出格式。
应理解，当模型信息中包括模型功能信息时，该模型功能信息中可以包括上述列举的模型功能中的至少一种。类似地，当模型信息中包括模型范围标识信息时，该模型范围标识信息中可以包括上述列举的模型功能中的至少一种。
还应理解，第一模型的模型信息指的是第一模型对应的模型信息。示例性地，当第一模型的模型信息包括模型类型信息时，该模型类型信息用于表示第一模型为单边模型还是双边模型。当第一模型的模型信息包括模型功能信息时，该模型功能信息用于表示第一模型的功能。类似地，当第一模型的模型信息包括模型范围标识信息、模型上传信息和模型运行信息时，均用于表示第一模型对应的模型范围标识信息、模型上传信息和模型运行信息，此处不再赘述。
需要说明的是，在本申请实施例中，模型是指AI模型。当终端设备向模型管理网元发送第一模型时，可以指的是终端设备向模型管理网元发送第一AI模型或者第一AI算子。
可选地，当模型管理网元为基站时，终端设备发送的第一请求消息为终端设备和基站之间交互的无线资源控制(radio resource control，RRC)消息或者RRC消息中的非接入层(non-access stratum，NAS)消息/信息。
S202，模型管理网元根据第一请求消息确定注册结果。
在一种可是实现的方式中，当第一请求消息中携带第一模型的模型信息时，模型管理网元根据第一请求消息确定注册结果的过程为：模型管理网元基于该第一模型的模型信息，在模型池中选择模型。若模型管理网元成功选择了一个或多个满足要求的模型，则该模型管理网元在多个满足要求的模型中任意选择一个作为第一模型。同时该模型管理网元确认该终端设备支持在线注册该第一模型。此时，该模型管理网元确认终端设备请求注册的第一模型注册成功。或者，模型管理网元没有选择出满足要求的模型，则该模型管理网元确认终端设备请求注册的第一模型注册失败。
在另一种可是实现的方式中，当模型管理网元接收到终端设备发送的第一模型时(通过第一请求消息或者通过其他信息接收)，模型管理网元根据第一请求消息确定注册结果的过程为：模型管理网元基于该第一模型确定该第一模型是否有效，即该模型管理网元确定第一模型能否进行在线注册。若该第一模型有效，即该第一模型可以进行在线注册。同时该模型管理网元确认该终端设备支持在线注册该第一模型。此时，模型管理网元确认该第一模型注册成功。或者，若该模型无效，即该第一模型不能被终端设备在线注册，模型管理网元确认该第一模型注册失败。
需要说明的是，模型管理网元确认该终端设备支持在线注册该第一模型，可以理解为模型管理网元验证该终端设备或者验证该终端设备的权限或者验证该终端设备的合法性。模型管理网元可以根据终端设备的标识信息识别终端设备，确定终端设备是否支持在线注册该第一模型。模型管理网元可以接收来自终端设备或者其它网络设备的终端设备的标识信息。示例性的，终端设备可以在S201的第一请求消息中携带终端设备的标识信息。
其中，在本申请实施例中，终端设备的标识信息可以包括：终端设备在基站或AMF为该终端设备 分配的用于在基站和核心网设备间的接口上识别该终端设备的标识，终端设备的默认凭证，终端设备的系统架构演进临时移动用户识别(System Architecture Evolution Temporary Mobile Subscriber Identity，S-TMSI)，终端设备的国际移动用户识别码(international mobile subscriber identity，IMSI)或者其他标识信息。其中，终端设备的默认凭证可以理解为用于对终端设备进行认证或验证终端设备的信息。
在一种可实现的方式中，模型管理网元确认该终端设备支持在线注册该第一模型可以是模型管理网元根据终端设备的标识信息对终端设备进行验证。此时，终端设备将终端设备的标识信息发送给模型管理网元。可选地，终端设备通过第一请求消息在S201中将终端设备的标识信息发送给模型管理网元。即模型管理网元根据终端设备的标识信息验证该终端设备的权限或者验证该终端设备的合法性或者验证该终端设备是否具有在线注册第一设备的权限。
在另一种可实现的方式中，模型管理网元确认该终端设备支持在线注册该第一模型可以是模型管理网元根据模型管理网元中存储的，或者通过其他消息携带的终端设备信息确定，此处不作限定。
S203，模型管理网元向终端设备发送第一响应消息，该第一响应消息用于通知终端设备第一模型注册是否成功。
具体地，当模型管理网元确定终端设备请求注册的第一模型注册成功时，该第一响应消息用于通知终端设备第一模型注册成功。当模型管理网元确定终端设备请求注册的第一模型注册失败时，该第一响应消息用于通知终端设备第一模型注册失败。
当终端设备请求注册的第一模型注册成功时，模型管理网元向终端设备发送注册成功的第一模型的模型配置信息。可选地，模型配置信息包括模型标识信息、模型信息和模型中的至少一种。其中，模型标识信息用于标识/识别/关联模型，模型标识信息可以是模型在模型池中的索引，或者是根据终端设备的标识信息或终端设备的标识信息的部分信息确定的，本申请对此不做限定。
应理解，第一模型的模型配置信息中可以包括第一模型的模型标识信息(下文也称为第一模型标识信息)、第一模型信息和第一模型中的至少一种。
在一种可实现的方式中，若模型管理网元确定终端设备存储有该第一模型，模型管理网元向终端设备发送的第一模型的模型配置信息中携带第一模型标识信息。终端设备接收到该第一模型的模型配置信息后，根据第一模型的模型配置信息中的第一模型标识信息，将第一模型与该第一模型标识信息关联。
或者，若模型管理网元确定终端设备存储有该第一模型，模型管理网元向终端设备发送的第一模型的模型配置信息中携带第一模型的模型信息。
或者，若模型管理网元确定终端设备存储有该第一模型，模型管理网元向终端设备发送的第一模型的模型配置信息中携带第一模型标识信息和第一模型的模型信息。终端设备接收到该第一模型的模型配置信息后，根据第一模型的模型配置信息中的第一模型标识信息，将第一模型与该第一模型标识信息关联。或者，终端设备将第一模型、第一模型的模型信息以及该第一模型标识信息三者关联。
在另一种可实现的方式中，若模型管理网元确定终端设备没有存储有该第一模型，模型管理网元向终端设备发送的第一模型的模型配置信息中携带第一模型。可选的，第一模型的配置信息中还可以携带第一模型标识信息和第一模型的模型信息中的至少一种。
具体地，模型管理网元向终端设备发送的第一模型的模型配置信息中携带第一模型标识信息和第一模型时，终端设备接收到该第一模型的模型配置信息后，将第一模型标识信息与该第一模型关联。
或者，模型管理网元向终端设备发送的第一模型的模型配置信息中携带第一模型的模型信息和第一模型时，终端设备接收到该第一模型的模型配置信息后，将第一模型的模型信息与该第一模型关联。
或者，模型管理网元向终端设备发送的第一模型的模型配置信息中携带第一模型的模型信息、第一模型标识信息和第一模型时，终端设备接收到该第一模型的模型配置信息后，将第一模型信息、第一模型和第一模型标识信息三者关联起来。
应理解，当模型管理网元向终端设备发送的模型配置信息中包括第一模型标识信息时，该方法还包括模型管理网元为第一模型分配该第一模型标识信息。
需要说明的是，当模型管理网元确定该第一模型注册成功时，若模型管理网元接收的是来自终端设备的第一模型的模型信息时，模型管理网元将该第一模型的模型信息与第一模型标识信息相关联。
或者，模型管理网元接收的是来自终端设备的第一模型时，模型管理网元存储该第一模型，并将 该第一模型与第一模型标识信息相关联。
或者，模型管理网元接收的是来自终端设备的第一模型的模型信息和第一模型时，模型管理网元存储该第一模型，并将该第一模型的模型信息、该第一模型、第一模型标识信息三者相关联。
可选地，模型配置信息可以部分或者全部携带在第一响应消息中。若模型管理网元不是基站，模型管理网元可以向基站发送第一模型的模型信息和第一模型标识信息的至少一种。模型管理网元可以向终端设备发送第一模型，第一模型可以通过NAS消息/信息或者以数据包的方式或者其它方式发送给终端设备。基站可以通过RRC消息向终端设备发送第一模型的模型信息和第一模型的标识信息给终端设备。可选的，基站还可以通过RRC消息中的NAS消息/信息或者以数据包的方式或者其它方式向终端设备发送第一模型。或者，若模型管理网元不是基站，模型管理网元可以向基站发送第一模型的模型信息和第一模型标识信息的至少一种。模型管理网元可以向终端设备发送模型配置信息，模型配置信息包括第一模型。模型配置信息还包括第一模型的模型信息和第一模型标识信息的至少一种。模型配置信息可以通过NAS消息/信息或者以数据包的方式或者其它方式发送给终端设备。基站可以通过RRC消息向终端设备发送第一模型的模型信息和第一模型的标识信息给终端设备，或者，基站可以通过RRC消息中的NAS消息/信息或者以数据包的方式或者其它方式向终端设备发送模型配置信息。应理解，携带上述模型配置信息的消息仅为示例，本申请实施例对此不做限定。
可选地，当模型管理网元确定终端设备请求注册的模型注册失败时，模型管理网元不向终端设备发送第一响应消息。即第一响应消息可以仅用于通知终端设备模型注册成功。示例性的，终端设备可以在发送完第一请求消息后，设置一个等待时间T1，当经过T1时间后，若终端设备未接收到第一响应消息，则终端设备默认模型注册失败。示例性地，终端设备可以在发送完第一请求消息后，设置一个T1时间的倒计时计时器，当计时器的是时间为0时，终端设备未接收到第一响应消息，则终端设备默认模型注册失败。
可选地，该方法200还包括如下步骤：
S2011，模型管理网元向终端设备发送第一消息，该第一消息用于指示终端设备是否可以进行模型注册。
具体地，在终端设备向模型管理网元发送第一请求消息请求模型注册之前，模型管理网元可以向终端设备先发送该第一消息，用于指示终端设备是否可以发起模型注册。
可选地，该第一消息可以是RRC消息，或者是NAS信令，或者是系统消息。
在一种可实现的方式中，该第一消息包括第三指示信息，该第三指示信息用于指示终是否支持终端设备进行模型注册。或者，若模型管理网元为基站时，该第三指示信息用于指示终端设备的服务小区或者主服务是否支持模型注册。当终端设备接收到该第一消息后，可以根据该第三指示信息确定其能够向模型管理网元发起模型注册。示例性地，该第三指示信息所占用的指示字段包括的比特数等于1，若该指示字段的值为0，用于指示基站所覆盖的小区不支持终端设备进行模型注册。若该指示字段的值为1时，用于指示基站所覆盖的小区支持终端设备进行模型注册。应理解，对于第三指示信息所占用的比特数以及不同指示字段的值所对应的含义仅为示例而非限定。
在另一种可实现的方式中，该第一消息用于指示终端设备支持注册的模型的模型信息(下文也称为模型注册支持信息)。当终端设备接收到该第一消息后，可以根据支持注册的模型的模型信息向模型管理网元发起模型注册。模型信息的描述可以参考步骤S201中的描述，此处不再赘述。
在又一种可实现的方式中，该第一消息包括上述第三指示信息和终端设备支持注册的模型的模型信息，此时，当终端设备接收到该第一消息后，可以根据第三指示信息和终端设备支持注册的模型的模型信息中的至少一个确定能够向模型管理网元发起模型注册。该第三指示信息和终端设备支持注册的模型的模型信息的作用可以分别参考上述说明，此处不再赘述。
可选的，S2012，模型管理网元向终端设备发送第四指示信息，该第四指示信息用于指示该第一模型是否支持在线更新。
具体地，当终端设备接收到模型管理网元发送的第四指示信息后，可以根据第四指示信息确定第一模型是否支持在线更新。示例性地，该第四指示信息所占用的指示字段包括的比特数等于1，若该指示字段的值为0，用于指示第一模型不支持在线更新。若该指示字段的值为1时，用于指示第一模型支持在线更新。
可选地，该第四指示信息可以携带在RRC消息中。
图3是从设备交互的角度示出的本申请实施例提供的一种处理模型方法300的示意性流程图。其中，在图3中，模型管理网元可以为AMF、MME或者为独立的与基站相连接的网元。应理解，模型管理网元为AMF或者MME是指，AMF或MME能够执行模型管理网元的操作，换句话说，AMF或MME能够实现模型管理网元的功能。模型管理网元可以为独立的与基站相连接的网元是指，该模型管理网元可以是此处未列举的其他与基站连接的网元或者该模型管理网元可以是单独定义的一个网元。如图所示，该方法300可以包括S301至S305。下面详细说明方法300中的各步骤。
S301，终端设备向基站发送第三请求消息，该第三请求消息用于请求注册第一模型。
具体地，当终端设备要进行模型注册时，终端设备向基站发送第三请求消息，该第三请求消息用于请求注册第一模型。相应地，基站接收终端设备的第三请求消息，并根据第三请求消息获知该终端设备发起了第一模型的模型注册请求。
可选地，该第三请求消息中携带第一模型的模型信息。或者，该第三请求消息中携带第一模型。或者该第三请求消息中携带第一模型的模型信息和第一模型。
作为一种可能的实现方式，终端设备在第三请消息中携带第一模型的模型信息。该第三请求消息还包括第一指示信息。第一指示信息用于指示终端设备向模型管理网元发送了第一模型。第一指示信息的描述可以参考步骤S201中的相关描述，此处不再赘述。第一模型可以携带在终端设备向基站发送的其他信息或消息中。
在一种可实现的方式中，当终端设备请求注册的第一模型为终端设备没有保存的模型时，该第三请求消息中携带第一模型的模型信息。换句话说，第三请求消息用于请求注册的第一模型为终端设备未保存的模型时，终端设备通过该第三请求消息向基站发送第一模型的模型信息。
在另一种可实现的方式中，当终端设备请求注册的第一模型为保存在终端设备中的模型时，该第三请求消息中携带第一模型的模型信息和第一模型中的至少一种。或者，该第三请求消息中携带第一模型的模型信息和第一指示信息，第一模型携带在其他信息或者消息中。换句话说，第三请求消息用于请求注册的第一模型为终端设备的本地模型时，终端设备通过该第三请求消息向基站发送第一模型的模型信息和第一模型中的至少一个。或者，终端设备通过该第三请求消息向基站发送第一指示信息，同时通过其他消息或信息向基站发送该第一模型。
应理解，若该第三请求消息中携带有第一模型的模型信息时，该第一模型的模型信息可以参考上述S201中的相关说明，此处不再赘述。同样的，若该第三请求消息中携带有第一模型时，该第一模型可以参考上述S201中的相关说明，此处不再赘述。
可选地，终端设备发送的第一请求消息为终端设备和基站之间交互的RRC消息或携带在RRC消息中的NAS信息。
S302，基站向模型管理网元发送第一请求消息。
第一请求消息用于请求模型管理网元进行模型的在线注册。
具体地，基站向模型管理网元发送第一请求消息，相应地，模型管理网元接收来自基站的第一请求消息。其中，该第一请求消息是根据S301中的第三请求消息确定的。
可选的，第一请求消息中可以携带第一模型的模型信息。
可选的，基站可以向模型管理网元发送请求注册的第一模型。
第一请求消息中还可以携带终端设备的标识信息。终端设备的标识信息的描述可以参考步骤S202中的描述，此处不再赘述。
S303，模型管理网元根据第一请求消息确定注册结果。
该步骤可以参考上述S202中的相关说明，此处不再赘述。
S304，模型管理网元向基站发送第一响应消息，该第一响应消息用于通知基站第一模型注册是否成功。
具体地，当模型管理网元确定终端设备请求注册的第一模型注册成功时，该第一响应消息用于通知基站第一模型注册成功。当模型管理网元确定终端设备请求注册的第一模型注册失败时，该第一响应消息用于通知基站第一模型注册失败。
当终端设备请求注册的第一模型注册成功时，模型管理网元可以向基站发送注册成功的第一模型 的模型配置信息。有关该第一模型的模型配置信息的相关说明，可以参考上述S203，此处不再赘述。
S305，基站向终端设备发送第三响应消息，该第三响应消息用于通知终端设备第一模型注册是否成功。
具体地，基站向终端设备发送第三响应消息，相应地，终端设备接收来自基站的第三响应消息。其中，该第三响应消息是根据S304中的第一响应消息确定的。具体地，当第一响应消息用于通知基站第一模型注册成功时，该第三响应消息用于通知终端设备第一模型注册成功。当第一响应消息用于通知基站第一模型注册失败时，该第三响应消息用于通知终端设备第一模型注册失败。
第三响应消息中可以包括模型配置信息，具体的描述参考步骤S203，此处不再赘述。
需要说明的是，当基站接收到来自终端设备的第三请求消息为NAS信令，同时基站向模型管理网元发送的第一请求消息为NAS信令时，该方法可以不包括S306。比如，此时该第三请求消息中携带AMF的标识信息，基站可以根据该AMF的标识信息将第三请求消息中携带的信息通过第一请求直接发送给模型管理网元。当基站接收到来自终端设备的第三请求消息为RRC信令，基站向模型管理网元发送的第一请求消息为RRC信令，同时该模型管理网元为AMF/MME时，在S302之前，该方法还包括S306。
S306，基站确定模型管理网元所在的AMF/MME。
具体地，基站根据终端设备发送的第三请求以及本地维护的AMF信息/MME信息，选择AMF/MME。示例性地，基站可以根据AMF/MME支持在线注册的模型的模型信息中的模型功能信息(例如，功能列表)中是否包括终端设备请求的模型功能信息，来进行选择。
可选地，该方法300还包括如下步骤：
S3011，模型管理网元向基站发送是否支持模型注册的指示信息。
具该指示信息的描述可以参考步骤S2011中第三指示信息和/或模型注册支持信息的描述，此处不再赘述。
作为有一种可能的实现方式，该指示信息可以包括模型注册支持信息和支持模型注册的模型信息中的至少一个。相应地，基站接收来自模型管理网元的模型注册支持信息和支持模型注册的模型信息中的至少一个，其中，模型注册支持信息是指支持终端设备注册的模型，支持模型注册的模型信息是指基站能够支持终端设备注册的模型对应的模型信息。
可选地，基站通过接收模型管理网元发送的模型注册支持信息和/或支持模型注册的模型信息。
S3012，基站向终端设备发送第一消息，该第一消息用于通知终端设备进行模型注册。
具体地，在终端设备向基站发送第三请求消息请求模型注册之前，基站可以向终端设备先发送该第一消息，用于通知终端设备该基站可以提供终端设备发起模型注册。
可选地，该第一消息可以是RRC消息，或者是用于透传NAS信令的RRC消息，或者是系统消息。
其中，该第一消息可以包括第三指示信息和终端设备支持注册的模型的模型信息中的至少一种。该第一消息的相关说明可以参考上述S2011，此处不再赘述。
S3013，模型管理网元向基站发送第四指示信息，该第四指示信息用于指示该第一模型是否支持在线更新。
可选地，该第四指示信息可以携带在RRC消息中或者NAS信令中。
S3014，基站向终端设备转发第四指示信息。
应理解，S3013和S3014中对第四指示信息的说明可以参考上述S2012，此处不再赘述。
可选的，S3015，模型管理网元向基站发送第五指示信息，该第五指示信息用于指示是否向基站发送第一模型和/或第一模型的模型信息。
具体地，当模型管理网元确定基站没有存储第一模型和/或第一模型的模型信息时，模型管理网元发送第五指示信息，用于指示终端设备是否向基站发送第一模型和/或第一模型的模型信息。
可选的，S3016，基站向终端设备转发第五指示信息。
具体地，基站向终端设备转发第五指示信息，相应地，终端设备接收第五指示信息，并根据第五指示信息确定是否向基站发送第一模型和/或第一模型的模型信息。示例性地，当该第五指示信息所占用的指示字段包括的比特数等于2时，若该指示字段的值为00，用于指示向基站发送第一模型。若该指示字段的值为01时，用于指示向基站发送第一模型的模型信息。若该指示字段的值为10时，用于 指示向基站同时发送了第一模型和第一模型的模型信息。若该指示字段的值为11时，用于指示不向基站发送第一模型和第一模型的模型信息中的任意一个。应理解，对于第五指示信息所占用的比特数以及不同指示字段的值所对应的含义仅为示例而非限定。
可选的，S3017，终端设备向基站发送模型配置信息。
具体地，当第一响应消息中携带模型配置信息时，终端设备的模型配置信息可以是S305中通过第三响应消息获得的。或者模型配置信息是模型管理网元通过其他消息发送给基站时，基站将该其他消息包括的模型配置信息转发给终端设备。
在一种可实现的方式中，若终端设备确定基站上未存储有第一模型时，终端设备可以向基站发送该第一模型。
在另一种可实现的方式中，若终端设备确定基站上未存储有该第一模型的模型信息，终端设备可以向基站发送该第一模型的模型信息。
在另一种可实现的方式中，若终端设备确定基站上未存储有第一模型同时也没有存储第一模型的模型信息，终端设备可以向基站发送该第一模型和该第一模型的模型信息。
在另一种可实现的方式中，若终端设备根据模型管理网元发送的第五指示信息确定需要向基站发送第一模型和/或第一模型的模型信息时，终端设备根据第五指示信息向基站发送第一模型和/或第一模型的模型信息。
可选的，S3018，模型管理网元向基站发送第一模型和/或第一模型的模型信息。
在一种可实现的方式中，当第一模型注册成功后，若模型管理网元确定基站没有存储第一模型时，模型管理网元向基站发送第一模型和/或第一模型的模型信息。
在另一种可实现的方式中，当第一模型注册成功后，若模型管理网元确定基站未存储有第一模型的模型信息，模型管理网元可以向基站发送该第一模型的模型信息。
在又一种可实现的方式中，若终端设备存储有该第一模型，但是未向基站发送该第一模型，同时终端设备向模型管理网元发送未向基站发送第一模型的第七指示信息，模型管理网元根据第七指示信息向基站发送第一模型。
在又一种可实现的方式中，若终端设备存储有该第一模型的模型信息，但是未向基站发送该第一模型的模型信息，同时终端设备向模型管理网元发送未向基站发送第一模型的的模型信息第八指示信息，模型管理网元根据第八指示信息向基站发送第一模型的模型信息。
可选的，S3019，模型管理网元向基站发送第六指示信息，该第六指示信息用于指示是否向终端设备发送第一模型和/或第一模型的模型信息。示例性地，该第六指示信息所占用的指示字段包括的比特数为2，若该指示字段的值为00，用于指示基站向终端设备发送第一模型。若该指示字段的值为01时，用于指示基站向终端设备发送第一模型的模型信息。若该指示字段的值为10时，用于指示基站向终端设备发送第一模型和第一模型的模型信息。若该指示字段的值为11时，用于指示基站不向终端设备发送第一模型和第一模型的模型信息中的任意一个。应理解，对于第六指示信息所占用的比特数以及不同指示字段的值所对应的含义仅为示例而非限定。
需要说明的是，若基站根据第六指示信息确定需要向终端设备发送的内容，此时，在S305中，基站根据第六指示信息确定第三响应消息中携带的信息。示例性地，当基站根据第六指示信息确定不需要向终端设备发送第一模型和第一模型的模型信息中的任意一个时，基站认为终端设备中存储有第一模型和第一模型的模型信息，此时基站仅在第三响应消息中向终端设备发送第一模型标识信息。
图4是从设备交互的角度示出的本申请实施例提供的一种处理模型方法400的示意性流程图。其中，在图4中，模型管理网元可以为SMF、OAM、OTT、NWDAF或者为独立的与AMF连接的的网元。应理解，模型管理网元为SMF、OAM、NWDAF或OTT是指，SMF、OAM、NWDAF或OTT能够执行模型管理网元的操作，换句话说，SMF、OAM、NWDAF或OTT能够实现模型管理网元的功能。模型管理网元可以为独立的与AMF相连接的网元是指，该模型管理网元可以是此处未列举的其他与AMF连接的网元或者该模型管理网元可以是单独定义的一个网元。如图所示，该方法400可以包括S401至S407。下面说明方法400中的各步骤。
应理解，该方法400与方法300相比的区别在于，在图4所示的方法400中，模型管理网元向基站发送的消息或信息等是通过AMF转发给基站的，而在图3所示的方法300中，模型管理网元向基站 发送的消息或信息等是直接发送给基站的。因此，该方法400与该方法300相比，存在以下多个区别步骤。
S402，基站向AMF发送第四请求消息。
具体地，当基站接收到终端设备发送的第三请求消息之后，基站向AMF发送第四请求消息。相应地，AMF接收第四请求消息。
应理解，由于第三请求为RRC信令或者NAS信令，因此，可选地，该第四请求消息可以为RRC信令或者NAS信令。此外，该第四请求消息是根据第三请求消息确定的。换句话说，基站接收到第三请求消息后可以不对接收到的第三请求消息做处理，而是直接将第三请求消息以第四请求消息的形式转发给AMF。或者基站接收到第三请求消息后根据第三请求消息中携带的信息确定第四请求消息。
在一种可实现的方式中，当该第四请求消息为NAS信令时，基站接收到第三请求消息之后，将来自终端设备的第三请求消息以第四请求消息的形式直接透传给AMF。
在另一种可实现的方式中，当该第四请求消息为RRC信令时，基站接收到第三请求消息之后，由于第三请求消息为RRC信令，此时，该方法还包括S408。
S408，基站确定AMF。
具体地，基站基于终端设备的第三请求消息和本地维护的AMF信息确定AMF。示例性地，基站可以根据AMF支持在线注册的模型的模型信息中的模型功能信息(例如，功能列表)中是否包括终端设备请求的模型功能信息来确定AMF。当基站确定任意一个AMF的功能列表中包括第一模型对应的功能信息时，基站可以将该AMF确定为目标AMF。
当基站确定目标AMF后，基站将第四请求消息发送给目标AMF。
S406，AMF向基站发送第四响应消息。
具体地，AMF接收到来自模型管理网元的第一响应消息后，AMF将接收到的第一响应消息转发给基站，该AMF向基站发送的响应消息为第四响应消息。示例性地，AMF可以通过与基站的N2接口向基站发送该第四响应消息。
应理解，图4所示的方法400中，包括的可选步骤，即S4012、S4015、S4018、S4022、S4024均是由AMF向基站转发的信息，应理解，该转发过程可以是通过AMF与基站的N2接口来实现的。此外，AMF向基站转发的各个信息。例如模型注册支持信息、支持模型注册的模型信息、第四指示信息、第五指示信息以及第六指示信息等，均可以参考上述图3中相关的说明，此处不再赘述。
还应理解，图4中未说明的步骤可以相应的参考图3中的步骤，以下简单说明对应的关系。
S401，终端设备向基站发送第三请求消息。该步骤可以参考S301，此处不再赘述。
S403，AMF向模型管理网元发送第一请求消息。该步骤可以参考S302，基站向模型管理网元发送第一请求消息，此处不再赘述。
S404，模型管理网元根据第一请求消息确定注册结果。该步骤可以参考S303，此处不再赘述。
S405，模型管理网元向AMF发送第一响应消息。该步骤可以参考S304，模型管理网元向基站发送第一响应消息，此处不再赘述。
S407，基站向终端设备发送第三响应消息。该步骤可以参考S305，此处不再赘述。
S4011，模型管理网元向AMF发送模型注册支持信息和/或支持模型注册的模型信息。该步骤可以参考S3011，模型管理网元向基站发送模型注册支持信息和/或支持模型注册的模型信息，此处不再赘述。
S4013，基站向终端设备发送第一消息。该步骤可以参考S3012，此处不再赘述。
S4014，模型管理网元向AMF发送第四指示信息。该步骤可以参考S3013，模型管理网元向基站发送第四指示信息，此处不再赘述。
S4016，基站向终端设备发送第四指示消息。该步骤可以参考S3014，此处不再赘述。
S4017，模型管理网元向AMF发送第五指示信息。该步骤可以参考S3015，模型管理网元向基站发送第四指示信息，此处不再赘述。
S4019，基站向终端设备发送第五指示消息。该步骤可以参考S3016，此处不再赘述。
S4020，终端设备向基站发送模型配置信息，该步骤可以参考S3017，此处不再赘述。
S4021，模型管理网元向AMF发送第一模型和/或第一模型的模型信息。该步骤可以参考S3018， 模型管理网元向基站发送第一模型和/或第一模型的模型信息，此处不再赘述。
S4023，模型管理网元向AMF发送第六指示信息。该步骤可以参考S3019，模型管理网元向基站发送第六指示信息，此处不再赘述。
图5是从设备交互的角度示出的本申请实施例提供的一种处理模型方法500的示意性流程图。其中，在图5中，模型管理网元为基站，如图所示，该方法500可以包括S501至S504。下面详细说明方法500中的各步骤。
S501，终端设备向模型管理网元(本申请实施例为基站)发送第二请求消息，该第二请求消息用于请求更新第三模型为第二模型。
具体地，当终端设备在本地完成第三模型更新后，生成第二模型。此时，终端设备向模型管理网元发送第二请求消息，该第二请求消息用于请求更新第三模型为第二模型。其中，该第二请求消息中携带第二指示信息，该第二指示信息用于指示终端设备请求的流程为更新流程。即该第二指示信息用于指示更新第三模型为第二模型。
示例性地，终端设备进行的模型更新过程可以是对本地的模型进行优化而非重新训练的过程。
可选地，该第二请求消息可以是RRC信令或NAS信令。
可选地，该第二请求消息携带终端设备的标识信息(默认凭证)，该终端设备的标识信息用于模型管理网元识别该终端设备，确定终端设备是否支持在线更新该第三模型。
S502，终端设备向模型管理网元发送第三模型的模型更新信息。
具体地，终端设备要将第三模型更新为第二模型时，终端设备向模型管理网元发送第三模型的模型更新信息。
可选地，该模型更新信息包括梯度信息和模型中的至少一个，以及模型标识信息。其中，该模型标识信息可以参考上文中的相关说明，此处不再赘述。梯度信息可以理解为第三模型更新为第二模型过程中产生的一个参数信息，当模型管理网元接收到梯度信息时，可以根据该梯度信息将第三模型更新为第二模型。模型可以参考上文中的相关说明，此处不再赘述。
应理解，终端设备向模型管理网元发送的第三模型的模型更新信息可以包括第三模型的模型标识信息和梯度信息。或者包括第三模型的模型标识信息和第二模型。或者包括第三模型的模型标识信息、梯度信息和第二模型。
在一种可实现的方式中，当终端设备请求更新的第三模型为模型管理网元保存的模型时，即终端设备请求被更新的第三模型为父模型时，该第三模型更新信息中携带第三模型的模型标识信息。换句话说，第三请求消息用于请求更新的第三模型为模型管理网元保存的模型时，终端设备向模型管理网元发送第三模型的模型标识信息。
在另一种可实现的方式中，当终端设备请求更新的第三模型为模型管理网元未保存的模型时，即终端设备请求被更新的第三模型为子模型时，该第三模型更新信息中携带梯度信息和第二模型中的至少一个，以及第三模型的模型标识信息。换句话说，第三请求消息用于请求更新的第三模型为模型管理网元未保存的模型时，终端设备向模型管理网元发送梯度信息和第二模型中的至少一个，以及第三模型的模型标识信息。
需要说明的是，在本申请实施例里中，父模型是指网络侧设备发送给终端设备的模型，或者指存储在网络设备中的模型，或者，可以指由网络设备进行管理的模型。子模型是指终端设备本地更新的模型，或者指仅存储在终端设备中的模型，或者，可以指仅有终端设备进行管理模型。
可选地，该第三模型的模型更新信息可以携带在第二请求消息中，或者可以通过其他终端设备向基站发送的消息或信息携带。示例性地，当第三模型的模型更新信息包括第二模型时，该第三模型的模型更新信息通过上行数据包发送给模型管理网元。
S503，模型管理网元根据第二请求消息和第三模型的模型更新信息确定更新结果。
在一种可是实现的方式中，当第三模型的模型更新信息中携带第三模型的模型标识信息和梯度信息，模型管理网元根据第二请求消息确定更新结果的过程为：模型管理网元基于该第三模型的模型标识信息和梯度信息，对第三模型的参数进行更新。同时该模型管理网元确认该终端设备支持在线更新该第三模型。此时，该模型管理网元确认终端设备请求更新的第三模型更新成功。否则，该模型管理网元确认终端设备请求更新的第三模型更新失败。
在另一种可是实现的方式中，当第三模型的模型更新信息中携带第三模型的模型标识信息和第二模型时，模型管理网元根据第二请求消息确定更新结果的过程为：模型管理网元确认该第三模型标识有效，并接收该第二模型，同时该模型管理网元确认该终端设备支持在线更新该第三模型。此时，模型管理网元确认终端设备请求更新的第三模型更新成功。否则，该模型管理网元确认终端设备请求更新的第三模型更新失败。
需要说明的是，模型管理网元确认该终端设备支持在线更新该第三模型，可以理解为模型管理网元验证该终端设备或者验证该终端设备的权限或者验证该终端设备的合法性。该过程可以通过模型管理网元根据终端设备的标识信息进行识别。其中，终端设备的标识信息可以参考S202中的相关说明，此处不再赘述。
应理解，当模型管理网元确认终端设备请求更新的第三模型更新成功时，模型管理网元生成第二模型，模型管理网元可以将该第二模型作为父模型加入到模型管理网元管理的父模型的模型池中。或者，将第二模型作为第三模型的更新版本进行管理。即模型管理网元基于第二模型进行模型版本管理。或者模型管理网元基于第二模型对父模型版本进行更显，更新模型池。
在一种可选的方式中，若模型管理网元将该第二模型作为父模型加入到模型管理网元管理的父模型的模型池时，该模型管理网元为该第二模型分配第二模型的模型标识信息(下文也称为第二模型标识信息)。此时模型管理网元更新其模型池中支持在线更新的功能类型或模型标识，并生成如下表1所示的模型版本管理列表。
表1
在另一种可选的方式中，若模型管理网元将该第二模型作为子模型版本管理，即将第二模型作为第三模型的更新版本进行管理时，此时，模型管理网元更新其模型池中第三模型对应的子模型版本，并生成如下表2所示的模型版本管理列表。
表2
在又一种可选的方式中，若模型管理网元更新的第三模型为非首次更新的模型时，即该第三模型在此次更新前还经过至少一次更新时，那么，可以对该第三模型的首次额更新采用父模型更新的方式，后续的模型更新采用子模型的模型更新方式，此时，模型管理网元更新其模型池中第三模型对应的子模型版本，并生成如下表3所示的模型版本管理列表。
表3
应理解，上述表1或表2或表3仅为示例而非限定。
S504，模型管理网元向终端设备发送第二响应消息，该第二响应消息用于通知更新结果。
具体地，当模型管理网元确定终端设备请求更新的第三模型更新成功时，该第二响应消息用于通知终端设备第三模型更新成功。当模型管理网元确定终端设备请求更新的第三模型更新失败时，该第二响应消息用于通知终端设备第三模型更新失败。
当终端设备更新的第三模型更新成功时，模型管理网元向终端设备发送第二模型的模型更新信息。换句话说，当模型管理网元将第三模型更新为第二模型后，模型管理网元向终端设备发送第二模型的模型更新信息。
应理解，第二模型的模型更新信息中可以包括第二模型的模型标识信息。该第二模型的模型标识信息用于与第二模型关联。
可选地，第二模型的模型更新信息可以部分或者全部携带在第二响应消息中，或者，第二模型的模型更新信息携带在基站发送给终端设备的其他消息中。
可选地，当模型管理网元确定终端设备请求更新的模型更新失败时，模型管理网元不向终端设备发送第二响应消息。即第二响应消息可以仅用于通知终端设备模型更新成功。示例性的，终端设备可以在发送完第二请求消息后，设置一个等待时间T2，当经过T2时间后，若终端设备未接收到第二响应消息，则终端设备默认模型更新失败。示例性地，终端设备可以在发送完第二请求消息后，设置一个T2时间的倒计时计时器，当计时器的是时间为0时，终端设备未接收到第二响应消息，则终端设备默认模型更新失败。
还需要说明的是，上述模型管理网元执行S503与执行S502不一定严格遵守S502在S503之前的顺序。
应理解，模型管理网元接收到第二请求消息后，可以根据第二请求消息先确定第三模型是否支持更新。当模型管理网元确定第三模型支持更新后，再根据S502中接收到的第三模型的模型更新信息将第三模型更新为第二模型。
可选地，该第二响应消息可以为RRC信令或者NAS信令。
可选地，该方法500还包括如下步骤：
S5011，终端设备更新第三模型为第二模型。
具体地，在终端设备发送第二请求消息之前，终端设备根据本地数据将第三模型更新为第二模型。
图6是从设备交互的角度示出的本申请实施例提供的一种处理模型方法600的示意性流程图。其中，在图6中，模型管理网元可以为AMF、MME或者为独立的与基站相连接的网元。应理解，模型管理网元为AMF或者MME是指，AMF或MME能够执行模型管理网元的操作，换句话说，AMF或MME能够实现模型管理网元的功能。模型管理网元可以为独立的与基站相连接的网元是指，该模型管理网元可以是此处未列举的其他与基站连接的网元或者该模型管理网元可以是单独定义的一个网元。如图所示，该方法600可以包括S601至S607。下面详细说明方法600中的各步骤。
S601，终端设备向基站发送第五请求消息，该第五请求消息用于请求更新第三模型为第二模型。
具体地，终端设备向基站发送第五请求消息，基站接收该第五请求消息。具体地，当终端设备在本地完成第三模型更新后，生成第二模型。此时，终端设备向基站发送第五请求消息，该第五请求消息用于请求更新第三模型为第二模型。其中，该第二请求消息中携带第二指示信息，该第二指示信息用于指示终端设备请求的流程为更新流程。即该第二指示信息用于指示更新第三模型为第二模型。
示例性地，终端设备进行的模型更新过程可以是对本地的模型进行优化而非重新训练的过程。
可选地，该第五请求消息可以是RRC信令或NAS信令。
可选地，该第五请求消息携带终端设备的标识信息，用于模型管理网元根据终端设备的标识信息识别终端设备。
在一种可实现的方式中，当该第五请求消息为NAS信令时，基站接收到第五请求消息之后，由该第五请求信令中携带AMF的标识信息，因此，基站根据AMF的标识信息将来自终端设备的第五请求消息以第二请求消息的形式直接透传给模型管理网元。
在另一种可实现的方式中，当该第五请求消息为RRC信令时，基站接收到第五请求消息之后，该方法还包括S608。
S608，基站确定AMF/MME。
具体地，以基站确定AMF为例说明，基站基于终端设备的第五请求消息和本地维护的AMF信息确定AMF。示例性地，基站可以根据AMF支持在线注册的模型的模型信息中的模型功能信息(例如，功能列表)中是否包括终端设备请求的模型功能信息来确定AMF。当基站确定任意一个AMF的功能列表中包括第一模型对应的功能信息时，基站可以将该AMF确定为目标AMF。
当基站确定目标AMF后，基站将第二请求消息发送给目标AMF。
需要说明的是，当基站接收到来自终端设备的第五请求消息为NAS信令，同时基站向模型管理网元发送的第二请求消息为NAS信令时，该方法不包括S608。当基站接收到来自终端设备的第五请求消息为RRC信令，基站向模型管理网元发送的第二请求消息为RRC信令，同时该模型管理网元为AMF时，方法包括S608。
应理解，模型管理网元为MME时，基站用于确定目标MME。该确定目标MME的过程与确定目标AMF的过程相同，此处不再赘述。
S602，基站向模型管理网元发送第二请求消息，该第二请求消息用于请求更新第三模型为第二模型。
具体地，当基站接收到终端设备的第五请求消息后，根据第五请求消息确定第二请求消息。
可选地，该第二请求消息中携带第二模型的模型信息。
S603，终端设备向基站发送第三模型的模型更新信息。
具体地，终端设备要将第三模型更新为第二模型时，终端设备向基站发送第三模型的模型更新信息。基站接收该第三模型的模型更新信息。其中，第三模型的模型更新信息可以参考上述S502中的说明，此处不再赘述。
可选地，终端设备向基站发送的第三模型的模型更新信息可以携带在第五请求消息中，或者可以通过其他终端设备向基站发送的消息或信息携带，例如上行数据包携带，此处不再赘述。
S604，基站向模型管理网元发送第三模型的模型更新信息。
具体地，基站接收到来自终端设备的第三模型的模型更新信息后，将第三模型的模型更新信息转发给模型管理网元。
可选地，基站向模型管理网元发送的第三模型的模型更新信息可以携带在第二请求消息中，或者可以通过其他终端设备向基站发送的消息或信息携带，例如上行数据包携带，此处不再赘述。
S605，模型管理网元根据第二请求消息和第三模型的模型更新信息确定更新结果。
具体地，该过程可以参考上述S503，此处不再赘述。
S606，模型管理网元向基站发送第二响应消息，该第二响应消息用于通知更新结果。
具体地，当模型管理网元确定终端设备请求更新的第三模型更新成功时，该第二响应消息用于通知终端设备第三模型更新成功。当模型管理网元确定终端设备请求更新的第三模型更新失败时，该第二响应消息用于通知终端设备第三模型更新失败。其中，第二响应消息中携带的信息可以参考上述S504，此处不再赘述。
需要说明的是，当模型管理网在更新过程中生成例如上述表1或表2或表3的模型版本管理列表时，可以将诸如表1或表2或表3通过该第二响应消息或其他消息发送给AMF。
S607，基站向终端设备发送第五响应消息，该第五响应消息用于通知更新结果。
具体地，该第五响应消息是根据第二响应消息确定的，其中，第五响应消息中携带的信息可以参考上述S504，此处不再赘述。
可选地，该第五响应消息可以为RRC信令或者NAS信令。
可选地，该方法600还包括如下步骤：
S6011，终端设备更新第三模型为第二模型。
具体地，在终端设备发送第五请求消息之前，终端设备根据本地数据将第三模型更新为第二模型。
图7是从设备交互的角度示出的本申请实施例提供的一种处理模型方法700的示意性流程图。其中，在图7中，模型管理网元可以为SMF、OAM、OTT、NWDAF或者为独立的与AMF连接的的网元。应理解，模型管理网元为SMF、OAM、NWDAF或OTT是指，SMF、OAM、NWDAF或OTT能够执行模型管理网元的操作，换句话说，SMF、OAM、NWDAF或OTT能够实现模型管理网元的功能。模型管理网元可以为独立的与AMF相连接的网元是指，该模型管理网元可以是此处未列举的其他与 AMF连接的网元或者该模型管理网元可以是单独定义的一个网元。如图7所示，该方法700可以包括S701至S710。下面说明方法700中的各步骤。
应理解，该方法700与方法600相比的区别在于，在图7所示的方法700中，模型管理网元向基站发送的消息或信息等是通过AMF转发给基站的，而在图6所示的方法600中，模型管理网元向基站发送的消息或信息等是直接发送给基站的。因此，该方法700与该方法600相比，存在以下多个区别步骤。
S702，基站向AMF发送第六请求消息。
具体地，当基站接收到终端设备发送的第五请求消息之后，基站向AMF发送第六请求消息。相应地，AMF接收第六请求消息。
应理解，由于第五请求为RRC信令或者NAS信令，因此，可选地，该第六请求消息可以为RRC信令或者NAS信令。此外，该第六请求消息是根据第五请求消息确定的。换句话说，基站接收到第五请求消息后可以不对接收到的第五请求消息做处理，而是直接将第五请求消息以第六请求消息的形式转发给AMF。或者基站接收到第五请求消息后根据第五请求消息中携带的信息确定第六请求消息。
在一种可实现的方式中，当该第六请求消息为NAS信令时，基站接收到第五请求消息之后，由于第五请求消息为NAS信令，此时，该第五请求信令中携带AMF的标识信息，因此，基站根据AMF的标识信息将来自终端设备的第五请求消息以第六请求消息的形式直接透传给AMF。
在另一种可实现的方式中，当该第六请求消息为RRC信令时，基站接收到第五请求消息之后，由于第五请求消息为RRC信令，此时，该方法还包括S711。
S711，基站确定AMF。
具体地，基站基于终端设备的第五请求消息和本地维护的AMF信息确定AMF。示例性地，基站可以根据AMF支持在线注册的模型的模型信息中的模型功能信息(例如，功能列表)中是否包括终端设备请求的模型功能信息来确定AMF。当基站确定任意一个AMF的功能列表中包括第一模型对应的功能信息时，基站可以将该AMF确定为目标AMF。
当基站确定目标AMF后，基站将第六请求消息发送给目标AMF。
S705，基站向AMF发送第三模型的模型更新信息。
具体地，终端设备要将第三模型更新为第二模型时，终端设备向基站发送第三模型的模型更新信息。基站将接收道德第三模型的模型更新信息转发给AMF。其中，第三模型的模型更新信息可以参考上述S502中的说明，此处不再赘述。
S709，AMF向基站发送第六响应消息。
具体地，AMF接收到来自模型管理网元的第二响应消息后，AMF将接收到的第二响应消息转发给基站，该AMF向基站发送的响应消息为第六响应消息。示例性地，AMF可以通过与基站的N2接口向基站发送该第六响应消息。
应理解，图7中未说明的步骤可以相应的参考图6中的步骤，此处不在赘述。
图8是本申请实施例提供的一种处理模型方法800的示意性流程框图。具体地，该方法800包括如下步骤。
S801，对第一模型进行注册。
S802，将第一模型更新为第二模型。
应理解，该方法800可以视为上述方法200至方法700进行合理的组合使用的场景。示例性的，S801可以是上述方法200至方法400中的一个，S802可以是上述方法500至方法700中的一个。应理解，S801与S802使用的场景相同。具体地，当模型管理网元为基站时，S801执行方法200，相应地，S801执行方法500。当模型管理网元为例如AMF时，S801执行方法300，相应地，S801执行方法600。当模型管理网元为例如OAM时，S801执行方法400，相应地，S801执行方法700。
此外，上述方法200至方法700实施例中出现的各个术语的解释或说明适用于该方法800。
图9为本申请实施例提供的一种处理模型的装置900的示意性框图。该装置900包括接收模块901，接收模块901可以用于实现相应的接收功能。接收模块901还可以称为接收单元。
该装置900还包括处理模块902，处理模块902可以用于实现相应的处理功能。
该装置900还包括发送模块903，发送模块903可以用于实现相应的发送功能，发送模块903还可 以称为发送单元。
可选地，该装置900还包括存储单元，该存储单元可以用于存储指令和/或数据，处理单元902可以读取存储单元中的指令和/或数据，以使得装置实现前述各个方法实施例中的相关装置的动作。
该装置900可以用于执行上文各个方法实施例中的终端设备或模型管理网元所执行的动作，这时，该装置900可以为终端设备或模型管理网元的组成部件，接收模块901用于执行上文方法实施例中终端设备或模型管理网元的接收相关的操作，处理模块902用于执行上文方法实施例中终端设备或模型管理网元的处理相关的操作，发送模块903用于执行上文方法实施例中终端设备或模型管理网元的发送相关的操作。
作为一种设计，该装置900用于执行上文各个方法实施例(方法200至方法800)中任意设备所执行的动作。在一个实施例中，该处理模型的装置可用于执行上述图2至图8中终端设备的操作。例如：
接收模块901，用于接收第一响应消息，第一响应消息用于通知第一模型的注册是否成功。
发送模块903，用于发送第一请求消息，第一请求消息包括第一模型的模型信息，第一请求消息用于请求注册第一模型。
可选地，第一请求消息还包括终端设备的标识信息，终端设备的标识信息用于模型管理网元识别终端设备。
应理解，各模块执行上述相应步骤的具体过程在上述方法实施例中已经详细说明，为了简洁，在此不再赘述。
此外，该处理模型的装置900中的接收模块901、处理模块902和发送模块903还可实现上述方法中终端设备的其他操作或功能，此处不再赘述。
可选地，该处理模型的装置900可以为包括终端设备的设备。或者，该处理模型的装置900可以为配置在终端设备中的部件，例如，终端设备中的芯片。这种情况下，接收模块901和发送模块903可以为接口电路、管脚等。具体地，接口电路可以包括输入电路和输出电路，其中，接收模块901可以包括输入电路、发送模块903可以包括输出电路，处理模块902可以包括处理电路。
在另一个实施例中，该处理模型的装置可用于执行上述图2至图8中模型管理网元的操作。例如：
接收模块901，用于接收第一请求消息，第一请求消息包括第一模型的模型信息，第一请求消息用于请求注册第一模型。
可选地，第一请求消息还包括终端设备的标识信息，终端设备的标识信息用于模型管理网元识别终端设备。
处理模块902，用于根据第一模型的模型信息，确定所述第一模型
发送模块903，用于发送第一响应消息，所述第一响应消息用于通知所述第一模型的注册是否成功。
此外，该处理模型的装置900中的接收模块901、处理模块902和发送模块903还可实现上述方法中模型管理网元的其他操作或功能，此处不再赘述。
可选地，该处理模型的装置900可以为包括模型管理网元的设备。或者，该处理模型的装置900可以为配置在模型管理网元中的部件，例如，模型管理网元中的芯片。这种情况下，接收模块901和发送模块903可以为接口电路、管脚等。具体地，接口电路可以包括输入电路和输出电路，其中，接收模块901可以包括输入电路、发送模块903可以包括输出电路，处理模块902可以包括处理电路。
应理解，各模块执行上述相应步骤的具体过程在上述方法实施例中已经详细说明，为了简洁，在此不再赘述。
图10为本申请实施例提供的另一种处理模型的装置的示意性结构图。该处理模型的装置1000包括处理器1001，如图10所示，处理模型的装置还可以包括至少一个存储器1002，用于存储计算机程序或指令或者和/或数据。存储器1002和处理器1001耦合，处理器1001用于执行存储器1002存储的计算机程序或指令和/或者数据，使得上文方法实施例中的方法(方法200至方法800)被执行。本申请实施例中的耦合是装置、单元或模块之间的间接耦合或通信连接，可以是电性，机械或其它的形式，用于装置、单元或模块之间的信息交互。处理器1001可能和存储器1002协同操作。至少一个存储器中1002的至少一个可以包括于处理器1001中。
可选地，该处理模型的装置1000包括的处理器1001为一个或多个。
可选地，该存储器1002可以与该处理器1001集成在一起，或者分离设置。
该处理模型的装置1000还可以包括收发器1003，用于通过传输介质和其它设备进行处理模型，从而用于装置可以和其它设备进行处理模型。可选地，收发器1003可以是接口、总线、电路或者能够实现收发功能的装置。
可选地，可以将收发器1003中用于实现接收功能的器件视为接收模块，将收发器1003中用于实现发送功能的器件视为发送模块，即收发器1003包括接收器和发送器。
本申请实施例中不限定上述处理器1001、存储器1002以及收发器1003之间的具体连接介质。本申请实施例在图10中以处理器1001、存储器1002以及收发器1003之间通过总线1004连接，总线在图10中以粗线表示，其它部件之间的连接方式，仅是进行示意性说明，并不引以为限。所述总线可以分为地址总线、数据总线、控制总线等。
应理解，为便于表示，图10中仅用一条粗线表示，但并不表示仅有一根总线或一种类型的总线。
可选地，如图10所示，该处理模型的装置1000还可以包括收发器1003和/或通信接口，收发器1003和/或通信接口用于信号的接收和/或发送。例如，处理器1001用于控制收发器1003和/或通信接口进行数据的接收和/或发送。
收发器有时也可以称为收发机、收发模块、或收发电路等。接收器有时也可以称为接收机、接收模块、或接收电路等。发送器有时也可以称为发射机、发射器、发射模块或者发射电路等。
例如，在一个实施例中，处理器1001被配置为终端设备的其他操作或功能。收发器1003用于实现该处理模型的装置与模型管理网元之间的处理模型。
在另一个实施例中，处理器1001被配置为模型管理网元的其他操作或功能。收发器1003用于实现该处理模型的装置与终端设备之间的处理模型。
以上模块或单元的一个或多个可以软件、硬件或二者结合来实现。当以上任一模块或单元以软件实现的时候，所述软件以计算机程序指令的方式存在，并被存储在存储器中，处理器可以用于执行所述程序指令并实现以上方法流程。所述处理器可以包括但不限于以下至少一种：中央处理单元(central processing unit，CPU)、微处理器、数字信号处理器(DSP)、微控制器(microcontroller unit，MCU)、或人工智能处理器等各类运行软件的计算设备，每种计算设备可包括一个或多个用于执行软件指令以进行运算或处理的核。该处理器可以内置于SoC(片上系统)或专用集成电路(application specific integrated circuit，ASIC)，也可是一个独立的半导体芯片。该处理器内处理用于执行软件指令以进行运算或处理的核外，还可进一步包括必要的硬件加速器，如现场可编程门阵列(field programmable gate array，FPGA)、PLD(可编程逻辑器件)、或者实现专用逻辑运算的逻辑电路。
当以上模块或单元以硬件实现的时候，该硬件可以是CPU、微处理器、DSP、MCU、人工智能处理器、ASIC、SoC、FPGA、PLD、专用数字电路、硬件加速器或非集成的分立器件中的任一个或任一组合，其可以运行必要的软件或不依赖于软件以执行以上方法流程。
当以上模块或单元使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。所述计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机程序指令时，全部或部分地产生按照本发明实施例所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，所述计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线(DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。所述计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。所述可用介质可以是磁性介质，(例如，软盘、硬盘、磁带)、光介质(例如，DVD)、或者半导体介质(例如固态硬盘(solid state disk，SSD))等。
以上所述的具体实施方式，对本申请的目的、技术方案和有益效果进行了进一步详细说明，所应理解的是，以上所述仅为本申请的具体实施方式而已，并不用于限定本申请的保护范围，凡在本申请的技术方案的基础之上，所做的任何修改、等同替换、改进等，均应包括在本申请的保护范围之内。
本申请实施例提供一种处理模型的装置1100，该装置1100可以是模型管理网元，也可以是芯片。该装置1100可以用于执行上述方法实施例(方法200至方法800)中由模型管理网元所执行的操作。
当该处理模型的装置1100为模型管理网元时，图11示出了一种简化的模型管理网元的结构示意 图。模型管理网元包括1110部分以及1120部分。1110部分包括天线和射频电路，天线主要用于射频信号的收发，射频电路主要用于射频信号与基带信号的转换。1120部分包括存储器和处理器，主要用于基带处理，对模型管理网元进行控制等。1110部分通常可以称为收发单元、收发机、收发电路、或者收发器等。1120部分通常是模型管理网元的控制中心，通常可以称为处理单元，用于控制模型管理网元执行上述方法实施例中模型管理网元侧的处理操作。
可选地，可以将1110部分中用于实现接收功能的器件视为接收单元，将用于实现发送功能的器件视为发送单元，即1110部分包括接收单元和发送单元。接收单元也可以称为接收机、接收器、或接收电路等，发送单元可以称为发射机、发射器或者发射电路等。
当需要发送数据时，处理器对待发送的数据进行基带处理后，输出基带信号至射频电路，射频电路将基带信号进行射频处理后将射频信号通过天线以电磁波的形式向外发送。当有数据发送到模型管理网元时，射频电路通过天线接收到射频信号，将射频信号转换为基带信号，并将基带信号输出至处理器，处理器将基带信号转换为数据并对该数据进行处理。
1120部分可以包括一个或多个单板，每个单板可以包括一个或多个处理器和一个或多个存储器。为便于说明，图11中仅示出了一个存储器和处理器。处理器用于读取和执行存储器中的程序以实现基带处理功能以及对模型管理网元的控制。若存在多个单板，各个单板之间可以互联以增强处理能力。作为一种可选的实施方式，也可以是多个单板共用一个或多个处理器，或者是多个单板共用一个或多个存储器。
应理解，图11仅为示例而非限定，上述包括收发单元和处理单元的模型管理网元可以不依赖于图11所示的结构。
当该处理模型的装置1100为芯片时，该芯片包括收发单元和处理单元。其中，收发单元可以是输入输出电路、通信接口；处理单元为该芯片上集成的处理器或者微处理器或者集成电路。
本申请实施例还提供另一种处理模型的装置1200，该装置1200可以是终端设备，也可以是芯片。该装置1200可以用于执行上述方法实施例(方法200至方法800)中由终端设备所执行的操作。
当该处理模型的装置1200为终端设备时，图12示出了一种简化的终端设备的结构示意图。如图12所示，终端设备包括处理器、存储器、射频电路、天线以及输入输出装置。处理器主要用于对通信协议以及通信数据进行处理，以及对终端设备进行控制，执行软件程序，处理软件程序的数据等。存储器主要用于存储软件程序和数据。射频电路主要用于基带信号与射频信号的转换以及对射频信号的处理。天线主要用于收发电磁波形式的射频信号。输入输出装置，例如触摸屏、显示屏，键盘等主要用于接收用户输入的数据以及对用户输出数据。需要说明的是，有些种类的终端设备可以不具有输入输出装置。
当需要发送数据时，处理器对待发送的数据进行基带处理后，输出基带信号至射频电路，射频电路将基带信号进行射频处理后将射频信号通过天线以电磁波的形式向外发送。当有数据发送到终端设备时，射频电路通过天线接收到射频信号，将射频信号转换为基带信号，并将基带信号输出至处理器，处理器将基带信号转换为数据并对该数据进行处理。为便于说明，图12中仅示出了一个存储器和处理器，在实际的终端设备产品中，可以存在一个或多个处理器和一个或多个存储器。存储器也可以称为存储介质或者存储设备等。存储器可以是独立于处理器设置，也可以是与处理器集成在一起，本申请实施例对此不做限制。
在本申请实施例中，可以将具有收发功能的天线和射频电路视为终端设备的收发单元，将具有处理功能的处理器视为终端设备的处理单元。
如图12所示，终端设备包括收发单元10和处理单元20。收发单元10也可以称为收发器、收发机、收发装置或收发电路等。处理单元20也可以称为处理器，处理单板，处理模块、处理装置等。
可选地，可以将收发单元10中用于实现接收功能的器件视为接收单元，将收发单元10中用于实现发送功能的器件视为发送单元，即收发单元10包括接收单元和发送单元。接收单元有时也可以称为接收机、接收器、接收装置或接收电路等。发送单元有时也可以称为发射机、发射器、发射装置或发射电路等。
应理解，图12仅为示例而非限定，上述包括收发单元和处理单元的终端设备可以不依赖于图12所示的结构。
当该处理模型的装置1200为芯片时，该芯片包括收发单元和处理单元。其中，收发单元可以是输入输出电路或通信接口；处理单元可以为该芯片上集成的处理器或者微处理器或者集成电路。
根据本申请实施例提供的方法，本申请还提供一种计算机程序产品，该计算机程序产品包括：计算机程序代码，当该计算机程序代码在计算机上运行时，使得该计算机执行前述方法实施例中终端设备的方法。
根据本申请实施例提供的方法，本申请还提供一种计算机程序产品，该计算机程序产品包括：计算机程序代码，当该计算机程序代码在计算机上运行时，使得该计算机执行前述方法实施例中模型管理网元的方法。
根据本申请实施例提供的方法，本申请还提供一种计算机可读介质，该计算机可读介质存储有程序代码，当该程序代码在计算机上运行时，使得该计算机执行前述方法实施例中终端设备的方法。
根据本申请实施例提供的方法，本申请还提供一种计算机可读介质，该计算机可读介质存储有程序代码，当该程序代码在计算机上运行时，使得该计算机执行前述方法实施例中模型管理网元的方法。
本申请实施例还提供了一种处理装置，包括处理器和接口；所述处理器用于执行上述任一方法实施例中的处理模型的方法。
本申请实施例还提供了一种处理模型的系统，该系统包括上述实施例中的多个模型管理网元和一个终端设备。
在本说明书中使用的术语“部件”、“模块”、“系统”等用于表示计算机相关的实体、硬件、固件、硬件和软件的组合、软件、或执行中的软件。例如，部件可以是但不限于，在处理器上运行的进程、处理器、对象、可执行文件、执行线程、程序和/或计算机。通过图示，在计算设备上运行的应用和计算设备都可以是部件。一个或多个部件可驻留在进程和/或执行线程中，部件可位于一个计算机上和/或分布在2个或更多个计算机之间。此外，这些部件可从在上面存储有各种数据结构的各种计算机可读介质执行。部件可例如根据具有一个或多个数据分组(例如来自与本地系统、分布式系统和/或网络间的另一部件交互的二个部件的数据，例如通过信号与其它系统交互的互联网)的信号通过本地和/或远程进程来通信。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各种说明性逻辑块(illustrative logical block)和步骤(step)，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统、装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
应理解，本申请实施例中的“至少一个”是指一个或者多个，“多个”是指两个或两个以上。“和/或”，描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B的情况，其中A，B可以是单数或者复数。字符“/”一般表示前后关联对象是一种“或”的关系。“以下至少一项(个)”或其类似表达，是指的这些项中的任意组合，包括单项(个)或复数项(个)的任意组合。例如，a、b和c中的至少一项(个)，可以表示：a，或，b，或，c，或，a和b，或，a和c，或，b和c，或，a、b和c。其中a、b和c分别可以是单个，也可以是多个。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统、装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是 或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外，在本申请各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。
所述功能如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(read-only memory，ROM)、随机存取存储器(random access memory，RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。

Claims (32)

  1. 一种处理模型的方法，其特征在于，包括：
    模型管理网元接收第一请求消息，所述第一请求消息包括第一模型的模型信息，所述第一请求消息用于请求注册所述第一模型；
    所述模型管理网元根据所述第一模型的模型信息，确定所述第一模型；
    所述模型管理网元发送第一响应消息，所述第一响应消息用于通知所述第一模型的注册是否成功。
  2. 根据权利要求1所述的方法，其特征在于，
    所述第一请求消息还包括终端设备的标识信息，所述终端设备的标识信息用于所述模型管理网元识别所述终端设备。
  3. 根据权利要求1或2所述的方法，其特征在于，所述方法还包括：
    当所述第一模型注册成功时，所述模型管理网元发送所述第一模型的模型配置信息，所述第一模型的模型配置信息包括第一模型的模型标识信息，所述第一模型的模型信息和所述第一模型中的至少一个，其中，所述第一模型的模型标识信息用于关联所述第一模型。
  4. 根据权利要求3所述的方法，其特征在于，所述第一模型的模型配置信息携带在所述第一响应消息中。
  5. 根据权利要求3或4所述的方法，其特征在于，所述方法还包括：
    当所述第一模型注册成功时，所述模型管理网元为所述第一模型分配所述第一模型的模型标识信息。
  6. 根据权利要求1至5中任一项所述的方法，其特征在于，当所述第一模型注册成功时，所述方法还包括：
    所述模型管理网元接收第二请求消息，所述第二请求消息包括第二指示信息，所述第二指示信息用于指示更新所述第一模型为第二模型；
    所述模型管理网元接收第一模型的模型更新信息；
    所述模型管理网元根据所述第二指示信息和所述第一模型的模型更新信息，确定更新结果；
    所述模型管理网元发送第二响应消息，所述第二响应消息包括所述更新结果，所述第二响应消息用于通知所述更新结果。
  7. 根据权利要求6所述的方法，其特征在于，
    所述第二请求消息包括终端设备的标识信息，所述终端设备的标识信息用于所述管理网元识别所述终端设备。
  8. 根据权利要求6或7所述的方法，其特征在于，所述第一模型的模型更新信息包括梯度信息和所述第二模型中的至少一个，以及第一模型的模型标识信息。
  9. 根据权利要求7或8所述的方法，其特征在于，
    所述第一模型的模型更新信息携带在所述第二请求消息中。
  10. 根据权利要求7至9中任一项所述的方法，其特征在于，所述方法还包括：
    当所述更新结果为成功时，所述模型管理网元发送第二模型的模型更新信息，所述第二模型的模型更新信息包括第二模型的模型标识信息，所述第二模型的模型标识信息用于关联所述第二模型。
  11. 根据权利要求10所述的方法，其特征在于，所述方法还包括：
    当所述更新结果为成功时，所述模型管理网元为所述第二模型分配所述第二模型的模型标识信息。
  12. 根据权利要求10或11所述的方法，其特征在于，所述第二模型的模型更新信息携带在所述第二响应消息中。
  13. 根据权利要求1至12中任一项所述的方法，其特征在于，所述模型管理网元接收第一请求消息，包括：
    所述模型管理网元接收来自基站的所述第一请求消息；
    或者，
    所述管理网元接收来自终端设备的所述第一请求消息。
  14. 根据权利要求6至13中任一项所述的方法，其特征在于，所述模型管理网元接收第二请求消息，包括：
    所述模型管理网元接收来自基站的所述第二请求消息；
    或者，
    所述模型管理网元接收来自终端设备的所述第二请求消息。
  15. 根据权利要求1至14中任一项所述的方法，其特征在于，在所述模型管理网元接收第一请求消息之前，所述方法还包括：
    所述模型管理网元发送模型注册支持信息和模型信息中的至少一个，所述模型注册支持信息用于指示接收所述模型注册支持信息的终端设备支持模型注册，所述模型信息用于指示接收所述模型信息的所述终端设备支持注册的模型。
  16. 一种处理模型的方法，其特征在于，包括：
    终端设备发送第一请求消息，所述第一请求消息包括第一模型的模型信息，所述第一请求消息用于请求注册所述第一模型；
    所述终端设备接收第一响应消息，所述第一响应消息用于通知所述第一模型的注册是否成功。
  17. 根据权利要求16所述的方法，其特征在于，
    所述第一请求消息还包括所述终端设备的标识信息，所述终端设备的标识信息用于模型管理网元识别所述终端设备。
  18. 根据权利要求17所述的方法，其特征在于，所述方法还包括：
    当所述第一模型注册成功时，所述终端设备接收所述第一模型的模型配置信息，所述第一模型的模型配置信息包括第一模型的模型标识信息、所述第一模型的模型信息和所述第一模型中的至少一个，其中，所述第一模型的模型标识信息用于关联所述第一模型。
  19. 根据权利要求18所述的方法，其特征在于，所述第一模型的模型配置信息携带在所述第一响应消息中。
  20. 根据权利要求16至19中任一项所述的方法，其特征在于，当所述第一模型注册成功时，所述方法还包括：
    所述终端设备发送第二请求消息，所述第二请求消息包括第二指示信息，所述第二指示信息用于指示更新所述第一模型为第二模型；
    所述终端设备发送第一模型的模型更新信息；
    所述终端设备接收第二响应消息，所述第二响应消息包括所述更新结果，所述第二响应消息用于通知所述更新结果。
  21. 根据权利要求20所述的方法，其特征在于，
    所述第二请求消息包括所述终端设备的标识信息，所述终端设备的标识信息用于管理网元识别所述终端设备。
  22. 根据权利要求20或21所述的方法，其特征在于，所述第一模型的模型更新信息包括梯度信息和所述第二模型中的至少一个，以及第一模型的模型标识信息。
  23. 根据权利要求20至22中任一项所述的方法，其特征在于，
    所述第一模型的模型更新信息携带在所述第二请求消息中。
  24. 根据权利要求20至23中任一项所述的方法，其特征在于，所述方法还包括：
    当所述更新结果为成功时，所述终端设备接收第二模型的模型更新信息，所述第二模型的模型更新信息包括第二模型的模型标识信息，所述第二模型的模型标识信息用于关联所述第二模型。
  25. 根据权利要求24所述的方法，其特征在于，所述第二模型的模型更新信息携带在所述第二响应消息中。
  26. 根据权利要求16至25中任一项所述的方法，其特征在于，所述终端设备接收第一响应消息，包括：
    所述终端设备接收来自基站的所述第一响应消息；
    或者，
    所述终端设备接收来自模型管理网元的所述第一响应消息。
  27. 根据权利要求20至26中任一项所述的方法，其特征在于，所述终端设备接收所述第二响应消息，包括：
    所述终端设备接收来自基站的所述第二响应消息；
    或者，
    所述终端设备接收来自模型管理网元的所述第二响应消息。
  28. 一种处理模型的装置，其特征在于，所述装置包括处理器，所述处理器与存储器耦合，所述存储器存储有指令，所述指令被所述处理器运行时，
    使得所述处理器执行如权利要求1至15中任意一项所述的方法，或者
    使得所述处理器执行如权利要求16至27中任意一项所述的方法。
  29. 一种处理模型的装置，其特征在于，所述装置包括逻辑电路和输入输出接口，所述逻辑电路用于与输入/输出接口耦合，通过所述输入/输出接口传输数据，以执行如权利要求1至15中任一项所述的方法，或者，以执行如权利要求16至27中任一项所述的方法。
  30. 一种计算机可读存储介质，其特征在于，所述计算机可读存储介质用于存储计算机程序，当所述计算机程序在计算机上运行时，使得所述计算机执行如权利要求1至15中任一项所述的方法，或使得所述计算机执行如权利要求16至27中任一项所述的方法。
  31. 一种计算机程序产品，其特征在于，所述计算机程序产品包括：计算机程序代码，当所述计算机程序代码被运行时，实现如权利要求1至15中任一项所述的方法，或实现如权利要求中16至27任一项所述的方法。
  32. 一种芯片，其特征在于，包括处理器和存储器，所述存储器用于存储计算机程序，所述处理器用于调用并运行所述存储器中存储的计算机程序，以执行如权利要求1至15中任一项所述的方法，或实现如权利要求中16至27中任一项所述的方法。
PCT/CN2023/116867 2022-11-03 2023-09-05 一种处理模型的方法和装置 WO2024093503A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211372686.8 2022-11-03
CN202211372686.8A CN118042477A (zh) 2022-11-03 2022-11-03 一种处理模型的方法和装置

Publications (1)

Publication Number Publication Date
WO2024093503A1 true WO2024093503A1 (zh) 2024-05-10

Family

ID=90929648

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/116867 WO2024093503A1 (zh) 2022-11-03 2023-09-05 一种处理模型的方法和装置

Country Status (2)

Country Link
CN (1) CN118042477A (zh)
WO (1) WO2024093503A1 (zh)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112130865A (zh) * 2020-09-30 2020-12-25 北京明略昭辉科技有限公司 一种模型管理方法、系统
CN112383891A (zh) * 2020-10-02 2021-02-19 中盈优创资讯科技有限公司 一种基于物模型自动匹配的设备注册方法和装置
WO2021163895A1 (zh) * 2020-02-18 2021-08-26 Oppo广东移动通信有限公司 网络模型的管理方法及建立或修改会话的方法、装置

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021163895A1 (zh) * 2020-02-18 2021-08-26 Oppo广东移动通信有限公司 网络模型的管理方法及建立或修改会话的方法、装置
CN112130865A (zh) * 2020-09-30 2020-12-25 北京明略昭辉科技有限公司 一种模型管理方法、系统
CN112383891A (zh) * 2020-10-02 2021-02-19 中盈优创资讯科技有限公司 一种基于物模型自动匹配的设备注册方法和装置

Also Published As

Publication number Publication date
CN118042477A (zh) 2024-05-14

Similar Documents

Publication Publication Date Title
WO2022022334A1 (zh) 基于人工智能的通信方法和通信装置
US20230209390A1 (en) Intelligent Radio Access Network
Zhou et al. Software-defined hyper-cellular architecture for green and elastic wireless access
CN108668381A (zh) 通信方法及装置
TWI745851B (zh) 藉由聯合網路及雲端資源管理之服務傳遞
CN107113188B (zh) 基于云的无线电接入网络中的层管理器装置及操作方法
CN113966592B (zh) 更新后台数据传输策略的方法和装置
US20230262478A1 (en) Model configuration method and apparatus
US20240015534A1 (en) Model processing method, communication apparatus, and system
Pedram et al. Energy efficiency in 5G cellular network systems
WO2024093503A1 (zh) 一种处理模型的方法和装置
Zhang et al. A framework for energy efficient control in heterogeneous cloud radio access networks
WO2023039905A1 (zh) Ai数据的传输方法、装置、设备及存储介质
US20230319597A1 (en) Network node and a method performed in a wireless communication network for handling configuration of radio network nodes using reinforcement learning
KR20230065806A (ko) 무선 통신 시스템에서 분할 컴퓨팅 서비스를 제공하는 방법 및 장치
WO2023246267A1 (zh) 通信方法、通信装置和系统
WO2024067351A1 (zh) 传输数据的方法和相关装置
CN108738121A (zh) 一种基于软件定义网络的无线网络资源分配方法
WO2024061125A1 (zh) 一种通信方法及装置
TWI812134B (zh) 行動通訊系統之決定上行鏈路方法、分散單元裝置及連接用戶平面功能之方法
WO2023280143A1 (zh) Ai任务的控制方法、终端、基站及存储介质
WO2024082274A1 (zh) Ai任务指示的方法、通信装置和系统
WO2024067408A1 (zh) 通信方法和装置
WO2024007156A1 (zh) 一种通信方法和装置
WO2023160560A1 (zh) 一种切换方法及相关装置