WO2022206567A1 - 管控模型训练的方法及装置、系统 - Google Patents

管控模型训练的方法及装置、系统 Download PDF

Info

Publication number
WO2022206567A1
WO2022206567A1 PCT/CN2022/082872 CN2022082872W WO2022206567A1 WO 2022206567 A1 WO2022206567 A1 WO 2022206567A1 CN 2022082872 W CN2022082872 W CN 2022082872W WO 2022206567 A1 WO2022206567 A1 WO 2022206567A1
Authority
WO
WIPO (PCT)
Prior art keywords
model
information
network
training
model training
Prior art date
Legal status (The legal status 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 status listed.)
Ceased
Application number
PCT/CN2022/082872
Other languages
English (en)
French (fr)
Inventor
许瑞岳
邹兰
张凯
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Priority to EP22778750.4A priority Critical patent/EP4311171A4/en
Publication of WO2022206567A1 publication Critical patent/WO2022206567A1/zh
Priority to US18/478,640 priority patent/US12506663B2/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/085—Retrieval of network configuration; Tracking network configuration history
    • H04L41/0853—Retrieval of network configuration; Tracking network configuration history by actively collecting configuration information or by backing up configuration information
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/16—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using machine learning or artificial intelligence

Definitions

  • the present application relates to the technical field of network management, and in particular, to a method, device, and system for managing and controlling model training.
  • AI artificial intelligence
  • ML machine learning
  • the intelligence of the management unit and the network element depends on the AI/ML model.
  • the management unit and the network element also introduce the network AI model training function to train the network AI model locally.
  • the network AI model is trained locally, operators need to manage and control the training of the network AI model. Therefore, the operator needs to inform the operation and maintenance personnel of the equipment manufacturer of the corresponding training management and control requirements, and the operation and maintenance personnel of the equipment manufacturer need to control the training of the network AI model locally through the local manual interface. Since the operator does not support the online real-time control of the local network AI model training function, the labor efficiency is low and labor-intensive.
  • the present application discloses a method, device and system for managing and controlling model training, which can realize flexible management and control of network AI model training functions.
  • an embodiment of the present application provides a method for managing and controlling model training, including: receiving configuration information from a first network management unit, the configuration information configures a model training function, and the configuration information includes at least one of the following information: status information, Used to activate or deactivate the model training function; trigger information, used to trigger model training; data information, used to indicate the data for model training; model training according to the configuration information to obtain network model information.
  • the second network management unit configures the model training function based on the received configuration information sent by the first network management unit, so as to perform model training according to the configuration information to obtain network model information.
  • operators can flexibly configure the corresponding network AI model training function on demand, which can realize the management and control of network model training, improve the efficiency of management and control model training, and save manpower.
  • the embodiment of the present application also provides a method for managing and controlling model training, including: receiving configuration information from a first network management unit, the configuration information configures a model training function, and the configuration information includes at least one of the following information: status information, used to activate Or deactivate the model training function; trigger information, used to trigger model training; data information, used to indicate data for model training; send configuration information to the network element, so that the network element can perform model training according to the configuration information and obtain network model information .
  • the second network management unit sends the configuration information to the network element based on the received configuration information sent by the first network management unit, so that the network element can perform model training according to the configuration information.
  • network model training can be managed and controlled, enabling operators to flexibly configure corresponding network AI model training functions as needed.
  • the configuration information further includes at least one of the following: data source information, used to indicate a data source for model training, where the data source is an entity that provides training data, which may be a network element list, management functional unit list, database list, etc.
  • the model information associated with the model training function is used to indicate the model for model training, for example, the associated network AI model information, indicating the output network AI model object, such as the identifier of the network AI model object, etc.
  • Model type information which is used to indicate the model type for model training.
  • the model type information includes at least one of the following: load information analysis model, service experience analysis model, network performance analysis model, congestion analysis model, service quality (Quality of Service, QoS) analysis model, energy saving analysis model, traffic flow analysis model, Massive MIMO analysis model, User Equipment (UE) trajectory analysis model.
  • load information analysis model service experience analysis model
  • network performance analysis model congestion analysis model
  • service quality (Quality of Service, QoS) analysis model energy saving analysis model
  • traffic flow analysis model Massive MIMO analysis model
  • UE User Equipment
  • the above data source information enables operators to configure network AI model training data sources on demand.
  • the training function of the network model can be trained according to the data of the data source specified by the operator; at the same time, compared with the method of local control, the use of this solution can improve the training efficiency of the management and control model, while saving the cost of manpower.
  • the model information associated with the above model training function enables operators to configure network AI models for model training on demand.
  • the training function of the network model can be trained according to the model information specified by the operator; at the same time, compared with the method of local control, the use of this solution can improve the training efficiency of the management and control model and save manpower.
  • the type information of the above models enables operators to flexibly configure the types supported by the network AI model training function.
  • the training function of the network model can be trained according to the model type specified by the operator; at the same time, compared with the method of local control, the use of this solution can improve the training efficiency of the control model and save manpower.
  • the network model information includes at least one of the following: a network model identifier, a network model version, a network model file storage address, and a network model file name.
  • the above-mentioned network model identifier is, for example, the network model 1I.
  • the above-mentioned network model version may be, for example, v1.0.0.
  • the above-mentioned network model file storage address may be, for example, a Uniform Resource Identifier (Uniform Resource Identifier, URI) or an Internet Protocol (Internet Protocol, IP) address.
  • the above-mentioned network model file name may be, for example, the name of the network model file.
  • the above method further includes: sending network model information to the first network management unit.
  • the first network management unit can acquire the network AI model information output by training in real time.
  • the above method further includes: determining association information between the configuration information and the network model information; and sending the association information to the first network management unit.
  • the correlation information between the above configuration information and the network model information can be understood as: the above network model information is obtained by training based on one or several configuration information, then the one or several configuration information and the network model information have connection relation.
  • the first network management unit can acquire the association information between the configuration information and the network model information in real time.
  • the above trigger information includes at least one of the following: a training period, used to indicate a model training period, for example, indicating that the network AI model training function needs to be trained according to a set period, such as a set period is every hour, every day, every week, etc.; training time, used to indicate the time of model training, for example, to instruct the network AI model training function to train at a certain point in time; training instruction information, used to indicate the start of model training, that is Start network model training now.
  • a training period used to indicate a model training period, for example, indicating that the network AI model training function needs to be trained according to a set period, such as a set period is every hour, every day, every week, etc.
  • training time used to indicate the time of model training, for example, to instruct the network AI model training function to train at a certain point in time
  • training instruction information used to indicate the start of model training, that is Start network model training now.
  • the above status information enables operators to flexibly activate and deactivate the corresponding network AI model training function as needed.
  • the training function of the network model can be trained according to the state information specified by the operator.
  • the above trigger information enables operators to configure network AI model training trigger information as needed.
  • the training function of the network model can be trained according to the trigger information specified by the operator.
  • the above-mentioned data information includes: input data, which is used to indicate the input data type of the model; and output data, which is used to indicate the output data type of the model.
  • input data can be Reference Signal Received Power (RSRP), Signal to Interference plus Noise Ratio (Signal to Interference plus) noise ratio, SINR), uplink and downlink throughput (DownLink/UpLink throughput, DL/UL throughput);
  • output data can be the tilt angle, azimuth angle, etc. under the Massive MIMO pattern coverage scenario.
  • This data information enables operators to configure network AI model training data types as needed.
  • the training function of the network model can be trained according to the data type specified by the operator.
  • an embodiment of the present application provides a method for managing and controlling model training, including: determining configuration information, the configuration information configures a model training function, and the configuration information includes at least one of the following information: status information, used for activation or deactivation Model training function; trigger information for triggering model training; data information for indicating data for model training; sending configuration information to the second network management unit.
  • the first network management unit determines the configuration information of the model training function, and sends the configuration information to the second network management unit.
  • network model training can be managed and controlled, enabling operators to flexibly configure corresponding network AI model training functions as needed.
  • the above configuration information further includes at least one of the following: data source information, used to indicate the data source for model training; model information associated with the model training function, used to indicate the model training Model; the type information of the model, which is used to indicate the model type for model training.
  • the above method further includes: receiving network model information sent by the second network management unit, where the network model information is obtained by the second network management unit after training according to the configuration information;
  • the configuration information is obtained by the network element after training according to the configuration information, and the configuration information is received by the network element from the second network management unit.
  • the network model information includes at least one of the following: a network model identifier, a network model version, a network model file storage address, and a network model file name.
  • the above method further includes: receiving association information sent by the second network management unit, where the association information is association information between configuration information and network model information.
  • the above trigger information includes at least one of the following: training period, used to indicate the model training period; training time, used to indicate the time of model training; training indication information, used to indicate the start of the training Model training.
  • the above-mentioned data information includes: input data, which is used to indicate the input data type of the model; and output data, which is used to indicate the output data type of the model.
  • an embodiment of the present application provides an apparatus for managing and controlling model training, including: a receiving module configured to receive configuration information from a first network management unit, the configuration information configures a model training function, and the configuration information includes at least one of the following information Type: status information, used to activate or deactivate the model training function; trigger information, used to trigger model training; data information, used to indicate the data for model training; processing module, used to perform model training according to the configuration information, get Network model information.
  • the configuration information further includes at least one of the following: data source information, used to indicate the data source for model training; model information associated with the model training function, used to indicate the model for model training; type information of the model , which indicates the type of model for model training.
  • the network model information includes at least one of the following: a network model identifier, a network model version, a network model file storage address, and a network model file name.
  • the above-mentioned apparatus further includes: a sending module, configured to send the network model information to the first network management unit.
  • the above-mentioned apparatus further includes: a determination module configured to determine association information between the configuration information and the network model information, and send the association information to the first network management unit.
  • the trigger information includes at least one of the following: training period, used to indicate the model training period; training time, used to indicate the time of model training; training indication information, used to indicate the start of model training.
  • the data information includes: input data, which is used to indicate the input data type of the model; and output data, which is used to indicate the output data type of the model.
  • an embodiment of the present application provides an apparatus for managing and controlling model training, including: a determination module configured to determine configuration information, the configuration information configures a model training function, and the configuration information includes at least one of the following information: for activating or deactivating the model training function; trigger information for triggering model training; data information for indicating data for model training; sending module for sending configuration information to the second network management unit.
  • the configuration information further includes at least one of the following: data source information, used to indicate a data source for model training; model information associated with the model training function, used to indicate a model for model training ; The type information of the model, which is used to indicate the model type for model training.
  • the above-mentioned apparatus further includes: a receiving module configured to receive network model information sent by the second network management unit, where the network model information is obtained by the second network management unit after training according to the configuration information; or, The network model information is obtained by the network element after training according to the configuration information, and the configuration information is received by the network element from the second network management unit.
  • the network model information includes at least one of the following: a network model identifier, a network model version, a network model file storage address, and a network model file name.
  • the receiving module is further configured to: receive the association information sent by the second network management unit, where the association information is association information between the configuration information and the network model information.
  • the trigger information includes at least one of the following: training period, used to indicate the model training period; training time, used to indicate the time of model training; training indication information, used to indicate the start of model training train.
  • the data information includes: input data, which is used to indicate the input data type of the model; and output data, which is used to indicate the output data type of the model.
  • the present application provides a computer storage medium, including computer instructions, which, when the computer instructions are executed on an electronic device, cause the electronic device to execute any one of the possible implementations and/or implementations of the first aspect.
  • the method provided by any possible implementation manner of the second aspect.
  • the embodiments of the present application provide a computer program product, which, when the computer program product runs on a computer, enables the computer to execute any possible implementation manner of the first aspect and/or any possible implementation manner of the second aspect.
  • the method provided by the embodiment is not limited to:
  • an embodiment of the present application provides an apparatus for managing and controlling model training, including a processor and a memory; wherein the memory is used to store program codes, and the processor is used to call the program codes to execute the first A method provided by any possible embodiment of the aspect and/or any possible embodiment of the second aspect.
  • an embodiment of the present application provides a system for managing and controlling model training, including the apparatus provided in any possible implementation manner of the third aspect and/or any possible implementation manner of the fourth aspect.
  • an embodiment of the present application provides a method for managing and controlling model training, including: a first network management unit determining configuration information, the configuration information configuring a model training function, and the configuration information including at least one of the following information: status information , used to activate or deactivate the model training function; trigger information, used to trigger model training; data information, used to indicate the data for model training; the first network management unit sends the configuration information; the second network management unit receives the configuration information; the second network management unit performs model training according to the configuration information to obtain network model information.
  • the device according to the third aspect, the device according to the fourth aspect, the computer storage medium according to the fifth aspect, the computer program product according to the sixth aspect, and the device according to the seventh aspect are provided above.
  • the systems described in the eighth aspect are all configured to perform any one of the methods provided in the first aspect and any one of the methods provided in the second aspect. Therefore, for the beneficial effects that can be achieved, reference may be made to the beneficial effects in the corresponding method, which will not be repeated here.
  • a chip is provided, the chip is coupled with a memory, and the method for implementing the management and control model training according to any one of the first aspect or the first aspect of the embodiments of the present application is implemented.
  • a chip is provided, the chip is coupled with a memory, and the method for implementing the management and control model training according to any one of the second aspect or the second aspect of the embodiments of the present application is implemented.
  • a twelfth aspect provides a chip, the chip is coupled with a memory, and performs the method for training a management and control model described in the ninth aspect of the embodiments of the present application.
  • Coupled in the embodiments of the present application means that two components are directly or indirectly combined with each other.
  • FIG. 1 is a schematic diagram of the architecture of a system for training a management and control model provided by an embodiment of the present application
  • FIG. 2 is a schematic diagram of the architecture of another system for training a management and control model provided by an embodiment of the present application;
  • FIG. 3 is a schematic diagram of the architecture of another system for training a management and control model provided by an embodiment of the present application
  • FIG. 4 is a schematic flowchart of a method for training a management and control model provided by an embodiment of the present application
  • FIG. 5 is a schematic flowchart of another method for training a control model provided by an embodiment of the present application.
  • FIG. 6 is a schematic flowchart of another method for training a control model provided by an embodiment of the present application.
  • FIG. 7 is a schematic flowchart of another method for managing and controlling model training provided by an embodiment of the present application.
  • FIG. 8 is a schematic structural diagram of an apparatus for training a control model provided by an embodiment of the present application.
  • FIG. 9 is a schematic structural diagram of another apparatus for managing and controlling model training provided by an embodiment of the present application.
  • FIG. 10 is a schematic structural diagram of another apparatus for managing and controlling model training provided by an embodiment of the present application.
  • the AI/ML model in the embodiments of the present application may be a mathematical model that generates predictions by finding patterns in data in machine learning.
  • network AI/ML models mainly generate data models that predict network performance by finding patterns in network data, such as network traffic prediction models.
  • the embodiments of the present application do not specifically limit the types of models.
  • the model training in the embodiment of the present application can be understood as: in order to achieve the goal of high recognition rate, a large amount of data is used to find the target configuration parameters and determine the process of the target model.
  • network model training mainly describes the relationship between network data, such as configuration parameter values in different network environments.
  • the system for training the management and control model includes a first network management unit 101 (there may be more, not shown in the figure), and a second network management unit 102 (there may be more, not shown in the figure) , and a network node set 103 managed by the second network management unit 102 (it is shown in the figure that the network node set 103 includes n network nodes).
  • the function of the first network management unit 101 may be deployed on an independent device/device, or may be deployed on a device/device with other functions; the device/device where the function of the first network management unit 101 is deployed is referred to as the first A network management device/first network management device; for the convenience of description, the first network management unit, the first network management device or the first network management device in the embodiments of the present application are uniformly referred to as the first network management unit.
  • the function of the second network management unit 102 can be deployed on an independent device/apparatus, or can be deployed on a device/apparatus with other functions; the device/apparatus on which the function of the second network management unit 102 is deployed is called
  • the second network management device/second network management device, the second network management unit, the second network management device or the second network management device are collectively referred to as the second network management unit.
  • the first network management unit may be a network management system (network management system, NMS), a cross-domain management function (Cross Domain management function, Cross-Domain MnF), which may also be referred to as network management Functional unit (network management function, NMF), or business support system (business support system, BSS).
  • the second network management unit may be a network element management system (element management system, EMS), or a domain management function (Domain management function, Domain MnF), also can be called a subnetwork management function (subnetwork management function, NMF) or Network element/function management function.
  • EMS network element management system
  • Domain MnF domain management function
  • NMF subnetwork management function
  • the network element may be a core network network element or a wireless network network element;
  • the core network network element includes but is not limited to: a mobile switching center (mobile switching center, MSC), a gateway mobile switching center (GMSC) ), GPRS (general packet radio service, general packet radio service) service support node (serving GPRS support node, SGSN), gateway GPRS support node (gateway GPRS support node, GGSN), mobility management entity (mobility management entity, MME) , serving gateway (serving gateway, SGW), packet gateway (packet gateway, PGW), access management function (access management function, AMF) equipment, user plane function (user plane function, UPF) equipment, session management function (session management) function, SMF) equipment;
  • wireless network elements include but are not limited to base stations and base station controllers, and base stations may be: global system for mobile communications (GSM) base stations, universal mobile telecommunications system (UMTS) ) base station, long term evolution (long term evolution, LTE) base station, new
  • the first network management unit determines configuration information and sends the configuration information to the second network management unit, and the second network management unit performs model training according to the configuration information to obtain network model information.
  • the second network management unit sends the configuration information to a network element, so that the network element performs model training according to the configuration information to obtain the network model information.
  • the configuration information is used to configure the model training function, and the configuration information includes at least one of the following information: status information, used to activate or deactivate the model training function; trigger information, used to trigger model training; data information , which indicates the data for model training.
  • FIG. 2 a schematic diagram of the architecture of another management and control model training system provided by an embodiment of the present application, wherein the cross-domain management functional unit 201 (eg, an operator-level network management unit) is the first network in FIG. 1 .
  • the management unit, the domain management function unit 202 (for example, a network management unit at the level of a subordinate unit of an operator) is the second network management unit in FIG. 1 .
  • the cross-domain management function unit 201 manages the domain management function units 202 (there may be multiple ones, only one domain management function unit is shown in the figure).
  • the domain management functional unit 202 manages the network element 203 .
  • the cross-domain management functional unit 201 determines the configuration information, and sends the configuration information to the domain management functional unit 202, and the domain management functional unit 202 performs model training according to the configuration information to obtain network model information; or, domain management The functional unit 202 sends the configuration information to the network element 203, so that the network element performs model training according to the configuration information.
  • FIG. 3 it is a schematic diagram of the architecture of another system for training a management and control model provided by an embodiment of the present application.
  • the management and control model training system includes a service operation unit 301 , a cross-domain management function unit 302 , a domain management function unit 303 and a network element 304 .
  • the cross-domain management function unit 302 is the first network management unit in FIG. 1 .
  • the management function unit 303 is the second network management unit in FIG. 1 .
  • the cross-domain management function unit 302 can manage one or more domain management function units 303 (one domain management function unit 303 is shown in the figure), and the domain management function unit 303 manages the network elements 304 connected to it.
  • the business operation unit 301 also known as the communication service management function unit (communication service management function), can provide billing, settlement, accounting, customer service, business, network monitoring, communication service life cycle management, business intent translation, etc. functionality and management services.
  • the above-mentioned business operation unit may include an operator's operation system or an operation system of a vertical industry (vertical operational technology system).
  • the service operation unit 301 can determine the configuration information, and deliver it to the cross-domain management function unit 302, and then the cross-domain management function unit 302 forwards it to the domain management function unit 303; or the user can input the configuration information through the service operation unit 301, and then the service The operation unit 301 delivers to the cross-domain management functional unit 302 , and the cross-domain management functional unit 302 forwards it to the domain management functional unit 303 .
  • the cross-domain management function unit 302 provides one or more of the following functions or management services: network life cycle management, network deployment, network fault management, network performance management, network configuration management, network assurance, and network optimization functions, translation of network intent from communication service provider (intent-CSP), translation of network intent from communication service consumer (intent-CSC), training of network AI models and Inference of network AI models, etc.
  • the network here may include one or more network elements, sub-networks or network slices.
  • the cross-domain management function unit 302 may be a network slice management function (NSMF), or a management data analytical function (MDAF), or a cross-domain self-organization network function (self-organization network function). , SON-function), or a cross-domain intent management functional unit.
  • the cross-domain management functional unit can also provide one or more of the following management functions or management services: lifecycle management of subnets, deployment of subnets, fault management of subnets, Sub-network performance management, sub-network configuration management, sub-network guarantee, sub-network optimization function, sub-network intent translation function, etc.
  • the sub-network may be composed of multiple small sub-networks or multiple network slice sub-networks.
  • an access network sub-network of an operator includes the access network sub-network of the first equipment manufacturer and the access network sub-network of the second equipment manufacturer. Access the network subnet.
  • the domain management functional unit 303 provides one or more of the following functions or management services: life cycle management of sub-networks or network elements, deployment of sub-networks or network elements, fault management of sub-networks or network elements, Performance management, assurance of subnets or network elements, optimized management of subnetworks or network elements, intent translation of subnetworks or network elements, training of network AI models, and reasoning of network AI models, etc.
  • the sub-network here includes one or more network elements.
  • the sub-network may also include one or more sub-networks, that is, one or more sub-networks form a sub-network with a larger coverage area.
  • the sub-network here may also include one or more network slice sub-networks. Subnets are described in one of the following ways:
  • a network in a certain technical domain such as a wireless access network, core network, transmission network, etc.
  • a certain standard network such as GSM network, LTE network, 5G network, etc.
  • the network provided by a certain equipment manufacturer such as the network provided by equipment manufacturer X, etc.
  • the network of a certain geographical area such as the network of factory A, the network of prefecture-level city B, etc.
  • a network element (Net Element, NE) 304 is an entity that provides network services, including core network network elements, access network network elements, and the like.
  • the network element NE in this solution can also provide at least one of the two functions of training the network AI model and inference of the network AI model.
  • core network elements may include but are not limited to access and mobility management function (AMF) entities, session management function (session management function, SMF) entities, policy control function (policy control function, PCF) entities Entity, network data analysis function (NWDAF) entity, network repository function (NRF), gateway, etc.
  • AMF access and mobility management function
  • SMF session management function
  • policy control function policy control function
  • PCF policy control function
  • NWDAF network data analysis function
  • NRF network repository function
  • the network elements of the access network may include, but are not limited to: various base stations (for example, a next-generation base station (generation node B, gNB), an evolved base station (evolved Node B, eNB), a centralized control unit (central unit control panel, CUCP), Centralized unit (CU), distributed unit (DU), centralized user plane unit (central unit user panel, CUUP), etc.
  • generation node B generation node B
  • eNB evolved base station
  • evolved Node B evolved Node B
  • eNB evolved base station
  • a centralized control unit central unit control panel, CUCP
  • Centralized unit CU
  • DU distributed unit
  • centralized user plane unit central unit user panel
  • CUUP central unit user panel
  • the net element data analysis function (NEDAF) in this solution can be an independent network element or a logical function in one of the above network elements, which is not limited in this solution. It should be noted that the network element data analysis function in this solution may also be called a network element reasoning function or an intelligent function, and the name is not limited.
  • FIG. 4 it is a schematic flowchart of a method for training a management and control model provided by an embodiment of the present application. As shown in FIG. 4 , the method is applied to the second network management unit, which may include steps 401-402, as follows:
  • the second network management unit receives configuration information from the first network management unit, the configuration information configures the model training function, and the configuration information includes at least one of the following information:
  • the above-mentioned first network management unit may be a cross-domain management functional unit Cross-Domain MnF.
  • the second network management unit may be a domain management functional unit Domain MnF.
  • the cross-domain management function unit Cross-Domain MnF determines the configuration information of the training function of the network AI model.
  • the configuration information of the training function of the network AI model can be manually input into the Cross-Domain MnF, or it can be generated by the internal calculation and analysis of the Cross-Domain MnF.
  • the above model training function may be a training function of a certain model, or may be a training function of multiple models.
  • the above configuration information may include A) status information, which is used to activate or deactivate the model training function. That is to say, the state information is used to describe the state of the training function of the network AI model, and specifically, it may include three states: activated state, deactivated state, and in progress. For example, the operator can activate or deactivate the network AI model training function.
  • the above status information enables operators to flexibly activate and deactivate the corresponding network AI model training function as needed.
  • the above configuration information may include B) trigger information, which is used to trigger model training.
  • the B) trigger information includes at least one of the following:
  • Training period used to indicate the model training period. For example, it is indicated that the AI model training function of the network needs to be trained according to a set period, such as the set period is every hour, every day, every week, etc.
  • Training time used to indicate when the model was trained. For example, instructing the network AI model training function to train at a certain point in time.
  • the training time is used as an example for description. It can also be other conditions, such as using the network AI model to perform model inference or analysis.
  • model reasoning also called intelligent analysis, can be understood as: using models to give analysis results. For example, a set of network configuration parameters is determined according to the current network environment.
  • Training instruction information which is used to instruct to start model training, that is, network model training immediately.
  • the above trigger information enables operators to configure network AI model training trigger information as needed.
  • the above configuration information may include C) data information, which is used to indicate data for model training.
  • the data information includes:
  • Output data which indicates the output data type of the model.
  • the input data can be Reference Signal Received Power (RSRP), Signal to Interference plus Noise Ratio (Signal to Interference plus) noise ratio, SINR), uplink and downlink throughput (DownLink/UpLink throughput, DL/UL throughput); the output data can be the tilt angle, azimuth angle, etc. under the Massive MIMO pattern coverage scenario.
  • RSRP Reference Signal Received Power
  • SINR Signal to Interference plus Noise Ratio
  • SINR Signal to Interference plus Noise Ratio
  • DownLink/UpLink throughput downlink and downlink throughput
  • the output data can be the tilt angle, azimuth angle, etc. under the Massive MIMO pattern coverage scenario.
  • the data information enables the operator to configure the network AI model training data type as needed.
  • the above data information also includes the granularity of the corresponding training data, that is, the training data type of the corresponding granularity.
  • the granularity includes cells, grids, tracking areas, network slices, services, etc.
  • the configuration information also includes at least one of the following:
  • Data source information which is used to indicate the data source for model training.
  • the data source is an entity that provides training data, which can be a list of network elements, a list of management functional units, a list of databases, and so on.
  • the above data source information enables operators to configure network AI model training data sources on demand.
  • Model information associated with the model training function used to indicate the model for model training.
  • the associated network AI model information indicates the output network AI model object, such as the identifier of the network AI model object.
  • the model information associated with the above model training function enables operators to configure network AI models for model training on demand.
  • the type information of the model which is used to indicate the model type for model training.
  • the type information of the model includes at least one of the following: load information analysis model, service experience analysis model, network performance analysis model, congestion analysis model, quality of service (Quality of Service, QoS) analysis model, energy saving analysis model , Traffic flow analysis model, Massive MIMO analysis model, User Equipment (UE) trajectory analysis model.
  • the type information of the above models enables operators to flexibly configure the types supported by the network AI model training function.
  • the model training function may also include G) intelligent analysis function information associated with the model training function, which is used to indicate that the intelligent analysis function of the network AI model can be used.
  • the associated intelligence analysis function information here refers to the information of the intelligence analysis function that can use the network model obtained by training.
  • the second network management unit receives configuration information from the first network management unit, which may be that the second network management unit receives a model training function control object creation request sent by the first network management unit, and the model training function control object creation request is Carrying model training function control object identification and configuration information.
  • the above-mentioned model training function control object is the management and control information of the model training function, as shown in Table 1.
  • Domain MnF writes the received management and control information of the network AI model training function in the corresponding fields.
  • the network AI model training function on Domain MnF trains the network AI model according to the information configured in the control object of the network AI model training function.
  • the domain management function unit configures the received network AI model training function management and control information in the network AI model training function management and control object TrainingFunction
  • the model training management and control object does not exist in the domain management function unit, configure the Before describing the network AI model training function control object, first create the network AI model training function control object TrainingFunction, and then configure the configuration information in the model training control object; if the model training control object exists in the domain management function unit, then The configuration information is configured in the existing model training control object.
  • the second network management unit performs model training according to the configuration information to obtain network model information.
  • step 402 may include steps 4021-4022, which are as follows:
  • the second network management unit configures the model training function according to the configuration information
  • the domain management functional unit can perform different configurations of the model training function.
  • the domain management function unit performs model training function configuration according to the configuration information, as follows:
  • the domain management function unit configures the state information of the corresponding network AI model training function as activated.
  • the domain management function unit configures the state information of the corresponding network AI model training function as de-activated.
  • the above-mentioned domain management function unit performs model training function configuration according to the configuration information, as follows:
  • the domain management functional unit configures the training period of the network AI model.
  • the domain management functional unit configures the training trigger condition of the network AI model.
  • the domain management functional unit configures the network AI model training instruction to be True.
  • the above-mentioned domain management functional unit carries out model training function configuration according to the configuration information, specifically as follows:
  • the above-mentioned domain management function unit performs model training function configuration according to the configuration information, as follows:
  • the domain management function unit configures the network AI model training data source.
  • the above-mentioned domain management function unit performs model training function configuration according to the configuration information, specifically as follows:
  • the domain management functional unit configures the network AI model information.
  • the above-mentioned domain management function unit performs model training function configuration according to the configuration information, as follows:
  • the domain management function unit adds the network to be added in the network AI model type list AI model type (ie, load information analysis model).
  • the domain management function unit deletes the network AI model type from the network AI model type list (ie, energy saving analysis model).
  • the above-mentioned domain management function unit configures the model training function according to the configuration information, specifically as follows:
  • the domain management functional unit configures the information of the intelligent analysis function in the associated intelligent analysis function information.
  • Domain MnF configures the state information of the network AI model training function.
  • the model training is performed on the Domain MnF, the configuration of the above state information is directly performed on the Domain MnF; if the model training is performed on the network element, the Domain MnF sends the state information of the network AI model training function to the network element. , so as to configure the state information of the network AI model training function on the network element.
  • Domain MnF configures the type of network AI model that the network AI model training function supports training. If model training is performed on Domain MnF, the network AI model training function is directly performed on Domain MnF. The configuration of the network AI model type that supports training; if the model training is performed on the network element, Domain MnF sends the network AI model type that the network AI model training function supports training to the network element, so that the network AI model can be trained on the network element. The AI model training function supports the configuration of the type of network AI model being trained.
  • the above network AI model type indicates that the network AI model training function can train the network AI model of the type.
  • Domain MnF configures the network AI model training trigger information of the network AI model training function. If the model training is performed on the Domain MnF, the network AI model training function is directly performed on the Domain MnF. The configuration of the network AI model training trigger information; if the model training is performed on the network element, the Domain MnF sends the network AI model training trigger information to the network element, so that the network AI model training function of the network element is performed. The configuration of AI model training trigger information. The above-mentioned network AI model training trigger information instructs the network AI model training function to perform network AI model training according to the training trigger information.
  • Domain MnF configures the training data of the network AI model training function. If model training is performed on Domain MnF, the training data configuration is directly performed on Domain MnF; Model training is performed on the network element, then Domain MnF sends the training data to the network element, so that the configuration of the training data is performed on the network element.
  • the above training data indicates that the network AI model training function can use the training data to perform network AI model training.
  • Domain MnF configures the training data source of the network AI model training function, if model training is performed on Domain MnF, the configuration of the training data source is directly performed on Domain MnF; Model training is performed on the network element, then Domain MnF sends the training data source to the network element, so that the configuration of the training data source is performed on the network element.
  • the above training data source indicates that the network AI model training function can use the data of the data source to perform network AI model training.
  • Domain MnF configures the associated intelligent analysis function information of the network AI model training function. If the model training is performed on the Domain MnF, the network AI model training function is directly performed on the Domain MnF. The configuration of the associated intelligent analysis function information; if model training is performed on the network element, then Domain MnF sends the associated intelligent analysis function information of the network AI model training function to the network element, so that the network element is performed on the network element. The configuration of the associated intelligent analysis function information of the AI model training function.
  • the associated intelligent analysis function information of the above-mentioned network AI model training function indicates that the network AI model trained by the network AI model training function can be used by the associated intelligent analysis function.
  • Domain MnF configures the network AI model information associated with the network AI model training function. If the model training is performed on the Domain MnF, the network AI model training function is directly performed on the Domain MnF. The configuration of the associated network AI model information; if the model training is performed on the network element, Domain MnF sends the associated network AI model information of the network AI model training function to the network element, so that the network AI model can be trained on the network element. The configuration of the associated network AI model information for the AI model training function. The associated network AI model information of the above network AI model training function indicates that the network AI model training function can update or retrain the network AI model.
  • the second network management unit also sends the configuration result to the first network management unit.
  • This operation is adopted so that the first network management unit knows the configuration situation, and if the configuration fails, the configuration operation can be performed again.
  • the second network management unit performs model training based on the configured model training function to obtain network model information.
  • the network AI model training function control object is configured with a network AI model training cycle
  • the corresponding cycle time is reached, and Domain MnF starts the network AI model training.
  • Network AI model training function control object When the network AI model training function control object is configured with the network AI model training trigger indication as True, Domain MnF starts network AI model training. Optionally, when the training is completed, configure the network AI model training trigger indication information in the network AI model training function control object to False.
  • the above-mentioned network model information includes at least one of the following: a network model identifier, a network model version, a network model file storage address, and a network model file name.
  • the above-mentioned network model identifier is, for example, the network model 1I.
  • the above-mentioned network model version may be, for example, v1.0.0.
  • the above-mentioned network model file storage address may be, for example, a Uniform Resource Identifier (Uniform Resource Identifier, URI) or an Internet Protocol (Internet Protocol, IP) address.
  • the above-mentioned network model file name may be, for example, the name of the network model file.
  • Domain MnF stores the above network model information.
  • the Domain MnF first creates the network AI model object, and then configures the network AI model information in the network AI model object.
  • the network AI model training function control object has been configured with network AI model object information, configure the network AI model information in the corresponding network AI model object.
  • Domain MnF configures the network model information of the network model object so as to use the model later.
  • the Domain MnF also sends the network model information to the Cross-Domain MnF.
  • Domain MnF sends a notification of network AI model addition or change to Cross Domain MnF, and the notification can carry network AI model information.
  • Cross-Domain MnF can obtain the network AI model information output by training in real time.
  • the method further includes:
  • the second network management unit determines the association information between the configuration information and the network model information
  • the second network management unit sends the association information to the first network management unit.
  • the correlation information between the above configuration information and the network model information can be understood as: the above network model information is obtained by training based on one or several configuration information, then the one or several configuration information and the network model information have connection relation.
  • the second network management unit configures the model training function based on the received configuration information sent by the first network management unit, so as to perform model training according to the configuration information to obtain network model information.
  • operators can flexibly configure the corresponding network AI model training function on demand, which can realize the management and control of network model training, improve the efficiency of management and control model training, and save manpower.
  • This solution also enables operators to uniformly manage and control the training of all supported network AI models, as well as achieve differentiated management and control of the training of different supported network AI models.
  • FIG. 5 it is a schematic flowchart of another method for training a management and control model provided by an embodiment of the present application. As shown in FIG. 5, the method is applied to the second network management unit, which may include steps 501-502, as follows:
  • the second network management unit receives configuration information from the first network management unit, the configuration information configures the model training function, and the configuration information includes at least one of the following information:
  • the above-mentioned first network management unit may be a cross-domain management functional unit Cross-Domain MnF.
  • the second network management unit may be a domain management functional unit Domain MnF.
  • the cross-domain management function unit Cross-Domain MnF determines the configuration information of the training function of the network AI model.
  • the configuration information of the training function of the network AI model can be manually input into the Cross-Domain MnF, or it can be generated by the internal calculation and analysis of the Cross-Domain MnF.
  • the above model training function may be a training function of a certain model, or may be a training function of multiple models.
  • the second network management unit sends the configuration information to a network element, so that the network element performs model training according to the configuration information.
  • the domain management function unit sends the configuration information to the network element, and the network element configures the model training function based on the above configuration information, and performs model training.
  • the second network management unit receives the network model information sent by the network element.
  • the above-mentioned network model information includes at least one of the following: network model identifier, network model version, network model file storage address, network model file name.
  • Domain MnF stores the above network model information.
  • the Domain MnF first creates the network AI model object, and then configures the network AI model information in the network AI model object.
  • the network AI model training function control object has been configured with network AI model object information, configure the network AI model information in the corresponding network AI model object.
  • Domain MnF configures the network model information of the network model object in order to use the model.
  • the Domain MnF also sends the network model information to the Cross-Domain MnF.
  • Domain MnF sends a notification of adding or changing a network AI model to Cross Domain MnF, and the notification may carry the network AI model information.
  • Cross-Domain MnF can obtain the network AI model information output by training in real time.
  • the method further includes:
  • the second network management unit determines the association information between the configuration information and the network model information
  • the second network management unit sends the association information to the first network management unit.
  • the correlation information between the above configuration information and the network model information can be understood as: the above network model information is obtained by training based on one or several configuration information, then the one or several configuration information and the network model information have connection relation.
  • the second network management unit sends the configuration information to the network element based on the received configuration information sent by the first network management unit, so that the network element performs model training according to the configuration information.
  • network model training can be managed and controlled, enabling operators to flexibly configure corresponding network AI model training functions as needed.
  • FIG. 6 it is a schematic flowchart of another method for managing and controlling model training provided by an embodiment of the present application. As shown in FIG. 6, the method is applied to the first network management unit, which may include steps 601-602, as follows:
  • the first network management unit determines configuration information, the configuration information configures the model training function, and the configuration information includes at least one of the following information:
  • the above-mentioned first network management unit may be a cross-domain management functional unit Cross-Domain MnF.
  • the second network management unit may be a domain management functional unit Domain MnF.
  • the cross-domain management function unit Cross-Domain MnF determines the configuration information of the training function of the network AI model.
  • the configuration information of the training function of the network AI model can be manually input into the Cross-Domain MnF, or it can be generated by the internal calculation and analysis of the Cross-Domain MnF.
  • the above model training function may be a training function of a certain model, or may be a training function of multiple models.
  • the first network management unit sends the configuration information to a second network management unit.
  • the method further includes:
  • the first network management unit receives network model information sent by the second network management unit, where the network model information is obtained by the second network management unit after training according to the configuration information.
  • the method further includes:
  • the first network management unit receives the network model information sent by the second network management unit, the network model information is obtained after the network element is trained according to the configuration information, and the configuration information is obtained by the network element from the received by the second network management unit.
  • the above-mentioned model training may be obtained by the training of the second network management unit, or it may be obtained by the training of the network element.
  • the network model information includes at least one of the following: a network model identifier, a network model version, a network model file storage address, and a network model file name.
  • the method further includes:
  • the first network management unit receives the association information sent by the second network management unit, where the association information is association information between the configuration information and the network model information.
  • the first network management unit determines the configuration information of the model training function, and sends the configuration information to the second network management unit.
  • network model training can be managed and controlled, enabling operators to flexibly configure corresponding network AI model training functions as needed.
  • FIG. 7 it is a schematic flowchart of a method for training a management and control model provided by an embodiment of the present application. As shown in FIG. 7, the method may include steps 701-705, as follows:
  • the first network management unit determines configuration information of a first model, the configuration information configures a model training function of the first model, and the configuration information includes at least one of the following information:
  • the above-mentioned first model may be any kind of model.
  • the above-mentioned first model is identified by adopting a Distinguish Name (DN). It may also adopt other methods, which are not specifically limited in this solution.
  • DN Distinguish Name
  • the first network management unit sends the configuration information of the first model to the second network management unit.
  • the first network management unit receives the configuration result sent by the second network management unit.
  • Domain MnF returns the configuration result of the model training function of the first network described above. This operation is adopted so that the first network management unit knows the configuration situation, and if the configuration fails, the configuration operation can be performed again.
  • the second network management unit performs model training according to the configuration information to obtain model information of the first model.
  • the first network management unit receives the model information of the first model sent by the second network management unit.
  • the first network management unit can flexibly control the specific network AI model training, and at the same time, can acquire the network AI model information output by the training in real time. Adopting this method enables operators to control the training of different network AI models on the network AI model training function in a differentiated manner.
  • FIG. 8 it is a schematic structural diagram of an apparatus for managing and controlling model training provided by an embodiment of the present application.
  • the device includes a receiving module 801 and a processing module 802, as follows:
  • the receiving module 801 is configured to receive configuration information from the first network management unit, the configuration information configures the model training function, and the configuration information includes at least one of the following information:
  • Trigger information used to trigger model training
  • Data information used to indicate the data for model training
  • the processing module 802 is configured to perform model training according to the configuration information to obtain network model information.
  • the configuration information further includes at least one of the following:
  • Data source information which is used to indicate the data source for model training
  • the model information associated with the model training function is used to indicate the model for model training
  • the type information of the model which is used to indicate the model type for model training.
  • the network model information includes at least one of the following: a network model identifier, a network model version, a network model file storage address, and a network model file name.
  • the device further includes:
  • a sending module configured to send the network model information to the first network management unit.
  • the device further includes:
  • a determination module configured to determine the association information between the configuration information and the network model information, and send the association information to the first network management unit.
  • the trigger information includes at least one of the following:
  • Training time used to indicate the time of model training
  • Training instruction information used to instruct to start model training.
  • the data information includes:
  • Output data which indicates the output data type of the model.
  • FIG. 9 it is a schematic structural diagram of another apparatus for managing and controlling model training provided by an embodiment of the present application.
  • the device includes a determining module 901 and a sending module 902, as follows:
  • a determination module 901 configured to determine configuration information, the configuration information configures the model training function, and the configuration information includes at least one of the following information:
  • Trigger information used to trigger model training
  • Data information used to indicate the data for model training
  • the sending module 902 is configured to send the configuration information to the second network management unit.
  • the configuration information further includes at least one of the following:
  • Data source information which is used to indicate the data source for model training
  • the model information associated with the model training function is used to indicate the model for model training
  • the type information of the model which is used to indicate the model type for model training.
  • the device further includes:
  • a receiving module configured to receive network model information sent by the second network management unit, where the network model information is obtained after the second network management unit is trained according to the configuration information; or, the network model information is The network element is obtained after training according to the configuration information, where the configuration information is received by the network element from the second network management unit.
  • the network model information includes at least one of the following: a network model identifier, a network model version, a network model file storage address, and a network model file name.
  • the receiving module is also used for:
  • association information sent by the second network management unit, where the association information is association information between the configuration information and the network model information.
  • the trigger information includes at least one of the following:
  • Training time used to indicate the time of model training
  • Training instruction information used to instruct to start model training.
  • the data information includes:
  • Output data which indicates the output data type of the model.
  • the apparatus 1000 for managing and controlling model training includes at least one processor 1001 , at least one memory 1002 and at least one communication interface 1003 .
  • the processor 1001, the memory 1002 and the communication interface 1003 are connected through the communication bus and complete the communication with each other.
  • the processor 1001 may be a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits used to control the execution of the above programs.
  • CPU central processing unit
  • ASIC application-specific integrated circuit
  • the communication interface 1003 is used to communicate with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area network (Wireless Local Area Networks, WLAN).
  • RAN radio access network
  • WLAN Wireless Local Area Networks
  • Memory 1002 may be read-only memory (ROM) or other type of static storage device that can store static information and instructions, random access memory (RAM), or other type of static storage device that can store information and instructions It can also be an electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, EEPROM), a compact disc read-only memory (CD-ROM) or other optical disk storage, optical disk storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or capable of carrying or storing desired program code in the form of instructions or data structures and capable of being executed by a computer Access any other medium without limitation.
  • the memory can exist independently and be connected to the processor through a bus.
  • the memory can also be integrated with the processor.
  • the memory 1002 is used for storing the application code for executing the above solution, and the execution is controlled by the processor 1001 .
  • the processor 1001 is configured to execute the application code stored in the memory 1002 .
  • the code stored in the memory 1002 can execute the training method of the management and control model provided above.
  • An embodiment of the present application further provides a chip system, the chip system is applied to an electronic device; the chip system includes one or more interface circuits and one or more processors; the interface circuit and the processor pass through line interconnection; the interface circuit is used to receive signals from the memory of the electronic device and send the signals to the processor, the signals include computer instructions stored in the memory; when the processor executes the When executing the computer instructions, the electronic device performs the method.
  • Embodiments of the present application also provide a computer-readable storage medium, where instructions are stored in the computer-readable storage medium, and when the computer or processor is run on a computer or a processor, the computer or the processor is made to execute any one of the above methods. or multiple steps.
  • Embodiments of the present application also provide a computer program product including instructions.
  • the computer program product when run on a computer or processor, causes the computer or processor to perform one or more steps of any of the above methods.
  • An embodiment of the present application further provides a system for managing and controlling model training, including one or more modules in any of the foregoing apparatuses.
  • At least one (a) of a, b, or c can represent: a, b, c, a-b, a-c, b-c, or a-b-c, where a, b, c may be single or multiple .
  • words such as “first” and “second” are used to distinguish the same or similar items with basically the same function and effect. Those skilled in the art can understand that the words “first”, “second” and the like do not limit the quantity and execution order, and the words “first”, “second” and the like are not necessarily different.
  • words such as “exemplary” or “for example” are used to represent examples, illustrations or illustrations. Any embodiments or designs described in the embodiments of the present application as “exemplary” or “such as” should not be construed as preferred or advantageous over other embodiments or designs. Rather, the use of words such as “exemplary” or “such as” is intended to present the related concepts in a specific manner to facilitate understanding.
  • the disclosed system, apparatus and method may be implemented in other manners.
  • the division of the unit is only a logical function division, and there may be other division methods in actual implementation, for example, multiple units or components may be combined or integrated into another system, or some features may be ignored, or not implement.
  • the shown or discussed mutual coupling, or direct coupling, or communication connection may be through some interfaces, indirect coupling or communication connection of devices or units, and may be in electrical, mechanical or other forms.
  • Units described as separate components may or may not be physically separated, and components shown as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution in this embodiment.
  • the above-mentioned embodiments it may be implemented in whole or in part by software, hardware, firmware or any combination thereof.
  • software it can be implemented in whole or in part in the form of a computer program product.
  • the computer program product includes one or more computer instructions.
  • the computer program instructions When the computer program instructions are loaded and executed on a computer, the procedures or functions according to the embodiments of the present application are generated in whole or in part.
  • the computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable device.
  • the computer instructions may be stored in or transmitted over a computer-readable storage medium.
  • the computer instructions can be sent from one website site, computer, server, or data center to another by wire (eg, coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (eg, infrared, wireless, microwave, etc.)
  • wire e.g. coaxial cable, fiber optic, digital subscriber line (DSL)
  • wireless e.g., infrared, wireless, microwave, etc.
  • the computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that includes one or more available media integrated.
  • the available media may be read-only memory (ROM), or random access memory (RAM), or magnetic media, such as floppy disks, hard disks, magnetic tapes, magnetic disks, or optical media, such as , digital versatile disc (digital versatile disc, DVD), or semiconductor media, for example, solid state disk (solid state disk, SSD) and the like.
  • ROM read-only memory
  • RAM random access memory
  • magnetic media such as floppy disks, hard disks, magnetic tapes, magnetic disks, or optical media, such as , digital versatile disc (digital versatile disc, DVD), or semiconductor media, for example, solid state disk (solid state disk, SSD) and the like.

Landscapes

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

Abstract

本申请实施例提供一种管控模型训练的方法及装置、系统，包括从第一网络管理单元接收配置信息，所述配置信息配置模型训练功能，所述配置信息包括以下信息中的至少一种：状态信息，用于激活或去激活所述模型训练功能；触发信息，用于触发进行模型训练；数据信息，用于指示进行模型训练的数据；根据所述配置信息进行模型训练，得到网络模型信息。通过本申请实施例，第二网络管理单元基于接收到的第一网络管理单元发送的配置信息，对模型训练功能进行配置，以根据所述配置信息进行模型训练，得到网络模型信息。采用该手段，可以实现对网络模型训练的管控，使能运营商可以灵活按需配置相应的网络AI模型训练功能。

Description

管控模型训练的方法及装置、系统
本申请要求于2021年3月30日提交中国专利局、申请号为202110343727.X、申请名称为“管控模型训练的方法及装置、系统”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及网络管理技术领域，尤其涉及一种管控模型训练的方法及装置、系统。
背景技术
随着垂直行业的引入，终端设备的增加和业务的多样性，运营商的网络越来越复杂，导致网络运维难度的增加。如何降低网络运维成本，简化网络运维流程，快速部署网络来满足多样性化地业务是网络运维迫切需要改善的点。
目前，电信系统中的网络管理单元和网元引入人工智能(Artificial Intelligence，AI)/机器学习(Machine learning，ML)来实现网络的智能化功能。管理单元和网元的智能化都依赖于AI/ML模型，管理单元和网元也都引入了网络AI模型训练功能，在本地对网络AI模型进行训练。网络AI模型虽然是在本地进行训练，但是运营商需要对网络AI模型的训练进行管控。因此，需要运营商告知设备商运维人员相应训练管控需求，设备商运维人员需要到本地通过本地人工界面对网络AI模型的训练进行控制。由于不支持运营商在线对本地的网络AI模型训练功能进行实时控制，因此人工效率低且耗人力。
发明内容
本申请公开了一种管控模型训练的方法及装置、系统，可以实现对网络AI模型训练功能的灵活管控。
第一方面，本申请实施例提供一种管控模型训练的方法，包括:从第一网络管理单元接收配置信息，配置信息配置模型训练功能，配置信息包括以下信息中的至少一种：状态信息，用于激活或去激活模型训练功能；触发信息，用于触发进行模型训练；数据信息，用于指示进行模型训练的数据；根据配置信息进行模型训练，得到网络模型信息。
通过本申请实施例，第二网络管理单元基于接收到的第一网络管理单元发送的配置信息，对模型训练功能进行配置，以根据配置信息进行模型训练，得到网络模型信息。采用该手段，使能运营商可以灵活按需配置相应的网络AI模型训练功能，可以实现对网络模型训练的管控，提高了管控模型训练的效率，同时节省了人力。
本申请实施例还提供一种管控模型训练的方法，包括:从第一网络管理单元接收配置信息，配置信息配置模型训练功能，配置信息包括以下信息中的至少一种：状态信息，用于激活或去激活模型训练功能；触发信息，用于触发进行模型训练；数据信息，用于指示进行模型训练的数据；向网元发送配置信息，以便网元根据配置信息进行模型训练，得到网络模型信息。
通过本申请实施例，第二网络管理单元基于接收到的第一网络管理单元发送的配置信息，并向网元发送配置信息，以便网元根据配置信息进行模型训练。采用该手段，可以实现对网络模型训练的管控，使能运营商可以灵活按需配置相应的网络AI模型训练功能。
作为一种可选的实现方式，配置信息还包括以下中的至少一种：数据源信息，用于指示进行模型训练的数据来源，其中，该数据来源即提供训练数据的实体，可以是网元列表，管理功能单元列表，数据库列表等。模型训练功能关联的模型信息，用于指示进行模型训练的模型，例如，关联的网络AI模型信息，指示输出的网络AI模型对象，如可以是网络AI模型对象的标识等。模型的类型信息，用于指示进行模型训练的模型类型，例如，模型的类型信息包括以下中的至少一种：负载信息分析模型，业务体验分析模型，网络性能分析模型，拥塞分析模型，服务质量(Quality of Service，QoS)分析模型，节能分析模型，话务流向分析模型，大规模多进多出Massive MIMO分析模型，用户设备(User Equipment，UE)轨迹分析模型。
上述数据源信息使能运营商按需配置网络AI模型训练数据源。采用该手段，可以使得网络模型的训练功能按照运营商指定的数据源的数据进行训练；同时，相较于到本地进行控制的手段，采用本方案，可以提高管控模型训练的效率，同时节省了人力。
上述模型训练功能关联的模型信息使能运营商按需配置进行模型训练的网络AI模型。采用该手段，可以使得网络模型的训练功能按照运营商指定的模型信息进行训练；同时，相较于到本地进行控制的手段，采用本方案，可以提高管控模型训练的效率，同时节省了人力。
上述模型的类型信息使能运营商灵活配置网络AI模型训练功能支持的类型。采用该手段，可以使得网络模型的训练功能按照运营商指定的模型类型进行训练；同时，相较于到本地进行控制的手段，采用本方案，可以提高管控模型训练的效率，同时节省了人力。
作为一种可选的实现方式，网络模型信息包括以下中的至少一种：网络模型标识，网络模型版本，网络模型文件存放地址，网络模型文件名。上述网络模型标识例如为网络模型1I。上述网络模型版本例如可以是v1.0.0。上述网络模型文件存放地址例如可以是统一资源标识符(Uniform Resource Identifier，URI)或者网际互连协议(Internet Protocol，IP)地址。上述网络模型文件名例如可以是网络模型文件的名称。
作为又一种可选的实现方式，上述方法还包括：向第一网络管理单元发送网络模型信息。
通过该手段，以便第一网络管理单元可以实时获取训练输出的网络AI模型信息。
作为另一种可选的实现方式，上述方法还包括：确定配置信息和网络模型信息之间的关联信息；向第一网络管理单元发送关联信息。
上述配置信息和网络模型信息的关联信息，可以理解为：上述网络模型信息是基于某个或某几个配置信息进行训练得到的，则该某个或某几个配置信息与该网络模型信息具备关联关系。
通过该手段，以便第一网络管理单元可以实时获取配置信息和网络模型信息之间的关联信息。
作为一种可选的实现方式，上述触发信息包括以下中的至少一种：训练周期，用于指示模型训练周期，例如，指示网络AI模型训练功能需要按设定周期进行训练，如设定周期为每小时，每天，每星期等；训练时间，用于指示模型训练的时间，例如，指示网络AI模型训练功能在某个时间点进行训练；训练指示信息，用于指示开始进行模型训练，即立即进行网络模型训练。
上述状态信息使能运营商灵活按需激活、去激活相应的网络AI模型训练功能。采用该手段，可以使得网络模型的训练功能按照运营商指定的状态信息进行训练。
上述触发信息使能运营商按需配置网络AI模型训练触发信息。采用该手段，可以使得网络模型的训练功能按照运营商指定的触发信息进行训练。
作为一种可选的实现方式，上述数据信息包括：输入数据，用于指示模型的输入数据类型；输出数据，用于指示模型的输出数据类型。例如，支持大规模多进多出(Massive multiple input multiple output，Massive MIMO)模式优化模型的，输入数据可以是参考信号接收功率(Reference Signal Received Power，RSRP)，信号干扰噪声比(Signal to Interference plus noise ratio，SINR)，上下行吞吐率(DownLink/UpLink throughput，DL/UL throughput)；输出数据可以是Massive MIMO pattern覆盖场景下的倾斜角，方位角等。
该数据信息使能运营商按需配置网络AI模型训练数据类型。采用该手段，可以使得网络模型的训练功能按照运营商指定的数据类型进行训练。
第二方面，本申请实施例提供一种管控模型训练的方法，包括:确定配置信息，配置信息配置模型训练功能，配置信息包括以下信息中的至少一种：状态信息，用于激活或去激活模型训练功能；触发信息，用于触发进行模型训练；数据信息，用于指示进行模型训练的数据；向第二网络管理单元发送配置信息。
通过本申请实施例，第一网络管理单元确定模型训练功能的配置信息，并将该配置信息发送给第二网络管理单元。采用该手段，可以实现对网络模型训练的管控，使能运营商可以灵活按需配置相应的网络AI模型训练功能。
作为一种可选的实现方式，上述配置信息还包括以下中的至少一种：数据源信息，用于指示进行模型训练的数据来源；模型训练功能关联的模型信息，用于指示进行模型训练的模型；模型的类型信息，用于指示进行模型训练的模型类型。
作为一种可选的实现方式，上述方法还包括：接收第二网络管理单元发送的网络模型信息，网络模型信息为第二网络管理单元根据配置信息训练后得到的；或者，网络模型信息为网元根据配置信息训练后得到的，配置信息为网元从第二网络管理单元接收到的。
可选的，网络模型信息包括以下中的至少一种：网络模型标识，网络模型版本，网络模型文件存放地址，网络模型文件名。
作为另一种可选的实现方式，上述方法还包括：接收第二网络管理单元发送的关联信息，其中，关联信息为配置信息和网络模型信息的关联信息。
作为一种可选的实现方式，上述触发信息包括以下中的至少一种：训练周期，用于指示模型训练周期；训练时间，用于指示模型训练的时间；训练指示信息，用于指示开始进行模型训练。
作为另一种可选的实现方式，上述数据信息包括：输入数据，用于指示模型的输入数据类型；输出数据，用于指示模型的输出数据类型。
第三方面，本申请实施例提供一种管控模型训练的装置，包括:接收模块，用于从第一网络管理单元接收配置信息，配置信息配置模型训练功能，配置信息包括以下信息中的至少一种：状态信息，用于激活或去激活模型训练功能；触发信息，用于触发进行模型训练；数据信息，用于指示进行模型训练的数据；处理模块，用于根据配置信息进行模型训练，得到网络模型信息。
可选的，配置信息还包括以下中的至少一种：数据源信息，用于指示进行模型训练的数据来源；模型训练功能关联的模型信息，用于指示进行模型训练的模型；模型的类型信息，用于指示进行模型训练的模型类型。
作为一种可选的实现方式，网络模型信息包括以下中的至少一种：网络模型标识，网络模型版本，网络模型文件存放地址，网络模型文件名。
作为一种可选的实现方式，上述装置还包括：发送模块，用于向第一网络管理单元发送 网络模型信息。
作为另一种可选的实现方式，上述装置还包括：确定模块，用于确定配置信息和网络模型信息之间的关联信息，并向第一网络管理单元发送关联信息。
其中，触发信息包括以下中的至少一种：训练周期，用于指示模型训练周期；训练时间，用于指示模型训练的时间；训练指示信息，用于指示开始进行模型训练。
作为一种可选的实现方式，数据信息包括：输入数据，用于指示模型的输入数据类型；输出数据，用于指示模型的输出数据类型。
第四方面，本申请实施例提供一种管控模型训练的装置，包括：确定模块，用于确定配置信息，配置信息配置模型训练功能，配置信息包括以下信息中的至少一种：状态信息，用于激活或去激活模型训练功能；触发信息，用于触发进行模型训练；数据信息，用于指示进行模型训练的数据；发送模块，用于向第二网络管理单元发送配置信息。
作为一种可选的实现方式，配置信息还包括以下中的至少一种：数据源信息，用于指示进行模型训练的数据来源；模型训练功能关联的模型信息，用于指示进行模型训练的模型；模型的类型信息，用于指示进行模型训练的模型类型。
作为一种可选的实现方式，上述装置还包括：接收模块，用于接收第二网络管理单元发送的网络模型信息，网络模型信息为第二网络管理单元根据配置信息训练后得到的；或者，网络模型信息为网元根据配置信息训练后得到的，配置信息为网元从第二网络管理单元接收到的。
作为一种可选的实现方式，网络模型信息包括以下中的至少一种：网络模型标识，网络模型版本，网络模型文件存放地址，网络模型文件名。
作为另一种可选的实现方式，接收模块还用于：接收第二网络管理单元发送的关联信息，其中，关联信息为配置信息和网络模型信息的关联信息。
作为一种可选的实现方式，触发信息包括以下中的至少一种：训练周期，用于指示模型训练周期；训练时间，用于指示模型训练的时间；训练指示信息，用于指示开始进行模型训练。
作为又一种可选的实现方式，数据信息包括：输入数据，用于指示模型的输入数据类型；输出数据，用于指示模型的输出数据类型。
第五方面，本申请提供了一种计算机存储介质，包括计算机指令，当所述计算机指令在电子设备上运行时，使得所述电子设备执行如第一方面任一种可能的实施方式和/或第二方面任一种可能的实施方式提供的方法。
第六方面，本申请实施例提供一种计算机程序产品，当计算机程序产品在计算机上运行时，使得计算机执行如第一方面任一种可能的实施方式和/或第二方面任一种可能的实施方式提供的方法。
第七方面，本申请实施例提供一种管控模型训练的装置，包括处理器和存储器；其中，所述存储器用于存储程序代码，所述处理器用于调用所述程序代码，以执行如第一方面任一种可能的实施方式和/或第二方面任一种可能的实施方式提供的方法。
第八方面，本申请实施例提供一种管控模型训练的系统，包括如第三方面任一种可能的实施方式和/或第四方面任一种可能的实施方式提供的所述装置。
第九方面，本申请实施例提供一种管控模型训练的方法，包括：第一网络管理单元确定配置信息，该配置信息配置模型训练功能，该配置信息包括以下信息中的至少一种：状态信息，用于激活或去激活模型训练功能；触发信息，用于触发进行模型训练；数据信息，用于 指示进行模型训练的数据；该第一网络管理单元发送该配置信息；第二网络管理单元接收该配置信息；第二网络管理单元根据该配置信息进行模型训练，得到网络模型信息。
可以理解地，上述提供的第三方面所述的装置、第四方面所述的装置、第五方面所述的计算机存储介质、第六方面所述的计算机程序产品、第七方面所述的装置或者第八方面所述的系统均用于执行第一方面中任一所提供的方法以及第二方面中任一所提供的方法。因此，其所能达到的有益效果可参考对应方法中的有益效果，此处不再赘述。
第十方面，提供了一种芯片，所述芯片与存储器耦合，执行本申请实施例第一方面或第一方面中任一实现所述的管控模型训练的方法。
第十一方面，提供了一种芯片，所述芯片与存储器耦合，执行本申请实施例第二方面或第二方面中任一实现所述的管控模型训练的方法。
第十二方面，提供了一种芯片，所述芯片与存储器耦合，执行本申请实施例第九方面所述的管控模型训练的方法。
需要说明的是，本申请实施例中“耦合”是指两个部件彼此直接或间接地结合。
附图说明
下面对本申请实施例用到的附图进行介绍。
图1是本申请实施例提供的一种管控模型训练的系统的架构示意图；
图2是本申请实施例提供的另一种管控模型训练的系统的架构示意图；
图3是本申请实施例提供的又一种管控模型训练的系统的架构示意图；
图4是本申请实施例提供的一种管控模型训练的方法的流程示意图；
图5是本申请实施例提供的另一种管控模型训练的方法的流程示意图；
图6是本申请实施例提供的又一种管控模型训练的方法的流程示意图；
图7是本申请实施例提供的另一种管控模型训练的方法的流程示意图；
图8是本申请实施例提供的一种管控模型训练的装置的结构示意图；
图9是本申请实施例提供的另一种管控模型训练的装置的结构示意图；
图10是本申请实施例提供的另一种管控模型训练的装置的结构示意图。
具体实施方式
下面结合本申请实施例中的附图对本申请实施例进行描述。本申请实施例的实施方式部分使用的术语仅用于对本申请的具体实施例进行解释，而非旨在限定本申请。
需要说明的是，本申请实施例中的AI/ML模型，可以是机器学习中通过查找数据中的模式来生成预测的数学模型。例如，网络AI/ML模型，主要是通过查找网络数据中的模式来生成预测网络性能的数据模型，比如网络话务预测模型等。其中，本申请实施例对于模型的种类不做具体限定。
本申请实施例中的模型训练，可以理解为：为达成高识别率的目标，使用大量数据，找出目标配置参数并确定目标模型的过程。例如，网络模型训练主要描述的是网络数据之间的关系，如可以是不同网络环境下的配置参数值等。
参照图1所示，为本申请实施例提供的一种管控模型训练的系统的架构示意图。如图1所示，该管控模型训练的系统包括第一网络管理单元101(可以有更多，图中未示出)、第二网络管理单元102(可以有更多，图中未示出)、以及由第二网络管理单元102管理的网络节点集合103(图中示出该网络节点集合103包括n个网络节点)。其中，第一网络管理单元101的功能可以部署在独立的设备/装置上，也可以部署在具备其他功能的设备/装置上；部署了第一网络管理单元101的功能的设备/装置称为第一网络管理设备/第一网络管理装置；为叙述方便，本申请实施例中第一网络管理单元、第一网络管理装置或第一网络管理设备统一用第一网络管理单元指代。同样地，第二网络管理单元102的功能可以部署在独立的设备/装置上，也可以部署在具备其他功能的设备/装置上；部署了第二网络管理单元102的功能的设备/装置称为第二网络管理设备/第二网络管理装置，第二网络管理单元、第二网络管理装置或第二网络管理设备统一用第二网络管理单元指代。在一种可能的方案中，第一网络管理单元可以是网络管理系统(network management system，NMS)、跨域管理功能单元(Cross Domain management function，Cross-Domain MnF)，也可以称之为网络管理功能单元(network management function，NMF)、或业务支撑系统(business support system，BSS)。第二网络管理单元可以是网元管理系统(element management system，EMS)、或域管理功能单元(Domain management function，Domain MnF)，也可以称之为子网络管理功能(subnetwork management function，NMF)或者网元管理功能单元(network element/function management function)。
其中，网元可以是核心网网元，也可以是无线网络网元；核心网网元包括但不限于：移动交换中心(mobile switching center，MSC)、关口移动交换中心(gateway mobile switching center，GMSC)、GPRS(general packet radio service，通用分组无线业务)业务支撑节点(serving GPRS support node，SGSN)、网关GPRS支撑节点(gateway GPRS support node，GGSN)、移动性管理实体(mobility management entity，MME)、服务网关(serving gateway，SGW)、分组网关(packet gateway，PGW)、接入管理功能(access management function，AMF)设备、用户面功能(user plane function，UPF)设备、会话管理功能(session management function，SMF)设备；无线网络网元包括但不限于基站和基站控制器，基站可以是：全球移动通信系统(global system for mobile communications，GSM)基站、通用移动通信系统(universal mobile telecommunications system，UMTS)基站、长期演进(long term evolution，LTE)基站、新空口(new radio，NR)基站，其中，LTE基站也称为演进型基站(evolved NodeB，eNB)，新空口基站也称为5G基站(gNodeB，gNB)；基站控制器可以是GSM基站控制器、UMTS基站控制器。
在该管控模型训练的系统中，第一网络管理单元确定配置信息，并向第二网络管理单元发送配置信息，第二网络管理单元根据所述配置信息进行模型训练，得到网络模型信息。可替代的，或者，第二网络管理单元向网元发送所述配置信息，以便所述网元根据所述配置信息进行模型训练得到所述网络模型信息。该配置信息用于配置模型训练功能，所述配置信息包括以下信息中的至少一种：状态信息，用于激活或去激活所述模型训练功能；触发信息，用于触发进行模型训练；数据信息，用于指示进行模型训练的数据。
如图2所示，为本申请实施例提供的又一种管控模型训练的系统的架构示意图，其中，跨域管理功能单元201(例如运营商级别的网络管理单元)为图1的第一网络管理单元，域管理功能单元202(例如运营商下属单位级别的网络管理单元)为图1的第二网络管理单元。跨域管理功能单元201对域管理功能单元202(可以为多个，图中仅示出了一个域管理功能单元)进行管理。域管理功能单元202管理网元203。在进行模型训练管控过程中，跨域管 理功能单元201确定配置信息，向域管理功能单元202发送配置信息，域管理功能单元202根据配置信息，进行模型训练，得到网络模型信息；或者，域管理功能单元202向网元203发送所述配置信息，以便所述网元根据所述配置信息进行模型训练。
如图3所示，为本申请实施例提供的又一种管控模型训练的系统的架构示意图。该管控模型训练的系统包括业务运营单元301、跨域管理功能单元302、域管理功能单元303和网元304。跨域管理功能单元302为图1的第一网络管理单元。管理功能单元303为图1的第二网络管理单元。跨域管理功能单元302可以管理一个或多个域管理功能单元303(图中示出了一个域管理功能单元303)，域管理功能单元303又对与之连接的网元304进行管理。
其中，业务运营单元301，也可以称为通信业务管理功能单元(communication service management function)，可以提供计费、结算、帐务、客服、营业、网络监控、通信业务生命周期管理，业务意图翻译等功能和管理服务。上述业务运营单元可包括运营商的运营系统或者垂直行业的运营系统(vertical operational technology system)。该业务运营单元301可以确定配置信息，并下发给跨域管理功能单元302，进而跨域管理功能单元302转发给域管理功能单元303；或者用户可以通过业务运营单元301输入配置信息，进而业务运营单元301下发给跨域管理功能单元302，跨域管理功能单元302转发给域管理功能单元303。
跨域管理功能单元302提供以下一项或几项功能或者管理服务：网络的生命周期管理，网络的部署，网络的故障管理，网络的性能管理，网络的配置管理，网络的保障，网络的优化功能，通信服务提供商的网络意图(intent from communication service provider，intent-CSP)的翻译，通信服务使用者的网络意图(intent from communication service consumer，intent-CSC)的翻译，网络AI模型的训练和网络AI模型的推理等。这里的网络可以包括一个或者多个网元，子网络或者网络切片。例如，跨域管理功能单元302可以是网络切片管理功能(network slice management function，NSMF)，或者管理数据分析功能(management data analytical function，MDAF)，或者跨域自组织网络功能(self-organization network function，SON-function)，或者跨域意图管理功能单元。
需要说明的是，在某些部署场景下，跨域管理功能单元也可以提供以下一项或几项管理功能或者管理服务：子网络的生命周期管理，子网络的部署，子网络的故障管理，子网络的性能管理，子网络的配置管理，子网络的保障，子网络的优化功能，子网意图翻译功能等。其中，子网络可以由多个小的子网络组成或者由多个网络切片子网络组成，比如运营商的一个接入网子网络包括第一设备商的接入网子网络和第二设备商的接入网子网络。
域管理功能单元303提供以下一项或者多项功能或者管理服务：子网络或者网元的生命周期管理，子网络或者网元的部署，子网络或者网元的故障管理，子网络或者网元的性能管理，子网络或者网元的保障，子网络或者网元的优化管理，子网络或者网元的意图翻译，网络AI模型的训练和网络AI模型的推理等。这里的子网络包括一个或者多个网元。或者，这里的子网络也可以包括一个或多个子网络，即一个或多个子网络组成一个更大覆盖范围的子网络。又或者，这里的子网络也可以包括一个或多个网络切片子网络。子网络包括以下几种描述方式之一：
某个技术域的网络，比如无线接入网，核心网，传输网等。
某个制式的网络，比如GSM网络，LTE网络，5G网络等。
某个设备商提供的网络，比如设备商X提供的网络等。
某个地理区域的网络，比如工厂A的网络，地级市B的网络等。
网元(Net Element，NE)304，为提供网络服务的实体，包括核心网网元、接入网网元 等。本方案中的网元NE还可以提供网络AI模型的训练和网络AI模型的推理两个功能中的至少一个。例如，核心网网元可以包括但不限于接入与移动管理功能(access and mobility management function，AMF)实体、会话管理功能(session management function，SMF)实体、策略控制功能(policy control function，PCF)实体、网络数据分析功能(network data analysis function，NWDAF)实体、网络存储功能(network repository function，NRF)、网关等。接入网网元可以包括但不限于：各类基站(例如下一代基站(generation node B，gNB)，演进型基站(evolved Node B，eNB)、集中控制单元(central unit control panel，CUCP)、集中单元(central unit，CU)、分布式单元(distributed unit，DU)、集中用户面单元(central unit user panel，CUUP)等。在本方案中，网络功能(Net function，NF)，也称为网元NE。
本方案中的网元数据分析功能(net element data analysis function，NEDAF)可以是一个独立的网元，也可以是以上某个网元中的一个逻辑功能，本方案不做限制。需要说明的是，本方案中的网元数据分析功能也可以称为网元推理功能或者智能化功能，名称不作限制。
下面结合图1～图3所示的系统架构，详细描述本方案的管控模型训练的方法。
参照图4所示，为本申请实施例提供的一种管控模型训练的方法的流程示意图。如图4所示，该方法应用于第二网络管理单元，其可包括步骤401-402，具体如下：
401、第二网络管理单元从第一网络管理单元接收配置信息，所述配置信息配置模型训练功能，所述配置信息包括以下信息中的至少一种：
A)状态信息，用于激活或去激活所述模型训练功能；
B)触发信息，用于触发进行模型训练；
C)数据信息，用于指示进行模型训练的数据；
可选的，上述第一网络管理单元可以是跨域管理功能单元Cross-Domain MnF。第二网络管理单元可以是域管理功能单元Domain MnF。其中，跨域管理功能单元Cross-Domain MnF确定网络AI模型的训练功能的配置信息。该网络AI模型的训练功能的配置信息可以是人工输入到Cross-Domain MnF，其也可以是Cross-Domain MnF内部计算分析生成的。
上述模型训练功能，可以是某一个模型的训练功能，也可以是多个模型的训练功能。
下面对上述配置信息进行介绍。
作为一种可选的实现方式，上述配置信息可包括A)状态信息，用于激活或去激活所述模型训练功能。也就是说，该状态信息用于描述网络AI模型训练功能的状态，具体地，可包括激活态，去激活态和进行中三个状态。例如，运营商可以激活或者去激活所述网络AI模型训练功能。
上述状态信息使能运营商灵活按需激活、去激活相应的网络AI模型训练功能。
作为另一种可选的实现方式，上述配置信息可包括B)触发信息，用于触发进行模型训练。
具体地，所述B)触发信息包括以下中的至少一种：
训练周期，用于指示模型训练周期。例如，指示网络AI模型训练功能需要按设定周期进行训练，如设定周期为每小时，每天，每星期等。
训练时间，用于指示模型训练的时间。例如，指示网络AI模型训练功能在某个时间点进行训练。此处仅以训练时间为例进行说明，其还可以是其他条件，如利用网络AI模型进行模型推理或者分析，得到的结果不符合预设需求时，即分析结果不准确，则触发进行训练等。其中，模型推理，也叫智能分析，可以理解为：利用模型给出分析结果。例如，根据当前网 络环境确定网络配置参数集合。
训练指示信息，用于指示开始进行模型训练，即立即进行网络模型训练。
上述触发信息使能运营商按需配置网络AI模型训练触发信息。
作为再一种可选的实现方式，上述配置信息可包括C)数据信息，用于指示进行模型训练的数据。具体地，所述数据信息包括：
输入数据，用于指示模型的输入数据类型；
输出数据，用于指示模型的输出数据类型。
例如，支持大规模多进多出(Massive multiple input multiple output，Massive MIMO)模式优化模型的，输入数据可以是参考信号接收功率(Reference Signal Received Power，RSRP)，信号干扰噪声比(Signal to Interference plus noise ratio，SINR)，上下行吞吐率(DownLink/UpLink throughput，DL/UL throughput)；输出数据可以是Massive MIMO pattern覆盖场景下的倾斜角，方位角等。
该数据信息使能运营商按需配置网络AI模型训练数据类型等。
进一步地，上述数据信息还包括相应训练数据的粒度，即对应粒度的训练数据类型。其中，粒度包括小区，栅格，跟踪区域，网络切片，业务等。
所述配置信息还包括以下中的至少一种：
D)数据源信息，用于指示进行模型训练的数据来源。该数据来源即提供训练数据的实体，可以是网元列表，管理功能单元列表，数据库列表等。
上述数据源信息使能运营商按需配置网络AI模型训练数据源。
E)模型训练功能关联的模型信息，用于指示进行模型训练的模型。例如，关联的网络AI模型信息，指示输出的网络AI模型对象，如可以是网络AI模型对象的标识等。
上述模型训练功能关联的模型信息使能运营商按需配置进行模型训练的网络AI模型。
F)模型的类型信息，用于指示进行模型训练的模型类型。其中，所述模型的类型信息包括以下中的至少一种：负载信息分析模型，业务体验分析模型，网络性能分析模型，拥塞分析模型，服务质量(Quality of Service，QoS)分析模型，节能分析模型，话务流向分析模型，大规模多进多出Massive MIMO分析模型，用户设备(User Equipment，UE)轨迹分析模型。
上述模型的类型信息使能运营商灵活配置网络AI模型训练功能支持的类型。
进一步地还可以包括G)模型训练功能关联的智能分析功能信息，用于指示可以使用网络AI模型的智能分析功能。这里的关联的智能分析功能信息表示可以使用训练得到的网络模型的智能分析功能的信息。
上述仅以部分配置信息为例进行说明，其还可以包括其他配置信息，本方案对此不做具体限定。
其中，第二网络管理单元从第一网络管理单元接收配置信息，可以是第二网络管理单元接收第一网络管理单元发送的模型训练功能管控对象创建请求，所述模型训练功能管控对象创建请求中携带模型训练功能管控对象标识和配置信息。
上述模型训练功能管控对象即为模型训练功能的管控信息，可参照表一所示。例如，Domain MnF将接收到的网络AI模型训练功能的管控信息写入在相应的字段上。Domain MnF上的网络AI模型训练功能根据所述网络AI模型训练功能管控对象中配置的信息进行网络AI模型的训练。
其中，域管理功能单元在网络AI模型训练功能管控对象TrainingFunction中配置接收到的网络AI模型训练功能管控信息时，若所述域管理功能单元中不存在所述模型训练管控对 象，则在配置所述网络AI模型训练功能管控对象之前先创建网络AI模型训练功能管控对象TrainingFunction，进而在模型训练管控对象中配置所述配置信息；若所述域管理功能单元中存在所述模型训练管控对象，则在现有模型训练管控对象中配置所述配置信息。
上述网络AI模型训练功能对象的信息如下表一所示：
表一
Figure PCTCN2022082872-appb-000001
402、所述第二网络管理单元根据所述配置信息进行模型训练，得到网络模型信息。
其中，上述步骤402可包括步骤4021-4022，具体如下：
4021、所述第二网络管理单元根据所述配置信息进行模型训练功能配置；
其中，基于上述不同的配置信息，域管理功能单元可进行模型训练功能的不同配置。
a)当配置信息包括A)状态信息时，域管理功能单元根据所述配置信息进行模型训练功能配置，具体如下：
当Cross-Domain MnF需要网络AI模型训练功能处于工作状态，即激活网络AI模型训练功能时，则域管理功能单元配置相应网络AI模型训练功能的状态信息为activated。
当Cross-Domain MnF需要禁止使用网络AI模型训练功能，即去激活网络AI模型训练功能时，则域管理功能单元配置相应网络AI模型训练功能的状态信息为de-activated。
b)当配置信息包括B)触发信息时，上述域管理功能单元根据所述配置信息进行模型训练功能配置，具体如下：
当Cross-Domain MnF需要网络AI模型训练功能进行周期性训练时，则域管理功能单元配置所述网络AI模型的训练周期。
当Cross-Domain MnF需要网络AI模型训练功能达到某个条件时进行训练，则域管理功能单元配置所述网络AI模型的训练触发条件。
当Cross-Domain MnF需要网络AI模型训练功能立即进行训练时，则域管理功能单元配置所述网络AI模型训练指示为True。
c)当配置信息包括C)数据信息时，上述域管理功能单元根据所述配置信息进行模型训 练功能配置，具体如下：
当Cross-Domain MnF需要修改网络AI模型训练数据类型，增加或者减少网络AI模型训练数据类型时，则需要配置所述网络AI模型训练数据类型。
d)当配置信息包括D)数据源信息时，上述域管理功能单元根据所述配置信息进行模型训练功能配置，具体如下：
当Cross-Domain MnF需要修改网络AI模型训练数据源，增加或者减少网络AI模型训练数据源时，则域管理功能单元配置网络AI模型训练数据源。
e)当配置信息包括E)关联的模型信息时，上述域管理功能单元根据所述配置信息进行模型训练功能配置，具体如下：
当Cross-Domain MnF需要修改，增加或者减少关联的网络AI模型时，则域管理功能单元配置所述网络AI模型信息。
f)当配置信息包括F)模型的类型信息时，上述域管理功能单元根据所述配置信息进行模型训练功能配置，具体如下：
当Cross-Domain MnF需要网络AI模型训练功能新增支持对某类网络AI模型的训练(比如负载信息分析模型)时，所述域管理功能单元在网络AI模型类型列表里增加需要新增的网络AI模型类型(即负载信息分析模型)。
当Cross-Domain MnF需要网络AI模型训练功能不再支持对某类网络AI模型的训练(比如节能分析模型)时，所述域管理功能单元在网络AI模型类型列表里删除所述网络AI模型类型(即节能分析模型)。
g)当配置信息包括G)模型训练功能关联的智能分析功能信息时，上述域管理功能单元根据所述配置信息进行模型训练功能配置，具体如下：
当Cross-Domain MnF需要修改，增加或者减少关联的智能分析功能时，则域管理功能单元在关联的智能分析功能信息里配置所述智能分析功能的信息。
上述仅以部分配置信息为例进行说明，其还可以包括其他配置信息，本方案对此不做具体限定。
作为一种可选的实现方式，Domain MnF配置所述网络AI模型训练功能的状态信息。其中，若在Domain MnF上进行模型训练，则直接在Domain MnF上进行上述状态信息的配置；若在网元上进行模型训练，则Domain MnF发送所述网络AI模型训练功能的状态信息给网元，以便在网元上进行所述网络AI模型训练功能的状态信息的配置。
作为一种可选的实现方式，Domain MnF配置所述网络AI模型训练功能支持训练的网络AI模型类型，若在Domain MnF上进行模型训练，则直接在Domain MnF上进行所述网络AI模型训练功能支持训练的网络AI模型类型的配置；若在网元上进行模型训练，则Domain MnF发送所述网络AI模型训练功能支持训练的网络AI模型类型给网元，以便在网元上进行所述网络AI模型训练功能支持训练的网络AI模型类型的配置。上述网络AI模型类型指示所述网络AI模型训练功能可以对所述类型的网络AI模型进行训练。
作为一种可选的实现方式，Domain MnF配置所述网络AI模型训练功能的网络AI模型训练触发信息，若在Domain MnF上进行模型训练，则直接在Domain MnF上进行所述网络AI模型训练功能的网络AI模型训练触发信息的配置；若在网元上进行模型训练，则Domain MnF发送所述网络AI模型训练触发信息给网元，以便在网元上进行所述网络AI模型训练功能的网络AI模型训练触发信息的配置。上述网络AI模型训练触发信息指示网络AI模型训练功能根据所述训练触发信息进行网络AI模型训练。
作为一种可选的实现方式，Domain MnF配置所述网络AI模型训练功能的训练数据，若在Domain MnF上进行模型训练，则直接在Domain MnF上进行所述训练数据的配置；若在网元上进行模型训练，则Domain MnF发送所述训练数据给网元，以便在网元上进行所述训练数据的配置。上述训练数据指示网络AI模型训练功能可以使用所述训练数据进行网络AI模型训练。
作为一种可选的实现方式，Domain MnF配置所述网络AI模型训练功能的训练数据源，若在Domain MnF上进行模型训练，则直接在Domain MnF上进行所述训练数据源的配置；若在网元上进行模型训练，则Domain MnF发送所述训练数据源给网元，以便在网元上进行所述训练数据源的配置。上述训练数据源指示网络AI模型训练功能可以使用所述数据源的数据进行网络AI模型训练。
作为一种可选的实现方式，Domain MnF配置所述网络AI模型训练功能的关联的智能分析功能信息，若在Domain MnF上进行模型训练，则直接在Domain MnF上进行所述网络AI模型训练功能的关联的智能分析功能信息的配置；若在网元上进行模型训练，则Domain MnF发送所述网络AI模型训练功能的关联的智能分析功能信息给网元，以便在网元上进行所述网络AI模型训练功能的关联的智能分析功能信息的配置。上述网络AI模型训练功能的关联的智能分析功能信息指示网络AI模型训练功能训练得到的网络AI模型可以给所述关联的智能分析功能使用。
作为一种可选的实现方式，Domain MnF配置所述网络AI模型训练功能的关联的网络AI模型信息，若在Domain MnF上进行模型训练，则直接在Domain MnF上进行所述网络AI模型训练功能的关联的网络AI模型信息的配置；若在网元上进行模型训练，则Domain MnF发送所述网络AI模型训练功能的关联的网络AI模型信息给网元，以便在网元上进行所述网络AI模型训练功能的关联的网络AI模型信息的配置。上述网络AI模型训练功能的关联的网络AI模型信息指示网络AI模型训练功能可以对所述网络AI模型进行更新或者重训练。
进一步地，所述第二网络管理单元还向所述第一网络管理单元发送配置结果。
采用该操作，以便第一网络管理单元知晓配置情况，若配置失败可再次进行配置操作。
4022、所述第二网络管理单元基于该配置后的模型训练功能进行模型训练，得到网络模型信息。
例如，当网络AI模型训练功能管控对象配置了网络AI模型训练周期，则相应周期时间达到了，Domain MnF启动网络AI模型训练。
当网络AI模型训练功能管控对象配置了网络AI模型训练触发条件，则当相应触发条件达到了，Domain MnF启动网络AI模型训练。
当网络AI模型训练功能管控对象配置了网络AI模型训练触发指示信息为True，则Domain MnF启动网络AI模型训练。可选的，当训练完成时，配置网络AI模型训练功能管控对象中的网络AI模型训练触发指示信息为False。
其中，上述网络模型信息包括以下中的至少一种：网络模型标识，网络模型版本，网络模型文件存放地址，网络模型文件名。
上述网络模型标识例如为网络模型1I。上述网络模型版本例如可以是v1.0.0。上述网络模型文件存放地址例如可以是统一资源标识符(Uniform Resource Identifier，URI)或者网际互连协议(Internet Protocol，IP)地址。上述网络模型文件名例如可以是网络模型文件的名称。
作为一种可选的实现方式，Domain MnF保存上述网络模型信息。
可选地，如果Domain MnF中不存在对应的网络AI模型对象，则Domain MnF先创建网 络AI模型对象，再在网络AI模型对象中配置所述网络AI模型信息。
如果网络AI模型训练功能管控对象中已经配置了网络AI模型对象信息，则在对应的网络AI模型对象中配置所述网络AI模型信息。
其中，Domain MnF在获取网络模型信息之后，配置网络模型对象的网络模型信息，以便后续使用该模型。
作为一种可选的实现方式，Domain MnF还向Cross-Domain MnF发送所述网络模型信息。例如，Domain MnF向Cross Domain MnF发送网络AI模型新增或者变更通知，该通知中可携带网络AI模型信息。通过该手段，以便Cross-Domain MnF可以实时获取训练输出的网络AI模型信息。
作为一种可选的实现方式，所述方法还包括：
所述第二网络管理单元确定所述配置信息和网络模型信息的关联信息；
所述第二网络管理单元向所述第一网络管理单元发送所述关联信息。
上述配置信息和网络模型信息的关联信息，可以理解为：上述网络模型信息是基于某个或某几个配置信息进行训练得到的，则该某个或某几个配置信息与该网络模型信息具备关联关系。
通过本申请实施例，第二网络管理单元基于接收到的第一网络管理单元发送的配置信息，对模型训练功能进行配置，以根据所述配置信息进行模型训练，得到网络模型信息。采用该手段，使能运营商可以灵活按需配置相应的网络AI模型训练功能，可以实现对网络模型训练的管控，提高了管控模型训练的效率，同时节省了人力。采用该方案，还可以使能运营商对所有支持的网络AI模型训练的统一管控，以及实现对支持的不同网络AI模型训练的差异化管控。
参照图5所示，为本申请实施例提供的又一种管控模型训练的方法的流程示意图。如图5所示，该方法应用于第二网络管理单元，其可包括步骤501-502，具体如下：
501、第二网络管理单元从第一网络管理单元接收配置信息，所述配置信息配置模型训练功能，所述配置信息包括以下信息中的至少一种：
A)状态信息，用于激活或去激活所述模型训练功能；
B)触发信息，用于触发进行模型训练；
C)数据信息，用于指示进行模型训练的数据；
可选的，上述第一网络管理单元可以是跨域管理功能单元Cross-Domain MnF。第二网络管理单元可以是域管理功能单元Domain MnF。其中，跨域管理功能单元Cross-Domain MnF确定网络AI模型的训练功能的配置信息。该网络AI模型的训练功能的配置信息可以是人工输入到Cross-Domain MnF，其也可以是Cross-Domain MnF内部计算分析生成的。
上述模型训练功能，可以是某一个模型的训练功能，也可以是多个模型的训练功能。
上述配置信息的介绍可参阅如图4所示实施例，在此不再赘述。
502、所述第二网络管理单元向网元发送所述配置信息，以便所述网元根据所述配置信息进行模型训练。
当在网元上进行模型训练时，域管理功能单元将所述配置信息发送给网元，进而网元基于上述配置信息进行模型训练功能配置，并进行模型训练。
作为一种可选的实现方式，所述第二网络管理单元接收所述网元发送的网络模型信息。其中，上述网络模型信息包括以下中的至少一种：网络模型标识，网络模型版本，网络模型 文件存放地址，网络模型文件名。
作为一种可选的实现方式，Domain MnF保存上述网络模型信息。
可选地，如果Domain MnF中不存在对应的网络AI模型对象，则Domain MnF先创建网络AI模型对象，再在网络AI模型对象中配置所述网络AI模型信息。
如果网络AI模型训练功能管控对象中已经配置了网络AI模型对象信息，则在对应的网络AI模型对象中配置所述网络AI模型信息。
其中，Domain MnF在获取网络模型信息之后，配置网络模型对象的网络模型信息，以便使用该模型。
作为一种可选的实现方式，Domain MnF还向Cross-Domain MnF发送所述网络模型信息。例如，Domain MnF向Cross Domain MnF发送网络AI模型新增或者变更通知，该通知中可携带所述网络AI模型信息。通过该手段，以便Cross-Domain MnF可以实时获取训练输出的网络AI模型信息。
作为一种可选的实现方式，所述方法还包括：
所述第二网络管理单元确定所述配置信息和网络模型信息的关联信息；
所述第二网络管理单元向所述第一网络管理单元发送所述关联信息。
上述配置信息和网络模型信息的关联信息，可以理解为：上述网络模型信息是基于某个或某几个配置信息进行训练得到的，则该某个或某几个配置信息与该网络模型信息具备关联关系。
通过本申请实施例，第二网络管理单元基于接收到的第一网络管理单元发送的配置信息，并向网元发送所述配置信息，以便所述网元根据所述配置信息进行模型训练。采用该手段，可以实现对网络模型训练的管控，使能运营商可以灵活按需配置相应的网络AI模型训练功能。
参照图6所示，为本申请实施例提供的另一种管控模型训练的方法的流程示意图。如图6所示，该方法应用于第一网络管理单元，其可包括步骤601-602，具体如下：
601、第一网络管理单元确定配置信息，所述配置信息配置模型训练功能，所述配置信息包括以下信息中的至少一种：
A)状态信息，用于激活或去激活所述模型训练功能；
B)触发信息，用于触发进行模型训练；
C)数据信息，用于指示进行模型训练的数据；
可选的，上述第一网络管理单元可以是跨域管理功能单元Cross-Domain MnF。第二网络管理单元可以是域管理功能单元Domain MnF。其中，跨域管理功能单元Cross-Domain MnF确定网络AI模型的训练功能的配置信息。该网络AI模型的训练功能的配置信息可以是人工输入到Cross-Domain MnF，其也可以是Cross-Domain MnF内部计算分析生成的。
上述模型训练功能，可以是某一个模型的训练功能，也可以是多个模型的训练功能。
其中，上述配置信息的相关介绍可参阅前述图4的描述，在此不再赘述。
602、所述第一网络管理单元向第二网络管理单元发送所述配置信息。
作为一种可选的实现方式，所述方法还包括：
所述第一网络管理单元接收所述第二网络管理单元发送的网络模型信息，所述网络模型信息为所述第二网络管理单元根据所述配置信息训练后得到的。
作为另一种可选的实现方式，所述方法还包括：
所述第一网络管理单元接收所述第二网络管理单元发送的网络模型信息，所述网络模型信息为网元根据所述配置信息训练后得到的，所述配置信息为所述网元从所述第二网络管理单元接收到的。
也就是说，上述模型训练可以是第二网络管理单元进行训练得到的，其也可以是网元进行训练得到的。
其中，所述网络模型信息包括以下中的至少一种：网络模型标识，网络模型版本，网络模型文件存放地址，网络模型文件名。
作为又一种可选的实现方式，所述方法还包括：
所述第一网络管理单元接收所述第二网络管理单元发送的关联信息，其中，所述关联信息为所述配置信息和所述网络模型信息的关联信息。
通过本申请实施例，第一网络管理单元确定模型训练功能的配置信息，并将该配置信息发送给第二网络管理单元。采用该手段，可以实现对网络模型训练的管控，使能运营商可以灵活按需配置相应的网络AI模型训练功能。
参照图7所示，为本申请实施例提供的一种管控模型训练的方法的流程示意图。如图7所示，该方法可包括步骤701-705，具体如下：
701、第一网络管理单元确定第一模型的配置信息，所述配置信息配置所述第一模型的模型训练功能，所述配置信息包括以下信息中的至少一种：
A)状态信息，用于激活或去激活所述模型训练功能；
B)触发信息，用于触发进行模型训练；
C)数据信息，用于指示进行模型训练的数据；
上述第一模型可以是任意种类的模型。可选的，如通过采用区分名(Distinguish Name，DN)来识别上述第一模型。其还可以采用其他方式，本方案对此不做具体限定。
702、所述第一网络管理单元向第二网络管理单元发送所述第一模型的配置信息。
703、所述第一网络管理单元接收所述第二网络管理单元发送的配置结果。
例如，Domain MnF返回上述第一网络的模型训练功能的配置结果。采用该操作，以便第一网络管理单元知晓配置情况，若配置失败可再次进行配置操作。
704、第二网络管理单元根据所述配置信息进行模型训练，得到所述第一模型的模型信息。
705、所述第一网络管理单元接收所述第二网络管理单元发送的所述第一模型的模型信息。
通过本申请实施例，第一网络管理单元可以灵活控制特进行定网络AI模型训练，同时能实时获取训练输出的网络AI模型信息。采用该手段，使能运营商差异化控制网络AI模型训练功能上的不同网络AI模型的训练。
参照图8所示，为本申请实施例提供的一种管控模型训练的装置的结构示意图。如图8所示，该装置包括接收模块801和处理模块802，具体如下:
接收模块801，用于从第一网络管理单元接收配置信息，所述配置信息配置模型训练功能，所述配置信息包括以下信息中的至少一种：
状态信息，用于激活或去激活所述模型训练功能；
触发信息，用于触发进行模型训练；
数据信息，用于指示进行模型训练的数据；
处理模块802，用于根据所述配置信息进行模型训练，得到网络模型信息。
作为一种可选的实现方式，所述配置信息还包括以下中的至少一种：
数据源信息，用于指示进行模型训练的数据来源；
所述模型训练功能关联的模型信息，用于指示进行模型训练的模型；
模型的类型信息，用于指示进行模型训练的模型类型。
作为一种可选的实现方式，所述网络模型信息包括以下中的至少一种：网络模型标识，网络模型版本，网络模型文件存放地址，网络模型文件名。
作为一种可选的实现方式，所述装置还包括：
发送模块，用于向所述第一网络管理单元发送所述网络模型信息。
作为另一种可选的实现方式，所述装置还包括：
确定模块，用于确定所述配置信息和网络模型信息之间的关联信息，并向所述第一网络管理单元发送所述关联信息。
其中，所述触发信息包括以下中的至少一种：
训练周期，用于指示模型训练周期；
训练时间，用于指示模型训练的时间；
训练指示信息，用于指示开始进行模型训练。
作为一种可选的实现方式，所述数据信息包括：
输入数据，用于指示模型的输入数据类型；
输出数据，用于指示模型的输出数据类型。
上述各模块的具体实现方式可参阅前述实施例，在此不再赘述。
参照图9所示，为本申请实施例提供的另一种管控模型训练的装置的结构示意图。如图9所示，该装置包括确定模块901和发送模块902，具体如下:
确定模块901，用于确定配置信息，所述配置信息配置模型训练功能，所述配置信息包括以下信息中的至少一种：
状态信息，用于激活或去激活所述模型训练功能；
触发信息，用于触发进行模型训练；
数据信息，用于指示进行模型训练的数据；
发送模块902，用于向第二网络管理单元发送所述配置信息。
作为一种可选的实现方式，所述配置信息还包括以下中的至少一种：
数据源信息，用于指示进行模型训练的数据来源；
所述模型训练功能关联的模型信息，用于指示进行模型训练的模型；
模型的类型信息，用于指示进行模型训练的模型类型。
作为一种可选的实现方式，所述装置还包括：
接收模块，用于接收所述第二网络管理单元发送的网络模型信息，所述网络模型信息为所述第二网络管理单元根据所述配置信息训练后得到的；或者，所述网络模型信息为网元根据所述配置信息训练后得到的，所述配置信息为所述网元从所述第二网络管理单元接收到的。
作为一种可选的实现方式，所述网络模型信息包括以下中的至少一种：网络模型标识，网络模型版本，网络模型文件存放地址，网络模型文件名。
作为另一种可选的实现方式，所述接收模块还用于：
接收所述第二网络管理单元发送的关联信息，其中，所述关联信息为所述配置信息和所述网络模型信息的关联信息。
作为另一种可选的实现方式，所述触发信息包括以下中的至少一种：
训练周期，用于指示模型训练周期；
训练时间，用于指示模型训练的时间；
训练指示信息，用于指示开始进行模型训练。
作为又一种可选的实现方式，所述数据信息包括：
输入数据，用于指示模型的输入数据类型；
输出数据，用于指示模型的输出数据类型。
上述各模块的具体实现方式可参阅前述实施例，在此不再赘述。
参照图10所示，为本申请实施例提供的又一种管控模型训练的装置的结构示意图。如图10所示，该管控模型训练的装置1000包括至少一个处理器1001，至少一个存储器1002以及至少一个通信接口1003。所述处理器1001、所述存储器1002和所述通信接口1003通过所述通信总线连接并完成相互间的通信。
处理器1001可以是通用中央处理器(CPU)，微处理器，特定应用集成电路(application-specific integrated circuit，ASIC)，或一个或多个用于控制以上方案程序执行的集成电路。
通信接口1003，用于与其他设备或通信网络通信，如以太网，无线接入网(RAN)，无线局域网(Wireless Local Area Networks，WLAN)等。
存储器1002可以是只读存储器(read-only memory，ROM)或可存储静态信息和指令的其他类型的静态存储设备，随机存取存储器(random access memory，RAM)或者可存储信息和指令的其他类型的动态存储设备，也可以是电可擦可编程只读存储器(Electrically Erasable Programmable Read-Only Memory，EEPROM)、只读光盘(Compact Disc Read-Only Memory，CD-ROM)或其他光盘存储、光碟存储(包括压缩光碟、激光碟、光碟、数字通用光碟、蓝光光碟等)、磁盘存储介质或者其他磁存储设备、或者能够用于携带或存储具有指令或数据结构形式的期望的程序代码并能够由计算机存取的任何其他介质，但不限于此。存储器可以是独立存在，通过总线与处理器相连接。存储器也可以和处理器集成在一起。
其中，所述存储器1002用于存储执行以上方案的应用程序代码，并由处理器1001来控制执行。所述处理器1001用于执行所述存储器1002中存储的应用程序代码。
存储器1002存储的代码可执行以上提供的一种管控模型训练方法。
本申请实施例还提供一种芯片系统，所述芯片系统应用于电子设备；所述芯片系统包括一个或多个接口电路，以及一个或多个处理器；所述接口电路和所述处理器通过线路互联；所述接口电路用于从所述电子设备的存储器接收信号，并向所述处理器发送所述信号，所述信号包括所述存储器中存储的计算机指令；当所述处理器执行所述计算机指令时，所述电子设备执行所述方法。
本申请实施例还提供了一种计算机可读存储介质，该计算机可读存储介质中存储有指令，当其在计算机或处理器上运行时，使得计算机或处理器执行上述任一个方法中的一个或多个步骤。
本申请实施例还提供了一种包含指令的计算机程序产品。当该计算机程序产品在计算机或处理器上运行时，使得计算机或处理器执行上述任一个方法中的一个或多个步骤。
本申请实施例还提供一种管控模型训练的系统，包括上述任一个装置中的一个或多个模块。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统、装置 和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
应理解，在本申请的描述中，除非另有说明，“/”表示前后关联的对象是一种“或”的关系，例如，A/B可以表示A或B；其中A，B可以是单数或者复数。并且，在本申请的描述中，除非另有说明，“多个”是指两个或多于两个。“以下至少一项(个)”或其类似表达，是指的这些项中的任意组合，包括单项(个)或复数项(个)的任意组合。例如，a，b，或c中的至少一项(个)，可以表示：a，b，c，a-b，a-c，b-c，或a-b-c，其中a，b，c可以是单个，也可以是多个。另外，为了便于清楚描述本申请实施例的技术方案，在本申请的实施例中，采用了“第一”、“第二”等字样对功能和作用基本相同的相同项或相似项进行区分。本领域技术人员可以理解“第一”、“第二”等字样并不对数量和执行次序进行限定，并且“第一”、“第二”等字样也并不限定一定不同。同时，在本申请实施例中，“示例性的”或者“例如”等词用于表示作例子、例证或说明。本申请实施例中被描述为“示例性的”或者“例如”的任何实施例或设计方案不应被解释为比其它实施例或设计方案更优选或更具优势。确切而言，使用“示例性的”或者“例如”等词旨在以具体方式呈现相关概念，便于理解。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统、装置和方法，可以通过其它的方式实现。例如，该单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如，多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。所显示或讨论的相互之间的耦合、或直接耦合、或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。该计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行该计算机程序指令时，全部或部分地产生按照本申请实施例的流程或功能。该计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。该计算机指令可以存储在计算机可读存储介质中，或者通过该计算机可读存储介质进行传输。该计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线(digital subscriber line，DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。该计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。该可用介质可以是只读存储器(read-only memory，ROM)，或随机存取存储器(random access memory，RAM)，或磁性介质，例如，软盘、硬盘、磁带、磁碟、或光介质，例如，数字通用光盘(digital versatile disc，DVD)、或者半导体介质，例如，固态硬盘(solid state disk，SSD)等。
以上所述，仅为本申请实施例的具体实施方式，但本申请实施例的保护范围并不局限于此，任何在本申请实施例揭露的技术范围内的变化或替换，都应涵盖在本申请实施例的保护范围之内。因此，本申请实施例的保护范围应以所述权利要求的保护范围为准。

Claims (33)

  1. 一种管控模型训练的方法，其特征在于，包括:
    从第一网络管理单元接收配置信息，所述配置信息配置模型训练功能，所述配置信息包括以下信息中的至少一种：
    状态信息，用于激活或去激活所述模型训练功能；
    触发信息，用于触发进行模型训练；
    数据信息，用于指示进行模型训练的数据；
    根据所述配置信息进行模型训练，得到网络模型信息。
  2. 根据权利要求1所述的方法，其特征在于，所述配置信息还包括以下中的至少一种：
    数据源信息，用于指示进行模型训练的数据来源；
    所述模型训练功能关联的模型信息，用于指示进行模型训练的模型；
    模型的类型信息，用于指示进行模型训练的模型类型。
  3. 根据权利要求1或2所述的方法，其特征在于，所述网络模型信息包括以下中的至少一种：网络模型标识，网络模型版本，网络模型文件存放地址，网络模型文件名。
  4. 根据权利要求1至3任一项所述的方法，其特征在于，所述方法还包括：
    向所述第一网络管理单元发送所述网络模型信息。
  5. 根据权利要求1至4任一项所述的方法，其特征在于，所述方法还包括：
    确定所述配置信息和网络模型信息之间的关联信息；
    向所述第一网络管理单元发送所述关联信息。
  6. 根据权利要求1至5任一项所述的方法，其特征在于，所述触发信息包括以下中的至少一种：
    训练周期，用于指示模型训练周期；
    训练时间，用于指示模型训练的时间；
    训练指示信息，用于指示开始进行模型训练。
  7. 根据权利要求1至6任一项所述的方法，其特征在于，所述数据信息包括：
    输入数据，用于指示模型的输入数据类型；
    输出数据，用于指示模型的输出数据类型。
  8. 一种管控模型训练的方法，其特征在于，包括:
    确定配置信息，所述配置信息配置模型训练功能，所述配置信息包括以下信息中的至少一种：
    状态信息，用于激活或去激活所述模型训练功能；
    触发信息，用于触发进行模型训练；
    数据信息，用于指示进行模型训练的数据；
    向第二网络管理单元发送所述配置信息。
  9. 根据权利要求8所述的方法，其特征在于，所述配置信息还包括以下中的至少一种：
    数据源信息，用于指示进行模型训练的数据来源；
    所述模型训练功能关联的模型信息，用于指示进行模型训练的模型；
    模型的类型信息，用于指示进行模型训练的模型类型。
  10. 根据权利要求8或9所述的方法，其特征在于，所述方法还包括：
    接收所述第二网络管理单元发送的网络模型信息，所述网络模型信息为所述第二网络管理单元根据所述配置信息训练后得到的；或者，所述网络模型信息为网元根据所述配置信息训练后得到的，所述配置信息为所述网元从所述第二网络管理单元接收到的。
  11. 根据权利要求10所述的方法，其特征在于，所述网络模型信息包括以下中的至少一种：网络模型标识，网络模型版本，网络模型文件存放地址，网络模型文件名。
  12. 根据权利要求10或11所述的方法，其特征在于，所述方法还包括：
    接收所述第二网络管理单元发送的关联信息，其中，所述关联信息为所述配置信息和所述网络模型信息的关联信息。
  13. 根据权利要求8至12任一项所述的方法，其特征在于，所述触发信息包括以下中的至少一种：
    训练周期，用于指示模型训练周期；
    训练时间，用于指示模型训练的时间；
    训练指示信息，用于指示开始进行模型训练。
  14. 根据权利要求8至13任一项所述的方法，其特征在于，所述数据信息包括：
    输入数据，用于指示模型的输入数据类型；
    输出数据，用于指示模型的输出数据类型。
  15. 一种管控模型训练的装置，其特征在于，包括:
    接收模块，用于从第一网络管理单元接收配置信息，所述配置信息配置模型训练功能，所述配置信息包括以下信息中的至少一种：
    状态信息，用于激活或去激活所述模型训练功能；
    触发信息，用于触发进行模型训练；
    数据信息，用于指示进行模型训练的数据；
    处理模块，用于根据所述配置信息进行模型训练，得到网络模型信息。
  16. 根据权利要求15所述的装置，其特征在于，所述配置信息还包括以下中的至少一种：
    数据源信息，用于指示进行模型训练的数据来源；
    所述模型训练功能关联的模型信息，用于指示进行模型训练的模型；
    模型的类型信息，用于指示进行模型训练的模型类型。
  17. 根据权利要求15或16所述的装置，其特征在于，所述网络模型信息包括以下中的至少一种：网络模型标识，网络模型版本，网络模型文件存放地址，网络模型文件名。
  18. 根据权利要求15至17任一项所述的装置，其特征在于，所述装置还包括：
    发送模块，用于向所述第一网络管理单元发送所述网络模型信息。
  19. 根据权利要求15至18任一项所述的装置，其特征在于，所述装置还包括：
    确定模块，用于确定所述配置信息和网络模型信息之间的关联信息，并向所述第一网络管理单元发送所述关联信息。
  20. 根据权利要求15至19任一项所述的装置，其特征在于，所述触发信息包括以下中的至少一种：
    训练周期，用于指示模型训练周期；
    训练时间，用于指示模型训练的时间；
    训练指示信息，用于指示开始进行模型训练。
  21. 根据权利要求15至20任一项所述的装置，其特征在于，所述数据信息包括：
    输入数据，用于指示模型的输入数据类型；
    输出数据，用于指示模型的输出数据类型。
  22. 一种管控模型训练的装置，其特征在于，包括:
    确定模块，用于确定配置信息，所述配置信息配置模型训练功能，所述配置信息包括以下信息中的至少一种：
    状态信息，用于激活或去激活所述模型训练功能；
    触发信息，用于触发进行模型训练；
    数据信息，用于指示进行模型训练的数据；
    发送模块，用于向第二网络管理单元发送所述配置信息。
  23. 根据权利要求22所述的装置，其特征在于，所述配置信息还包括以下中的至少一种：
    数据源信息，用于指示进行模型训练的数据来源；
    所述模型训练功能关联的模型信息，用于指示进行模型训练的模型；
    模型的类型信息，用于指示进行模型训练的模型类型。
  24. 根据权利要求22或23所述的装置，其特征在于，所述装置还包括：
    接收模块，用于接收所述第二网络管理单元发送的网络模型信息，所述网络模型信息为所述第二网络管理单元根据所述配置信息训练后得到的；或者，所述网络模型信息为网元根据所述配置信息训练后得到的，所述配置信息为所述网元从所述第二网络管理单元接收到的。
  25. 根据权利要求24所述的装置，其特征在于，所述网络模型信息包括以下中的至少一 种：网络模型标识，网络模型版本，网络模型文件存放地址，网络模型文件名。
  26. 根据权利要求24或25所述的装置，其特征在于，所述接收模块还用于：
    接收所述第二网络管理单元发送的关联信息，其中，所述关联信息为所述配置信息和所述网络模型信息的关联信息。
  27. 根据权利要求22至26任一项所述的装置，其特征在于，所述触发信息包括以下中的至少一种：
    训练周期，用于指示模型训练周期；
    训练时间，用于指示模型训练的时间；
    训练指示信息，用于指示开始进行模型训练。
  28. 根据权利要求22至27任一项所述的装置，其特征在于，所述数据信息包括：
    输入数据，用于指示模型的输入数据类型；
    输出数据，用于指示模型的输出数据类型。
  29. 一种管控模型训练的装置，其特征在于，包括处理器和存储器；其中，所述存储器用于存储程序代码，所述处理器用于调用所述程序代码，以执行如权利要求1至7任意一项所述的方法和/或8至14任意一项所述的方法。
  30. 一种计算机可读存储介质，其特征在于，所述计算机可读存储介质存储有计算机程序，所述计算机程序被处理器执行以实现权利要求1至7任意一项所述的方法和/或8至14任意一项所述的方法。
  31. 一种计算机程序产品，其特征在于，当计算机程序产品在计算机上运行时，使得所述计算机执行如权利要求1至7任意一项所述的方法和/或8至14任意一项所述的方法。
  32. 一种管控模型训练的系统，其特征在于，包括如权利要求15至21中任一项所述的装置和如权利要求22至28中任一项所述的装置。
  33. 一种管控模型训练的方法，其特征在于，包括：
    第一网络管理单元确定配置信息，所述配置信息配置模型训练功能，所述配置信息包括以下信息中的至少一种：
    状态信息，用于激活或去激活所述模型训练功能；
    触发信息，用于触发进行模型训练；
    数据信息，用于指示进行模型训练的数据；
    所述第一网络管理单元发送所述配置信息；
    第二网络管理单元接收所述配置信息；
    所述第二网络管理单元根据所述配置信息进行模型训练，得到网络模型信息。
PCT/CN2022/082872 2021-03-30 2022-03-24 管控模型训练的方法及装置、系统 Ceased WO2022206567A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP22778750.4A EP4311171A4 (en) 2021-03-30 2022-03-24 Method and apparatus for training management and control model, and system
US18/478,640 US12506663B2 (en) 2021-03-30 2023-09-29 Method, apparatus, and system for training management and control model

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110343727.XA CN115146691A (zh) 2021-03-30 2021-03-30 管控模型训练的方法及装置、系统
CN202110343727.X 2021-03-30

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US18/478,640 Continuation US12506663B2 (en) 2021-03-30 2023-09-29 Method, apparatus, and system for training management and control model

Publications (1)

Publication Number Publication Date
WO2022206567A1 true WO2022206567A1 (zh) 2022-10-06

Family

ID=83403916

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/082872 Ceased WO2022206567A1 (zh) 2021-03-30 2022-03-24 管控模型训练的方法及装置、系统

Country Status (4)

Country Link
US (1) US12506663B2 (zh)
EP (1) EP4311171A4 (zh)
CN (1) CN115146691A (zh)
WO (1) WO2022206567A1 (zh)

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20230403588A1 (en) * 2022-06-10 2023-12-14 Qualcomm Incorporated Machine learning data collection, validation, and reporting configurations
CN118042476A (zh) * 2022-11-01 2024-05-14 华为技术有限公司 切换或更新ai模型的方法以及通信装置
CN117993516A (zh) * 2022-11-04 2024-05-07 华为技术有限公司 一种通信方法及装置
CN118041764A (zh) * 2022-11-11 2024-05-14 华为技术有限公司 一种机器学习的管控方法和装置
CN116095750B (zh) * 2023-01-13 2023-10-31 广州爱浦路网络技术有限公司 数据面转发方法、装置、电子设备及可读存储介质
CN118520282A (zh) * 2023-02-17 2024-08-20 华为技术有限公司 一种通信方法和装置
CN118695273A (zh) * 2023-03-24 2024-09-24 华为技术有限公司 一种通信方法及装置
EP4690597A1 (en) * 2023-04-05 2026-02-11 QUALCOMM Incorporated Meta information signaling for network devices
CN118945670A (zh) * 2023-05-12 2024-11-12 华为技术有限公司 一种通信方法和装置
WO2024234386A1 (zh) * 2023-05-18 2024-11-21 Oppo广东移动通信有限公司 信息配置方法和设备
WO2025079100A1 (en) * 2023-10-11 2025-04-17 Jio Platforms Limited System and method for identifying customers for migrating from existing network to new network
CN119907131A (zh) * 2023-10-26 2025-04-29 华为技术有限公司 一种通信方法、相关设备及系统
WO2025137801A1 (zh) * 2023-12-25 2025-07-03 北京小米移动软件有限公司 通信方法、终端、网络设备及存储介质
GB2638222A (en) * 2024-02-16 2025-08-20 Nokia Technologies Oy Capturing additional conditions

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105809248A (zh) * 2016-03-01 2016-07-27 中山大学 一种分布式人工神经网络在sdn上的配置和交互方法
EP3418821A1 (en) * 2017-06-19 2018-12-26 Nokia Technologies Oy Method and device for configuring a data transmission system
CN109460396A (zh) * 2018-10-12 2019-03-12 中国平安人寿保险股份有限公司 模型处理方法及装置、存储介质和电子设备
CN109934512A (zh) * 2019-03-28 2019-06-25 努比亚技术有限公司 一种预测模型的训练方法及系统
CN112534864A (zh) * 2018-07-19 2021-03-19 诺基亚技术有限公司 用于认知功能的网络状态的环境建模和抽象

Family Cites Families (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110121180B (zh) * 2018-02-05 2020-12-15 华为技术有限公司 一种数据分析装置、系统及方法
CN113159145B (zh) * 2018-04-28 2024-10-22 华为技术有限公司 一种特征工程编排方法及装置
US11026103B2 (en) * 2019-05-31 2021-06-01 At&T Intellectual Property I, L.P. Machine learning deployment in radio access networks
US20210012239A1 (en) * 2019-07-12 2021-01-14 Microsoft Technology Licensing, Llc Automated generation of machine learning models for network evaluation
CN112396070B (zh) * 2019-08-13 2025-07-18 中兴通讯股份有限公司 一种模型训练方法、装置和系统，以及预测方法和装置
CN110569288A (zh) * 2019-09-11 2019-12-13 中兴通讯股份有限公司 一种数据分析方法、装置、设备和存储介质
CN111126621B (zh) * 2019-12-17 2021-02-09 北京九章云极科技有限公司 在线模型训练方法及装置
CN111126626A (zh) * 2019-12-24 2020-05-08 上海器魂智能科技有限公司 一种训练方法、装置、服务器、平台及存储介质
CN111222563B (zh) * 2020-01-02 2023-10-10 腾讯科技（深圳）有限公司 一种模型训练方法、数据获取方法以及相关装置
CN111327692A (zh) * 2020-02-05 2020-06-23 北京百度网讯科技有限公司 模型训练方法、装置及集群系统
CN111399958B (zh) * 2020-03-17 2023-04-28 青岛创新奇智科技集团股份有限公司 具有用户交互界面的模型训练系统、方法
CN112512058B (zh) * 2020-05-24 2025-07-25 中兴通讯股份有限公司 网络优化方法、服务器、客户端设备、网络设备和介质
EP4179754A1 (en) * 2020-07-13 2023-05-17 Telefonaktiebolaget LM Ericsson (publ) Managing a wireless device that is operable to connect to a communication network
CN116034381A (zh) * 2020-11-06 2023-04-28 华为技术有限公司 通信方法和通信装置
US20240306011A1 (en) * 2021-01-08 2024-09-12 Lenovo (Beijing) Limited Method and apparatus for determining prediction for status of wireless network
US11818806B2 (en) * 2021-05-18 2023-11-14 Qualcomm Incorporated ML model training procedure

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105809248A (zh) * 2016-03-01 2016-07-27 中山大学 一种分布式人工神经网络在sdn上的配置和交互方法
EP3418821A1 (en) * 2017-06-19 2018-12-26 Nokia Technologies Oy Method and device for configuring a data transmission system
CN112534864A (zh) * 2018-07-19 2021-03-19 诺基亚技术有限公司 用于认知功能的网络状态的环境建模和抽象
CN109460396A (zh) * 2018-10-12 2019-03-12 中国平安人寿保险股份有限公司 模型处理方法及装置、存储介质和电子设备
CN109934512A (zh) * 2019-03-28 2019-06-25 努比亚技术有限公司 一种预测模型的训练方法及系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP4311171A4

Also Published As

Publication number Publication date
EP4311171A4 (en) 2024-08-21
EP4311171A1 (en) 2024-01-24
CN115146691A (zh) 2022-10-04
US20240031240A1 (en) 2024-01-25
US12506663B2 (en) 2025-12-23

Similar Documents

Publication Publication Date Title
WO2022206567A1 (zh) 管控模型训练的方法及装置、系统
US20230095706A1 (en) Radio access network control with deep reinforcement learning
EP4099635A1 (en) Method and device for selecting service in wireless communication system
US20220311671A1 (en) Network intent synthesis
CN114402654B (zh) 用于无线电接入网数据收集的装置
US11672001B2 (en) Systems and methods for interference management in a radio access network
US11647457B2 (en) Systems and methods for performance-aware energy saving in a radio access network
US11706642B2 (en) Systems and methods for orchestration and optimization of wireless networks
US20260040149A1 (en) Methods and mechanisms for slice coverage enhancement
US11638171B2 (en) Systems and methods for dynamic wireless network configuration based on mobile radio unit location
US20250219923A1 (en) Data analysis apparatus management and control method and communication apparatus
US20160192202A1 (en) Methods And Apparatus For Small Cell Deployment In Wireless Network
US12207130B2 (en) Systems and methods for network design and configuration based on user-level usage modeling
US20250294433A1 (en) Network data analytics function slice load enhancement
US12581333B2 (en) Systems and methods for granular network configuration via network exposure function
US12549996B2 (en) Cell outage compensation optimization
US20250208909A1 (en) Actions application programming interface for wireless network configuration
US20260059343A1 (en) Systems and methods for automated deployment of cell sites in a wireless network
US20260101248A1 (en) Systems and methods for ai/ml-based handover thresholds for heterogeneous wireless networks
US20260121935A1 (en) Cellular network operator-defined network data analytics enhancement
US20260006457A1 (en) Digital twin enablers for open radio access networks
CN119233279A (zh) 一种自治功能管理方法及装置
Alotaibi et al. A distributed approach to improving epc controller performance
Kakadia et al. Self organizing networks (SON)
Auroux et al. Flexible reassignment of flow processing-aware controllers in future wireless networks

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: 22778750

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 2022778750

Country of ref document: EP

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2022778750

Country of ref document: EP

Effective date: 20231019