WO2023213288A1 - 模型获取方法及通信设备 - Google Patents

模型获取方法及通信设备 Download PDF

Info

Publication number
WO2023213288A1
WO2023213288A1 PCT/CN2023/092194 CN2023092194W WO2023213288A1 WO 2023213288 A1 WO2023213288 A1 WO 2023213288A1 CN 2023092194 W CN2023092194 W CN 2023092194W WO 2023213288 A1 WO2023213288 A1 WO 2023213288A1
Authority
WO
WIPO (PCT)
Prior art keywords
model
communication device
information
request
models
Prior art date
Application number
PCT/CN2023/092194
Other languages
English (en)
French (fr)
Inventor
程思涵
吴晓波
崇卫微
Original Assignee
维沃移动通信有限公司
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
Priority claimed from CN202210482186.3A external-priority patent/CN117062047A/zh
Priority claimed from CN202210483893.4A external-priority patent/CN117062098A/zh
Application filed by 维沃移动通信有限公司 filed Critical 维沃移动通信有限公司
Publication of WO2023213288A1 publication Critical patent/WO2023213288A1/zh

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/02—Arrangements for optimising operational condition

Definitions

  • This application belongs to the field of communication technology, and specifically relates to a model acquisition method and communication equipment.
  • NWDAF Network Data Analytics Function
  • NWDAF can include two parts, Analysis Logical Function (Analytics Logical Function, AnLF) and Model Training Logical Function (MTLF).
  • AnLF Analysis Logical Function
  • MTLF Model Training Logical Function
  • AnLF When AnLF receives a task initiated by a consumer, it needs to obtain the model to use the model for reasoning and generate analysis results. Therefore, for those skilled in the art, there is an urgent need to implement a model acquisition method.
  • Embodiments of the present application provide a model acquisition method and communication device, which can solve the problem of how to implement the model acquisition method.
  • the first aspect provides a model acquisition method, including:
  • the first communication device sends a first request to the second communication device, the first request is used to request to obtain information of at least one third communication device, wherein each of the third communication devices is capable of providing the first communication device Model information for at least one model required;
  • the first communication device receives information from at least one third communication device sent by the second communication device;
  • the first communication device acquires model information of multiple models based on the information of the at least one third communication device; the multiple models are used to generate data analysis result information.
  • the second aspect provides a model acquisition method, including:
  • the second communication device receives the first request sent by the first communication device, the first request is used to request to obtain information of at least one third communication device, wherein each of the third communication devices is capable of providing the first communication Model information of at least one model required by the device;
  • the second communication device determines at least one third communication device based on the first request
  • the second communication device sends the information of the at least one third communication device to the first communication device; the information of the at least one third communication device is used by the first communication device to obtain model information of multiple models. ; The multiple models are used to generate data analysis result information.
  • the third aspect provides a model acquisition method, including:
  • the third communication device receives a second request sent by the first communication device, where the second request is used to request acquisition of at least one model;
  • the third communication device sends model information of at least one model to the first communication device based on the second request;
  • the second request includes at least one of the following:
  • An analysis task identifier which is used to identify data analysis tasks applicable to the required model
  • the fourth aspect provides a model acquisition method, including:
  • the fourth communication device sends a task request to the first communication device
  • the fourth communication device receives the data analysis result information sent by the first communication device.
  • the data analysis result information is obtained by the first communication device based on analysis and processing based on multiple models.
  • the multiple models are the The first communication device obtains the model information based on the information of at least one third communication device, and each of the third communication devices can provide model information of at least one model required by the first communication device.
  • a model acquisition device including:
  • a sending module configured to send a first request to a second communication device, the first request being used to request to obtain information of at least one third communication device, wherein each of the third communication devices can provide the first communication Model information of at least one model required by the device;
  • a receiving module configured to receive information from at least one third communication device sent by the second communication device
  • An acquisition module is configured to acquire model information of multiple models based on the information of the at least one third communication device; the multiple models are used to generate data analysis result information.
  • a model acquisition device including:
  • a receiving module configured to receive a first request sent by a first communication device, the first request being used to request to obtain information of at least one third communication device, wherein each of the third communication devices can provide the first Model information of at least one model required by the communication device;
  • a processing module configured to determine at least one third communication device based on the first request
  • a sending module configured to send the information of the at least one third communication device to the first communication device; the information of the at least one third communication device is used by the first communication device to obtain model information of multiple models; The multiple models are used to generate data analysis result information.
  • a model acquisition device including:
  • a receiving module configured to receive a second request sent by the first communication device, where the second request is used to request acquisition of at least one model
  • a sending module configured to send model information of at least one model to the first communication device based on the second request; the second request includes at least one of the following:
  • An analysis task identifier which is used to identify data analysis tasks applicable to the required model
  • a model acquisition device including:
  • a sending module used to send a task request to the first communication device
  • a receiving module configured to receive data analysis result information sent by the first communication device, where the data analysis result information is obtained by analysis and processing by the first communication device based on multiple models, and the multiple models are the The first communication device obtains based on the information of at least one third communication device, and each of the third communication devices can provide model information of at least one model required by the first communication device.
  • a first communication device in a ninth aspect, includes a processor and a memory.
  • the memory stores a program or instructions executable on the processor.
  • the program or instructions are processed by the processor.
  • the processor When the processor is executed, the steps of the method described in the first aspect are implemented.
  • a first communication device including a processor and a communication interface, wherein the communication interface is used for the first communication device to send a first request to the second communication device, and the first request is used to request Obtain information of at least one third communication device, wherein each third communication device can provide model information of at least one model required by the first communication device; receive at least one first communication device sent by the second communication device. Information of three communication devices; based on the information of the at least one third communication device, obtain model information of multiple models.
  • a second communication device comprising a processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions being The processor, when executed, implements the steps of the method described in the second aspect.
  • a second communication device including a processor and a communication interface, wherein the communication interface is used to receive a first request sent by the first communication device, and the first request is used to request to obtain at least Information of a third communication device, wherein each of the third communication devices is capable of providing model information of at least one model required by the first communication device; sending the at least one third communication device to the first communication device information of the communication device; the information of the at least one third communication device is used by the first communication device to obtain model information of multiple models; a processor configured to determine at least one third communication device based on the first request .
  • a third communication device in a thirteenth aspect, includes a processor and a memory.
  • the memory stores a program or instructions executable on the processor.
  • the program or instructions are executed by the processor.
  • the processor when executed, implements the steps of the method described in the first aspect.
  • a third communication device including a processor and a communication interface, wherein the communication interface is used to receive a second request sent by the first communication device, and the second request is used to request to obtain at least A model; sending model information of at least one model to the first communication device based on the second request; the second request includes at least one of the following: an analysis task identifier, the analysis task identifier is used to identify a required model Applicable data analysis tasks; identification of required models; required number of models; model attribute information that the model needs to satisfy.
  • a fourth communication device in a fifteenth aspect, includes a processor and a memory.
  • the memory stores a program or instructions executable on the processor.
  • the program or instructions are executed by the processor.
  • the processor when executed, implements the steps of the method described in the first aspect.
  • a fourth communication device including a processor and a communication interface, wherein the communication interface is used to send a task request to a first communication device; and to receive a data analysis result sent by the first communication device.
  • Information the data analysis result information is obtained by the first communication device based on analysis and processing of multiple models, and the multiple models are obtained by the first communication device based on information from at least one third communication device,
  • Each of the third communication devices is capable of providing model information of at least one model required by the first communication device.
  • a communication system including: a first communication device, a second communication device, a third communication device and a fourth communication device.
  • the first communication device can be used to perform the method described in the first aspect.
  • the steps of the model acquisition method can be used to perform the steps of the model acquisition method as described in the second aspect
  • the third communication device can be used to perform the steps of the model acquisition method as described in the third aspect
  • the fourth communication device may be used to perform the steps of the model acquisition method described in the fourth aspect.
  • a readable storage medium is provided. Programs or instructions are stored on the readable storage medium. When the programs or instructions are executed by a processor, the steps of the method described in the first aspect are implemented, or the steps of the method are implemented. The steps of the method described in the second aspect, or the steps of implementing the method described in the third aspect, or the steps of the method described in the fourth aspect.
  • a chip in a nineteenth aspect, includes a processor and a communication interface.
  • the communication interface is coupled to the processor.
  • the processor is used to run programs or instructions to implement the method described in the first aspect. method, Or implement the method described in the second aspect, or implement the method described in the third aspect, or implement the method described in the fourth aspect.
  • a computer program/program product is provided, the computer program/program product is stored in a storage medium, and the computer program/program product is executed by at least one processor to implement the first aspect to the third aspect.
  • the first communication device sends a first request to the second communication device.
  • the first request is used to request to obtain information of at least one third communication device, where each third communication device can provide the first communication device.
  • the first communication device can obtain model information of multiple models based on the information of at least one third communication device sent by the second communication device, thereby realizing a model
  • the acquisition method has low implementation complexity and high efficiency, and multiple models are used to generate data analysis result information, making the data analysis results more accurate.
  • Figure 1 is a structural diagram of a wireless communication system applicable to the embodiment of the present application.
  • Figure 2 is one of the flow diagrams of the model acquisition method provided by the embodiment of the present application.
  • Figure 3 is one of the interactive flow diagrams of the model acquisition method provided by the embodiment of the present application.
  • Figure 4 is the second schematic flow chart of the model acquisition method provided by the embodiment of the present application.
  • Figure 5 is the third schematic flowchart of the model acquisition method provided by the embodiment of the present application.
  • Figure 6 is the fourth schematic flowchart of the model acquisition method provided by the embodiment of the present application.
  • Figure 7 is one of the structural schematic diagrams of the model acquisition device provided by the embodiment of the present application.
  • Figure 8 is the second structural schematic diagram of the model acquisition device provided by the embodiment of the present application.
  • Figure 9 is the third structural schematic diagram of the model acquisition device provided by the embodiment of the present application.
  • Figure 10 is the fourth structural schematic diagram of the model acquisition device provided by the embodiment of the present application.
  • Figure 11 is a schematic structural diagram of a communication device provided by an embodiment of the present application.
  • Figure 12 is a schematic structural diagram of a network side device according to an embodiment of the present application.
  • Figure 13 is one of the flow diagrams of the model processing method provided by the embodiment of the present application.
  • Figure 14 is the second schematic flowchart of the model processing method provided by the embodiment of the present application.
  • first, second, etc. in the description and claims of this application are used to distinguish similar objects and are not used to describe a specific order or sequence. It is to be understood that the terms so used are interchangeable under appropriate circumstances so that the embodiments of the present application can be practiced in sequences other than those illustrated or described herein, and that "first" and “second” are distinguished objects It is usually one type, and the number of objects is not limited.
  • the first object can be one or multiple.
  • “and/or” in the description and claims indicates at least one of the connected objects, and the character “/" generally indicates that the related objects are in an "or” relationship.
  • LTE Long Term Evolution
  • LTE-Advanced, LTE-A Long Term Evolution
  • LTE-A Long Term Evolution
  • CDMA Code Division Multiple Access
  • TDMA Time Division Multiple Access
  • FDMA Frequency Division Multiple Access
  • OFDMA Orthogonal Frequency Division Multiple Access
  • SC-FDMA Single-carrier Frequency Division Multiple Access
  • NR New Radio
  • FIG. 1 shows a block diagram of a wireless communication system to which embodiments of the present application are applicable.
  • the wireless communication system includes a terminal 11 and a network side device 12.
  • the terminal 11 may be a mobile phone, a tablet computer (Tablet Personal Computer), a laptop computer (Laptop Computer), or a notebook computer, a personal digital assistant (Personal Digital Assistant, PDA), a palmtop computer, a netbook, or a super mobile personal computer.
  • Tablet Personal Computer Tablet Personal Computer
  • laptop computer laptop computer
  • PDA Personal Digital Assistant
  • PDA Personal Digital Assistant
  • UMPC ultra-mobile personal computer
  • UMPC mobile Internet device
  • Mobile Internet Device MID
  • AR augmented reality
  • VR virtual reality
  • robots wearable devices
  • VUE vehicle-mounted equipment
  • PUE pedestrian terminal
  • smart home home equipment with wireless communication functions, such as refrigerators, TVs, washing machines or furniture, etc.
  • PC personal computers
  • teller machines or self-service Terminal devices such as mobile phones
  • wearable devices include: smart watches, smart bracelets, smart headphones, smart glasses, smart jewelry (smart bracelets, smart bracelets, smart rings, smart necklaces, smart anklets, smart anklets, etc.), Smart wristbands, smart clothing, etc.
  • the network side device 12 may include an access network device or a core network device, where the access network device 12 may also be called a radio access network device, a radio access network (Radio Access Network, RAN), a radio access network function or Wireless access network unit.
  • the access network device 12 may include a base station, a WLAN access point or a WiFi node, etc.
  • the base station may be called a Node B, an evolved Node B (eNB), an access point, a Base Transceiver Station (BTS), or wireless Wire base station, radio transceiver, Basic Service Set (BSS), Extended Service Set (ESS), home B-node, home evolved B-node, Transmitting Receiving Point (TRP) or some other appropriate terminology in the field.
  • BSS Basic Service Set
  • ESS Extended Service Set
  • TRP Transmitting Receiving Point
  • the base station is not limited to specific technical terms. It should be noted that in the embodiment of this application, only the base station in the NR system is used as an example for introduction. , does not limit the specific type of base station.
  • the core network equipment may include but is not limited to at least one of the following: core network node, core network function, mobility management entity (Mobility Management Entity, MME), access mobility management function (Access and Mobility Management Function, AMF), session management function (Session Management Function, SMF), User Plane Function (UPF), Policy Control Function (PCF), Policy and Charging Rules Function (PCRF), Edge Application Service Discovery function (Edge Application Server Discovery Function, EASDF), Unified Data Management (UDM), Unified Data Repository (UDR), Home Subscriber Server (HSS), centralized network configuration ( Centralized network configuration (CNC), Network Repository Function (NRF), Network Exposure Function (NEF), Local NEF (Local NEF, or L-NEF), Binding Support Function (Binding Support Function, BSF), application function (Application Function, AF), etc.
  • MME mobility management entity
  • AMF Access and Mobility Management Function
  • SMF Session Management Function
  • UPF User Plane Function
  • PCF Policy Control Function
  • NWDAF Network Data Analytics Function
  • 3GPP 3rd Generation Partnership Project
  • NWDAF has a certain artificial intelligence analysis function. NWDAF collects some data and uses built-in algorithms and analysis capabilities to analyze the results and provide them to the core network equipment to do some operation optimization or statistical analysis.
  • the network data analysis function automatically perceives and analyzes the network based on network data, and participates in the entire life cycle of network planning, construction, operation and maintenance, network optimization, and operation, making the network easy to maintain and control, and improving the efficiency of network resource usage. Improve user business experience.
  • NWDAF can provide a network data analysis function related to observed service experience (Observed Service Experience related network data analytics). With this function, network elements can request the collection terminal to access a certain server IP by initiating requests to other network elements.
  • the quality of service (QoS) information of the address such as uplink and downlink rates, packet loss rate, etc., outputs a statistical information, which includes the user experience of the terminal accessing the server.
  • NWDAF can also predict based on the terminal's historical experience of accessing the server. For example, if this terminal accesses the server in a certain period of time in the future, what is its possible user experience.
  • the analysis or prediction content that NWDAF can provide is distinguished by an analytical ID.
  • NWDAF can provide the analysis or prediction content. Analysis and prediction results corresponding to the analysis identifier.
  • Analytic ID Service Experience; therefore, as long as the network element of the NWDAF service is obtained and the Analytic ID is entered when requesting analysis, NWDAF can provide the corresponding analysis.
  • NWDAF can include two parts, the analysis logic function AnLF and the model training logic function MTLF.
  • AnLF is responsible for the inference function
  • MTLF is responsible for the training function.
  • AnLF performs inference through models and generates analysis results.
  • Figure 2 is one of the flow diagrams of the model acquisition method provided by the embodiment of the present application. As shown in Figure 2, the method provided by this embodiment includes:
  • Step 101 The first communication device sends a first request to the second communication device.
  • the first request is used to request to obtain information of at least one third communication device, where each third communication device can provide the information required by the first communication device.
  • Model information for at least one model is used to obtain information of at least one third communication device, where each third communication device can provide the information required by the first communication device.
  • the first communication device and the second communication device may be network side devices (such as core network devices).
  • the first communication device may be, for example, AnLF of NWDAF
  • the second communication device may be a network storage function (NF Repository Function, NRF).
  • the third communication device can be NWDAF's MTLF, or Analytics Data Repository Function (ADRF).
  • the first communication device sends a first request to the second communication device when it receives a task request sent by another communication device or when other triggering conditions are met.
  • the first request is used to request to obtain information of at least one third communication device.
  • the third communication device can provide model information of at least one model required by the first communication device.
  • the model information may be the model itself (for example, a file including the model) or a download address of the model.
  • Step 102 The first communication device receives the information of at least one third communication device sent by the second communication device;
  • the second communication device determines at least one third communication device based on the first request, and sends information about the at least one third communication device to the first communication device, and the first communication device receives the at least one third communication device sent by the second communication device.
  • Communication equipment information The information of the third communication device includes, for example, the analysis task identifier supported by the third communication device, the identifier and address of the third communication device, the number of models that can be used by the analysis task corresponding to the supported analysis task identifier, and the model of the supported model. Attribute information, etc.
  • Step 103 The first communication device obtains model information of multiple models based on the information of at least one third communication device; the multiple models are used to generate data analysis result information.
  • each third communication device can provide model information of at least one model, and can obtain model information of multiple models from at least one third communication device. Furthermore, the first communication device can also obtain multiple models, based on multiple models. Performing analysis and processing operations on a model can make the data analysis results more accurate.
  • the first communication device sends a first request to the second communication device.
  • the first request is used to request to obtain information of at least one third communication device, where each third communication device can provide the first communication device.
  • the first communication device can obtain the model information of at least one model required by the second communication device based on the information of at least one third communication device sent by the second communication device, thereby realizing a model
  • the acquisition method has low implementation complexity and high efficiency, and multiple models are used to generate data analysis result information, which can make the data analysis results more accurate.
  • the first request includes: an analysis task identifier, which is used to identify the data analysis task applicable to the required model; the first request also includes at least one of the following:
  • the first indication information is used to indicate a request to obtain multiple models
  • the second indication information is used to indicate a request to obtain information of multiple third communication devices
  • a sorting method is used to indicate the sorting method of the acquired information of multiple third communication devices
  • the analysis task identifier may indicate a data analysis task applicable to the model required by the first communication device.
  • the information of multiple third communication devices is sorted from large to small by model number. For example, it includes 3 MTLF, MTLF1, MTLF2, and MTLF3 to meet the needs of AnLF.
  • MTLF1, MTLF2, and MTLF3 have 2 models, 1 model, and 3 models respectively.
  • the MTLF information finally returned by NRF to AnLF is arranged in the order of MTLF3, MTLF1, and MTLF2.
  • the required number of models can be a numerical value or a range of quantities, such as limiting the minimum quantity and the maximum quantity.
  • the number of third communication devices required is similar to the number of models required.
  • the information of each third communication device includes at least one of the following:
  • the analysis task identifier supported by the third communication device the identifier of the third communication device, the address of the third communication device, the number of models corresponding to the analysis task identifier supported by the third communication device, and the model attributes of at least one model supported by the third communication device information.
  • the number of models corresponding to the analysis task identifier supported by the third communication device refers to the number of models capable of processing the analysis task corresponding to the analysis task identifier.
  • model attribute information includes at least one of the following: model usage range information, model identification, model training result evaluation information, model usage result evaluation information, model size, model inference duration, and training data information.
  • the range information used by the model includes at least one of the following: target object targeted by the model, target time range, target location range, single network slice selection auxiliary information S-NSSAI, data network name DNN; the target time range is used for indication The time range used by the model, and the target location range is used to indicate the location range used by the model.
  • the result evaluation information of model training includes at least one of the following: accuracy, error information, model training duration, and data volume of model training; the result evaluation information of model use includes at least one of the following: accuracy, error Information, how long the model was run.
  • the target object targeted by the model is, for example, mobility analysis of a certain terminal.
  • the target time range is that the model can be used within the target time range.
  • the target location range is that the model can be used within the target location range.
  • the target object is within the target location.
  • the result evaluation information of model training is used to describe the result evaluation information after the model is trained, such as the accuracy of recognition or decision-making, error information, duration, data volume, etc.
  • the error information can be mean absolute error (Mean Absolute Error, MAE), root mean square error (Root Mean Square Error, RMSE), etc.
  • the result evaluation information used by the model is similar to the result evaluation information of model training.
  • the model size refers to the amount of storage space required to store the model or run the model.
  • the model inference time is used to indicate the time it takes for the model to run to obtain converged data inference results.
  • the training data information may include, for example, at least one of the following: source information of the training data, such as the source location information of the data, information of the source communication device; time information of the training data, such as the training data is from three months ago to two months ago. Data from months ago.
  • the model attribute information that the model needs to meet is, for example: the target object of the model is a terminal in a certain area, the model size is within 500MB, the accuracy rate used by the model is greater than a certain threshold, and the source communication device of the training data used by the model is communication device a, and the training data is data from the previous week or month, etc.
  • the method also includes:
  • the first communication device acquires multiple models based on model information of the multiple models
  • the first communication device performs analysis and processing based on multiple models to obtain data analysis result information.
  • the model information of multiple models is, for example, the addresses of the models
  • multiple models can be acquired based on the addresses of the models, and analysis and processing can be performed based on the multiple acquired models to obtain data analysis result information.
  • the first communication device performs a model inference operation based on each of the multiple models to obtain multiple data inference results;
  • the first communication device processes multiple data inference results to obtain data analysis result information.
  • the above processing such as aggregation, voting and other operations, generates the final data analysis result information.
  • the first communication device processes multiple data inference results to obtain data analysis result information, which can be implemented in at least one of the following ways:
  • the first communication device performs a weighted average on multiple data inference results to obtain data analysis result information
  • the first communication device performs an average operation on multiple inference result data to obtain analysis result information
  • the first communication device accumulates multiple inference result data to obtain analysis result information
  • the first communication device accumulates multiple inference result data taking performance into account to obtain analysis result information.
  • weighted average, averaging, etc. are aggregation operations, and accumulation and accumulation considering performance are the specific expressions of cumulative voting.
  • the performance of the model can be considered when accumulating. For example, if the accuracy of the model is 0.6, the weight of the inference result can be 0.6.
  • model inference, analysis and other processing are performed based on multiple models, and the obtained results have better performance and higher accuracy, and are not easily affected by changes in network conditions and network data.
  • step 101 it also includes:
  • the first communication device receives the task request sent by the fourth communication device
  • Step 101 may specifically be:
  • the first communication device sends the first request to the second communication device based on the task request;
  • the first communication device performs analysis and processing based on multiple models, and after obtaining the data analysis result information, it also includes:
  • the first communication device sends data analysis result information to the fourth communication device.
  • the fourth communication device may be a consumer device, such as an Application Function (AF).
  • the fourth communication device may send a task request to the first communication device to request the acquisition of data corresponding to certain data analysis tasks. Analysis result information.
  • AF Application Function
  • the content included in the task request may be the same as the content included in the first request.
  • step 103 can be implemented in the following ways:
  • the first communication device determines at least one target communication device based on the information of at least one third communication device and the model attribute information that the model needs to satisfy, and obtains model information of multiple models from the at least one target communication device, wherein each target communication device The device can provide model information of the model that matches the model attribute information.
  • the analysis task identifiers supported by a certain third communication device are identifier 1 and identifier 2.
  • the number of models corresponding to identifier 1 is 5, the number of models corresponding to identifier 2 is 10, and there are 3 out of the 5 models corresponding to identifier 1.
  • the usage scope of each model is consistent with the usage scope in the model attribute information that needs to be met.
  • the result evaluation information of model training and use matches the result evaluation information in the model attribute information required above.
  • the model size matches the model size in the model attribute information. also matches, so the above-mentioned third communication device can be used as the target communication device.
  • obtain model information of multiple models from at least one target communication device which can be implemented in the following ways:
  • the first communication device sends a second request to at least one target communication device, and the second request is used to obtain model information of a plurality of models from at least one target communication device;
  • the first communication device receives model information of a plurality of models sent by at least one target communication device;
  • the second request includes at least one of the following:
  • Analysis task identifier which is used to identify the data analysis tasks applicable to the required model
  • the first communication device receives model information of multiple models sent by at least one target communication device, and each target communication device can send model information of one or more models.
  • the number of at least one target communication device may be one, and the first communication device sends the second request to the at least one target communication device, including:
  • the first communication device sends a second request to one of the at least one target communication device
  • the first communication device receives model information of multiple models sent by at least one target communication device, including:
  • the first communication device receives model information of multiple models sent by a target communication device; at this time, the target communication device can provide model information of multiple models.
  • the method before the first communication device sends the second request to one of the at least one target communication device, the method further includes:
  • the first communication device determines that one target communication device can provide model information of a plurality of models.
  • the number of at least one target communication device may be multiple, and the first communication device sends a second request to at least one target communication device, including:
  • the first communication device sends a second request to a plurality of target communication devices in the at least one target communication device;
  • the first communication device receives model information of multiple models sent by at least one target communication device, including:
  • the first communication device receives model information of multiple models sent by multiple target communication devices
  • the method before the first communication device sends the second request to the plurality of target communication devices in the at least one target communication device, the method further includes:
  • the first communication device determines that each of the at least one target communication device is unable to provide model information for all models of the plurality of models.
  • each target communication device can send model information of one or more models.
  • Each target communication device cannot provide model information of all the multiple models required by the first communication device, and therefore needs to be obtained from multiple target communication devices.
  • the first communication device sends a second request to one target communication device or multiple target communication devices among the at least one target communication device, thereby achieving the acquisition of model information of multiple models, with lower implementation complexity and higher efficiency. high.
  • step 1 MTLF (or NWDAF, or NWDAF including MTLF) can send a capability registration message to the NRF second communication device.
  • the capability registration message includes at least one of the following: identification , supported analysis task identifiers, network function types, network function instance identifiers, the number of models corresponding to the supported analysis task identifiers, and model attribute information of at least one supported model.
  • the capability registration message can be carried through signaling Nnrf_NFManagement_NFRegister Register.
  • Network function instance ID refers to the identification information of the communication device registered this time, such as its fully qualified domain name (Fully Qualified Domain Name, FQDN), used to indicate the location of this communication device and the connection to this communication device) information or IP address information (another type of identifying information).
  • FQDN Fully Qualified Domain Name
  • the supported task analysis ID indicates the type of tasks that the NWDAF can perform.
  • Steps 2 and 3 NRF stores the MTLF information and sends a response message.
  • the response message can be carried through Nnrf_NFManagement_NFRegister response signaling.
  • the response message is used to notify the third communication device that the registration is successful.
  • Step 4 The consumer device sends a task request to AnLF (or NWDAF, or AnLF including MTLF), for example, carried through Nnwdaf_AnalyticsInfo_Request signaling.
  • AnLF or NWDAF, or AnLF including MTLF
  • Step 5 Send a first request to NRF.
  • the first request may be carried, for example, through Nnrf_NFDiscovery_Request signaling.
  • the first request includes a task analysis identifier and may also include at least one of the following:
  • the first indication information is used to indicate a request to obtain multiple models
  • the second indication information is used to indicate a request to obtain information of multiple third communication devices
  • a sorting method is used to indicate the sorting method of the acquired information of multiple third communication devices
  • Step 6 The NRF determines the MTLF information to be fed back according to the first request of AnLF.
  • NRF searches for all registered MTLFs, selects the final MTLF by matching information, and generates a MTLF list.
  • the MTLF information is fed back through the MTLF list.
  • AnLF's first request includes the required If the number of models and the indication information of multiple MTLF information are required, then NRF will search for all MTLFs that support the analysis task identifier included in the first request and feedback the MTLF information of multiple models corresponding to the analysis task identifier. If no multiple MTLFs meet the conditions, other MTLFs that do not meet the model quantity requirements will be fed back. At this time, the number of models owned by the multiple MTLFs needs to meet the required number of models in the first request.
  • NRF will select the information of multiple MTLFs with models corresponding to the analysis task identifier and provide feedback.
  • NRF will match the MTLF information with multiple models corresponding to the analysis task identifier and provide feedback. If not, a failure message is returned.
  • Step 7 NRF returns qualified MTLF information to AnLF.
  • Step 8 AnLF determines the target MTLF of the model to be obtained.
  • Step 10 AnLF uses multiple models for analysis and processing.
  • Step 11 AnLF returns data analysis result information to the consumer device. For example, carried through Nnwdaf_AnalyticsInfo_Request response signaling.
  • the consumer device can also send data analysis result information to user devices such as terminals.
  • the execution subject may be a model acquisition device.
  • the model acquisition device executing the model acquisition method is used as an example to illustrate the model acquisition device provided by the embodiment of the present application.
  • Figure 4 is the second schematic flow chart of the model acquisition method provided by this application. As shown in Figure 4, the model acquisition method provided by this embodiment includes:
  • Step 201 The second communication device receives a first request sent by the first communication device.
  • the first request is used to request to obtain information of at least one third communication device, where each third communication device can provide the information required by the first communication device.
  • Model information of at least one model is used to obtain information of at least one third communication device, where each third communication device can provide the information required by the first communication device.
  • Step 202 The second communication device determines at least one third communication device based on the first request
  • Step 203 The second communication device sends information of at least one third communication device to the first communication device; the information of at least one third communication device is used by the first communication device to obtain model information of multiple models; multiple models use To generate data analysis result information.
  • the first request includes: an analysis task identifier, which is used to identify the data analysis task for which the required model is applicable; the first request also includes at least one of the following:
  • the first indication information is used to indicate a request to obtain multiple models
  • the second indication information is used to indicate a request to obtain information of multiple third communication devices
  • a sorting method is used to indicate the sorting method of the acquired information of multiple third communication devices
  • the method also includes:
  • the second communication device determines the number of the at least one third communication device to return based on the first request sent by the first communication device.
  • the second communication device determines the number of the at least one third communication device based on the first request sent by the first communication device, including at least one of the following:
  • the second communication device determines that the number of the at least one third communication device is one based on the first indication information and/or the required number of models, and one third communication device can provide multiple models or Model information of the required number of models;
  • the second communication device determines a plurality of the at least one third communication device based on the first indication information and/or the required model number, and each third communication device in the plurality of third communication devices None of the communication devices can provide model information for all models of the plurality of models or the required number of models;
  • the second communication device determines that the number of the at least one third communication device is multiple based on the second indication information
  • the second communication device determines the number of the at least one third communication device to be the number of required third communication devices based on the number of the required third communication devices.
  • the model attribute information includes at least one of the following: model usage range information, model identification, model training result evaluation information, model usage result evaluation information, model size, model execution time, and source information of training data. , time information of training data; model size is used to represent the storage space required for model storage or operation, and model inference duration is used to represent the time it takes for the model to run to obtain data inference results.
  • the result evaluation information of model training includes at least one of the following: accuracy, error information, model training duration, and data volume of model training; the result evaluation information of model use includes at least one of the following: accuracy. , error information, model running time.
  • the information of each third communication device includes at least one of the following:
  • the analysis task identifier supported by the third communication device the identifier of the third communication device, the address of the third communication device, the number of models corresponding to the analysis task identifier supported by the third communication device, or the third communication device Model attribute information of at least one model supported by the communication device.
  • the method also includes:
  • the second communication device receives a capability registration message sent by the third communication device.
  • the capability registration message includes at least one of the following: an analysis task identifier supported by the third communication device, and a network function type of the third communication device. , the network function instance identifier of the third communication device, the number of models corresponding to the analysis task identifier supported by the third communication device, and the model attribute information of at least one model supported by the third communication device.
  • Figure 5 is the third schematic flow chart of the model acquisition method provided by this application. As shown in Figure 5, the model acquisition method provided by this embodiment includes:
  • Step 301 The third communication device receives a second request sent by the first communication device, where the second request is used to request acquisition of at least one model;
  • Step 302 The third communication device sends model information of at least one model to the first communication device based on the second request;
  • the second request includes at least one of the following:
  • An analysis task identifier which is used to identify data analysis tasks applicable to the required model
  • the method also includes:
  • the third communication device sends a capability registration message to the second communication device.
  • the capability registration message includes at least one of the following: an analysis task identifier supported by the third communication device, a network function type of the third communication device, The network function instance identifier of the third communication device, the number of models corresponding to the analysis task identifier supported by the third communication device, and the model attribute information of at least one model supported by the third communication device.
  • the model attribute information includes at least one of the following: model usage range information, model identification, model training result evaluation information, model usage result evaluation information, model size, model execution time, and source information of training data. , time information of training data; model size is used to represent the storage space required for model storage or operation, and model inference duration is used to represent the time it takes for the model to run to obtain data inference results.
  • the result evaluation information of model training includes at least one of the following: accuracy, error information, model training duration, and data volume of model training; the result evaluation information of model use includes at least one of the following: accuracy. , error information, model running time.
  • the information of each third communication device includes at least one of the following:
  • the analysis task identifier supported by the third communication device the identifier of the third communication device, the address of the third communication device, the number of models corresponding to the analysis task identifier supported by the third communication device, or the third communication device Model attribute information of at least one model supported by the communication device.
  • Figure 6 is the fourth schematic flow chart of the model acquisition method provided by this application. As shown in Figure 6, the model acquisition method provided by this embodiment includes:
  • Step 401 The fourth communication device sends a task request to the first communication device
  • Step 402 The fourth communication device receives the data analysis result information sent by the first communication device.
  • the data analysis result information is obtained by the first communication device based on analysis and processing based on multiple models.
  • the multiple models are obtained by the first communication device based on at least one first communication device.
  • each third communication device can provide model information of at least one model required by the first communication device.
  • the model attribute information includes at least one of the following: model usage range information, model identification, model training result evaluation information, model usage result evaluation information, model size, model execution time, and source information of training data. , time information of training data; model size is used to represent the storage space required for model storage or operation, and model inference duration is used to represent the time it takes for the model to run to obtain data inference results.
  • the result evaluation information of model training includes at least one of the following: accuracy, error information, model training duration, and data volume of model training; the result evaluation information of model use includes at least one of the following: accuracy. , error information, model running time.
  • the information of each third communication device includes at least one of the following:
  • the analysis task identifier supported by the third communication device the identifier of the third communication device, the address of the third communication device, the number of models corresponding to the analysis task identifier supported by the third communication device, or the third communication device Model attribute information of at least one model supported by the communication device.
  • Figure 7 is one of the structural schematic diagrams of the model acquisition device provided by this application.
  • the model acquisition device provided in this embodiment can be applied to the first communication device.
  • the model acquisition device provided by this embodiment includes:
  • Sending module 110 configured to send a first request to a second communication device, where the first request is used to request to obtain information of at least one third communication device, wherein each of the third communication devices can provide the first Model information of at least one model required by the communication device;
  • the receiving module 120 is configured to receive information from at least one third communication device sent by the second communication device;
  • the acquisition module 130 is configured to acquire model information of multiple models based on the information of the at least one third communication device; the multiple models are used to generate data analysis result information.
  • the acquisition module 130 is also used to:
  • the device further includes: a processing module, configured to perform analysis and processing based on the multiple models to obtain the data analysis result information.
  • a processing module configured to perform analysis and processing based on the multiple models to obtain the data analysis result information.
  • the acquisition processing module is specifically used for:
  • the plurality of data inference results are processed to obtain the data analysis result information.
  • the acquisition module 130 is specifically configured to perform at least one of the following:
  • the multiple inference result data are accumulated taking performance into account to obtain the analysis result information.
  • the receiving module 120 is also used to:
  • the sending module 110 is specifically used for:
  • the acquisition module 130 is specifically used to:
  • At least one target communication device is determined, and model information of the plurality of models is obtained from the at least one target communication device, wherein each The target communication device can provide model information of a model matching the model attribute information.
  • the first request includes:
  • An analysis task identifier which is used to identify data analysis tasks applicable to the required model
  • the first request also includes at least one of the following:
  • the first indication information is used to indicate a request to obtain multiple models
  • the second indication information is used to indicate a request to obtain information of multiple third communication devices
  • a sorting method is used to indicate the sorting method of the acquired information of multiple third communication devices
  • the information of each third communication device includes at least one of the following:
  • the analysis task identifier supported by the third communication device the identifier of the third communication device, the address of the third communication device, the number of models corresponding to the analysis task identifier supported by the third communication device, the third Model attribute information of at least one model supported by the communication device.
  • the sending module 110 is also used to:
  • the receiving module 120 is also configured to receive model information of the multiple models sent by the at least one target communication device;
  • the second request includes at least one of the following:
  • An analysis task identifier which is used to identify data analysis tasks applicable to the required model
  • the sending module 110 is specifically used to:
  • the receiving module 120 is specifically used for:
  • Processing module also used for:
  • the one target communication device is capable of providing model information of the plurality of models.
  • the sending module 110 is specifically used to:
  • the receiving module 120 is specifically used for:
  • Processing module also used for:
  • each of the at least one target communication device is unable to provide model information for all models of the plurality of models.
  • the model attribute information includes at least one of the following: model usage range information, model identification, model training result evaluation information, model usage result evaluation information, model size, model inference duration, and source information of training data.
  • model usage range information time information of training data
  • model size is used to represent the storage space required for model storage or operation
  • model inference duration is used to represent the time it takes for the model to run to obtain data inference results.
  • the result evaluation information of model training includes at least one of the following: accuracy, error information, model training duration, and data volume of model training; the result evaluation information of model use includes at least one of the following: accuracy. , error information, model running time.
  • the device of this embodiment can be used to execute the method of any of the foregoing first communication device side method embodiments. Its specific implementation process and technical effects are the same as those in the first communication device side method embodiment. For details, please refer to Chapter 1 The detailed introduction in the communication device side method embodiment will not be described again here.
  • Figure 8 is the second structural schematic diagram of the model acquisition device provided by this application.
  • the model acquisition device provided in this embodiment can be applied to the second communication device.
  • the model acquisition device provided by this embodiment includes:
  • the receiving module 210 is configured to receive a first request sent by a first communication device, where the first request is used to request acquisition of information of at least one third communication device, wherein each of the third communication devices can provide the third communication device.
  • the processing module 220 is configured to determine at least one third communication device based on the first request
  • Sending module 230 configured to send the information of the at least one third communication device to the first communication device; the information of the at least one third communication device is used by the first communication device to obtain model information of multiple models. ; The multiple models are used to generate data analysis result information.
  • the first request includes: an analysis task identifier, which is used to identify the data analysis task for which the required model is applicable; the first request also includes at least one of the following:
  • the first indication information is used to indicate a request to obtain multiple models
  • the second indication information is used to indicate a request to obtain information of multiple third communication devices
  • a sorting method is used to indicate the sorting method of the acquired information of multiple third communication devices
  • processing module 220 is also used to:
  • the number of the at least one third communication device is determined based on the first request sent by the first communication device.
  • the second communication device determines the number of the at least one third communication device based on the first request sent by the first communication device, including at least one of the following:
  • the second communication device determines that the number of the at least one third communication device is one based on the first indication information and/or the required number of models, and one third communication device can provide multiple models or Model information of the required number of models;
  • the second communication device determines a plurality of the at least one third communication device based on the first indication information and/or the required model number, and each third communication device in the plurality of third communication devices None of the communication devices can provide model information for all models of the plurality of models or the required number of models;
  • the second communication device determines that the number of the at least one third communication device is multiple based on the second indication information
  • the second communication device determines the number of the at least one third communication device to be the number of required third communication devices based on the number of the required third communication devices.
  • the model attribute information includes at least one of the following: model usage range information, model identification, model training result evaluation information, model usage result evaluation information, model size, model execution time, and source information of training data. , time information of training data; model size is used to represent the storage space required for model storage or operation, and model inference duration is used to represent the time it takes for the model to run to obtain data inference results.
  • the result evaluation information of model training includes at least one of the following: accuracy, error information, model training duration, and data volume of model training; the result evaluation information of model use includes at least one of the following: accuracy. , error information, model running time.
  • the information of each third communication device includes at least one of the following:
  • the analysis task identifier supported by the third communication device the identifier of the third communication device, the address of the third communication device, the number of models corresponding to the analysis task identifier supported by the third communication device, or the third communication device Model attribute information of at least one model supported by the communication device.
  • the receiving module 210 is also used to:
  • a capability registration message sent by a third communication device including at least one of the following: an analysis task identifier supported by the third communication device, a network function type of the third communication device, the third communication device The network function instance identifier of the device, the number of models corresponding to the analysis task identifier supported by the third communication device, and the model attribute information of at least one model supported by the third communication device.
  • the device of this embodiment can be used to execute the method of any of the foregoing second communication device side method embodiments. Its specific implementation process and technical effects are the same as those in the second communication device side method embodiment. For details, please refer to Chapter 1 The detailed introduction in the second communication device side method embodiment will not be described again here.
  • Figure 9 is the third structural schematic diagram of the model acquisition device provided by this application.
  • the model acquisition device provided in this embodiment can be applied to the third communication device.
  • the model acquisition device provided by this embodiment includes:
  • the receiving module 310 is configured to receive a second request sent by the first communication device, where the second request is used to request acquisition of at least one model;
  • Sending module 320 configured to send model information of at least one model to the first communication device based on the second request; the second request includes at least one of the following:
  • An analysis task identifier which is used to identify data analysis tasks applicable to the required model
  • the model attribute information includes at least one of the following: model usage range information, model identification, model training result evaluation information, model usage result evaluation information, model size, model execution time, and source information of training data. , time information of training data; model size is used to represent the storage space required for model storage or operation, and model inference duration is used to represent the time it takes for the model to run to obtain data inference results.
  • the result evaluation information of model training includes at least one of the following: accuracy, error information, model training duration, and data volume of model training; the result evaluation information of model use includes at least one of the following: accuracy. , error information, model running time.
  • the information of each third communication device includes at least one of the following:
  • the analysis task identifier supported by the third communication device the identifier of the third communication device, the address of the third communication device, the number of models corresponding to the analysis task identifier supported by the third communication device, or the third communication device Model attribute information of at least one model supported by the communication device.
  • the sending module 320 is also used to:
  • the capability registration message including at least one of the following: an analysis task identifier supported by the third communication device, a network function type of the third communication device, The network function instance identifier, the number of models corresponding to the analysis task identifier supported by the third communication device, and the model attribute information of at least one model supported by the third communication device.
  • the device of this embodiment can be used to execute the method of any of the foregoing third communication device side method embodiments. Its specific implementation process and technical effects are the same as those in the third communication device side method embodiment. For details, see Chapter 1 The detailed introduction of the three communication device side method embodiments will not be repeated here.
  • Figure 10 is the third structural schematic diagram of the model acquisition device provided by this application.
  • the model acquisition device provided in this embodiment can be applied to the fourth communication device.
  • the model acquisition device provided by this embodiment includes:
  • Sending module 410 used to send a task request to the first communication device
  • the receiving module 420 is configured to receive data analysis result information sent by the first communication device.
  • the data analysis result information is obtained by analysis and processing by the first communication device based on multiple models.
  • the multiple models are the The first communication device obtains the model information based on the information of at least one third communication device, and each of the third communication devices can provide model information of at least one model required by the first communication device.
  • the task request includes: an analysis task identifier, which is used to identify the data analysis task applicable to the required model; the task request also includes at least one of the following:
  • the first indication information is used to indicate a request to obtain multiple models
  • the second indication information is used to indicate a request to obtain information of multiple third communication devices
  • a sorting method is used to indicate the sorting method of the acquired information of multiple third communication devices
  • the model attribute information includes at least one of the following: model usage range information, model identification, model training result evaluation information, model usage result evaluation information, model size, model execution time, and source information of training data. , time information of training data; model size is used to represent the storage space required for model storage or operation, and model inference duration is used to represent the time it takes for the model to run to obtain data inference results.
  • the result evaluation information of model training includes at least one of the following: accuracy, error information, model training duration, and data volume of model training; the result evaluation information of model use includes at least one of the following: accuracy. , error information, model running time.
  • the information of each third communication device includes at least one of the following:
  • the analysis task identifier supported by the third communication device the identifier of the third communication device, the address of the third communication device, the number of models corresponding to the analysis task identifier supported by the third communication device, or the third communication device Model attribute information of at least one model supported by the communication device.
  • the device of this embodiment can be used to execute the method of any of the foregoing fourth communication device side method embodiments. Its specific implementation process and technical effects are the same as those in the fourth communication device side method embodiment. For details, see Chapter 1 The detailed introduction of the fourth communication device side method embodiment will not be described again here.
  • the model acquisition device in the embodiment of the present application may be an electronic device, such as an electronic device with an operating system, or may be a component in the electronic device, such as an integrated circuit or chip.
  • the electronic device may be a terminal or other devices other than the terminal.
  • terminals may include but are not limited to the types of terminals 11 listed above, and other devices may be servers, network attached storage (Network Attached Storage, NAS), etc., which are not specifically limited in the embodiment of this application.
  • NAS Network Attached Storage
  • the model acquisition device provided by the embodiment of the present application can implement each process implemented by the method embodiments of Figures 2 to 6, and achieve the same technical effect. To avoid duplication, the details will not be described here.
  • this embodiment of the present application also provides a communication device 1100, which includes a processor 1101 and a memory 1102.
  • the memory 1102 stores programs or instructions that can be run on the processor 1101, such as , when the communication device 1100 is a terminal, when the program or instruction is executed by the processor 1101, each step of the above model acquisition method embodiment is implemented, and the same technical effect can be achieved.
  • the communication device 1100 is a network-side device, when the program or instruction is executed by the processor 1101, each step of the above model acquisition method embodiment is implemented, and the same technical effect can be achieved. To avoid duplication, the details are not repeated here.
  • An embodiment of the present application also provides a first communication device, including a processor and a communication interface.
  • the communication interface is used for the first communication device to send a first request to the second communication device.
  • the first request is used to request to obtain at least Information of a third communication device, wherein each third communication device is capable of providing model information of at least one model required by the first communication device; receiving at least one third communication sent by the second communication device Information of the device; based on the information of the at least one third communication device, obtain model information of multiple models; the multiple models are used to generate data analysis result information.
  • This first communication device embodiment corresponds to the above-mentioned first communication device side method embodiment.
  • Each implementation process and implementation manner of the above-mentioned method embodiment can be applied to this first communication device embodiment, and can achieve the same technical effect. .
  • An embodiment of the present application also provides a second communication device, including a processor and a communication interface.
  • the communication interface is used to receive a first request sent by the first communication device.
  • the first request is used to request to obtain at least one third party.
  • Information of communication devices wherein each of the third communication devices is capable of providing model information of at least one model required by the first communication device; sending the information of the at least one third communication device to the first communication device information; the information of the at least one third communication device is used by the first communication device to obtain model information of multiple models; the multiple models are used to generate data analysis result information; a processor, configured to based on the first A request to identify at least one third communications device.
  • This second communication device embodiment corresponds to the above-mentioned second communication device-side method embodiment. Each implementation process and implementation manner of the above-mentioned method embodiment can be applied to this second communication device embodiment, and can achieve the same technical effect. .
  • Embodiments of the present application also provide a third communication device, including a processor and a communication interface, the communication interface being used to receive a second request sent by the first communication device, where the second request is used to request acquisition of at least one model; Send model information of at least one model to the first communication device based on the second request; the second request includes at least one of the following: an analysis task identifier, the analysis task identifier is used to identify data to which the required model is applicable Analysis task; identification of the required model; required number of models; model attribute information that the model needs to satisfy.
  • This third communication device embodiment corresponds to the above-mentioned third communication device-side method embodiment. Each implementation process and implementation manner of the above-mentioned method embodiment can be applied to this third communication device embodiment, and can achieve the same technical effect. .
  • Embodiments of the present application also provide a fourth communication device, including a processor and a communication interface, the communication interface being used to send a task request to the first communication device; and receiving the data analysis result information sent by the first communication device.
  • Information the data analysis result information is obtained by analysis and processing by the first communication device based on multiple models, and the multiple models are obtained by the first communication device based on information of at least one third communication device,
  • Each of the third communication devices is capable of providing model information of at least one model required by the first communication device.
  • This fourth communication device embodiment corresponds to the above-mentioned fourth communication device-side method embodiment.
  • Each implementation process and implementation manner of the above-mentioned method embodiment can be applied to this fourth communication device embodiment, and can achieve the same technical effect. .
  • the embodiment of the present application also provides a network side device.
  • the network side device 1200 includes: a processor 1201, a network interface 1202, and a memory 1203.
  • the network interface 1202 is, for example, a common public radio interface (CPRI).
  • CPRI common public radio interface
  • the network side device 1200 in this embodiment of the present invention also includes: instructions or programs stored in the memory 1203 and executable on the processor 1201.
  • the processor 1201 calls the instructions or programs in the memory 1203 to execute Figures 7-10
  • the execution methods of each module are shown and achieve the same technical effect. To avoid repetition, they will not be described in detail here.
  • the first communication device, the second communication device, the third communication device and the fourth communication device may adopt the structure of the network side device.
  • Embodiments of the present application also provide a readable storage medium.
  • Programs or instructions are stored on the readable storage medium.
  • the program or instructions are executed by a processor, each process of the above model acquisition method embodiment is implemented, and the same can be achieved. The technical effects will not be repeated here to avoid repetition.
  • the processor is the processor in the terminal described in the above embodiment.
  • the readable storage medium includes computer readable storage media, such as computer read-only memory ROM, random access memory RAM, magnetic disk or optical disk, etc.
  • An embodiment of the present application further provides a chip.
  • the chip includes a processor and a communication interface.
  • the communication interface is coupled to the processor.
  • the processor is used to run programs or instructions to implement the above model acquisition method embodiment. Each process can achieve the same technical effect. To avoid duplication, it will not be described again here.
  • chips mentioned in the embodiments of this application may also be called system-on-chip, system-on-a-chip, system-on-chip or system-on-chip, etc.
  • Embodiments of the present application further provide a computer program/program product.
  • the computer program/program product is stored in a storage medium.
  • the computer program/program product is executed by at least one processor to implement the above model acquisition method embodiment.
  • Each process can achieve the same technical effect. To avoid repetition, we will not go into details here.
  • An embodiment of the present application also provides a communication system, including: a first communication device, a second communication device, a third communication device, and a fourth communication device.
  • the first communication device, the second communication device, the third communication device And the fourth communication device may be used to perform the steps of the model acquisition method as described above.
  • the embodiment of the present application also provides a model processing method, as shown in Figure 13.
  • the model processing method includes:
  • Step 1301 The first network element sends a registration request message to the second network element, where the registration request message includes capability information of the first network element, where the capability information includes at least one of model quantity information and model information;
  • Step 1302 The first network element receives a registration request response message from the second network element
  • the model quantity information is used to indicate the number of models corresponding to the analysis task identifier supported by the first network element; the model information includes at least one of the following models supported by the first network element corresponding to the analysis task identifier. kind of information:
  • Model size which is used to indicate the storage space required to store or run the model
  • Inference duration which is used to indicate the time required to perform model inference operations based on the model
  • Training data source information which is used to indicate at least one of the location information and network element information of the training data source used by the model in the training phase;
  • Training data time information which is used to indicate the generation time of the training data used by the model in the training phase.
  • the above-mentioned first network element can be understood as a model training network element.
  • it can be a model training logical function (MTLF), or it can be called a network data analysis function (MTLF).
  • MTLF contained in Network Data Analytics Function (NWDAF).
  • NWDAAF Network Data Analytics Function
  • the above-mentioned second network element can be understood as NRF.
  • registration request message may also include other information in addition to capability information.
  • it may include:
  • the NF type is used to indicate what kind of network element the registered network element is, such as NWDAF type or MTLF type.
  • Network element instance identification information used to indicate the network element representation information of the registered network, such as fully qualified domain name (Fully Qualified Domain Name, FQDN) or IP address information;
  • the supported analysis task identifier (analytic ID) is used to indicate the type of tasks that the NWDAF network element can perform.
  • the above model performance information can be used to represent the accuracy and error value of the model output results, etc.
  • the model performance information may include at least one of the following: first performance information, the first performance information is used to indicate the performance that the model can achieve during the training phase; second performance information, the second performance The information is used to indicate the performance that the model can achieve during the inference phase.
  • the second network element can store the information carried in the registration request message, and after confirming that the registration is passed, send a registration request response message to the first network element.
  • the above model size can be expressed by the parameters of the model.
  • the above training data source information and training data time information can be understood as model training data information.
  • the registration request message may also include other model training data information, which is not further limited here.
  • the third network element when it needs to call the model in the first network element to perform the target analysis task, it can first send a second request message to the second network element to request to query the first network that can perform the target analysis task. Yuan. Since capability information is stored in the second network element, N first network elements that can be used to perform the model of the target analysis task can be determined based on the capability information. Then the third network element selects the target network element from the N first network elements, and calls the model in the target network element to perform the target analysis task, that is, the model in the target network element is used to perform model inference on the target task.
  • the embodiment of the present application can further improve the query efficiency of the first network element.
  • the second network element can match the first network element to obtain the model number requirement that meets the target analysis task, thereby improving the reliability of model inference for the target analysis task.
  • the first network element can be matched to obtain the first network element that meets the model requirements (such as model performance requirements) for the target analysis task, thereby improving model performance for the target analysis task. Reliability of reasoning.
  • the above-mentioned third network element can be understood as a data analysis logical function (Analytics logical function, AnLF), or it can be called AnLF included in NWDAF.
  • AnLF Analytics logical function
  • the first network element sends a registration request message to the second network element.
  • the registration request message includes the capability information of the first network element.
  • the capability information includes at least one of model quantity information and model information.
  • Item The first network element receives a registration request response message from the second network element.
  • the method further includes:
  • the first network element receives a first request message from the third network element, where the first request message is used to obtain a target model that can be used to perform a target analysis task;
  • the first network element sends a first request response message to the third network element, where the first request response message includes at least one of the target model and address information used to obtain the target model.
  • the third network element may send a first request message to the first network element to request to obtain a target model that can perform the target analysis task.
  • the above target analysis task request may be information received by the third network element from the fourth network element.
  • the fourth network element may send a task request message to the third network element, and the task request message may include the information of the target analysis task.
  • the analysis task identification may further include task limitation information.
  • the task limitation information may include the limitation information of the data analysis task model and the analysis purpose. (analytic target), where the limited information of the data analysis task model can be called the limited information of the machine learning model (Machine Learning model filter info).
  • task limitation information can be used to limit the scope of the task, such as the analysis object targeted by the model (such as mobility analysis for a certain UE), task target time, area of interest (AOI), single network slice selection Auxiliary information (Single Network Slice Selection Assistance Information, S-NSSAI) and data network name (Data Network Name, DNN), etc.
  • the analysis object targeted by the model such as mobility analysis for a certain UE
  • task target time area of interest
  • AOI area of interest
  • single network slice selection Auxiliary information Single Network Slice Selection Assistance Information
  • S-NSSAI Single Network Slice Selection Assistance Information
  • DNN Data Network Name
  • the first network element may determine a target model that can be used to perform the target analysis task based on the first request message, and then indicate the target model through the first request response message.
  • the third network element can use the target model to perform a model inference operation for the target task, generate a target analysis report, and finally send the target analysis report to the fourth network element.
  • the first request message includes the analysis task identification and model definition information of the target analysis task, and the model definition information includes at least one of the following:
  • the above model definition information may be determined based on the above task definition information, may be determined by a protocol (for example, the protocol stipulates model definition information corresponding to different tasks), or may be determined independently by the third network element. Among them, the target model fed back by the first network element should meet the above model qualification information.
  • the specific situation is as follows:
  • the above-mentioned limiting information on the number of models can be understood as limiting information on the number of models corresponding to the analysis task identifier of the target analysis task.
  • the limiting information on the number of models may include a quantity threshold.
  • the model limiting information includes limiting information on the number of models, the number of target models that the first network element needs to feed back to the third network element should be is greater than or equal to the quantity threshold, or the number of feedback target models should be less than or equal to the quantity threshold.
  • the limitation information for the model identification may include one or more preset model identifications.
  • the model limitation information includes the limitation information of the model identification
  • the first network element needs to feed back the target model from the model corresponding to the preset model identification to The third network element, or the third network element needs to feed back the target model to the third network element from a model other than the model corresponding to the preset model identifier.
  • the limited information for the model performance information may include performance indicators.
  • the model limited information includes limited information for the performance information
  • the first network element needs to feed back the target model that satisfies the performance indicators to the third network element.
  • the definition information of the type performance information may include at least one of the definition information of the first performance information and the definition information of the second performance information.
  • the limiting information for the model size may include a model size threshold.
  • the model limiting information includes limiting information for the model size
  • the first network element needs to feed back a target model that is greater than or equal to the model size threshold to the third network element, or feedback Target models less than or equal to this model size threshold.
  • the limiting information for the inference time length may include an inference time threshold.
  • the model limiting information includes the limiting information for the inference time length
  • the first network element needs to feed back to the third network element a target model that is greater than or equal to the inference time threshold, or feedback Target models less than or equal to this inference time threshold.
  • the limited information for the training data source may include training data source information.
  • the model limited information includes limited information for the training data source
  • the first network element needs to feed back the target model from the model trained by the training data source information to the third network element.
  • the third network element, or the third network element needs to feed back the target model to the third network element from a model other than the model trained by the training data source information.
  • the limiting information for the training data time information may include a training data time threshold.
  • the time range may be determined based on the time threshold.
  • the first network element needs to report to the third network element.
  • the generation time of the training data used by the feedback target model in the training phase is within this time range.
  • the first performance information includes at least one of the following information of the model during training: the first performance index, the calculation method of the first performance index, the first time information, the first numerical value and the first result; wherein, the first performance index includes at least one of accuracy during training and error value during training, the first time information includes time information corresponding to the calculation of the first performance index, and the first numerical value is used to represent calculation The number of data used by the first performance indicator, and the first result is a result value calculated based on multiple first performance indicators.
  • the above-mentioned first performance indicator can also be understood as the model's performance in training (performance in training). That is, a value calculated based on some statistics, for example, it can be at least one of the above-mentioned accuracy in training (AiT) and training error (Mean Absolute Error in Training, MAEiT).
  • the accuracy during training can be called the accuracy of the model during training. This accuracy can be obtained by dividing the number of correct decision-making results of the model into the total number of decisions.
  • the first network element can set up a verification data set for evaluating model accuracy.
  • the verification set includes data used for model input and real label data.
  • the first network element inputs the verification input data into the trained model to obtain the output. data, the first network element then compares whether the output data is consistent with the real label data, and then uses the above calculation method to obtain the value of model accuracy.
  • the calculation method of the above-mentioned first performance indicator may include at least one of the following: the ratio of the accurate number of model predictions to the total number of model predictions, MAE, root mean square error (Root Mean Square Error), recall (Recall), F1 score (F1 score), etc. .
  • the above-mentioned first time information is represented as a time node or a period of time (for example, including the start time of calculating the first performance index and the end time of calculating the first performance index).
  • the above-mentioned first result may represent the distribution of multiple first performance indicators, and may be calculated using a preset calculation method.
  • the first result may be an average, a median, a variance, etc.
  • the second performance information includes at least one of the following information when the model is actually used: second performance index, calculation method of the second performance index, second time information, second numerical value and a second result; wherein, the second performance index includes at least one of accuracy during actual use and error value during actual use, and the second time information includes time information corresponding to calculating the second performance index,
  • actual use can be understood as using the model for model inference.
  • the above-mentioned second performance information corresponds to the above-mentioned first performance information.
  • the second performance indicator can also be understood as the performance of the model in actual use (performance in Use).
  • the calculation method of the above-mentioned second performance indicator may include at least one of the following: the ratio of the accurate number of model predictions to the total number of model predictions, MAE, root mean square error (Root Mean Square Error), recall (Recall) and F1 score (F1 score) )wait.
  • the above-mentioned second time information is represented as a time node or a period of time (for example, including the start time of calculating the second performance index and the end time of calculating the second performance index).
  • the above-mentioned second result may represent the distribution of multiple second performance indicators, and may be calculated using a preset calculation method.
  • the second result may be an average, a median, a variance, etc.
  • one network element uses the model in another network element to perform model inference on the target analysis task, including the following process:
  • Step 1401 MTLF sends a registration request message to NRF.
  • the registration request message may be called a capability registration message for capability registration.
  • the registration request message may include, in addition to MTLF's own identification information and supported analytic ID information, the above-mentioned capability information, that is, at least one of model quantity information and model information.
  • the accuracy and error values in the model information can determine which of the two models is more suitable for the target analysis task; the distribution of accuracy can determine whether the performance of the model is stable; by judging the source information and time information of the model training data, It can be judged whether the model matches the target analysis task (for example, a model trained with old data has a greater probability of being affected by changes in network data).
  • Step 1402 NRF stores the information carried in the registration request message
  • Step 1403 NRF sends a registration request response message.
  • Step 1404 The task consumer sends a task request message to AnLF.
  • the task request message includes the analysis task identifier of the target analysis task, the definition information of the data analysis task model, and the analysis target.
  • step 1403 is located before step 1404.
  • Step 1405 AnLF sends a second request message to NRF.
  • This second request message is used to find a suitable MTLF.
  • it may further include the target. Request information.
  • the target requirement information includes at least one of the following:
  • Requirement information on model size which is used to indicate the storage space required to store or run the model
  • Requirement information for inference duration which is used to indicate the duration required for model inference operations based on the model
  • the training data source information is used to indicate at least one of the location information and network element information of the training data source used by the model in the training phase;
  • the training data time requirement information is used to indicate the generation time of the training data used by the model in the training phase.
  • the target requirement information can be understood as a requirement for the model or a limited requirement for the model.
  • the limiting information on the number of models mentioned above may be a quantity threshold.
  • the number of models that a suitable MTLF can provide that can be used to perform the target analysis task needs to be greater than or equal to the quantity threshold, or less than or equal to the quantity threshold.
  • the above-mentioned model identification requirement information may include at least one model identification.
  • appropriate MTLF can provide a model that can be used to perform the target analysis task, including or not including a model corresponding to the at least one model identification.
  • the requirement information of the model performance information may include performance indicators (for example, it may include at least one of the first performance indicator and the second performance indicator).
  • the appropriate MTLF can provide a model that can be used to perform the target analysis task that needs to meet the requirements. Performance.
  • the above-mentioned model size limitation information may include a model size threshold.
  • a suitable MTLF can provide a model that can be used to perform the target analysis task. It needs to include a model that is greater than or equal to the model size threshold or it needs to include a model that is less than or equal to the model size threshold. model.
  • the above-mentioned limited information about the training data source may include the training data source information.
  • the appropriate MTLF can provide a model that can be used to perform the target analysis task. It needs to include or exclude the model trained by the training data source information.
  • the above training data time information may include a training data time threshold, and the time range may be determined based on the time threshold. In this case, the appropriate MTLF needs to include models whose training data used in the training phase are generated within this time range.
  • Step 1406 NRF feeds back a second request response message to AnLF, where the second request response message needs to include the determined N MTLFs. Furthermore, at least one of the task analysis identification, model quantity information and model information supported by each MTLF may also be included.
  • each MTLF includes a model that supports the analysis task identification of the above target analysis task, and the model (or the MTLF) needs to meet the above target requirement information and the definition information of the data analysis task model.
  • the second request response message may also include the validity time corresponding to each of the N first network elements.
  • This valid time can be understood as the valid time of the capability information registered by MTLF. Beyond this time, the capability information of MTLF may change. Outside the valid time, requesting the model in MTLF may cause the final model inference to be unreliable. Be protected. Therefore, AnLF can first request models in MTLF that are within the valid time.
  • Step 1407 AnLF determines the target MTLF among N MTLFs.
  • AnLF may also select at least two MTLFs as target MTLFs.
  • Step 1408 AnLF sends a first request message (ie, model acquisition request) to the target MTLF, which may be, for example, Nnwdaf_MLModelInfo_Request or Nnwdaf_MLModelProvision_Request.
  • a first request message ie, model acquisition request
  • the first request message may include the analysis task identifier and model definition information of the target analysis task.
  • AnLF can match the appropriate target model according to the analysis task identifier and model definition information of the target analysis task.
  • Step 1409 The target MTLF sends a first request response message to the AnLF.
  • the first request response message includes the target model and address information used to obtain the target model.
  • the first request information includes the target model, which can be understood as including the configuration file of the model and the description information of the model.
  • the address information of the target model may include the Uniform Resource Locator (URL), FQDN information, IP address, etc. After AnLF obtains the address information, it can directly download the target model.
  • Step 1410 AnLF performs model inference for the target analysis task based on the target model, and obtains the target task report.
  • AnLF can use one target model to obtain inference results and use the inference results as a target task report;
  • AnLF can also use multiple target models to obtain different inference results, and finally generate a target task report based on multiple inference results, for example , multiple inference results can be aggregated or voted to obtain the final inference result as a target task report.
  • the target task report can be understood as data analysis result information.
  • Step 1411 AnLF feeds back a task request response message to the task consumer, where the task request response message includes the target analysis report.
  • this application protects the following solutions:
  • a model processing method including:
  • the first network element sends a registration request message to the second network element, the registration request message includes capability information of the first network element, and the capability information includes at least one of model quantity information and model information;
  • the first network element receives a registration request response message from the second network element
  • the model quantity information is used to indicate the number of models corresponding to the analysis task identifier supported by the first network element; the model information includes at least one of the following models supported by the first network element corresponding to the analysis task identifier. kind of information:
  • Model size which is used to indicate the storage space required to store or run the model
  • Inference duration which is used to indicate the time required to perform model inference operations based on the model
  • Training data source information which is used to indicate at least one of the location information and network element information of the training data source used by the model in the training phase;
  • Training data time information which is used to indicate the generation time of the training data used by the model in the training phase.
  • the method further includes:
  • the first network element receives a first request message from the third network element, where the first request message is used to obtain a target model that can be used to perform a target analysis task;
  • the first network element sends a first request response message to the third network element, where the first request response message includes at least one of the target model and address information used to obtain the target model.
  • the first request message includes the analysis task identification and model definition information of the target analysis task, and the model definition information includes at least one of the following:
  • the model performance information includes at least one of the following:
  • the first performance information is used to indicate the performance that the model can achieve during the training phase
  • Second performance information the second performance information is used to indicate the performance that the model can achieve in the inference phase.
  • the first performance information includes at least one of the following information of the model during training: the first performance index, the calculation method of the first performance index, the first time information, the first numerical value and The first result; wherein, the first performance index includes at least one of accuracy during training and error value during training, the first time information includes time information corresponding to calculating the first performance index, and the third A numerical value is used to represent the amount of data used to calculate the first performance indicator, and the first result is a result value calculated based on multiple first performance indicators.
  • the second performance information includes at least one of the following information when the model is actually used: the second performance index, the calculation method of the second performance index, the second time information, and the second numerical value. and a second result; wherein, the second performance index includes at least one of accuracy during actual use and error value during actual use, and the second time information includes time information corresponding to calculating the second performance index,
  • the second numerical value is used to represent the amount of data used to calculate the second performance index, and the second result is a result value calculated based on multiple second performance indexes.
  • a model processing method including:
  • the second network element receives a registration request message from the first network element, where the registration request message includes capability information of the first network element;
  • the second network element stores the capability information and sends a registration request response message to the first network element
  • the capability information includes at least one of model quantity information and model information;
  • the model quantity information is used to indicate the number of models corresponding to the analysis task identifier supported by the first network element;
  • the model information includes At least one of the following information of the model corresponding to the analysis task identifier supported by the first network element:
  • Model size which is used to indicate the storage space required to store or run the model
  • Inference duration which is used to indicate the time required to perform model inference operations based on the model
  • Training data source information which is used to indicate at least one of the location information and network element information of the training data source used by the model in the training phase;
  • Training data time information which is used to indicate the generation time of the training data used by the model in the training phase.
  • the method further includes:
  • the second network element receives a second request message from the third network element, where the second request message includes an analysis task identifier of the target analysis task;
  • the second network element determines N first network elements based on the second request message.
  • the N first network elements are first network elements that can provide a model that can be used to perform the target analysis task, and N is positive integer;
  • the second network element sends a second request response message to the third network element, where the second request response message is used to indicate the N first network elements.
  • the second request message also includes target requirement information, and any first network element among the N first network elements satisfies the target requirement information, wherein the target
  • the required information includes at least one of the following:
  • the second request response message includes at least one of the identification information of the N first network elements and the address information of the N first network elements.
  • the second request response message further includes the validity time corresponding to each of the N first network elements.
  • the model performance information includes at least one of the following:
  • the first performance information is used to indicate the performance that the model can achieve during the training phase
  • Second performance information the second performance information is used to indicate the performance that the model can achieve in the inference stage.
  • the first performance information includes at least one of the following information of the model during training: the first performance index, the calculation method of the first performance index, the first time information, the first numerical value and The first result; wherein, the first performance index includes at least one of accuracy during training and error value during training, the first time information includes time information corresponding to calculating the first performance index, and the third A numerical value is used to represent the amount of data used to calculate the first performance indicator, and the first result is a result value calculated based on multiple first performance indicators.
  • the second performance information includes at least one of the following information when the model is actually used: the second performance index, the calculation method of the second performance index, the second time information, and the second numerical value. and a second result; wherein, the second performance index includes at least one of accuracy during actual use and error value during actual use, and the second time information includes time information corresponding to calculating the second performance index,
  • the second numerical value is used to represent the amount of data used to calculate the second performance index, and the second result is a result value calculated based on multiple second performance indexes.
  • a model processing method including:
  • the third network element sends a second request message to the second network element, where the second request message includes the analysis task identifier of the target analysis task;
  • the third network element receives a second request response message from the second network element.
  • the second request response message is used to indicate N first network elements.
  • the N first network elements are capable of providing The first network element of the model that performs the target analysis task, N is a positive integer;
  • the third network element sends a first request message to the target network element among the N first network elements
  • the third network element receives a first request response message from the target network element.
  • the first request response message includes at least one of a target model and address information used to obtain the target model.
  • the target model Can be used to perform the target analysis tasks described.
  • the first request message includes the analysis task identification and model definition information of the target analysis task, and the model definition information includes at least one of the following:
  • Qualification information of the training data source which is used to indicate at least one of the location information and network element information of the training data source used by the model in the training phase;
  • the training data time information is used to indicate the generation time of the training data used by the model in the training phase.
  • the model performance information includes at least one of the following:
  • the first performance information is used to indicate the performance that the model can achieve during the training phase
  • Second performance information the second performance information is used to indicate the performance that the model can achieve in the inference stage.
  • the first performance information includes at least one of the following information of the model during training: the first performance index, the calculation method of the first performance index, the first time information, the first numerical value and The first result; wherein, the first performance index includes at least one of accuracy during training and error value during training, the first time information includes time information corresponding to calculating the first performance index, and the third A numerical value is used to represent the amount of data used to calculate the first performance indicator, and the first result is a result value calculated based on multiple first performance indicators.
  • the second performance information includes at least one of the following information when the model is actually used: the second performance index, the calculation method of the second performance index, the second time information, and the second numerical value. and a second result; wherein, the second performance index includes at least one of accuracy during actual use and error value during actual use, and the second time information includes time information corresponding to calculating the second performance index,
  • the second numerical value is used to represent the amount of data used to calculate the second performance index, and the second result is a result value calculated based on multiple second performance indexes.
  • the second request message further includes target requirement information, and any first network element among the N first network elements satisfies the target requirement information, and the target requirement information Include at least one of the following:
  • Requirement information on model size which is used to indicate the storage space required to store or run the model
  • Requirement information for inference duration which is used to indicate the duration required for model inference operations based on the model
  • the training data source information is used to indicate at least one of the location information and network element information of the training data source used by the model in the training phase;
  • the training data time requirement information is used to indicate the generation time of the training data used by the model in the training phase.
  • the method before the third network element sends the second request message to the second network element, the method further includes:
  • the third network element receives a task request message from the fourth network element, where the task request message includes an analysis task identifier of the target analysis task.
  • the method further includes:
  • the third network element uses the target model to perform the target analysis task and obtain a target analysis report
  • the third network element sends a task request response message to the fourth network element, where the task request response message includes the target analysis report.
  • the number of the target models is M1, and M1 is a positive integer.
  • the third network element uses the target model to perform the target analysis task, and obtaining the target analysis report includes:
  • the third network element uses M2 target models to perform model inference for the target analysis task, and obtains M2 inference results, where M2 is a positive integer less than or equal to M1;
  • the third network element generates the target analysis report based on the M2 inference results.
  • the target network element is one of the first network elements.
  • the second request response message includes at least one of the identification information of the N first network elements and the address information of the N first network elements.
  • the second request response message further includes the validity time corresponding to each of the N first network elements.
  • a model processing device comprising:
  • a first sending module configured to send a registration request message to the second network element, where the registration request message includes capability information of the first network element, where the capability information includes at least one of model quantity information and model information;
  • a first receiving module configured to receive a registration request response message from the second network element
  • the model quantity information is used to indicate the number of models corresponding to the analysis task identifier supported by the first network element; the model information includes at least one of the following information about the models supported by the first network element and corresponding to the analysis task identifier. :
  • Model size which is used to indicate the storage space required to store or run the model
  • Inference duration which is used to indicate the time required to perform model inference operations based on the model
  • Training data source information which is used to indicate at least one of the location information and network element information of the training data source used by the model in the training phase;
  • Training data time information which is used to indicate the generation time of the training data used by the model in the training phase.
  • the first receiving module is further configured to receive a first request message from a third network element, where the first request message is used to obtain a target model that can be used to perform a target analysis task;
  • the first sending module is further configured to send a first request response message to the third network element, where the first request response message includes at least one of the target model and address information used to obtain the target model. item.
  • the first request message includes the analysis task identifier of the target analysis task and model definition information
  • the model definition information includes at least one of the following:
  • model performance information includes at least one of the following:
  • the first performance information is used to indicate the performance that the model can achieve during the training phase
  • Second performance information the second performance information is used to indicate the performance that the model can achieve in the inference phase.
  • the first performance information includes at least one of the following information of the model during training: first performance index, calculation method of the first performance index, first time information, first numerical value and The first result; wherein, the first performance index includes at least one of training accuracy and training error value, so
  • the first time information includes time information corresponding to the calculation of the first performance indicator, the first value is used to represent the amount of data used to calculate the first performance indicator, and the first result is based on a plurality of the first performance indicators.
  • a result value obtained by calculating a performance indicator is obtained by calculating a performance indicator.
  • the second performance information includes at least one of the following information when the model is actually used: the second performance index, the calculation method of the second performance index, the second time information, and the second numerical value. and a second result; wherein, the second performance index includes at least one of accuracy during actual use and error value during actual use, and the second time information includes time information corresponding to calculating the second performance index,
  • the second numerical value is used to represent the amount of data used to calculate the second performance index, and the second result is a result value calculated based on multiple second performance indexes.
  • a model processing device characterized in that it includes:
  • a second receiving module configured to receive a registration request message from the first network element, where the registration request message includes capability information of the first network element;
  • a storage module used to store the capability information
  • a second sending module configured to send a registration request response message to the first network element
  • the capability information includes at least one of model quantity information and model information;
  • the model quantity information is used to indicate the number of models corresponding to the analysis task identifier supported by the first network element;
  • the model information includes At least one of the following information of the model corresponding to the analysis task identifier supported by the first network element:
  • Model size which is used to indicate the storage space required to store or run the model
  • Inference duration which is used to indicate the time required to perform model inference operations based on the model
  • Training data source information which is used to indicate at least one of the location information and network element information of the training data source used by the model in the training phase;
  • Training data time information which is used to indicate the generation time of the training data used by the model in the training phase.
  • the second receiving module 701 is also configured to receive a second request message from the third network element, where the second request message includes the analysis task identifier of the target analysis task;
  • the first determination module is configured to determine N first network elements based on the second request message, where the N first network elements are first network elements that can provide a model that can be used to perform the target analysis task, N is a positive integer;
  • the second sending module 703 is also configured to send a second request response message to the third network element, where the second request response message is used to indicate the N first network elements.
  • the second request message further includes target requirement information, and any first network element among the N first network elements satisfies the target requirement information, wherein the The above-mentioned goal requires information to include at least one of the following:
  • the second request response message includes at least one of the identification information of the N first network elements and the address information of the N first network elements.
  • the second request response message further includes the valid time corresponding to each of the N first network elements.
  • the model performance information includes at least one of the following:
  • the first performance information is used to indicate the performance that the model can achieve during the training phase
  • Second performance information the second performance information is used to indicate the performance that the model can achieve in the inference stage.
  • the first performance information includes at least one of the following information of the model during training: the first performance index, the calculation method of the first performance index, the first time information, the first numerical value and The first result; wherein, the first performance index includes at least one of accuracy during training and error value during training, the first time information includes time information corresponding to calculating the first performance index, and the third A numerical value is used to represent the amount of data used to calculate the first performance indicator, and the first result is a result value calculated based on multiple first performance indicators.
  • the second performance information includes at least one of the following information when the model is actually used: the second performance index, the calculation method of the second performance index, the second time information, and the second numerical value. and a second result; wherein, the second performance index includes at least one of accuracy during actual use and error value during actual use, and the second time information includes time information corresponding to calculating the second performance index,
  • the second numerical value is used to represent the amount of data used to calculate the second performance index, and the second result is a result value calculated based on multiple second performance indexes.
  • a model processing device characterized in that it includes:
  • the third sending module is configured to send a second request message to the second network element, where the second request message includes the analysis task identifier of the target analysis task;
  • the third receiving module is further configured to receive a second request response message from the second network element, where the second request response message is used to indicate N first network elements, and the N first network elements are capable of providing The first network element of the model that can be used to perform the target analysis task, N is a positive integer;
  • the third sending module is also configured to send a first request message to the target network element among the N first network elements;
  • the third receiving module is further configured to receive a first request response message from the target network element, where the first request response message includes at least one of a target model and address information used to obtain the target model,
  • the target model can be used to perform the target analysis task.
  • the first request message includes the analysis task identification of the target analysis task and model definition information
  • the model definition information includes at least one of the following:
  • Qualification information of the training data source which is used to indicate at least one of the location information and network element information of the training data source used by the model in the training phase;
  • the training data time information is used to indicate the generation time of the training data used by the model in the training phase.
  • the model performance information includes at least one of the following:
  • the first performance information is used to indicate the performance that the model can achieve during the training phase
  • Second performance information the second performance information is used to indicate the performance that the model can achieve in the inference phase.
  • the first performance information includes at least one of the following information of the model during training: the first performance index, the calculation method of the first performance index, the first time information, the first numerical value and The first result; wherein, the first performance index includes at least one of accuracy during training and error value during training, the first time information includes time information corresponding to calculating the first performance index, and the third A numerical value is used to represent the amount of data used to calculate the first performance indicator, and the first result is a result value calculated based on multiple first performance indicators.
  • the second performance information includes at least one of the following information when the model is actually used: the second performance index, the calculation method of the second performance index, the second time information, and the second numerical value. and a second result; wherein, the second performance index includes at least one of accuracy during actual use and error value during actual use, and the second time information includes time information corresponding to calculating the second performance index,
  • the second The numerical value is used to represent the amount of data used to calculate the second performance index, and the second result is a result value calculated based on multiple second performance indexes.
  • the second request message further includes target requirement information, and any first network element among the N first network elements satisfies the target requirement information.
  • the target requirement information include at least one of the following:
  • Requirement information on model size which is used to indicate the storage space required to store or run the model
  • Requirement information for inference duration which is used to indicate the duration required for model inference operations based on the model
  • the training data source information is used to indicate at least one of the location information and network element information of the training data source used by the model in the training phase;
  • the training data time requirement information is used to indicate the generation time of the training data used by the model in the training phase.
  • the method before the third network element sends the second request message to the second network element, the method further includes:
  • the third network element receives a task request message from the fourth network element, where the task request message includes an analysis task identifier of the target analysis task.
  • the number of the target models is M1, and M1 is a positive integer.
  • the third network element uses the target model to perform the target analysis task, and obtaining the target analysis report includes:
  • the third network element uses M2 target models to perform model inference for the target analysis task, and obtains M2 inference results, where M2 is a positive integer less than or equal to M1;
  • the third network element generates the target analysis report based on the M2 inference results.
  • the target network element is one of the first network elements.
  • the second request response message includes at least one of the identification information of the N first network elements and the address information of the N first network elements.
  • the second request response message further includes the validity time corresponding to each of the N first network elements.
  • a network-side device including a processor and a memory.
  • the memory stores programs or instructions that can be run on the processor. When the program or instructions are executed by the processor, any of the above 1 to 26 are implemented. The steps of the model processing method described in one item.
  • a readable storage medium on which a program or instructions are stored. When the program or instructions are executed by a processor, the steps of the model processing method described in any one of 1 to 26 above are implemented.
  • the methods of the above embodiments can be implemented by means of software plus the necessary general hardware platform. Of course, it can also be implemented by hardware, but in many cases the former is better. implementation.
  • the technical solution of the present application can be embodied in the form of a computer software product that is essentially or contributes to the existing technology.
  • the computer software product is stored in a storage medium (such as ROM/RAM, disk , CD), including several instructions to cause a terminal (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in various embodiments of this application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

本申请公开了一种模型获取方法及通信设备，属于通信技术领域，本申请实施例的模型获取方法包括：第一通信设备向第二通信设备发送第一请求，所述第一请求用于请求获取至少一个第三通信设备的信息，其中，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息；所述第一通信设备接收所述第二通信设备发送的至少一个第三通信设备的信息；所述第一通信设备基于所述至少一个第三通信设备的信息，获取多个模型的模型信息；所述多个模型用于生成数据分析结果信息。

Description

模型获取方法及通信设备
相关申请的交叉引用
本申请要求享有于2022年05月05日提交的名称为“模型获取方法及通信设备”的中国专利申请202210483893.4的优先权和2022年5月5日提交的名称为“模型处理方法、装置、网络侧设备及可读存储介质”的中国专利申请202210482186.3的优先权，该申请的全部内容通过引用并入本文中。
技术领域
本申请属于通信技术领域，具体涉及一种模型获取方法及通信设备。
背景技术
网络数据分析功能(Network Data Analytics Function，NWDAF)以网络数据为基础对网络进行自动感知和分析，并参与到网络规划、建设、运维、网优、运营全生命周期中，使得网络易于维护和控制，提高网络资源使用效率，提升用户业务体验。NWDAF可包括两部分，分析逻辑功能(Analytics Logical Function，AnLF)和模型训练逻辑功能(Model Training Logical Function，MTLF)，前者AnLF负责推理功能，后者MTLF负责训练功能。
AnLF在收到消费者(consumer)发起的任务时，需要获取模型，以使用模型进行推理并生成分析结果。因此对于本领域技术人员来说，亟需实现一种模型获取方法。
发明内容
本申请实施例提供一种模型获取方法及通信设备，能够解决如何实现模型获取方法的问题。
第一方面，提供了一种模型获取方法，包括：
第一通信设备向第二通信设备发送第一请求，所述第一请求用于请求获取至少一个第三通信设备的信息，其中，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息；
所述第一通信设备接收所述第二通信设备发送的至少一个第三通信设备的信息；
所述第一通信设备基于所述至少一个第三通信设备的信息，获取多个模型的模型信息；所述多个模型用于生成数据分析结果信息。
第二方面，提供了一种模型获取方法，包括：
第二通信设备接收第一通信设备发送的第一请求，所述第一请求用于请求获取至少一个第三通信设备的信息，其中，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息；
所述第二通信设备基于所述第一请求，确定至少一个第三通信设备；
所述第二通信设备向所述第一通信设备发送所述至少一个第三通信设备的信息；所述至少一个第三通信设备的信息用于所述第一通信设备获取多个模型的模型信息；所述多个模型用于生成数据分析结果信息。
第三方面，提供了一种模型获取方法，包括：
第三通信设备接收第一通信设备发送的第二请求，所述第二请求用于请求获取至少一个模型；
所述第三通信设备基于所述第二请求向所述第一通信设备发送至少一个模型的模型信息；所述第二请求包括以下至少一项：
分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；
所需模型的标识；
所需模型数量；
模型需满足的模型属性信息。
第四方面，提供了一种模型获取方法，包括：
第四通信设备向第一通信设备发送任务请求；
所述第四通信设备接收所述第一通信设备发送的数据分析结果信息，所述数据分析结果信息为所述第一通信设备基于多个模型进行分析处理得到的，所述多个模型为所述第一通信设备基于至少一个第三通信设备的信息获取到的，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息。
第五方面，提供了一种模型获取装置，包括：
发送模块，用于向第二通信设备发送第一请求，所述第一请求用于请求获取至少一个第三通信设备的信息，其中，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息；
接收模块，用于接收所述第二通信设备发送的至少一个第三通信设备的信息；
获取模块，用于基于所述至少一个第三通信设备的信息，获取多个模型的模型信息；所述多个模型用于生成数据分析结果信息。
第六方面，提供了一种模型获取装置，包括：
接收模块，用于接收第一通信设备发送的第一请求，所述第一请求用于请求获取至少一个第三通信设备的信息，其中，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息；
处理模块，用于基于所述第一请求，确定至少一个第三通信设备；
发送模块，用于向所述第一通信设备发送所述至少一个第三通信设备的信息；所述至少一个第三通信设备的信息用于所述第一通信设备获取多个模型的模型信息；所述多个模型用于生成数据分析结果信息。
第七方面，提供了一种模型获取装置，包括：
接收模块，用于接收第一通信设备发送的第二请求，所述第二请求用于请求获取至少一个模型；
发送模块，用于基于所述第二请求向所述第一通信设备发送至少一个模型的模型信息；所述第二请求包括以下至少一项：
分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；
所需模型的标识；
所需模型数量；
模型需满足的模型属性信息。
第八方面，提供了一种模型获取装置，包括：
发送模块，用于向第一通信设备发送任务请求；
接收模块，用于接收所述第一通信设备发送的数据分析结果信息，所述数据分析结果信息为所述第一通信设备基于多个模型进行分析处理得到的，所述多个模型为所述第一通信设备基于至少一个第三通信设备的信息获取到的，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息。
第九方面，提供了一种第一通信设备，该第一通信设备包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如第一方面所述的方法的步骤。
第十方面，提供了一种第一通信设备，包括处理器及通信接口，其中，所述通信接口用于第一通信设备向第二通信设备发送第一请求，所述第一请求用于请求获取至少一个第三通信设备的信息，其中，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息；接收所述第二通信设备发送的至少一个第三通信设备的信息；基于所述至少一个第三通信设备的信息，获取多个模型的模型信息。
第十一方面，提供了一种第二通信设备，该第二通信设备包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如第二方面所述的方法的步骤。
第十二方面，提供了一种第二通信设备，包括处理器及通信接口，其中，所述通信接口用于接收第一通信设备发送的第一请求，所述第一请求用于请求获取至少一个第三通信设备的信息，其中，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息；向所述第一通信设备发送所述至少一个第三通信设备的信息；所述至少一个第三通信设备的信息用于所述第一通信设备获取多个模型的模型信息；处理器，用于基于所述第一请求，确定至少一个第三通信设备。
第十三方面，提供了一种第三通信设备，该第三通信设备包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如第一方面所述的方法的步骤。
第十四方面，提供了一种第三通信设备，包括处理器及通信接口，其中，所述通信接口用于接收第一通信设备发送的第二请求，所述第二请求用于请求获取至少一个模型；基于所述第二请求向所述第一通信设备发送至少一个模型的模型信息；所述第二请求包括以下至少一项：分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；所需模型的标识；所需模型数量；模型需满足的模型属性信息。
第十五方面，提供了一种第四通信设备，该第四通信设备包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如第一方面所述的方法的步骤。
第十六方面，提供了一种第四通信设备，包括处理器及通信接口，其中，所述通信接口用于向第一通信设备发送任务请求；接收所述第一通信设备发送的数据分析结果信息，所述数据分析结果信息为所述第一通信设备基于多个模型进行分析处理得到的，所述多个模型为所述第一通信设备基于至少一个第三通信设备的信息获取到的，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息。
第十七方面，提供了一种通信系统，包括：第一通信设备、第二通信设备、第三通信设备及第四通信设备，所述第一通信设备可用于执行如第一方面所述的模型获取方法的步骤，所述第二通信设备可用于执行如第二方面所述的模型获取方法的步骤，所述第三通信设备可用于执行如第三方面所述的模型获取方法的步骤，所述第四通信设备可用于执行如第四方面所述的模型获取方法的步骤。
第十八方面，提供了一种可读存储介质，所述可读存储介质上存储程序或指令，所述程序或指令被处理器执行时实现如第一方面所述的方法的步骤，或者实现如第二方面所述的方法的步骤，或者实现如第三方面所述的方法的步骤，或者实现如第四方面所述的方法的步骤。
第十九方面，提供了一种芯片，所述芯片包括处理器和通信接口，所述通信接口和所述处理器耦合，所述处理器用于运行程序或指令，实现如第一方面所述的方法， 或实现如第二方面所述的方法，或实现如第三方面所述的方法，或实现如第四方面所述的方法。
第二十方面，提供了一种计算机程序/程序产品，所述计算机程序/程序产品被存储在存储介质中，所述计算机程序/程序产品被至少一个处理器执行以实现如第一方面至第四方面任一项所述的模型获取方法的步骤。
在本申请实施例中，第一通信设备向第二通信设备发送第一请求，第一请求用于请求获取至少一个第三通信设备的信息，其中，每个第三通信设备能够提供第一通信设备所需的至少一个模型的模型信息，进一步地，第一通信设备基于第二通信设备发送的至少一个第三通信设备的信息，可以获取到多个模型的模型信息，从而实现了一种模型获取方法，实现复杂度较低，效率较高，而且多个模型用于生成数据分析结果信息，使得数据分析结果更加准确。
附图说明
图1是本申请实施例可应用的无线通信系统的结构图；
图2是本申请实施例提供的模型获取方法的流程示意图之一；
图3是本申请实施例提供的模型获取方法的交互流程示意图之一；
图4是本申请实施例提供的模型获取方法的流程示意图之二；
图5是本申请实施例提供的模型获取方法的流程示意图之三；
图6是本申请实施例提供的模型获取方法的流程示意图之四；
图7是本申请实施例提供的模型获取装置的结构示意图之一；
图8是本申请实施例提供的模型获取装置的结构示意图之二；
图9是本申请实施例提供的模型获取装置的结构示意图之三；
图10是本申请实施例提供的模型获取装置的结构示意图之四；
图11是本申请实施例提供的通信设备的结构示意图；
图12是本申请实施例的网络侧设备的结构示意图；
图13是本申请实施例提供的模型处理处理方法的流程示意图之一；
图14是本申请实施例提供的模型处理方法的流程示意图之二。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚描述，显然，所描述的实施例是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员所获得的所有其他实施例，都属于本申请保护的范围。
本申请的说明书和权利要求书中的术语“第一”、“第二”等是用于区别类似的对象，而不用于描述特定的顺序或先后次序。应该理解这样使用的术语在适当情况下可以互换，以便本申请的实施例能够以除了在这里图示或描述的那些以外的顺序实施，且“第一”、“第二”所区别的对象通常为一类，并不限定对象的个数，例如第一对象可以是一个，也可以是多个。此外，说明书以及权利要求中“和/或”表示所连接对象的至少其中之一，字符“/”一般表示前后关联对象是一种“或”的关系。
值得指出的是，本申请实施例所描述的技术不限于长期演进型(Long Term Evolution，LTE)/LTE的演进(LTE-Advanced，LTE-A)系统，还可用于其他无线通信系统，诸如码分多址(Code Division Multiple Access，CDMA)、时分多址(Time Division Multiple Access，TDMA)、频分多址(Frequency Division Multiple Access，FDMA)、正交频分多址(Orthogonal Frequency Division Multiple Access，OFDMA)、单载波频分多址(Single-carrier Frequency Division Multiple Access，SC-FDMA)和其他系统。本申请实施例中的术语“系统”和“网络”常被可互换地使用，所描述的技术既可用于以上提及的系统和无线电技术，也可用于其他系统和无线电技术。以下描述出于示例目的描述了新空口(New Radio，NR)系统，并且在以下大部分描述中使用NR术语，但是这些技术也可应用于NR系统应用以外的应用，如第6代(6th Generation，6G)通信系统。
图1示出本申请实施例可应用的一种无线通信系统的框图。无线通信系统包括终端11和网络侧设备12。其中，终端11可以是手机、平板电脑(Tablet Personal Computer)、膝上型电脑(Laptop Computer)或称为笔记本电脑、个人数字助理(Personal Digital Assistant，PDA)、掌上电脑、上网本、超级移动个人计算机(ultra-mobile personal computer，UMPC)、移动上网装置(Mobile Internet Device，MID)、增强现实(augmented reality，AR)/虚拟现实(virtual reality，VR)设备、机器人、可穿戴式设备(Wearable Device)、车载设备(VUE)、行人终端(PUE)、智能家居(具有无线通信功能的家居设备，如冰箱、电视、洗衣机或者家具等)、游戏机、个人计算机(personal computer，PC)、柜员机或者自助机等终端侧设备，可穿戴式设备包括：智能手表、智能手环、智能耳机、智能眼镜、智能首饰(智能手镯、智能手链、智能戒指、智能项链、智能脚镯、智能脚链等)、智能腕带、智能服装等。除了上述终端设备，也可以是终端内的芯片，例如调制解调器(Modem)芯片，系统级芯片(System on Chip,SoC)。需要说明的是，在本申请实施例并不限定终端11的具体类型。网络侧设备12可以包括接入网设备或核心网设备，其中，接入网设备12也可以称为无线接入网设备、无线接入网(Radio Access Network,RAN)、无线接入网功能或无线接入网单元。接入网设备12可以包括基站、WLAN接入点或WiFi节点等，基站可被称为节点B、演进节点B(eNB)、接入点、基收发机站(Base Transceiver Station，BTS)、无 线电基站、无线电收发机、基本服务集(Basic Service Set，BSS)、扩展服务集(Extended Service Set，ESS)、家用B节点、家用演进型B节点、发送接收点(Transmitting Receiving Point，TRP)或所述领域中其他某个合适的术语，只要达到相同的技术效果，所述基站不限于特定技术词汇，需要说明的是，在本申请实施例中仅以NR系统中的基站为例进行介绍，并不限定基站的具体类型。核心网设备可以包含但不限于如下至少一项：核心网节点、核心网功能、移动管理实体(Mobility Management Entity，MME)、接入移动管理功能(Access and Mobility Management Function，AMF)、会话管理功能(Session Management Function，SMF)、用户平面功能(User Plane Function，UPF)、策略控制功能(Policy Control Function，PCF)、策略与计费规则功能单元(Policy and Charging Rules Function，PCRF)、边缘应用服务发现功能(Edge Application Server Discovery Function，EASDF)、统一数据管理(Unified Data Management，UDM)，统一数据仓储(Unified Data Repository，UDR)、归属用户服务器(Home Subscriber Server，HSS)、集中式网络配置(Centralized network configuration，CNC)、网络存储功能(Network Repository Function，NRF)，网络开放功能(Network Exposure Function，NEF)、本地NEF(Local NEF，或L-NEF)、绑定支持功能(Binding Support Function，BSF)、应用功能(Application Function，AF)等。需要说明的是，在本申请实施例中仅以NR系统中的核心网设备为例进行介绍，并不限定核心网设备的具体类型。
目前，在第三代合作伙伴计划(3rd Generation Partnership Project，3GPP)中，引入网络数据分析功能(Network Data Analytics Function，NWDAF)来进行一些智能化的分析。NWDAF具有一定的人工智能分析的功能，NWDAF通过采集一些数据，并使用内置的算法和分析能力，分析出结果，提供给核心网设备，做一些操作优化或者统计分析。网络数据分析功能以网络数据为基础对网络进行自动感知和分析，并参与到网络规划、建设、运维、网优、运营全生命周期中，使得网络易于维护和控制，提高网络资源使用效率，提升用户业务体验。
举例来说，NWDAF能够提供一种观测服务体验相关的网络数据分析分析功能(Observed Service Experience related network data analytics)，这个功能网元可以通过向其他网元发起请求，请求采集终端访问某个服务器IP地址的服务质量(Quality of Service，QoS)信息，比如上下行速率，丢包率等等，输出一个统计信息，该信息包括了该终端访问该服务器的用户体验情况。NWDAF也可以根据该终端访问该服务器的历史体验情况进行预测，比如，在未来某个时间段，某个区域内，这个终端如果访问该服务器，那么它可能的用户体验情况是怎样的。
通常，对于NWDAF所能提供的分析或者预测内容，以分析标识(Analytic ID)来区分，通过指示给NWDAF不同的Analytic ID以及一些参数，NWDAF即可提供该 分析标识所对应的分析和预测结果。比如，上述NWDAF所提供的观测服务体验分析，则使用Analytic ID＝Service Experience；因此，只要获取NWDAF服务的网元，在请求分析的时候输入Analytic ID，那么NWDAF就可以提供对应的分析了。
NWDAF可包括两部分，分析逻辑功能AnLF和模型训练逻辑功能MTLF，前者AnLF负责推理功能，后者MTLF负责训练功能。通常，AnLF通过模型进行推理并生成分析结果。
下面结合附图，通过一些实施例及其应用场景对本申请实施例提供的模型获取方法进行详细地说明。
图2是本申请实施例提供的模型获取方法的流程示意图之一。如图2所示，本实施例提供的方法，包括：
步骤101、第一通信设备向第二通信设备发送第一请求，第一请求用于请求获取至少一个第三通信设备的信息，其中，每个第三通信设备能够提供第一通信设备所需的至少一个模型的模型信息；
具体地，第一通信设备和第二通信设备可以是网络侧设备(如核心网设备)，第一通信设备例如可以是NWDAF的AnLF，第二通信设备可以是网络存储功能(NF Repository Function，NRF)，第三通信设备可以是NWDAF的MTLF，或分析数据存储功能(Analytics Data Repository Function，ADRF)。
第一通信设备例如在接收到其他通信设备发送的任务请求，或在满足其他触发条件时向第二通信设备发送第一请求，第一请求用于请求获取至少一个第三通信设备的信息，每个第三通信设备能够提供第一通信设备所需的至少一个模型的模型信息，模型信息可以是模型本身(例如包括模型的文件)或模型的下载地址等。
步骤102、第一通信设备接收第二通信设备发送的至少一个第三通信设备的信息；
具体地，第二通信设备基于第一请求，确定至少一个第三通信设备，并向第一通信设备发送至少一个第三通信设备的信息，第一通信设备接收第二通信设备发送的至少一个第三通信设备的信息。其中，第三通信设备的信息例如包括该第三通信设备支持的分析任务标识、第三通信设备的标识、地址、支持的分析任务标识对应的分析任务能够使用的模型数量、支持的模型的模型属性信息等。
步骤103、第一通信设备基于至少一个第三通信设备的信息，获取多个模型的模型信息；多个模型用于生成数据分析结果信息。
具体地，每个第三通信设备可以提供至少一个模型的模型信息，可以从至少一个第三通信设备中获取多个模型的模型信息，进一步第一通信设备还可以获取到多个模型，基于多个模型进行分析处理等操作，可以使得成数据分析结果更加准确。
本实施例的方法，第一通信设备向第二通信设备发送第一请求，第一请求用于请求获取至少一个第三通信设备的信息，其中，每个第三通信设备能够提供第一通信设 备所需的至少一个模型的模型信息，进一步地，第一通信设备基于第二通信设备发送的至少一个第三通信设备的信息，可以获取到多个模型的模型信息，从而实现了一种模型获取方法，实现复杂度较低，效率较高，而且多个模型用于生成数据分析结果信息，可以使得成数据分析结果更加准确。
可选地，第一请求包括：分析任务标识，分析任务标识用于标识所需模型适用的数据分析任务；第一请求还包括以下至少一项：
第一指示信息，第一指示信息用于指示请求获取多个模型；
所需模型数量；
第二指示信息，所述第二指示信息用于指示请求获取多个第三通信设备的信息；
所需第三通信设备的数量；
排序方式，所述排序方式用于指示获取到的多个第三通信设备的信息的排序方式；
模型需满足的模型属性信息。
其中，分析任务标识可以指示第一通信设备所需模型适用的数据分析任务。
例如，多个第三通信设备的信息以模型数量从大到小进行排序。例如，包括3个MTLF，MTLF1，MTLF2，MTLF3满足AnLF的需求，MTLF1，MTLF2，MTLF3分别有2个模型，1个模型，3个模型。那么最终NRF返回给AnLF的MTLF信息按照MTLF3，MTLF1，MTLF2的顺序排列。
其中，所需模型数量可以为一个数值或数量范围，例如限定最小数量，最大数量。所需第三通信设备的数量与所需模型数量类似。
可选地，每个第三通信设备的信息包括以下至少一项：
第三通信设备支持的分析任务标识、第三通信设备的标识、第三通信设备的地址、第三通信设备支持的分析任务标识对应的模型数量、第三通信设备支持的至少一个模型的模型属性信息。
其中，第三通信设备支持的分析任务标识对应的模型数量，指的是能够处理分析任务标识对应的分析任务的模型的模型数量。
可选地，模型属性信息包括以下至少一项：模型使用的范围信息、模型标识、模型训练的结果评价信息、模型使用的结果评价信息、模型大小、模型推理时长、训练数据信息。
可选地，模型使用的范围信息包括以下至少一项：模型针对的目标对象、目标时间范围、目标位置范围、单个网络切片选择辅助信息S-NSSAI、数据网络名称DNN；目标时间范围用于指示模型使用的时间范围，目标位置范围用于指示模型使用的位置范围。
可选地，模型训练的结果评价信息包括以下至少一项：准确率、误差信息、模型训练时长、模型训练的数据量；所述模型使用的结果评价信息包括以下至少一项：准确率、误差信息，模型运行时长。
具体地，模型针对的目标对象，例如对某个终端进行移动性分析，目标时间范围即，在该目标时间范围可以使用该模型，目标位置范围，即在该目标位置范围内可以使用该模型，例如目标对象处于该目标位置范围内。
其中，模型训练的结果评价信息用于描述模型训练好后的结果评价信息，例如识别或决策的准确程度、误差信息、时长、数据量等。
例如误差信息可以是平均绝对值误差(Mean Absolute Error，MAE)、均方根误差(Root Mean Square Error，RMSE)等。
模型使用的结果评价信息与模型训练的结果评价信息类似。
其中，模型大小是指存储该模型或运行该模型时所需的存储空间大小。
其中，模型推理时长用于表示模型运行得到收敛的数据推理结果的时长。
其中，训练数据信息例如可以包括以下至少一项：训练数据的来源信息，如数据的来源位置信息，来源通信设备的信息；训练数据的时间信息，如训练数据为时间从三个月前到两个月前的数据。
其中，模型需满足的模型属性信息例如：模型针对的目标对象为某个区域内的终端、模型大小为500MB之内、模型使用的准确率大于某个阈值、模型所用的训练数据的来源通信设备为通信设备a，训练数据为前一周或一个月内的数据等。
可选地，该方法还包括：
第一通信设备基于多个模型的模型信息获取多个模型；
第一通信设备基于多个模型进行分析处理，得到数据分析结果信息。
示例性地，多个模型地模型信息例如为模型的地址，则可以基于模型的地址，获取多个模型，并基于获取到的多个模型进行分析处理，得到数据分析结果信息。
可选地，第一通信设备基于多个模型中的每个模型进行模型推理操作，获取多个数据推理结果；
第一通信设备对多个数据推理结果进行处理，得到数据分析结果信息。
上述处理，例如聚合(aggregation)、投票(voting)等操作，从而生成最终的数据分析结果信息。
示例性地，第一通信设备对多个数据推理结果进行处理，得到数据分析结果信息，可以通过以下至少一种方式实现：
第一通信设备对多个数据推理结果进行加权平均，得到数据分析结果信息；
第一通信设备对多个推理结果数据进行平均运算，得到分析结果信息；
第一通信设备对多个推理结果数据进行累加，得到分析结果信息；
第一通信设备对多个推理结果数据进行考虑性能的累加，得到分析结果信息。
具体地，加权平均、平均等为聚合操作，累加、考虑性能的累加，即为累加投票的具体表现方式。在进行累加时可以考虑模型的性能，例如模型的准确度为0.6，则对推理结果的权值可以为0.6。
上述实施方式中，根据多个模型进行模型推理、分析等处理，所获得的结果表现较好，准确性较高，并且不容易受到网络状况、网络数据变化所带来的影响。
可选地，步骤101之前还包括：
第一通信设备接收第四通信设备发送的任务请求；
步骤101具体可以为：
第一通信设备基于任务请求，向第二通信设备发送第一请求；
可选地，第一通信设备基于多个模型进行分析处理，得到数据分析结果信息之后，还包括：
第一通信设备向第四通信设备发送数据分析结果信息。
具体地，第四通信设备可以是消费者设备，例如应用功能(Application Function，AF)，该第四通信设备可以向第一通信设备发送任务请求，用于请求获取某些数据分析任务对应的数据分析结果信息。
可选地，任务请求包括的内容可以与第一请求包括的内容相同。
可选地，步骤103可以通过如下方式实现：
第一通信设备基于至少一个第三通信设备的信息以及模型需满足的模型属性信息，确定至少一个目标通信设备，并从至少一个目标通信设备获取多个模型的模型信息，其中，每个目标通信设备能够提供与模型属性信息匹配的模型的模型信息。
示例性地，某个第三通信设备支持的分析任务标识为标识1和标识2、标识1对应的模型数量为5，标识2对应的模型数量为10，标识1对应的5个模型中有3个模型的使用范围与需满足的模型属性信息中的使用范围一致、模型训练和使用的结果评价信息与上述要求的模型属性信息中的结果评价信息匹配、模型大小与模型属性信息中的模型大小也匹配，因此上述的第三通信设备可以作为目标通信设备。
可选地，从至少一个目标通信设备获取多个模型的模型信息，具体可以通过如下方式实现：
第一通信设备向至少一个目标通信设备发送第二请求，第二请求用于向至少一个目标通信设备获取多个模型的模型信息；
第一通信设备接收至少一个目标通信设备发送的多个模型的模型信息；
第二请求包括以下至少一项：
分析任务标识，分析任务标识用于标识所需模型适用的数据分析任务；
所需模型的标识；
所需模型数量；
模型需满足的模型属性信息。
具体地，第一通信设备接收至少一个目标通信设备发送的多个模型的模型信息，每个目标通信设备可以发一个或多个模型的模型信息。
可选地，至少一个目标通信设备的数量可以是一个，第一通信设备向至少一个目标通信设备发送第二请求，包括：
第一通信设备向至少一个目标通信设备中的一个目标通信设备发送第二请求；
第一通信设备接收至少一个目标通信设备发送的多个模型的模型信息，包括：
第一通信设备接收一个目标通信设备发送的多个模型的模型信息；此时该目标通信设备可以提供多个模型的模型信息。
其中，在第一通信设备向至少一个目标通信设备中的一个目标通信设备发送第二请求之前，方法还包括：
第一通信设备确定一个目标通信设备能够提供多个模型的模型信息。
可选地，至少一个目标通信设备的数量可以是多个，第一通信设备向至少一个目标通信设备发送第二请求，包括：
第一通信设备向至少一个目标通信设备中的多个目标通信设备发送第二请求；
第一通信设备接收至少一个目标通信设备发送的多个模型的模型信息，包括：
第一通信设备接收多个目标通信设备发送的多个模型的模型信息；
其中，在第一通信设备向至少一个目标通信设备中的多个目标通信设备发送第二请求之前，方法还包括：
第一通信设备确定至少一个目标通信设备中每个目标通信设备均不能提供多个模型的全部模型的模型信息。
具体地，此时，每个目标通信设备可以发一个或多个模型的模型信息。每个目标通信设备均不能提供第一通信设备所需的多个模型的全部模型的模型信息，因此需要从多个目标通信设备获取。
上述实施方式中，第一通信设备向至少一个目标通信设备中的一个目标通信设备或多个目标通信设备发送第二请求，实现了获取多个模型的模型信息，实现复杂度较低，效率较高。
示例性地，如图3所示，步骤1：MTLF(或NWDAF，或包括MTLF的NWDAF)可以向NRF第二通信设备发送能力注册消息，可选地，能力注册消息包括以下至少一项：标识、支持的分析任务标识、网络功能类型、网络功能实例标识、支持的分析任务标识对应的模型数量、支持的至少一个模型的模型属性信息。
例如，能力注册消息可以通过信令Nnrf_NFManagement_NFRegister Register携带。
其中，网络功能类型例如为NF type＝NWDAF type或MTLF type。
网络功能实例标识(NF instance ID)，指此次注册的通信设备的标识信息，如其全限定域名(Fully Qualified Domain Name，FQDN)，用于指示此通信设备的位置和连接此通信设备)信息或者IP地址信息(另一种标识信息)。
所支持的任务分析标识(analytic ID)，指示该NWDAF所能进行的任务类型。
步骤2、3：NRF储存该MTLF的信息并发送响应消息，例如响应消息可通过Nnrf_NFManagement_NFRegister response信令携带，响应消息用于通知第三通信设备注册成功。
步骤4、消费者设备向AnLF(或NWDAF，或包括MTLF的AnLF)发送任务请求，例如通过Nnwdaf_AnalyticsInfo_Request信令携带。
步骤5、向NRF发送第一请求，第一请求例如可以通过Nnrf_NFDiscovery_Request信令携带，第一请求包括任务分析标识，还可以包括以下至少一项：
第一指示信息，所述第一指示信息用于指示请求获取多个模型；
所需模型数量；
第二指示信息，所述第二指示信息用于指示请求获取多个第三通信设备的信息；
所需第三通信设备的数量；
排序方式，所述排序方式用于指示获取到的多个第三通信设备的信息的排序方式；
模型需满足的模型属性信息。
步骤6、NRF根据AnLF的第一请求确定要反馈的MTLF信息。
具体地，NRF在接收到第一请求后，查找所有注册的MTLF，并通过匹配信息以选择出最终的MTLF并生成MTLF列表，通过该MTLF列表反馈MTLF信息，AnLF的第一请求中包括所需模型数量和需要多个MTLF的信息的指示信息，那么NRF会查找所有支持该第一请求包括的分析任务标识的MTLF并反馈拥有该分析任务标识对应的多个模型的MTLF信息。如果没有多个MTLF满足条件，则会反馈其他不满足模型数量要求的MTLF，此时多个MTLF所拥有的模型数量需满足第一请求中的所需模型数量的要求。
如果第一请求只包括所需MTLF的数量，则NRF会选择多个拥有该分析任务标识对应的模型的MTLF的信息，进行反馈。
如果第一请求只包括所需模型数量，则NRF会匹配MTLF的信息中拥有该分析任务标识对应的多个模型的MTLF的信息，进行反馈。若没有则返回失败信息。
步骤7、NRF向AnLF返回符合条件的MTLF信息。
步骤8、AnLF确定所要获取模型的目标MTLF。
步骤9a、步骤9b、AnLF向目标MTLF请求满足任务需求的模型，例如包括任务分析标识、模型需满足的模型属性信息。
步骤10、AnLF使用多个模型进行分析处理。
步骤11、AnLF向消费者设备返回数据分析结果信息。例如，通过Nnwdaf_AnalyticsInfo_Request response信令携带。
可选地，消费者设备还可以向终端等用户设备发送数据分析结果信息。
本申请实施例提供的模型获取方法，执行主体可以为模型获取装置。本申请实施例中以模型获取装置执行模型获取方法为例，说明本申请实施例提供的模型获取装置。
图4是本申请提供的模型获取方法的流程示意图之二。如图4所示，本实施例提供的模型获取方法，包括：
步骤201、第二通信设备接收第一通信设备发送的第一请求，第一请求用于请求获取至少一个第三通信设备的信息，其中，每个第三通信设备能够提供第一通信设备所需的至少一个模型的模型信息；
步骤202、第二通信设备基于第一请求，确定至少一个第三通信设备；
步骤203、第二通信设备向所述第一通信设备发送至少一个第三通信设备的信息；至少一个第三通信设备的信息用于第一通信设备获取多个模型的模型信息；多个模型用于生成数据分析结果信息。
可选地，所述第一请求包括：分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；所述第一请求还包括以下至少一项：
第一指示信息，所述第一指示信息用于指示请求获取多个模型；
所需模型数量；
第二指示信息，所述第二指示信息用于指示请求获取多个第三通信设备的信息；
所需第三通信设备的数量；
排序方式，所述排序方式用于指示获取到的多个第三通信设备的信息的排序方式；
模型需满足的模型属性信息。
可选地，所述方法还包括：
所述第二通信设备基于所述第一通信设备发送的第一请求，确定要返回的所述至少一个第三通信设备的数量。
可选地，所述第二通信设备基于所述第一通信设备发送的第一请求，确定所述至少一个第三通信设备的数量，包括以下至少一项：
所述第二通信设备基于所述第一指示信息和/或所述所需模型数量，确定所述至少一个第三通信设备的数量为一个，一个所述第三通信设备能够提供多个模型或所述所需模型数量个模型的模型信息；
所述第二通信设备基于所述第一指示信息和/或所述所需模型数量，确定所述至少一个第三通信设备的数量多个，多个所述第三通信设备中每个第三通信设备均不能提供所述多个模型的全部模型或所述所需模型数量个模型的模型信息；
所述第二通信设备基于所述第二指示信息，确定所述至少一个第三通信设备的数量为多个；
所述第二通信设备基于所述所需第三通信设备的数量，确定所述至少一个第三通信设备的数量为所需第三通信设备的数量。
可选地，所述模型属性信息包括以下至少一项：模型使用的范围信息、模型标识、模型训练的结果评价信息、模型使用的结果评价信息、模型大小、模型执行时长、训练数据的来源信息、训练数据的时间信息；模型大小用于表示模型存储或运行所需的存储空间大小，模型推理时长用于表示模型运行得到数据推理结果的时长。
可选地，所述模型训练的结果评价信息包括以下至少一项：准确率、误差信息、模型训练时长、模型训练的数据量；所述模型使用的结果评价信息包括以下至少一项：准确率、误差信息，模型运行时长。
可选地，每个第三通信设备的信息包括以下至少一项：
所述第三通信设备支持的分析任务标识、所述第三通信设备的标识、所述第三通信设备的地址、所述第三通信设备支持的分析任务标识对应的模型数量或所述第三通信设备支持的至少一个模型的模型属性信息。
可选地，所述方法还包括：
所述第二通信设备接收第三通信设备发送的能力注册消息，所述能力注册消息包括以下至少一项：所述第三通信设备支持的分析任务标识、所述第三通信设备的网络功能类型、所述第三通信设备的网络功能实例标识、所述第三通信设备支持的分析任务标识对应的模型数量、所述第三通信设备支持的至少一个模型的模型属性信息。
本实施例的方法，其具体实现过程与技术效果与第一通信设备侧方法实施例中相同，具体可以参见第一通信设备侧方法实施例中的详细介绍，此处不再赘述。
图5是本申请提供的模型获取方法的流程示意图之三。如图5所示，本实施例提供的模型获取方法，包括：
步骤301、第三通信设备接收第一通信设备发送的第二请求，第二请求用于请求获取至少一个模型；
步骤302、第三通信设备基于第二请求向第一通信设备发送至少一个模型的模型信息；
所述第二请求包括以下至少一项：
分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；
所需模型的标识；
所需模型数量；
模型需满足的模型属性信息。
可选地，所述方法还包括：
所述第三通信设备向第二通信设备发送能力注册消息，所述能力注册消息包括以下至少一项：所述第三通信设备支持的分析任务标识、所述第三通信设备的网络功能类型、所述第三通信设备的网络功能实例标识、所述第三通信设备支持的分析任务标识对应的模型数量、所述第三通信设备支持的至少一个模型的模型属性信息。
可选地，所述模型属性信息包括以下至少一项：模型使用的范围信息、模型标识、模型训练的结果评价信息、模型使用的结果评价信息、模型大小、模型执行时长、训练数据的来源信息、训练数据的时间信息；模型大小用于表示模型存储或运行所需的存储空间大小，模型推理时长用于表示模型运行得到数据推理结果的时长。
可选地，所述模型训练的结果评价信息包括以下至少一项：准确率、误差信息、模型训练时长、模型训练的数据量；所述模型使用的结果评价信息包括以下至少一项：准确率、误差信息，模型运行时长。
可选地，每个第三通信设备的信息包括以下至少一项：
所述第三通信设备支持的分析任务标识、所述第三通信设备的标识、所述第三通信设备的地址、所述第三通信设备支持的分析任务标识对应的模型数量或所述第三通信设备支持的至少一个模型的模型属性信息。
本实施例的方法，其具体实现过程与技术效果与第一通信设备侧方法实施例中相同，具体可以参见第一通信设备侧方法实施例中的详细介绍，此处不再赘述。
图6是本申请提供的模型获取方法的流程示意图之四。如图6所示，本实施例提供的模型获取方法，包括：
步骤401、第四通信设备向第一通信设备发送任务请求；
步骤402、第四通信设备接收第一通信设备发送的数据分析结果信息，数据分析结果信息为第一通信设备基于多个模型进行分析处理得到的，多个模型为第一通信设备基于至少一个第三通信设备的信息获取到的，每个第三通信设备能够提供第一通信设备所需的至少一个模型的模型信息。
可选地，所述模型属性信息包括以下至少一项：模型使用的范围信息、模型标识、模型训练的结果评价信息、模型使用的结果评价信息、模型大小、模型执行时长、训练数据的来源信息、训练数据的时间信息；模型大小用于表示模型存储或运行所需的存储空间大小，模型推理时长用于表示模型运行得到数据推理结果的时长。
可选地，所述模型训练的结果评价信息包括以下至少一项：准确率、误差信息、模型训练时长、模型训练的数据量；所述模型使用的结果评价信息包括以下至少一项：准确率、误差信息，模型运行时长。
可选地，每个第三通信设备的信息包括以下至少一项：
所述第三通信设备支持的分析任务标识、所述第三通信设备的标识、所述第三通信设备的地址、所述第三通信设备支持的分析任务标识对应的模型数量或所述第三通信设备支持的至少一个模型的模型属性信息。
本实施例的方法，其具体实现过程与技术效果与第一通信设备侧方法实施例中相同，具体可以参见第一通信设备侧方法实施例中的详细介绍，此处不再赘述。
图7是本申请提供的模型获取装置的结构示意图之一。本实施例提供的模型获取装置可应用于第一通信设备。如图7所示，本实施例提供的模型获取装置，包括：
发送模块110，用于向第二通信设备发送第一请求，所述第一请求用于请求获取至少一个第三通信设备的信息，其中，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息；
接收模块120，用于接收所述第二通信设备发送的至少一个第三通信设备的信息；
获取模块130，用于基于所述至少一个第三通信设备的信息，获取多个模型的模型信息；所述多个模型用于生成数据分析结果信息。
可选地，所述获取模块130还用于：
基于所述多个模型的模型信息获取所述多个模型；
可选地，所述装置还包括：处理模块，用于基于所述多个模型进行分析处理，得到所述数据分析结果信息。
可选地，所述获处理模块具体用于：
基于所述多个模型中的每个模型进行模型推理操作，获取多个数据推理结果；
对所述多个数据推理结果进行处理，得到所述数据分析结果信息。
可选地，所述获取模块130具体用于执行以下至少一项：
对所述多个数据推理结果进行加权平均，得到所述数据分析结果信息；
对所述多个推理结果数据进行平均运算，得到所述分析结果信息；
对所述多个推理结果数据进行累加，得到所述分析结果信息；
对所述多个推理结果数据进行考虑性能的累加，得到所述分析结果信息。
可选地，接收模块120，还用于：
接收第四通信设备发送的任务请求；
所述发送模块110，具体用于：
基于所述任务请求，向所述第二通信设备发送所述第一请求；
向所述第四通信设备发送所述数据分析结果信息。
可选地，所述获取模块130具体用于：
基于所述至少一个第三通信设备的信息以及模型需满足的模型属性信息，确定至少一个目标通信设备，并从所述至少一个目标通信设备获取所述多个模型的模型信息，其中，每个所述目标通信设备能够提供与所述模型属性信息匹配的模型的模型信息。
可选地，所述第一请求包括：
分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；
所述第一请求还包括以下至少一项：
第一指示信息，所述第一指示信息用于指示请求获取多个模型；
所需模型数量；
第二指示信息，所述第二指示信息用于指示请求获取多个第三通信设备的信息；
所需第三通信设备的数量；
排序方式，所述排序方式用于指示获取到的多个第三通信设备的信息的排序方式；
模型需满足的模型属性信息。
可选地，每个第三通信设备的信息包括以下至少一项：
所述第三通信设备支持的分析任务标识、所述第三通信设备的标识、所述第三通信设备的地址、所述第三通信设备支持的分析任务标识对应的模型数量、所述第三通信设备支持的至少一个模型的模型属性信息。
可选地，所述发送模块110还用于：
向所述至少一个目标通信设备发送第二请求，所述第二请求用于向所述至少一个目标通信设备获取所述多个模型的模型信息；
所述接收模块120还用于接收所述至少一个目标通信设备发送的所述多个模型的模型信息；
所述第二请求包括以下至少一项：
分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；
所需模型的标识；
所需模型数量；
模型需满足的模型属性信息。
可选地，所述发送模块110具体用于：
向所述至少一个目标通信设备中的一个目标通信设备发送所述第二请求；
所述接收模块120具体用于：
接收所述一个目标通信设备发送的所述多个模型的模型信息；
处理模块，还用于：
确定所述一个目标通信设备能够提供所述多个模型的模型信息。
可选地，所述发送模块110具体用于：
向所述至少一个目标通信设备中的多个目标通信设备发送所述第二请求；
所述接收模块120具体用于：
接收所述多个目标通信设备发送的所述多个模型的模型信息；
处理模块，还用于：
确定所述至少一个目标通信设备中每个所述目标通信设备均不能提供所述多个模型的全部模型的模型信息。
可选地，所述模型属性信息包括以下至少一项：模型使用的范围信息、模型标识、模型训练的结果评价信息、模型使用的结果评价信息、模型大小、模型推理时长、训练数据的来源信息、训练数据的时间信息；模型大小用于表示模型存储或运行所需的存储空间大小，模型推理时长用于表示模型运行得到数据推理结果的时长。
可选地，所述模型训练的结果评价信息包括以下至少一项：准确率、误差信息、模型训练时长、模型训练的数据量；所述模型使用的结果评价信息包括以下至少一项：准确率、误差信息，模型运行时长。
本实施例的装置，可以用于执行前述第一通信设备侧方法实施例中任一实施例的方法，其具体实现过程与技术效果与第一通信设备侧方法实施例中相同，具体可以参见第一通信设备侧方法实施例中的详细介绍，此处不再赘述。
图8是本申请提供的模型获取装置的结构示意图之二。本实施例提供的模型获取装置可应用于第二通信设备。如图8所示，本实施例提供的模型获取装置，包括：
接收模块210，用于接收第一通信设备发送的第一请求，所述第一请求用于请求获取至少一个第三通信设备的信息，其中，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息；
处理模块220，用于基于所述第一请求，确定至少一个第三通信设备；
发送模块230，用于向所述第一通信设备发送所述至少一个第三通信设备的信息；所述至少一个第三通信设备的信息用于所述第一通信设备获取多个模型的模型信息；所述多个模型用于生成数据分析结果信息。
可选地，所述第一请求包括：分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；所述第一请求还包括以下至少一项：
第一指示信息，所述第一指示信息用于指示请求获取多个模型；
所需模型数量；
第二指示信息，所述第二指示信息用于指示请求获取多个第三通信设备的信息；
所需第三通信设备的数量；
排序方式，所述排序方式用于指示获取到的多个第三通信设备的信息的排序方式；
模型需满足的模型属性信息。
可选地，所述处理模块220还用于：
基于所述第一通信设备发送的第一请求，确定所述至少一个第三通信设备的数量。
可选地，所述第二通信设备基于所述第一通信设备发送的第一请求，确定所述至少一个第三通信设备的数量，包括以下至少一项：
所述第二通信设备基于所述第一指示信息和/或所述所需模型数量，确定所述至少一个第三通信设备的数量为一个，一个所述第三通信设备能够提供多个模型或所述所需模型数量个模型的模型信息；
所述第二通信设备基于所述第一指示信息和/或所述所需模型数量，确定所述至少一个第三通信设备的数量多个，多个所述第三通信设备中每个第三通信设备均不能提供所述多个模型的全部模型或所述所需模型数量个模型的模型信息；
所述第二通信设备基于所述第二指示信息，确定所述至少一个第三通信设备的数量为多个；
所述第二通信设备基于所述所需第三通信设备的数量，确定所述至少一个第三通信设备的数量为所需第三通信设备的数量。
可选地，所述模型属性信息包括以下至少一项：模型使用的范围信息、模型标识、模型训练的结果评价信息、模型使用的结果评价信息、模型大小、模型执行时长、训练数据的来源信息、训练数据的时间信息；模型大小用于表示模型存储或运行所需的存储空间大小，模型推理时长用于表示模型运行得到数据推理结果的时长。
可选地，所述模型训练的结果评价信息包括以下至少一项：准确率、误差信息、模型训练时长、模型训练的数据量；所述模型使用的结果评价信息包括以下至少一项：准确率、误差信息，模型运行时长。
可选地，每个第三通信设备的信息包括以下至少一项：
所述第三通信设备支持的分析任务标识、所述第三通信设备的标识、所述第三通信设备的地址、所述第三通信设备支持的分析任务标识对应的模型数量或所述第三通信设备支持的至少一个模型的模型属性信息。
可选地，接收模块210，还用于：
接收第三通信设备发送的能力注册消息，所述能力注册消息包括以下至少一项：所述第三通信设备支持的分析任务标识、所述第三通信设备的网络功能类型、所述第三通信设备的网络功能实例标识、所述第三通信设备支持的分析任务标识对应的模型数量、所述第三通信设备支持的至少一个模型的模型属性信息。
本实施例的装置，可以用于执行前述第二通信设备侧方法实施例中任一实施例的方法，其具体实现过程与技术效果与第二通信设备侧方法实施例中相同，具体可以参见第二通信设备侧方法实施例中的详细介绍，此处不再赘述。
图9是本申请提供的模型获取装置的结构示意图之三。本实施例提供的模型获取装置可应用于第三通信设备。如图9所示，本实施例提供的模型获取装置，包括：
接收模块310，用于接收第一通信设备发送的第二请求，所述第二请求用于请求获取至少一个模型；
发送模块320，用于基于所述第二请求向所述第一通信设备发送至少一个模型的模型信息；所述第二请求包括以下至少一项：
分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；
所需模型的标识；
所需模型数量；
模型需满足的模型属性信息。
可选地，所述模型属性信息包括以下至少一项：模型使用的范围信息、模型标识、模型训练的结果评价信息、模型使用的结果评价信息、模型大小、模型执行时长、训练数据的来源信息、训练数据的时间信息；模型大小用于表示模型存储或运行所需的存储空间大小，模型推理时长用于表示模型运行得到数据推理结果的时长。
可选地，所述模型训练的结果评价信息包括以下至少一项：准确率、误差信息、模型训练时长、模型训练的数据量；所述模型使用的结果评价信息包括以下至少一项：准确率、误差信息，模型运行时长。
可选地，每个第三通信设备的信息包括以下至少一项：
所述第三通信设备支持的分析任务标识、所述第三通信设备的标识、所述第三通信设备的地址、所述第三通信设备支持的分析任务标识对应的模型数量或所述第三通信设备支持的至少一个模型的模型属性信息。
可选地，发送模块320，还用于：
向第二通信设备发送能力注册消息，所述能力注册消息包括以下至少一项：所述第三通信设备支持的分析任务标识、所述第三通信设备的网络功能类型、所述第三通信设备的网络功能实例标识、所述第三通信设备支持的分析任务标识对应的模型数量、所述第三通信设备支持的至少一个模型的模型属性信息。
本实施例的装置，可以用于执行前述第三通信设备侧方法实施例中任一实施例的方法，其具体实现过程与技术效果与第三通信设备侧方法实施例中相同，具体可以参见第三通信设备侧方法实施例中的详细介绍，此处不再赘述。
图10是本申请提供的模型获取装置的结构示意图之三。本实施例提供的模型获取装置可应用于第四通信设备。如图10所示，本实施例提供的模型获取装置，包括：
发送模块410，用于向第一通信设备发送任务请求；
接收模块420，用于接收所述第一通信设备发送的数据分析结果信息，所述数据分析结果信息为所述第一通信设备基于多个模型进行分析处理得到的，所述多个模型为所述第一通信设备基于至少一个第三通信设备的信息获取到的，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息。
所述任务请求包括：分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；所述任务请求还包括以下至少一项：
第一指示信息，所述第一指示信息用于指示请求获取多个模型；
所需模型数量；
第二指示信息，所述第二指示信息用于指示请求获取多个第三通信设备的信息；
所需第三通信设备的数量；
排序方式，所述排序方式用于指示获取到的多个第三通信设备的信息的排序方式；
模型需满足的模型属性信息。
可选地，所述模型属性信息包括以下至少一项：模型使用的范围信息、模型标识、模型训练的结果评价信息、模型使用的结果评价信息、模型大小、模型执行时长、训练数据的来源信息、训练数据的时间信息；模型大小用于表示模型存储或运行所需的存储空间大小，模型推理时长用于表示模型运行得到数据推理结果的时长。
可选地，所述模型训练的结果评价信息包括以下至少一项：准确率、误差信息、模型训练时长、模型训练的数据量；所述模型使用的结果评价信息包括以下至少一项：准确率、误差信息，模型运行时长。
可选地，每个第三通信设备的信息包括以下至少一项：
所述第三通信设备支持的分析任务标识、所述第三通信设备的标识、所述第三通信设备的地址、所述第三通信设备支持的分析任务标识对应的模型数量或所述第三通信设备支持的至少一个模型的模型属性信息。
本实施例的装置，可以用于执行前述第四通信设备侧方法实施例中任一实施例的方法，其具体实现过程与技术效果与第四通信设备侧方法实施例中相同，具体可以参见第四通信设备侧方法实施例中的详细介绍，此处不再赘述。
本申请实施例中的模型获取装置可以是电子设备，例如具有操作系统的电子设备，也可以是电子设备中的部件，例如集成电路或芯片。该电子设备可以是终端，也可以为除终端之外的其他设备。示例性的，终端可以包括但不限于上述所列举的终端11的类型，其他设备可以为服务器、网络附属存储器(Network Attached Storage，NAS)等，本申请实施例不作具体限定。
本申请实施例提供的模型获取装置能够实现图2至图6的方法实施例实现的各个过程，并达到相同的技术效果，为避免重复，这里不再赘述。
可选地，如图11所示，本申请实施例还提供一种通信设备1100，包括处理器1101和存储器1102，存储器1102上存储有可在所述处理器1101上运行的程序或指令，例如，该通信设备1100为终端时，该程序或指令被处理器1101执行时实现上述模型获取方法实施例的各个步骤，且能达到相同的技术效果。该通信设备1100为网络侧设备时，该程序或指令被处理器1101执行时实现上述模型获取方法实施例的各个步骤，且能达到相同的技术效果，为避免重复，这里不再赘述。
本申请实施例还提供一种第一通信设备，包括处理器和通信接口，所述通信接口用于第一通信设备向第二通信设备发送第一请求，所述第一请求用于请求获取至少一个第三通信设备的信息，其中，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息；接收所述第二通信设备发送的至少一个第三通信设备的信息；基于所述至少一个第三通信设备的信息，获取多个模型的模型信息；所述多个模型用于生成数据分析结果信息。该第一通信设备实施例与上述第一通信设备侧方法实施例对应，上述方法实施例的各个实施过程和实现方式均可适用于该第一通信设备实施例中，且能达到相同的技术效果。
本申请实施例还提供一种第二通信设备，包括处理器和通信接口，所述通信接口用于接收第一通信设备发送的第一请求，所述第一请求用于请求获取至少一个第三通信设备的信息，其中，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息；向所述第一通信设备发送所述至少一个第三通信设备的信息；所述至少一个第三通信设备的信息用于所述第一通信设备获取多个模型的模型信息；所述多个模型用于生成数据分析结果信息；处理器，用于基于所述第一请求，确定至少一个第三通信设备。该第二通信设备实施例与上述第二通信设备侧方法实施例对应，上述方法实施例的各个实施过程和实现方式均可适用于该第二通信设备实施例中，且能达到相同的技术效果。
本申请实施例还提供一种第三通信设备，包括处理器和通信接口，所述通信接口用于接收第一通信设备发送的第二请求，所述第二请求用于请求获取至少一个模型；基于所述第二请求向所述第一通信设备发送至少一个模型的模型信息；所述第二请求包括以下至少一项：分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；所需模型的标识；所需模型数量；模型需满足的模型属性信息。该第三通信设备实施例与上述第三通信设备侧方法实施例对应，上述方法实施例的各个实施过程和实现方式均可适用于该第三通信设备实施例中，且能达到相同的技术效果。
本申请实施例还提供一种第四通信设备，包括处理器和通信接口，所述通信接口用于向第一通信设备发送任务请求；接收所述第一通信设备发送的数据分析结果信 息，所述数据分析结果信息为所述第一通信设备基于多个模型进行分析处理得到的，所述多个模型为所述第一通信设备基于至少一个第三通信设备的信息获取到的，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息。该第四通信设备实施例与上述第四通信设备侧方法实施例对应，上述方法实施例的各个实施过程和实现方式均可适用于该第四通信设备实施例中，且能达到相同的技术效果。
具体地，本申请实施例还提供了一种网络侧设备。如图12所示，该网络侧设备1200包括：处理器1201、网络接口1202和存储器1203。其中，网络接口1202例如为通用公共无线接口(common public radio interface，CPRI)。
具体地，本发明实施例的网络侧设备1200还包括：存储在存储器1203上并可在处理器1201上运行的指令或程序，处理器1201调用存储器1203中的指令或程序执行图7-图10所示各模块执行的方法，并达到相同的技术效果，为避免重复，故不在此赘述。
可选地，第一通信设备、第二通信设备、第三通信设备和第四通信设备均可以采用该网络侧设备的结构。
本申请实施例还提供一种可读存储介质，所述可读存储介质上存储有程序或指令，该程序或指令被处理器执行时实现上述模型获取方法实施例的各个过程，且能达到相同的技术效果，为避免重复，这里不再赘述。
其中，所述处理器为上述实施例中所述的终端中的处理器。所述可读存储介质，包括计算机可读存储介质，如计算机只读存储器ROM、随机存取存储器RAM、磁碟或者光盘等。
本申请实施例另提供了一种芯片，所述芯片包括处理器和通信接口，所述通信接口和所述处理器耦合，所述处理器用于运行程序或指令，实现上述模型获取方法实施例的各个过程，且能达到相同的技术效果，为避免重复，这里不再赘述。
应理解，本申请实施例提到的芯片还可以称为系统级芯片，系统芯片，芯片系统或片上系统芯片等。
本申请实施例另提供了一种计算机程序/程序产品，所述计算机程序/程序产品被存储在存储介质中，所述计算机程序/程序产品被至少一个处理器执行以实现上述模型获取方法实施例的各个过程，且能达到相同的技术效果，为避免重复，这里不再赘述。
本申请实施例还提供了一种通信系统，包括：第一通信设备、第二通信设备、第三通信设备和第四通信设备，所述第一通信设备、第二通信设备、第三通信设备及第四通信设备可用于执行如上所述的模型获取方法的步骤。
本申请实施例还提供了一种模型处理方法，如图13所示，该模型处理方法包括：
步骤1301，第一网元向第二网元发送注册请求消息，所述注册请求消息包括所述第一网元的能力信息，所述能力信息包括模型数量信息和模型信息中的至少一项；
步骤1302，所述第一网元从所述第二网元接收注册请求响应消息；
其中，所述模型数量信息用于指示所述第一网元支持的与分析任务标识对应的模型的数量；所述模型信息包括第一网元支持的与分析任务标识对应的模型的以下至少一种信息：
模型标识；
模型性能信息；
模型大小，所述模型大小用于指示存储或运行模型需要的存储空间；
推理时长，所述推理时长用于指示基于模型进行模型推理操作所需的时长；
训练数据来源信息，所述训练数据来源信息用于指示模型在训练阶段所使用的训练数据来源的位置信息和网元信息中的至少一项；
训练数据时间信息，所述训练数据时间信息用于指示模型在训练阶段所使用的训练数据的产生的时间。
本申请实施例中，上述第一网元可以理解为模型训练网元，例如，在一些实施例中可以为模型训练逻辑功能(Model Training logical function，MTLF)，或者称之为网络数据分析功能(Network Data Analytics Function，NWDAF)中包含的MTLF。上述第二网元可以理解为NRF。
应理解，上述注册请求消息除了包括能力信息之外还可以包括其他的信息，例如，可以包括：
NF类型，NF类型用于指示此次注册的网元是何种网元，例如为NWDAF类型或者MTLF类型。
网元实例标识信息(NF instance ID)，用于指示此次注册网络的网元表示信息，例如全限定域名(Fully Qualified Domain Name，FQDN)或者IP地址信息；
所支持的分析任务标识(analytic ID)，用于表示该NWDAF网元所能进行的任务类型。
可选地，上述模型性能信息可以用于表示模型输出结果的准确度和误差值等。在一些实施例中，该模型性能信息可以包括以下至少一项：第一性能信息，所述第一性能信息用于指示模型在训练阶段能达到的性能；第二性能信息，所述第二性能信息用于指示模型在推理阶段能达到的性能。
可选地，第二网元接收到上述注册请求消息后，可以储存注册请求消息中携带的信息，并确认注册通过后，向第一网元发送注册请求响应消息。
可选地，上述模型大小可以通过模型的参数量(parameter)表示。上述训练数据来源信息和训练数据时间信息可以理解为模型训练的数据信息，当然在其他实施例中注册请求消息还可以包括其他模型训练的数据信息，在此不做进一步的限定。
需要说明的是，在第三网元需要调用第一网元中的模型执行目标分析任务时，可以首先向第二网元发送第二请求消息，以请求查询可以执行目标分析任务的第一网元。由于第二网元中储存有能力信息，从而可以根据能力信息确定可用于执行所述目标分析任务的模型的N个第一网元。然后由第三网元从N个第一网元中选择目标网元，并调用目标网元中的模型执行目标分析任务，即利用目标网元中的模型对目标任务进行模型推理。由于在第二网元中可以基于能力信息查询到匹配目标分析任务的第一网元，从而可以提高后续针对目标分析任务进行模型推理的可靠性。与此同时，可以避免由于针对目标分析任务进行模型推理的准确度较低和/或误差值较大，导致需要重新查询新的第一网元重新执行目标分析任务的模型推理。因此本申请实施例还可以进一步提高第一网元的查询效率。
例如，在一些实施例中，当目标信息包括模型数量信息，第二网元可以匹配获得满足目标分析任务的模型数量要求的第一网元，从而可以提高针对目标分析任务进行模型推理的可靠性。在一些实施例中，当目标信息包括模型信息时，第一网元可以匹配获得满足对目标分析任务的模型要求(如模型性能要求)的第一网元，从而可以提高针对目标分析任务进行模型推理的可靠性。
可选地，上述第三网元可以理解为数据分析逻辑功能(Analytics logical function，AnLF)，或者称之为NWDAF中包含的AnLF。
本申请实施例通过第一网元向第二网元发送注册请求消息，所述注册请求消息包括所述第一网元的能力信息，所述能力信息包括模型数量信息和模型信息中的至少一项；所述第一网元从所述第二网元接收注册请求响应消息。这样，由于在注册时携带了上述能力信息，使得第二网元可以根据能力信息确定与目标分析任务匹配的第一网元，从而可以通过调用匹配的第一网元中的模型针对目标分析任务进行模型推理。因此本申请实施例可以提高了模型推理的可靠性。
可选地，在一些实施例中，所述第一网元从所述第二网元接收注册请求响应消息之后，所述方法还包括：
所述第一网元从第三网元接收第一请求消息，所述第一请求消息用于获取可用于执行目标分析任务的目标模型；
所述第一网元向所述第三网元发送第一请求响应消息，所述第一请求响应消息包括所述目标模型和用于获取所述目标模型的地址信息中的至少一项。
本申请实施例中，上述第三网元获得N个第一网元后，可以向第一网元发送第一请求消息，以请求获取可执行目标分析任务的目标模型。其中，上述目标分析任务请求可以为第三网元从第四网元接收到的信息，例如，第四网元可以向第三网元发送任务请求消息，该任务请求消息可以包括目标分析任务的分析任务标识，进一步还可以包括任务限定信息，该任务限定信息可以包括数据分析任务模型的限定信息和分析目 标(analytic target)，其中，数据分析任务模型的限定信息可以称之为机器学习模型的限定信息(Machine Learning model filter info)。其中，任务限定信息可以用于限定任务范围，如模型所针对的分析对象(如对某个UE进行移动性分析)、任务目标时间、感兴趣区域(Area of Interest，AOI)、单一网络切片选择辅助信息(Single Network Slice Selection Assistance Information，S-NSSAI)和数据网络名(Data Network Name，DNN)等
可选地，所述第一网元接收到第一请求消息后，可以基于第一请求消息确定可用于执行目标分析任务的目标模型，然后通过第一请求响应消息指示目标模型。第三网元接收到第一请求响应消息后，可以利用目标模型针对目标任务执行模型推理操作，并生成目标分析报告，最后将目标分析报告发送给第四网元。
可选地，在一些实施例中，所述第一请求消息包括所述目标分析任务的分析任务标识和模型限定信息，所述模型限定信息包括以下至少一项：
模型的数量的限定信息；
模型标识的限定信息；
模型性能信息的限定信息；
模型大小的限定信息；
推理时长的限定信息；
训练数据来源的限定信息；
训练数据时间的限定信息。
在本申请实施例中，上述模型限定信息可以基于上述任务限定信息确定，也可以由协议约定(例如，协议约定不同的任务对应的模型限定信息)，或者由第三网元自主确定。其中，第一网元反馈的目标模型应当满足上述模型限定信息，具体情况如下：
上述模型的数量的限定信息可以理解为，与目标分析任务的分析任务标识对应的模型的数量的限定信息。在一些实施例中，模型的数量的限定信息可以包括一个数量阈值，在模型限定信息包括模型的数量的限定信息的情况下，第一网元需要向第三网元反馈的目标模型的数量应大于或等于该数量阈值，或者反馈的目标模型的数量应小于或等于该数量阈值。
针对模型标识的限定信息可以包括一个或者多个预设模型标识，在模型限定信息包括模型标识的限定信息的情况下，第一网元需要从该预设模型标识对应的模型中反馈目标模型至第三网元，或者，第三网元需要从除该预设模型标识对应的模型之外的模型中反馈目标模型至第三网元。
针对模型性能信息的限定信息可以包括性能指标，在模型限定信息包括性能信息的限定信息的情况下，第一网元需要向第三网元反馈满足性能指标的目标模型。该模 型性能信息的限定信息可以包括第一性能信息的限定信息和第二性能信息的限定信息中的至少一项。
针对模型大小的限定信息可以包括模型大小阈值，在模型限定信息包括模型大小的限定信息的情况下，第一网元需要向第三网元反馈大于或等于该模型大小阈值的目标模型，或者反馈小于或等于该模型大小阈值的目标模型。
针对推理时长的限定信息可以包括推理时间阈值，在模型限定信息包括推理时长的限定信息的情况下，第一网元需要向第三网元反馈大于或等于该推理时间阈值的目标模型，或者反馈小于或等于该推理时间阈值的目标模型。
针对训练数据来源的限定信息可以包括训练数据来源信息，在模型限定信息包括训练数据来源的限定信息的情况下，第一网元需要从通过该训练数据来源信息训练的模型中反馈目标模型至第三网元，或者，第三网元需要从除通过该训练数据来源信息训练的模型之外的模型中反馈目标模型至第三网元。
针对训练数据时间信息的限定信息可以包括训练数据时间阈值，在模型限定信息包括训练数据时间信息的限定信息的情况下，可以基于该时间阈值确定时间范围，第一网元需要向第三网元反馈的目标模型在训练阶段所使用的训练数据的产生的时间位于该时间范围内。
可选地，所述第一性能信息包括模型在训练时的以下至少一项信息：第一性能指标、第一性能指标的计算方法、第一时间信息、第一数值和第一结果；其中，所述第一性能指标包括训练时准确度和训练时误差值中的至少一项，所述第一时间信息包括计算所述第一性能指标对应的时间信息，所述第一数值用于表示计算所述第一性能指标所用的数据数量，所述第一结果为基于多个所述第一性能指标计算获得的结果值。
本申请实施例中，上述第一性能指标也可以理解为模型在训练时表现(performance in Training)。即基于某种统计计算得到的值，例如可以为上述训练时准确度(accuracy in Training，AiT)和训练时误差值(Mean Absolute Error in Training，MAEiT)中的至少一项。其中，训练时准确度可以称之为模型在训练时的准确度，该准确度可以将模型决策结果正确的次数处于总决策次数得到准确度。例如，第一网元可以设置一个验证数据集用于评估模型准确度，该验证集中包括用于模型输入的数据和真实的标签数据，第一网元将验证输入数据输入训练后的模型得到输出数据，第一网元再比较输出数据与真实标签数据是否一致，进而利用上述计算方法获得模型准确度的值。
上述第一性能指标的计算方法可包括以下至少一项：模型预测准确数和模型预测总数的比、MAE、均方根误差(Root Mean Square Error)、召回(Recall)和F1分数(F1score)等。
上述第一时间信息表示为一个时间节点或者一段时间(如包括计算第一性能指标的开始时间和计算第一性能指标的结束时间)。
上述第一结果可以表示多个所述第一性能指标的分布情况，具体可以通过预设的计算方法进行计算得到，例如该第一结果可以为平均值、中位数或者方差等。
可选地，在一些实施例中，所述第二性能信息包括模型在实际使用时的以下至少一项信息：第二性能指标、第二性能指标的计算方法、第二时间信息、第二数值和第二结果；其中，所述第二性能指标包括实际使用时准确度和实际使用时误差值中的至少一项，所述第二时间信息包括计算所述第二性能指标对应的时间信息，
本申请实施例中，实际使用可以理解为使用模型进行模型推理，上述第二性能信息与上述第一性能信息对应，第二性能指标也可以理解为模型在实际使用时表现(performance in Use)。例如，上述第二性能指标的计算方法可包括以下至少一项：模型预测准确数和模型预测总数的比、MAE、均方根误差(Root Mean Square Error)、召回(Recall)和F1分数(F1score)等。上述第二时间信息表示为一个时间节点或者一段时间(如包括计算第二性能指标的开始时间和计算第二性能指标的结束时间)。上述第二结果可以表示多个所述第二性能指标的分布情况，具体可以通过预设的计算方法进行计算得到，例如该第二结果可以为平均值、中位数或者方差等。如图14所示，一个网元利用另一网元中的模型对目标分析任务进行模型推理包括以下流程：
步骤1401，MTLF向NRF发送注册请求消息。该注册请求消息可以称之为能力注册消息进行能力注册。
可选地，注册请求消息可以包括MTLF自身标识信息和支持的analytic ID等信息之外，还可以包括上述能力信息，即包括模型数量信息和模型信息中的至少一项。
其中，模型信息中的准确度和误差值可以判断两个模型中哪个模型更适合目标分析任务；准确度的分布可以判断该模型的表现是否稳定；通过判断模型训练数据的来源信息和时间信息，可以判断是否该模型与目标分析任务匹配(例如老的数据所训练的模型有更大概率会受到网络数据变化的影响)。
步骤1402，NRF储存该注册请求消息中携带的信息；
步骤1403，NRF发送注册请求响应消息。
步骤1404，任务消费者向AnLF发送任务请求消息，该任务请求消息包括目标分析任务的分析任务标识、数据分析任务模型的限定信息和分析目标。
可选地，上述步骤1403和步骤1404的顺序在此不做约定，通常的步骤1403位于步骤1404之前。
步骤1405，AnLF向NRF发送第二请求消息，该第二请求消息用于寻找合适的MTLF，除了携带上述目标分析任务的分析任务标识和数据分析任务模型的限定信息之外，还可以进一步包括目标要求信息。
可选地，该目标要求信息包括以下至少一项：
模型的数量的要求信息；
模型标识的要求信息；
模型性能信息的要求信息；
模型大小的要求信息，所述模型大小用于指示存储或运行模型需要的存储空间；
推理时长的要求信息，所述推理时长用于指示基于模型进行模型推理操作所需的时长；
训练数据来源的要求信息，所述训练数据来源信息用于指示模型在训练阶段所使用的训练数据来源的位置信息和网元信息中的至少一项；
训练数据时间的要求信息，所述训练数据时间信息用于指示模型在训练阶段所使用的训练数据的产生的时间。
本申请实施例中，目标要求信息可以理解为对模型的要求或者对模型的限定要求。其中，上述模型的数量的限定信息可以数量阈值，此时合适的MTLF能够提供可用于执行所述目标分析任务的模型的数量需要大于或等于该数量阈值，或者小于或等于该数量阈值。上述模型标识要求信息可以包括至少一个模型标识，此时合适的MTLF能够提供可用于执行所述目标分析任务的模型包括或者不包括该至少一个模型标识对应的模型。模型性能信息的要求信息可以包括性能指标(例如可以包括第一性能指标和第二性能指标中至少一项)，此时合适的MTLF能够提供可用于执行所述目标分析任务的模型需要满足要求的性能指标。上述模型大小的限定信息可以包括模型大小阈值，此时合适的MTLF能够提供可用于执行所述目标分析任务的模型需要包括大于或等于该模型大小阈值的模型或者需要包括小于或等于该模型大小阈值的模型。上述训练数据来源的限定信息可以包括训练数据来源信息，此时合适的MTLF能够提供可用于执行所述目标分析任务的模型需要包括或者不包括通过该训练数据来源信息训练的模型。上述训练数据时间信息可以包括训练数据时间阈值，可以基于该时间阈值确定时间范围，此时合适的MTLF需要包括训练阶段所使用的训练数据的产生的时间位于该时间范围内的模型。
步骤1406，NRF向AnLF反馈第二请求响应消息，其中该第二请求响应消息需包括确定的N个MTLF。进一步地，还可以包括每一个MTLF支持的任务分析标识、模型数量信息和模型信息中的至少一项。
应理解，NRF基于第二请求消息确定N个MTLF。可选地，每一MTLF包括支持上述目标分析任务的分析任务标识的模型，且该模型(或者说该MTLF)需要满足上述目标要求信息和数据分析任务模型的限定信息。
可选地，在一些实施例中，第二请求响应消息还可以包括N个第一网元中每一第一网元对应的有效时间。该有效时间可以理解为MTLF注册的能力信息的有效时间，超过该时间，MTLF的能力信息可能会发生变化，在有效时间之外，请求MTLF中的模型可能会导致最终的模型推理的可靠性无法得到保障。因此，AnLF优先可以去请求处于有效时间内的MTLF中的模型。
步骤1407，AnLF在N个MTLF中确定目标MTLF。
例如，若NRF反馈一个MTLF，则将该MTLF确定为目标MTLF；若NRF反馈多个MTLF，则还需从多个MTLF中选择一个作为目标MTLF。当然在其他实施例中，AnLF还可以选择至少两个MTLF作为目标MTLF。
步骤1408，AnLF向目标MTLF发送第一请求消息(即模型获取请求)，例如可以为Nnwdaf_MLModelInfo_Request或者Nnwdaf_MLModelProvision_Request。
在该第一请求消息中可以包括所述目标分析任务的分析任务标识和模型限定信息。
其中，AnLF接收到第一请求消息后，可以根据目标分析任务的分析任务标识和模型限定信息匹配出合适的目标模型。
步骤1409，目标MTLF向AnLF发送第一请求响应消息，该第一请求响应消息包括目标模型和用于获取目标模型的地址信息。
第一请求信息包括目标模型可以理解为包括模型的配置文件和模型的描述信息等，上述目标模型的地址信息可以包括统一资源定位符(Uniform Resource Locator，URL)、FQDN信息和IP地址等。AnLF获取到地址信息后，可以直接下载目标模型。
步骤1410，AnLF基于目标模型针对目标分析任务执行模型推理，获得目标任务报告。
可选地，AnLF可以使用一个目标模型获得推理结果，并将该推理结果作为目标任务报告；AnLF也可以使用多个目标模型获得不同的推理结果，最后基于多个推理结果生成目标任务报告，例如，可以对多个推理结果进行聚合(aggregation)或者投票(voting)等操作获得最终的推理结果作为目标任务报告。该目标任务报告可以理解为数据分析结果信息。
步骤1411，AnLF向任务消费者反馈任务请求响应消息，所述任务请求响应消息包括所述目标分析报告。
一种实施方式中，本申请保护以下方案：
1.一种模型处理方法，包括：
第一网元向第二网元发送注册请求消息，所述注册请求消息包括所述第一网元的能力信息，所述能力信息包括模型数量信息和模型信息中的至少一项；
所述第一网元从所述第二网元接收注册请求响应消息；
其中，所述模型数量信息用于指示所述第一网元支持的与分析任务标识对应的模型的数量；所述模型信息包括第一网元支持的与分析任务标识对应的模型的以下至少一种信息：
模型标识；
模型性能信息；
模型大小，所述模型大小用于指示存储或运行模型需要的存储空间；
推理时长，所述推理时长用于指示基于模型进行模型推理操作所需的时长；
训练数据来源信息，所述训练数据来源信息用于指示模型在训练阶段所使用的训练数据来源的位置信息和网元信息中的至少一项；
训练数据时间信息，所述训练数据时间信息用于指示模型在训练阶段所使用的训练数据的产生的时间。
2.根据上述1所述的方法，所述第一网元从所述第二网元接收注册请求响应消息之后，所述方法还包括：
所述第一网元从第三网元接收第一请求消息，所述第一请求消息用于获取可用于执行目标分析任务的目标模型；
所述第一网元向所述第三网元发送第一请求响应消息，所述第一请求响应消息包括所述目标模型和用于获取所述目标模型的地址信息中的至少一项。
3.根据上述2所述的方法，所述第一请求消息包括所述目标分析任务的分析任务标识和模型限定信息，所述模型限定信息包括以下至少一项：
模型的数量的限定信息；
模型标识的限定信息；
模型性能信息的限定信息；
模型大小的限定信息；
推理时长的限定信息；
训练数据来源的限定信息；
训练数据时间的限定信息。
4.根据上述1至3中任一项所述的方法，所述模型性能信息包括以下至少一项：
第一性能信息，所述第一性能信息用于指示模型在训练阶段能达到的性能；
第二性能信息，所述第二性能信息用于指示模型在推理阶段能达到的性能。
5.根据上述4所述的方法，所述第一性能信息包括模型在训练时的以下至少一项信息：第一性能指标、第一性能指标的计算方法、第一时间信息、第一数值和第一结果；其中，所述第一性能指标包括训练时准确度和训练时误差值中的至少一项，所述第一时间信息包括计算所述第一性能指标对应的时间信息，所述第一数值用于表示计算所述第一性能指标所用的数据数量，所述第一结果为基于多个所述第一性能指标计算获得的结果值。
6.根据上述4所述的方法，所述第二性能信息包括模型在实际使用时的以下至少一项信息：第二性能指标、第二性能指标的计算方法、第二时间信息、第二数值和第二结果；其中，所述第二性能指标包括实际使用时准确度和实际使用时误差值中的至少一项，所述第二时间信息包括计算所述第二性能指标对应的时间信息，所述第二数值用于表示计算所述第二性能指标所用的数据数量，所述第二结果为基于多个所述第二性能指标计算获得的结果值。
7.一种模型处理方法，包括：
第二网元从第一网元接收注册请求消息，所述注册请求消息包括所述第一网元的能力信息；
所述第二网元储存所述能力信息，并向所述第一网元发送注册请求响应消息；
其中，所述能力信息包括模型数量信息和模型信息中的至少一项；所述模型数量信息用于指示所述第一网元支持的与分析任务标识对应的模型的数量；所述模型信息包括第一网元支持的与分析任务标识对应的模型的以下至少一种信息：
模型标识；
模型性能信息；
模型大小，所述模型大小用于指示存储或运行模型需要的存储空间；
推理时长，所述推理时长用于指示基于模型进行模型推理操作所需的时长；
训练数据来源信息，所述训练数据来源信息用于指示模型在训练阶段所使用的训练数据来源的位置信息和网元信息至少一项；
训练数据时间信息，所述训练数据时间信息用于指示模型在训练阶段所使用的训练数据的产生的时间。
8.根据上述7所述的方法，所述第二网元储存所述能力信息，并向所述第一网元发送注册请求响应消息之后，所述方法还包括：
所述第二网元从第三网元接收第二请求消息，所述第二请求消息包括目标分析任务的分析任务标识；
所述第二网元基于所述第二请求消息确定N个第一网元，所述N个第一网元为能够提供可用于执行所述目标分析任务的模型的第一网元，N为正整数；
所述第二网元向所述第三网元发送第二请求响应消息，所述第二请求响应消息用于指示所述N个第一网元。
9.根据上述8所述的方法，所述第二请求消息还包括目标要求信息，所述N个第一网元中的任一个第一网元满足所述目标要求信息，其中，所述目标要求信息包括以下至少一项：
模型的数量的要求信息；
模型标识的要求信息；
模型性能信息的要求信息；
模型大小的要求信息；
推理时长的要求信息；
训练数据来源的要求信息；
训练数据时间的要求信息。
10.根据上述8所述的方法，所述第二请求响应消息包括所述N个第一网元的标识信息和所述N个第一网元的地址信息中的至少一项。
11.根据上述8所述的方法，所述第二请求响应消息还包括所述N个第一网元中每一第一网元对应的有效时间。
12.根据上述7至11中任一项所述的方法，所述模型性能信息包括以下至少一项：
第一性能信息，所述第一性能信息用于指示模型在训练阶段能达到的性能；
第二性能信息，所述第二性能信息用于指示模型在推理阶段能达到的性能。
13.根据上述12所述的方法，所述第一性能信息包括模型在训练时的以下至少一项信息：第一性能指标、第一性能指标的计算方法、第一时间信息、第一数值和第一结果；其中，所述第一性能指标包括训练时准确度和训练时误差值中的至少一项，所述第一时间信息包括计算所述第一性能指标对应的时间信息，所述第一数值用于表示计算所述第一性能指标所用的数据数量，所述第一结果为基于多个所述第一性能指标计算获得的结果值。
14.根据上述12所述的方法，所述第二性能信息包括模型在实际使用时的以下至少一项信息：第二性能指标、第二性能指标的计算方法、第二时间信息、第二数值和第二结果；其中，所述第二性能指标包括实际使用时准确度和实际使用时误差值中的至少一项，所述第二时间信息包括计算所述第二性能指标对应的时间信息，所述第二数值用于表示计算所述第二性能指标所用的数据数量，所述第二结果为基于多个所述第二性能指标计算获得的结果值。
15.一种模型处理方法，包括：
第三网元向第二网元发送第二请求消息，所述第二请求消息包括目标分析任务的分析任务标识；
所述第三网元从所述第二网元接收第二请求响应消息，所述第二请求响应消息用于指示N个第一网元，所述N个第一网元为能够提供可用于执行所述目标分析任务的模型的第一网元，N为正整数；
所述第三网元向所述N个第一网元中的目标网元发送第一请求消息；
所述第三网元从所述目标网元接收第一请求响应消息，所述第一请求响应消息包括目标模型和用于获取所述目标模型的地址信息中的至少一项，所述目标模型可用于执行所述目标分析任务。
16.根据上述15所述的方法，所述第一请求消息包括所述目标分析任务的分析任务标识和模型限定信息，所述模型限定信息包括以下至少一项：
模型的数量的限定信息；
模型标识的限定信息；
模型性能信息；
模型大小的限定信息，所述模型大小用于指示存储或运行模型需要的存储空间；
推理时长的限定信息，所述推理时长用于指示基于模型进行模型推理操作所需的时长；
训练数据来源的限定信息，所述训练数据来源信息用于指示模型在训练阶段所使用的训练数据来源的位置信息和网元信息中的至少一项；
训练数据时间的限定信息，所述训练数据时间信息用于指示模型在训练阶段所使用的训练数据的产生的时间。
17.根据上述16所述的方法，所述模型性能信息包括以下至少一项：
第一性能信息，所述第一性能信息用于指示模型在训练阶段能达到的性能；
第二性能信息，所述第二性能信息用于指示模型在推理阶段能达到的性能。
18.根据上述17所述的方法，所述第一性能信息包括模型在训练时的以下至少一项信息：第一性能指标、第一性能指标的计算方法、第一时间信息、第一数值和第一结果；其中，所述第一性能指标包括训练时准确度和训练时误差值中的至少一项，所述第一时间信息包括计算所述第一性能指标对应的时间信息，所述第一数值用于表示计算所述第一性能指标所用的数据数量，所述第一结果为基于多个所述第一性能指标计算获得的结果值。
19.根据上述17所述的方法，所述第二性能信息包括模型在实际使用时的以下至少一项信息：第二性能指标、第二性能指标的计算方法、第二时间信息、第二数值和第二结果；其中，所述第二性能指标包括实际使用时准确度和实际使用时误差值中的至少一项，所述第二时间信息包括计算所述第二性能指标对应的时间信息，所述第二数值用于表示计算所述第二性能指标所用的数据数量，所述第二结果为基于多个所述第二性能指标计算获得的结果值。
20.根据上述15所述的方法，所述第二请求消息还包括目标要求信息，所述N个第一网元中的任一个第一网元满足所述目标要求信息，所述目标要求信息包括以下至少一项：
模型的数量的要求信息；
模型标识的要求信息；
模型性能信息的要求信息；
模型大小的要求信息，所述模型大小用于指示存储或运行模型需要的存储空间；
推理时长的要求信息，所述推理时长用于指示基于模型进行模型推理操作所需的时长；
训练数据来源的要求信息，所述训练数据来源信息用于指示模型在训练阶段所使用的训练数据来源的位置信息和网元信息中的至少一项；
训练数据时间的要求信息，所述训练数据时间信息用于指示模型在训练阶段所使用的训练数据的产生的时间。
21.根据上述15所述的方法，所述第三网元向第二网元发送第二请求消息之前，所述方法还包括：
所述第三网元从第四网元接收任务请求消息，所述任务请求消息包括目标分析任务的分析任务标识。
22.根据上述21所述的方法，所述第三网元从所述目标网元接收第一请求响应消息之后，所述方法还包括：
所述第三网元利用所述目标模型执行所述目标分析任务，获得目标分析报告；
所述第三网元向所述第四网元发送任务请求响应消息，所述任务请求响应消息包括所述目标分析报告。
23.根据上述22所述的方法，所述目标模型的数量为M1个，M1为正整数，所述第三网元利用所述目标模型执行所述目标分析任务，获得目标分析报告包括：
所述第三网元利用M2个所述目标模型执行针对所述目标分析任务的模型推理，获得M2个推理结果，M2为小于或等于M1的正整数；
所述第三网元基于所述M2个推理结果，生成所述目标分析报告。
24.根据上述15所述的方法，所述目标网元为一个所述第一网元。
25.根据上述15所述的方法，所述第二请求响应消息包括所述N个第一网元的标识信息和所述N个第一网元的地址信息中的至少一项。
26.根据上述15所述的方法，所述第二请求响应消息还包括所述N个第一网元中每一第一网元对应的有效时间。
27.一种模型处理装置，包括：
第一发送模块，用于向第二网元发送注册请求消息，所述注册请求消息包括第一网元的能力信息，所述能力信息包括模型数量信息和模型信息中的至少一项；
第一接收模块，用于从所述第二网元接收注册请求响应消息；
其中，所述模型数量信息用于指示第一网元支持的与分析任务标识对应的模型的数量；所述模型信息包括第一网元支持的与分析任务标识对应的模型的以下至少一种信息：
模型标识；
模型性能信息；
模型大小，所述模型大小用于指示存储或运行模型需要的存储空间；
推理时长，所述推理时长用于指示基于模型进行模型推理操作所需的时长；
训练数据来源信息，所述训练数据来源信息用于指示模型在训练阶段所使用的训练数据来源的位置信息和网元信息中的至少一项；
训练数据时间信息，所述训练数据时间信息用于指示模型在训练阶段所使用的训练数据的产生的时间。
28、根据上述27所述的装置，所述第一接收模块还用于从第三网元接收第一请求消息，所述第一请求消息用于获取可用于执行目标分析任务的目标模型；
所述第一发送模块还用于向所述第三网元发送第一请求响应消息，所述第一请求响应消息包括所述目标模型和用于获取所述目标模型的地址信息中的至少一项。
29、根据上述28所述的装置，所述第一请求消息包括所述目标分析任务的分析任务标识和模型限定信息，所述模型限定信息包括以下至少一项：
模型的数量的限定信息；
模型标识的限定信息；
模型性能信息的限定信息；
模型大小的限定信息；
推理时长的限定信息；
训练数据来源的限定信息；
训练数据时间的限定信息。
30、根据上述27至29任一项所述的装置，所述模型性能信息包括以下至少一项：
第一性能信息，所述第一性能信息用于指示模型在训练阶段能达到的性能；
第二性能信息，所述第二性能信息用于指示模型在推理阶段能达到的性能。
31、根据上述30所述的装置，所述第一性能信息包括模型在训练时的以下至少一项信息：第一性能指标、第一性能指标的计算方法、第一时间信息、第一数值和第一结果；其中，所述第一性能指标包括训练时准确度和训练时误差值中的至少一项，所 述第一时间信息包括计算所述第一性能指标对应的时间信息，所述第一数值用于表示计算所述第一性能指标所用的数据数量，所述第一结果为基于多个所述第一性能指标计算获得的结果值。
32、根据上述30所述的装置，所述第二性能信息包括模型在实际使用时的以下至少一项信息：第二性能指标、第二性能指标的计算方法、第二时间信息、第二数值和第二结果；其中，所述第二性能指标包括实际使用时准确度和实际使用时误差值中的至少一项，所述第二时间信息包括计算所述第二性能指标对应的时间信息，所述第二数值用于表示计算所述第二性能指标所用的数据数量，所述第二结果为基于多个所述第二性能指标计算获得的结果值。
33.一种模型处理装置，其特征在于，包括：
第二接收模块，用于从第一网元接收注册请求消息，所述注册请求消息包括所述第一网元的能力信息；
存储模块，用于储存所述能力信息；
第二发送模块，用于向所述第一网元发送注册请求响应消息；
其中，所述能力信息包括模型数量信息和模型信息中的至少一项；所述模型数量信息用于指示所述第一网元支持的与分析任务标识对应的模型的数量；所述模型信息包括第一网元支持的与分析任务标识对应的模型的以下至少一种信息：
模型标识；
模型性能信息；
模型大小，所述模型大小用于指示存储或运行模型需要的存储空间；
推理时长，所述推理时长用于指示基于模型进行模型推理操作所需的时长；
训练数据来源信息，所述训练数据来源信息用于指示模型在训练阶段所使用的训练数据来源的位置信息和网元信息至少一项；
训练数据时间信息，所述训练数据时间信息用于指示模型在训练阶段所使用的训练数据的产生的时间。
34、根据上述33所述的装置，其中，所述装置还包括第一确定模块，
所述第二接收模块701还用于从第三网元接收第二请求消息，所述第二请求消息包括目标分析任务的分析任务标识；
所述第一确定模块用于基于所述第二请求消息确定N个第一网元，所述N个第一网元为能够提供可用于执行所述目标分析任务的模型的第一网元，N为正整数；
所述第二发送模块703还用于向所述第三网元发送第二请求响应消息，所述第二请求响应消息用于指示所述N个第一网元。
35、根据上述34所述的装置，其中，所述第二请求消息还包括目标要求信息，所述N个第一网元中的任一个第一网元满足所述目标要求信息，其中，所述目标要求信息包括以下至少一项：
模型的数量的要求信息；
模型标识的要求信息；
模型性能信息的要求信息；
模型大小的要求信息；
推理时长的要求信息；
训练数据来源的要求信息；
训练数据时间的要求信息。
36、根据上述34所述的装置，所述第二请求响应消息包括所述N个第一网元的标识信息和所述N个第一网元的地址信息中的至少一项。
37、根据上述34至36任一项所述的装置，所述第二请求响应消息还包括所述N个第一网元中每一第一网元对应的有效时间。
38、根据上述33至37任一项所述的装置，所述模型性能信息包括以下至少一项：
第一性能信息，所述第一性能信息用于指示模型在训练阶段能达到的性能；
第二性能信息，所述第二性能信息用于指示模型在推理阶段能达到的性能。
39、根据上述38所述的装置，所述第一性能信息包括模型在训练时的以下至少一项信息：第一性能指标、第一性能指标的计算方法、第一时间信息、第一数值和第一结果；其中，所述第一性能指标包括训练时准确度和训练时误差值中的至少一项，所述第一时间信息包括计算所述第一性能指标对应的时间信息，所述第一数值用于表示计算所述第一性能指标所用的数据数量，所述第一结果为基于多个所述第一性能指标计算获得的结果值。
40、根据上述39所述的装置，所述第二性能信息包括模型在实际使用时的以下至少一项信息：第二性能指标、第二性能指标的计算方法、第二时间信息、第二数值和第二结果；其中，所述第二性能指标包括实际使用时准确度和实际使用时误差值中的至少一项，所述第二时间信息包括计算所述第二性能指标对应的时间信息，所述第二数值用于表示计算所述第二性能指标所用的数据数量，所述第二结果为基于多个所述第二性能指标计算获得的结果值。
41.一种模型处理装置，其特征在于，包括：
第三发送模块，用于向第二网元发送第二请求消息，所述第二请求消息包括目标分析任务的分析任务标识；
第三接收模块，还用于从所述第二网元接收第二请求响应消息，所述第二请求响应消息用于指示N个第一网元，所述N个第一网元为能够提供可用于执行所述目标分析任务的模型的第一网元，N为正整数；
所述第三发送模块，还用于向所述N个第一网元中的目标网元发送第一请求消息；
所述第三接收模块，还用于从所述目标网元接收第一请求响应消息，所述第一请求响应消息包括目标模型和用于获取所述目标模型的地址信息中的至少一项，所述目标模型可用于执行所述目标分析任务。
42、根据上述41所述的装置，所述第一请求消息包括所述目标分析任务的分析任务标识和模型限定信息，所述模型限定信息包括以下至少一项：
模型的数量的限定信息；
模型标识的限定信息；
模型性能信息；
模型大小的限定信息，所述模型大小用于指示存储或运行模型需要的存储空间；
推理时长的限定信息，所述推理时长用于指示基于模型进行模型推理操作所需的时长；
训练数据来源的限定信息，所述训练数据来源信息用于指示模型在训练阶段所使用的训练数据来源的位置信息和网元信息中的至少一项；
训练数据时间的限定信息，所述训练数据时间信息用于指示模型在训练阶段所使用的训练数据的产生的时间。
43、根据上述42所述的装置，所述模型性能信息包括以下至少一项：
第一性能信息，所述第一性能信息用于指示模型在训练阶段能达到的性能；
第二性能信息，所述第二性能信息用于指示模型在推理阶段能达到的性能。
44、根据上述43所述的装置，所述第一性能信息包括模型在训练时的以下至少一项信息：第一性能指标、第一性能指标的计算方法、第一时间信息、第一数值和第一结果；其中，所述第一性能指标包括训练时准确度和训练时误差值中的至少一项，所述第一时间信息包括计算所述第一性能指标对应的时间信息，所述第一数值用于表示计算所述第一性能指标所用的数据数量，所述第一结果为基于多个所述第一性能指标计算获得的结果值。
45、根据上述43所述的装置，所述第二性能信息包括模型在实际使用时的以下至少一项信息：第二性能指标、第二性能指标的计算方法、第二时间信息、第二数值和第二结果；其中，所述第二性能指标包括实际使用时准确度和实际使用时误差值中的至少一项，所述第二时间信息包括计算所述第二性能指标对应的时间信息，所述第二 数值用于表示计算所述第二性能指标所用的数据数量，所述第二结果为基于多个所述第二性能指标计算获得的结果值。
46、根据上述41所述的装置，所述第二请求消息还包括目标要求信息，所述N个第一网元中的任一个第一网元满足所述目标要求信息，所述目标要求信息包括以下至少一项：
模型的数量的要求信息；
模型标识的要求信息；
模型性能信息的要求信息；
模型大小的要求信息，所述模型大小用于指示存储或运行模型需要的存储空间；
推理时长的要求信息，所述推理时长用于指示基于模型进行模型推理操作所需的时长；
训练数据来源的要求信息，所述训练数据来源信息用于指示模型在训练阶段所使用的训练数据来源的位置信息和网元信息中的至少一项；
训练数据时间的要求信息，所述训练数据时间信息用于指示模型在训练阶段所使用的训练数据的产生的时间。
47、根据上述46所述的装置，所述第三网元向第二网元发送第二请求消息之前，所述方法还包括：
所述第三网元从第四网元接收任务请求消息，所述任务请求消息包括目标分析任务的分析任务标识。
48、根据上述47所述的装置，所述目标模型的数量为M1个，M1为正整数，所述第三网元利用所述目标模型执行所述目标分析任务，获得目标分析报告包括：
所述第三网元利用M2个所述目标模型执行针对所述目标分析任务的模型推理，获得M2个推理结果，M2为小于或等于M1的正整数；
所述第三网元基于所述M2个推理结果，生成所述目标分析报告。
49、根据上述40所述的装置，所述目标网元为一个所述第一网元。
50、根据上述40所述的装置，所述第二请求响应消息包括所述N个第一网元的标识信息和所述N个第一网元的地址信息中的至少一项。
51.根据上述40所述的装置，所述第二请求响应消息还包括所述N个第一网元中每一第一网元对应的有效时间。
52.一种网络侧设备，包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如上述1至26任一项所述的模型处理方法的步骤。
53.一种可读存储介质，所述可读存储介质上存储程序或指令，所述程序或指令被处理器执行时实现如上述1至26任一项所述的模型处理方法的步骤。
需要说明的是，在本文中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者装置不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者装置所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、方法、物品或者装置中还存在另外的相同要素。此外，需要指出的是，本申请实施方式中的方法和装置的范围不限按示出或讨论的顺序来执行功能，还可包括根据所涉及的功能按基本同时的方式或按相反的顺序来执行功能，例如，可以按不同于所描述的次序来执行所描述的方法，并且还可以添加、省去、或组合各种步骤。另外，参照某些示例所描述的特征可在其他示例中被组合。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。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可以以计算机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(如ROM/RAM、磁碟、光盘)中，包括若干指令用以使得一台终端(可以是手机，计算机，服务器，空调器，或者网络设备等)执行本申请各个实施例所述的方法。
上面结合附图对本申请的实施例进行了描述，但是本申请并不局限于上述的具体实施方式，上述的具体实施方式仅仅是示意性的，而不是限制性的，本领域的普通技术人员在本申请的启示下，在不脱离本申请宗旨和权利要求所保护的范围情况下，还可做出很多形式，均属于本申请的保护之内。

Claims (33)

  1. 一种模型获取方法，包括：
    第一通信设备向第二通信设备发送第一请求，所述第一请求用于请求获取至少一个第三通信设备的信息，其中，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息；
    所述第一通信设备接收所述第二通信设备发送的至少一个第三通信设备的信息；
    所述第一通信设备基于所述至少一个第三通信设备的信息，获取多个模型的模型信息；所述多个模型用于生成数据分析结果信息。
  2. 根据权利要求1所述的模型获取方法，所述方法还包括:
    第一通信设备基于所述多个模型的模型信息获取所述多个模型；
    所述第一通信设备基于所述多个模型进行分析处理，得到所述数据分析结果信息。
  3. 根据权利要求2所述的模型获取方法，其中，所述第一通信设备基于所述多个模型进行分析处理，得到数据分析结果信息，包括：
    所述第一通信设备基于所述多个模型中的每个模型进行模型推理操作，获取多个数据推理结果；
    所述第一通信设备对所述多个数据推理结果进行处理，得到所述数据分析结果信息。
  4. 根据权利要求3所述的模型获取方法，其中，所述第一通信设备对所述多个数据推理结果进行处理，得到所述数据分析结果信息，包括以下至少一项：
    所述第一通信设备对所述多个数据推理结果进行加权平均，得到所述数据分析结果信息；
    所述第一通信设备对所述多个推理结果数据进行平均运算，得到所述分析结果信息；
    所述第一通信设备对所述多个推理结果数据进行累加，得到所述分析结果信息；
    所述第一通信设备对所述多个推理结果数据进行考虑性能的累加，得到所述分析结果信息。
  5. 根据权利要求1-4任一项所述的模型获取方法，其中，所述第一通信设备向第二通信设备发送第一请求之前，还包括：
    所述第一通信设备接收第四通信设备发送的任务请求；
    所述第一通信设备向第二通信设备发送第一请求，包括：
    所述第一通信设备基于所述任务请求，向所述第二通信设备发送所述第一请求；
    所述第一通信设备基于所述多个模型进行分析处理，得到数据分析结果信息之后，还包括：
    所述第一通信设备向所述第四通信设备发送所述数据分析结果信息。
  6. 根据权利要求1-4任一项所述的模型获取方法，其中，所述第一通信设备基于所述至少一个第三通信设备的信息，获取多个模型的模型信息，包括：
    所述第一通信设备基于所述至少一个第三通信设备的信息以及模型需满足的模型属性信息，确定至少一个目标通信设备，并从所述至少一个目标通信设备获取所述多个模型的模型信息，其中，每个所述目标通信设备能够提供与所述模型属性信息匹配的模型的模型信息。
  7. 根据权利要求1-4任一项所述的模型获取方法，其中，所述第一请求包括：
    分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；
    所述第一请求还包括以下至少一项：
    第一指示信息，所述第一指示信息用于指示请求获取多个模型；
    所需模型数量；
    第二指示信息，所述第二指示信息用于指示请求获取多个第三通信设备的信息；
    所需第三通信设备的数量；
    排序方式，所述排序方式用于指示获取到的多个第三通信设备的信息的排序方式；
    模型需满足的模型属性信息。
  8. 根据权利要求1-3任一项所述的模型获取方法，其中，
    每个第三通信设备的信息包括以下至少一项：
    所述第三通信设备支持的分析任务标识、所述第三通信设备的标识、所述第三通信设备的地址、所述第三通信设备支持的分析任务标识对应的模型数量、所述第三通信设备支持的至少一个模型的模型属性信息。
  9. 根据权利要求6所述的模型获取方法，其中，所述从所述至少一个目标通信设备获取所述多个模型的模型信息，包括：
    所述第一通信设备向所述至少一个目标通信设备发送第二请求，所述第二请求用于向所述至少一个目标通信设备获取所述多个模型的模型信息；
    所述第一通信设备接收所述至少一个目标通信设备发送的所述多个模型的模型信息；
    所述第二请求包括以下至少一项：
    分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；
    所需模型的标识；
    所需模型数量；
    模型需满足的模型属性信息。
  10. 根据权利要求9所述的模型获取方法，其中，所述第一通信设备向所述至少一个目标通信设备发送第二请求，包括：
    所述第一通信设备向所述至少一个目标通信设备中的一个目标通信设备发送所述第二请求；
    所述第一通信设备接收所述至少一个目标通信设备发送的所述多个模型的模型信息，包括：
    所述第一通信设备接收所述一个目标通信设备发送的所述多个模型的模型信息；
    其中，在所述第一通信设备向所述至少一个目标通信设备中的一个目标通信设备发送所述第二请求之前，所述方法还包括：
    所述第一通信设备确定所述一个目标通信设备能够提供所述多个模型的模型信息。
  11. 根据权利要求9所述的模型获取方法，其中，所述第一通信设备向所述至少一个目标通信设备发送第二请求，包括：
    所述第一通信设备向所述至少一个目标通信设备中的多个目标通信设备发送所述第二请求；
    所述第一通信设备接收所述至少一个目标通信设备发送的所述多个模型的模型信息，包括：
    所述第一通信设备接收所述多个目标通信设备发送的所述多个模型的模型信息；
    其中，在所述第一通信设备向所述至少一个目标通信设备中的多个目标通信设备发送所述第二请求之前，所述方法还包括：
    所述第一通信设备确定所述至少一个目标通信设备中每个所述目标通信设备均不能提供所述多个模型的全部模型的模型信息。
  12. 根据权利要求6-9任一项所述的模型获取方法，其中，所述模型属性信息包括以下至少一项：模型使用的范围信息、模型标识、模型训练的结果评价信息、模型使用的结果评价信息、模型大小、模型推理时长、训练数据的来源信息、训练数据的时间信息；模型大小用于表示模型存储或运行所需的存储空间大小，模型推理时长用于表示模型运行得到数据推理结果的时长。
  13. 根据权利要求12所述的模型获取方法，其中，
    所述模型训练的结果评价信息包括以下至少一项：准确率、误差信息、模型训练时长、模型训练的数据量；所述模型使用的结果评价信息包括以下至少一项：准确率、误差信息，模型运行时长。
  14. 一种模型获取方法，包括：
    第二通信设备接收第一通信设备发送的第一请求，所述第一请求用于请求获取至少一个第三通信设备的信息，其中，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息；
    所述第二通信设备基于所述第一请求，确定至少一个第三通信设备；
    所述第二通信设备向所述第一通信设备发送所述至少一个第三通信设备的信息；所述至少一个第三通信设备的信息用于所述第一通信设备获取多个模型的模型信息；所述多个模型用于生成数据分析结果信息。
  15. 根据权利要求14所述的模型获取方法，其中，所述第一请求包括：分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；所述第一请求还包括以下至少一项：
    第一指示信息，所述第一指示信息用于指示请求获取多个模型；
    所需模型数量；
    第二指示信息，所述第二指示信息用于指示请求获取多个第三通信设备的信息；
    所需第三通信设备的数量；
    排序方式，所述排序方式用于指示获取到的多个第三通信设备的信息的排序方式；
    模型需满足的模型属性信息。
  16. 根据权利要求15所述的模型获取方法，所述方法还包括：
    所述第二通信设备基于所述第一通信设备发送的第一请求，确定所述至少一个第三通信设备的数量。
  17. 根据权利要求16所述的模型获取方法，其中，所述第二通信设备基于所述第一通信设备发送的第一请求，确定所述至少一个第三通信设备的数量，包括以下至少一项：
    所述第二通信设备基于所述第一指示信息和/或所述所需模型数量，确定所述至少一个第三通信设备的数量为一个，一个所述第三通信设备能够提供多个模型或所述所需模型数量个模型的模型信息；
    所述第二通信设备基于所述第一指示信息和/或所述所需模型数量，确定所述至少一个第三通信设备的数量多个，多个所述第三通信设备中每个第三通信设备均不能提供所述多个模型的全部模型或所述所需模型数量个模型的模型信息；
    所述第二通信设备基于所述第二指示信息，确定所述至少一个第三通信设备的数量为多个；
    所述第二通信设备基于所述所需第三通信设备的数量，确定所述至少一个第三通信设备的数量为所需第三通信设备的数量。
  18. 根据权利要求15所述的模型获取方法，其中，
    所述模型属性信息包括以下至少一项：模型使用的范围信息、模型标识、模型训练的结果评价信息、模型使用的结果评价信息、模型大小、模型执行时长、训练数据的来源信息、训练数据的时间信息；模型大小用于表示模型存储或运行所需的存储空间大小，模型推理时长用于表示模型运行得到数据推理结果的时长。
  19. 根据权利要求18所述的模型获取方法，其中，
    所述模型训练的结果评价信息包括以下至少一项：准确率、误差信息、模型训练时长、模型训练的数据量；所述模型使用的结果评价信息包括以下至少一项：准确率、误差信息，模型运行时长。
  20. 根据权利要求14-19任一项所述的模型获取方法，其中，
    每个第三通信设备的信息包括以下至少一项：
    所述第三通信设备支持的分析任务标识、所述第三通信设备的标识、所述第三通信设备的地址、所述第三通信设备支持的分析任务标识对应的模型数量或所述第三通信设备支持的至少一个模型的模型属性信息。
  21. 根据权利要求14-19任一项所述的模型获取方法，所述方法还包括：
    所述第二通信设备接收第三通信设备发送的能力注册消息，所述能力注册消息包括以下至少一项：所述第三通信设备支持的分析任务标识、所述第三通信设备的网络功能类型、所述第三通信设备的网络功能实例标识、所述第三通信设备支持的分析任务标识对应的模型数量、所述第三通信设备支持的至少一个模型的模型属性信息。
  22. 一种模型获取方法，包括：
    第三通信设备接收第一通信设备发送的第二请求，所述第二请求用于请求获取至少一个模型；
    所述第三通信设备基于所述第二请求向所述第一通信设备发送至少一个模型的模型信息；所述第二请求包括以下至少一项：
    分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；
    所需模型的标识；
    所需模型数量；
    模型需满足的模型属性信息。
  23. 根据权利要求22所述的模型获取方法，所述方法还包括：
    所述第三通信设备向第二通信设备发送能力注册消息，所述能力注册消息包括以下至少一项：所述第三通信设备支持的分析任务标识、所述第三通信设备的网络功能类型、所述第三通信设备的网络功能实例标识、所述第三通信设备支持的分析任务标识对应的模型数量、所述第三通信设备支持的至少一个模型的模型属性信息。
  24. 一种模型获取方法，包括：
    第四通信设备向第一通信设备发送任务请求；
    所述第四通信设备接收所述第一通信设备发送的数据分析结果信息，所述数据分析结果信息为所述第一通信设备基于多个模型进行分析处理得到的，所述多个模型为所述第一通信设备基于至少一个第三通信设备的信息获取到的，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息。
  25. 一种模型获取装置，包括：
    发送模块，用于向第二通信设备发送第一请求，所述第一请求用于请求获取至少一个第三通信设备的信息，其中，每个所述第三通信设备能够提供第一通信设备所需的至少一个模型的模型信息；
    接收模块，用于接收所述第二通信设备发送的至少一个第三通信设备的信息；
    获取模块，用于基于所述至少一个第三通信设备的信息，获取多个模型的模型信息；所述多个模型用于生成数据分析结果信息。
  26. 一种模型获取装置，包括：
    接收模块，用于接收第一通信设备发送的第一请求，所述第一请求用于请求获取至少一个第三通信设备的信息，其中，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息；
    处理模块，用于基于所述第一请求，确定至少一个第三通信设备；
    发送模块，用于向所述第一通信设备发送所述至少一个第三通信设备的信息；所述至少一个第三通信设备的信息用于所述第一通信设备获取多个模型的模型信息；所述多个模型用于生成数据分析结果信息。
  27. 一种模型获取装置，包括：
    接收模块，用于接收第一通信设备发送的第二请求，所述第二请求用于请求获取至少一个模型；
    发送模块，用于基于所述第二请求向所述第一通信设备发送至少一个模型的模型信息；所述第二请求包括以下至少一项：
    分析任务标识，所述分析任务标识用于标识所需模型适用的数据分析任务；
    所需模型的标识；
    所需模型数量；
    模型需满足的模型属性信息。
  28. 一种模型获取装置，包括：
    发送模块，用于向第一通信设备发送任务请求；
    接收模块，用于接收所述第一通信设备发送的数据分析结果信息，所述数据分析结果信息为所述第一通信设备基于多个模型进行分析处理得到的，所述多个模型为所述第一通信设备基于至少一个第三通信设备的信息获取到的，每个所述第三通信设备能够提供所述第一通信设备所需的至少一个模型的模型信息。
  29. 一种第一通信设备，包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如权利要求1至13任一项所述的模型获取方法的步骤。
  30. 一种第二通信设备，包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如权利要求14至21任一项所述的模型获取方法的步骤。
  31. 一种第三通信设备，包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如权利要求22或23任一项所述的模型获取方法的步骤。
  32. 一种第四通信设备，包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如权利要求24所述的模型获取方法的步骤。
  33. 一种可读存储介质，所述可读存储介质上存储程序或指令，所述程序或指令被处理器执行时实现如权利要求1至13任一项所述的模型获取方法，或者实现如权利要求14至21任一项所述的模型获取方法，或者实现如权利要求22至23任一项所述的模型获取方法，或者实现如权利要求24所述的模型获取方法的步骤。
PCT/CN2023/092194 2022-05-05 2023-05-05 模型获取方法及通信设备 WO2023213288A1 (zh)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN202210482186.3A CN117062047A (zh) 2022-05-05 2022-05-05 模型处理方法、装置、网络侧设备及可读存储介质
CN202210483893.4 2022-05-05
CN202210483893.4A CN117062098A (zh) 2022-05-05 2022-05-05 模型获取方法及通信设备
CN202210482186.3 2022-05-05

Publications (1)

Publication Number Publication Date
WO2023213288A1 true WO2023213288A1 (zh) 2023-11-09

Family

ID=88646285

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/092194 WO2023213288A1 (zh) 2022-05-05 2023-05-05 模型获取方法及通信设备

Country Status (1)

Country Link
WO (1) WO2023213288A1 (zh)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021062740A1 (zh) * 2019-09-30 2021-04-08 华为技术有限公司 一种确定设备信息的方法、装置以及系统
WO2022062362A1 (zh) * 2020-09-25 2022-03-31 华为技术有限公司 通信方法、装置及系统
US20220108214A1 (en) * 2020-08-13 2022-04-07 Electronics And Telecommunications Research Institute Management method of machine learning model for network data analytics function device

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021062740A1 (zh) * 2019-09-30 2021-04-08 华为技术有限公司 一种确定设备信息的方法、装置以及系统
US20220108214A1 (en) * 2020-08-13 2022-04-07 Electronics And Telecommunications Research Institute Management method of machine learning model for network data analytics function device
WO2022062362A1 (zh) * 2020-09-25 2022-03-31 华为技术有限公司 通信方法、装置及系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
HUAWEI, HISILICON: "NWDAF discovery and selection based on ML model information", 3GPP DRAFT; S2-2102473, 3RD GENERATION PARTNERSHIP PROJECT (3GPP), MOBILE COMPETENCE CENTRE ; 650, ROUTE DES LUCIOLES ; F-06921 SOPHIA-ANTIPOLIS CEDEX ; FRANCE, vol. SA WG2, no. e-meeting; 20210412 - 20210416, 6 April 2021 (2021-04-06), Mobile Competence Centre ; 650, route des Lucioles ; F-06921 Sophia-Antipolis Cedex ; France , XP051993849 *

Similar Documents

Publication Publication Date Title
WO2021032498A1 (en) Methods, apparatus and machine-readable media relating to machine-learning in a communication network
WO2019195690A1 (en) Mechanisms for service layer resource ranking and enhanced resource discovery
WO2023207980A1 (zh) 模型信息获取方法、发送方法、装置、节点和储存介质
WO2023213288A1 (zh) 模型获取方法及通信设备
WO2023213246A1 (zh) 模型选择方法、装置及网络侧设备
WO2024061254A1 (zh) 数据获取方法、装置、系统和设备
WO2023179709A1 (zh) 信息处理方法、装置、通信设备及可读存储介质
CN117062098A (zh) 模型获取方法及通信设备
WO2024153013A1 (zh) 信息传输方法、装置及通信设备
WO2024146097A1 (zh) 网元注册方法、模型请求方法、装置、网元、通信系统及存储介质
WO2023169425A1 (zh) 通信网络中的数据处理方法及网络侧设备
WO2024208348A1 (zh) 信息处理方法、装置及网元
WO2023185788A1 (zh) 候选成员的确定方法、装置及设备
WO2024099187A1 (zh) Ai模型策略确定方法、装置、第一设备及第二设备
WO2023169402A1 (zh) 模型的准确度确定方法、装置及网络侧设备
WO2023185818A1 (zh) 设备确定方法、装置及通信设备
WO2023185822A1 (zh) 网元注册方法、模型确定方法、装置、网元、通信系统及存储介质
WO2023216960A1 (zh) 数据处理方法、装置、核心网节点、电子设备和存储介质
WO2024061252A1 (zh) 信息获取方法、装置、网络侧设备及存储介质
WO2023179558A1 (zh) 数据采集方法及通信设备
WO2024032588A1 (zh) 用户意愿验证方法、装置及网络侧设备
WO2023185725A1 (zh) 数据获取方法、装置及网络侧设备
WO2023185726A1 (zh) 模型获取方法、信息发送方法、信息接收方法、装置及网元
WO2024078400A1 (zh) 模型请求方法、装置、通信设备及可读存储介质
CN116776988A (zh) 通信网络中的数据处理方法及网络侧设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23799274

Country of ref document: EP

Kind code of ref document: A1