Model life cycle management method and device, communication equipment and storage medium

Abstract

The invention relates to a model life cycle management method and device, communication equipment, a storage medium and a computer program product. The method is applied to a first network device, and the method comprises the following steps: receiving a signaling sent by a second network device; wherein the signaling at least comprises model configuration information; according to the model configuration information, executing an artificial intelligence model training and / or data acquisition process; and sending a model configuration response message to the second network device based on an execution result of the artificial intelligence model training and / or data acquisition process. By adopting the method, the artificial intelligence model can be deployed on the wireless access network side, the spectrum efficiency, the user experience rate and the positioning precision can be improved, the time delay is reduced, the security privacy protection is improved, and the AI related indexes are optimized.

Description

Technical Field

[0001] The present application relates to the field of wireless communication technology, and in particular to a model lifecycle management method, apparatus, communication equipment, storage medium and computer program product. Background Art

[0002] With the development of wireless communication technology, artificial intelligence native (AI native) design has emerged. The integration of artificial intelligence (AI) and communication technology will affect the relevant capability indicators in the key performance indicator (KPI) capability indicator system. For example, AI native can improve spectrum efficiency, user experience rate and positioning accuracy, reduce latency, enhance security and privacy protection, and optimize AI-related indicators. AI native is a necessary and important technical feature to meet and improve multiple capability indicators in future mobile communication systems.

[0003] Therefore, there is an urgent need for a model lifecycle management method for deploying artificial intelligence models on the Radio Access Network (RAN) side. Summary of the invention

[0004] The embodiments of the present application provide a model lifecycle management method, apparatus, communication equipment, storage medium and computer program product, which can deploy an artificial intelligence model on the wireless access network side.

[0005] A model lifecycle management method is applied to a first network device, and the method includes:

[0006] Receiving a signaling sent by a second network device; wherein the signaling at least includes model configuration information;

[0007] Execute artificial intelligence model training and / or data collection process according to the model configuration information;

[0008] Based on the execution results of the artificial intelligence model training and / or data collection process, a model configuration response message is sent to the second network device.

[0009] In one of the embodiments, the model configuration information includes at least one of indication information and training configuration information; the indication information is used to indicate whether to perform artificial intelligence model training; the training configuration information includes at least one of the model construction geographical scope, storage space information, computing power information, model structure configuration information, training parameter information and data collection configuration information.

[0010] In one of the embodiments, the storage space information represents the storage space required for the artificial intelligence model training and / or data collection process; the computing power information represents the computing power required for the artificial intelligence model training and / or data collection process.

[0011] In one of the embodiments, the model structure configuration information includes at least one of pre-training indication information, model type and model structure parameters; the pre-training indication information indicates whether it is a pre-training model.

[0012] In one of the embodiments, the training parameter information includes at least one of a data set size, a loss function type, a training cycle, a deployment condition, a maximum number of model training times, and a training interval.

[0013] In one of the embodiments, the data collection configuration information includes at least one of a data set number, a data set type, a data set function, a data collection pilot signal configuration, a data collection period, a data collection sample type, a data collection accuracy, and a data collection amount.

[0014] In one of the embodiments, when the model configuration information includes data collection configuration information, performing artificial intelligence model training and / or data collection process according to the model configuration information includes at least one of the following:

[0015] Execute a data collection process based on the data collection configuration information; and,

[0016] Acquire training sample data, and train the artificial intelligence model based on the training sample data and the model configuration information; the training sample data includes at least one of the collected sample data and / or pre-stored sample data.

[0017] In one embodiment, the obtaining of training sample data includes at least one of the following:

[0018] If the data volume of the pre-stored sample data is less than a preset data volume threshold, the collected sample data is obtained, and when the total data volume of the collected sample data and the pre-stored sample data reaches the data volume threshold, the step of training the artificial intelligence model based on the training sample data and the model configuration information is executed;

[0019] If the data volume of the pre-stored sample data is greater than or equal to the preset data volume threshold, the pre-stored sample data is used as the training sample data.

[0020] In one embodiment, the obtaining of training sample data includes:

[0021] After the data collection is completed, the collected sample data is used as training sample data.

[0022] In one of the embodiments, the sending of a model configuration response message to the second network device based on the execution result of the artificial intelligence model training and / or data collection process includes at least one of the following:

[0023] If the artificial intelligence model training or the data collection process fails to start, sending a startup failure response message to the second network device;

[0024] If the artificial intelligence model training is successfully started, sending a model training start response message to the second network device;

[0025] If the artificial intelligence model training is completed, a model training completion message is sent to the second network device.

[0026] In one of the embodiments, after sending the model training completion message to the second network device, the method further includes at least one of the following:

[0027] Execute a model deployment process on the trained artificial intelligence model;

[0028] Receive the model deployment message sent by the second network device, and execute the model deployment process on the trained artificial intelligence model.

[0029] In one embodiment, the method further comprises:

[0030] When the artificial intelligence model training fails, the cumulative model training times of the artificial intelligence model are determined.

[0031] In one embodiment, the method further comprises at least one of the following:

[0032] If the cumulative model training times are less than the maximum model training times, the artificial intelligence model is retrained and the cumulative model training times are updated;

[0033] If the cumulative number of model training times is greater than or equal to the maximum number of model training times, the model training of the artificial intelligence model is stopped, the model training timer is started, and the artificial intelligence model is retrained when the timing length of the model training timer is greater than or equal to the training interval.

[0034] In one embodiment, before retraining the artificial intelligence model, the method further includes:

[0035] Update the artificial intelligence model according to a preset model update strategy.

[0036] In one embodiment, updating the artificial intelligence model according to a preset model update strategy includes at least one of the following:

[0037] According to a preset model update strategy, the model structure of the artificial intelligence model is changed;

[0038] According to a preset model update strategy, changing the impulse function or loss function of the artificial intelligence model;

[0039] According to a preset model update strategy, the model type of the artificial intelligence model is changed.

[0040] In one embodiment, before retraining the artificial intelligence model, the method further includes:

[0041] The training sample data is changed according to a preset training sample data update strategy.

[0042] In one of the embodiments, the first network device includes at least one of a terminal, a base station, a main control board, an intelligent computing board, a computing power board, a baseband board and a wireless access network element.

[0043] In one of the embodiments, the signaling includes at least one of user plane signaling, intelligent plane signaling, data plane signaling, computing plane signaling, radio resource control signaling, media access control-control element, data center interconnect, management plane signaling, control plane signaling, system message, non-access layer signaling, access layer signaling and dedicated configuration signaling.

[0044] A model lifecycle management method, the method being applied to a second network device, the method comprising:

[0045] Sending a signaling to a first network device; wherein the signaling includes at least model configuration information, and the model configuration information instructs the first network device to perform an artificial intelligence model training and / or data collection process;

[0046] Receive a model configuration response message sent by the first network device; the model configuration response message is generated based on the execution results of the artificial intelligence model training and / or data collection process.

[0047] In one of the embodiments, the receiving the model configuration response message sent by the first network device includes at least one of the following:

[0048] Receiving a startup failure response message sent by the first network device;

[0049] Receiving a model training start response message sent by the first network device;

[0050] Receive a model training completion message sent by the first network device.

[0051] In one of the embodiments, after receiving the model training completion message sent by the first network device, the method further includes:

[0052] Send a model deployment message to the first network device; the model deployment message is used to instruct the first network device to perform a model deployment process on the trained artificial intelligence model.

[0053] In one of the embodiments, the second network device includes at least one of a base station, a core network element, and a network management operation and maintenance management.

[0054] A model lifecycle management device, the device is applied to a first network device, the device comprising:

[0055] A first receiving module, configured to receive a signaling sent by a second network device; wherein the signaling at least includes model configuration information;

[0056] A first execution module, used to execute artificial intelligence model training and / or data collection process according to the model configuration information;

[0057] The first sending module is used to send a model configuration response message to the second network device based on the execution results of the artificial intelligence model training and / or data collection process.

[0058] A model lifecycle management device, the device is applied to a second network device, the device comprising:

[0059] A second sending module, configured to send a signaling to the first network device; wherein the signaling includes at least model configuration information, and the model configuration information instructs the first network device to perform an artificial intelligence model training and / or data collection process;

[0060] The second receiving module is used to receive a model configuration response message sent by the first network device; the model configuration response message is generated based on the execution results of the artificial intelligence model training and / or data collection process.

[0061] A communication device, comprising: a receiver, a processor and a transmitter;

[0062] The receiver is used to receive signaling sent by the second network device; wherein the signaling at least includes model configuration information;

[0063] The processor is used to perform artificial intelligence model training and / or data collection process according to the model configuration information;

[0064] The transmitter is used to send a model configuration response message to the second network device based on the execution results of the artificial intelligence model training and / or data collection process.

[0065] A communication device, comprising: a transmitter and a receiver;

[0066] The transmitter is used to send a signaling to the first network device; wherein the signaling includes at least model configuration information, and the model configuration information instructs the first network device to perform artificial intelligence model training and / or data collection process;

[0067] The receiver is used to receive a model configuration response message sent by the first network device; the model configuration response message is generated based on the execution results of the artificial intelligence model training and / or data collection process.

[0068] A computer-readable storage medium stores a computer program, which implements the steps of the above method when executed by a processor.

[0069] A chip includes a programmable logic circuit and / or program instructions, and can execute the steps of the above method when the chip is running.

[0070] A computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0071] The above-mentioned model lifecycle management method, device, communication equipment, storage medium and computer program product, the method is applied to the first network device, the method includes: receiving signaling sent by the second network device; wherein the signaling at least includes model configuration information; executing artificial intelligence model training and / or data collection process according to the model configuration information; based on the execution result of the artificial intelligence model training and / or data collection process, sending a model configuration response message to the second network device. In this way, the first network device executes artificial intelligence model training and / or data collection process according to the model configuration information in the signaling sent by the second network device, and sends a model configuration response message to the second network device based on the execution result, so as to realize the deployment of artificial intelligence model on the first network device, that is, the wireless access network side, which can improve spectrum efficiency, user experience rate and positioning accuracy, reduce latency, enhance security and privacy protection, and optimize AI-related indicators. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] FIG1 is an application environment diagram of a model lifecycle management method in one embodiment;

[0073] FIG2 is a schematic diagram of a flow chart of a model lifecycle management method in one embodiment;

[0074] FIG3 is a flowchart of executing artificial intelligence model training and / or data collection process steps according to the model configuration information when the model configuration information includes data collection configuration information in one embodiment;

[0075] FIG4 is a schematic diagram of a flow chart of steps for obtaining training sample data in one embodiment;

[0076] FIG5 is a flowchart of the steps of sending a model configuration response message to a second network device based on the execution results of an artificial intelligence model training and / or data collection process in one embodiment;

[0077] FIG6 is a flowchart of steps that may be further included in the model lifecycle management method after sending a model training completion message to the second network device in one embodiment;

[0078] FIG7 is a flowchart of steps that may be further included in the model lifecycle management method in one embodiment;

[0079] FIG8 is a schematic diagram of a flow chart of the steps of updating an artificial intelligence model according to a preset model update strategy in one embodiment;

[0080] FIG9 is a schematic diagram of a flow chart of a model lifecycle management method in another embodiment;

[0081] 10 is a schematic flow diagram of the steps of receiving a model configuration response message sent by a first network device in one embodiment;

[0082] FIG11 is a signaling interaction flow chart of a model lifecycle management method in one embodiment;

[0083] FIG12 is a structural block diagram of a model lifecycle management device in one embodiment;

[0084] FIG13 is a structural block diagram of a model lifecycle management device in another embodiment;

[0085] FIG14 is a diagram showing the internal structure of a communication device in one embodiment;

[0086] FIG. 15 is a diagram showing the internal structure of a communication device in another embodiment. DETAILED DESCRIPTION

[0087] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0088] Fig. 1 is a schematic diagram of an application scenario of a model lifecycle management method provided by an embodiment of the present application. As shown in Fig. 1, in the scenario, there is a first network device 100 and a second network device 200. The first network device 100 and the second network device 200 perform data transmission via a network.

[0089] The first network device 100 may include at least one of a terminal, a base station, a main control board, an intelligent computing board, a computing power board, a baseband board, and a wireless access network element. The second network device 200 may include at least one of a base station, a core network element, and a network management operation and maintenance (Operation Administration and Maintenance, OAM).

[0090] The base station can be a base station (Base TransceiverStation, BTS) in Global System of Mobile communication (GSM) or Code Division Multiple Access (CDMA), a base station (NodeB, NB) in Wideband Code Division Multiple Access (WCDMA), an evolutionary base station (eNB or eNodeB) in LTE, a relay station or access point, or a base station in a 5G network, etc., and is not limited here.

[0091] The terminal may be a wireless terminal, which may be a device that provides voice and / or other business data connectivity to users, or a handheld device with wireless connection function, or other processing equipment connected to a wireless modem. The wireless terminal may communicate with one or more core networks via a radio access network (Radio Access Network, referred to as RAN). The wireless terminal may be a mobile terminal, such as a mobile phone (or "cellular" phone) and a computer with a mobile terminal, for example, a portable, pocket-sized, handheld, computer-built-in or vehicle-mounted mobile device, which exchanges language and / or data with the radio access network. The wireless terminal may also be referred to as a system, a subscriber unit, a subscriber station, a mobile station, a mobile station, a remote station, a remote terminal, an access terminal, a user terminal, a user agent, a user device or user equipment, without limitation here.

[0092] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0093] In one embodiment, as shown in FIG. 2 , a model lifecycle management method is provided, which is described by taking the method applied to the first network device 100 in FIG. 1 as an example, and includes the following steps:

[0094] Step 201: Receive signaling sent by a second network device.

[0095] The signaling at least includes model configuration information.

[0096] In an embodiment of the present application, the life cycle of a model refers to the entire process from model creation to abandonment / invalidation. Model life cycle management may include at least one of model generation, model construction, model training, and model deployment. It is understandable that model training may represent model construction or model generation. The first network device is a network device to be deployed with an artificial intelligence model. The artificial intelligence model may be an artificial intelligence / machine learning (AI / ML) model. The artificial intelligence model may be a resource scheduling model and / or a decision model. Resources may be time domain resources, frequency domain resources, and airspace resources. The artificial intelligence model may be used in multi-carrier coordination scenarios, multi-frequency coordination scenarios, Industrial Internet of Things (IIoT) scenarios, telecommunications customized network scenarios, and millimeter wave deployment scenarios. The model configuration information instructs the first network device to perform artificial intelligence model training and / or data collection processes.

[0097] Optionally, the signaling may include at least one of user plane signaling, intelligent plane signaling, data plane signaling, computing plane signaling, radio resource control (RCC) signaling, medium access control-control element (MAC CE), data center interconnect (DCI), management plane signaling, control plane signaling, system messages, non-access-stratum (NAS) signaling, access-stratum (AS) signaling and dedicated configuration signaling.

[0098] Specifically, the second network device 200 sends a signaling including at least the model configuration information to the first network device 100. The first network device 100 receives the signaling sent by the second network device 200.

[0099] Step 202: Execute artificial intelligence model training and / or data collection process according to the model configuration information.

[0100] Optionally, the model configuration information may include at least one of indication information and training configuration information. The indication information is used to indicate whether to perform artificial intelligence model training. The training configuration information may include at least one of the geographical scope of model construction, storage space information, computing power information, model structure configuration information, training parameter information, and data collection configuration information.

[0101] Optionally, the geographical scope of model construction can be determined in one or more of the following ways: according to the neighboring cell relationship or tracking area (TA), indicated by an array of cell identity identification numbers (Identity, ID), indicated by a base station identifier (including multiple carriers supported by the base station), determined according to a preset neighboring cell relationship group or cell group (the preset neighboring cell relationship group or cell group is determined by the configuration of the second network device 200), and according to the number of base stations or cells in the configured geographical scope.

[0102] Optionally, the storage space information represents the storage space required for the artificial intelligence model training and / or data collection process. The computing power information represents the computing power required for the artificial intelligence model training and / or data collection process. The computing power information may include at least one of computing power requirement information (also known as central processing unit (CPU) capability requirement information) and intelligent computing capability requirement information (also known as graphics processing unit (GPU) capability requirement information).

[0103] Optionally, the model structure configuration information includes at least one of pre-training indication information, model type and model structure parameters. The pre-training indication information indicates whether it is a pre-trained model. The model type includes at least one of transformer (Transformer), deep neural network (Deep Neural Networks, DNN), convolution neural network (Convolution Neural Network, CNN) and multilayer perceptron (Multilayer Perceptron, MLP). The model structure parameters represent the model structure configuration. For example, the model structure parameters adopt a standardized representation format or a mainstream model representation format.

[0104] Optionally, the training parameter information may include at least one of a data set size, a loss function type, a training cycle, a deployment condition (also called a deployment threshold), a maximum number of model training times, and a training interval (also called a training duration or a training waiting window). The data set size may be a batch size used by the data set.

[0105] Optionally, the data collection configuration information may include at least one of a data set number, a data set type, a data set function, a data collection pilot signal configuration, a data collection period, a data collection sample type, a data collection accuracy, and a data collection amount.

[0106] Step 203: Send a model configuration response message to the second network device based on the execution results of the artificial intelligence model training and / or data collection process.

[0107] In an exemplary embodiment, the base station receives a signaling sent by the network management operation and maintenance management. The signaling includes at least model configuration information. The model configuration information may include the start indication information of the time domain artificial intelligence model, the shut down indication information of the time domain artificial intelligence model, the start indication information of the frequency domain artificial intelligence model, the shut down indication information of the frequency domain artificial intelligence model, the start indication information of the airspace artificial intelligence model, the shut down indication information of the airspace artificial intelligence model, the model activation threshold, the performance detection threshold, the model training cycle, the data collection duration, the service cell frequency and the neighboring cell frequency. Then, the base station performs the artificial intelligence model training and / or data collection process according to the model configuration information. Specifically, the base station collects the resource scheduling management data of the wireless access network according to the model configuration information. The artificial intelligence model is a resource scheduling management model of the wireless access network. Resource scheduling management data may include reference signal receiving power (RSRP) of the service cell frequency and the neighboring cell frequency in the time domain, reference signal receiving power of the service cell frequency and the neighboring cell frequency in the frequency domain, reference signal receiving power of the service cell frequency and the neighboring cell frequency in the airspace, user transmission rate, cell spectrum efficiency, user transmission signal to noise ratio (SINR), first layer (Layer 1, L1) or third layer (Layer 3, L3) reference signal receiving power. Then, the base station sends a model configuration response message to the network management operation and maintenance management based on the execution results of the artificial intelligence model training and / or data collection process.

[0108] In another exemplary embodiment, the base station receives a signaling sent by the network management operation and maintenance management. The signaling includes at least model configuration information. The model configuration information may include at least one of the measurement quantity input by the model, the base station identifier, the transmission and receiving point (TRP) identifier, and the beam identifier. The measurement quantity input by the model includes at least one of the channel impulse response (CIR), the reference signal received power, the arrival time (TOA) and the arrival time difference (TDOA). Then, the base station performs artificial intelligence model training and / or data collection process according to the model configuration information. Specifically, the base station collects AI positioning enhancement information or AI-based beam management information according to the model configuration information. Among them, the artificial intelligence model is an AI positioning enhancement model or an AI-based beam management model. AI positioning enhancement information may include at least one of the reference signal received power of different cells, arrival time, arrival time difference, channel impulse response, beam identifier, cell identifier and frequency point identifier. The AI-based beam management information may include at least one of the reference signal received power, arrival time, arrival time difference, channel impulse response, beam identification, cell identification and frequency identification of different cells. Then, the base station sends a model configuration response message to the network management operation and maintenance management based on the execution results of the artificial intelligence model training and / or data collection process.

[0109] Through this solution, the first network device executes the artificial intelligence model training and / or data collection process according to the model configuration information in the signaling sent by the second network device, and based on the execution result, sends a model configuration response message to the second network device, so as to realize the deployment of the artificial intelligence model on the first network device, that is, the wireless access network side, which can improve the spectrum efficiency, user experience rate and positioning accuracy, reduce latency, improve security and privacy protection, and optimize AI-related indicators. In addition, this solution can realize the signaling configuration of AI / ML model deployment between the wireless network and the base station on the RAN side, realize the network's controllable training of AI model generation, and enable AI to improve the performance of the wireless air interface system. In addition, this solution defines a specific configuration scheme for wireless access network model construction and model training, and defines a specific design scheme for related signaling interaction and configuration information for model construction and model training for RAN side network elements such as base stations through base stations, gateways or core network elements, which can support base stations, gateways or core network elements to build models on various network element devices such as base stations, intelligent computing boards, main control boards or new network elements on the RAN side through model configuration information, so as to achieve strong flexibility.

[0110] In one embodiment, as shown in FIG3 , when the model configuration information includes data collection configuration information, executing the artificial intelligence model training and / or data collection process according to the model configuration information may include at least one of the following:

[0111] Step 301, executing a data collection process based on data collection configuration information. And,

[0112] Optionally, when the data collection configuration information includes a data collection cycle, the first network device executes the data collection process according to the data collection cycle.

[0113] Optionally, if the artificial intelligence model meets the preset model deployment conditions, the data collection process is stopped.

[0114] Step 302: Obtain training sample data, and train the artificial intelligence model based on the training sample data and model configuration information.

[0115] The training sample data includes at least one of the collected sample data and / or pre-stored sample data.

[0116] Optionally, when the model configuration information includes storage space information and / or computing power information, the first network device determines whether to start artificial intelligence model training based on the storage space information and / or computing power information. If it is determined to start artificial intelligence model training, the first network device trains the artificial intelligence model based on the training sample data and the model configuration information.

[0117] Through this solution, the model configuration information includes data collection configuration information, and the data collection process is executed based on the data collection configuration information, and training sample data is obtained. Based on the training sample data and the model configuration information, the artificial intelligence model is trained, and the first network device executes the artificial intelligence model training and / or data collection process, thereby realizing the deployment of the artificial intelligence model on the first network device, i.e., the wireless access network side.

[0118] In one embodiment, as shown in FIG4 , obtaining training sample data may include at least one of the following:

[0119] Step 401, if the data volume of the pre-stored sample data is less than the preset data volume threshold, the collected sample data is obtained, and when the total data volume of the collected sample data and the pre-stored sample data reaches the data volume threshold, the step of training the artificial intelligence model based on the training sample data and the model configuration information is executed.

[0120] In the embodiment of the present application, different artificial intelligence models correspond to different data volume thresholds. The data volume of the pre-stored sample data can be 0.

[0121] Specifically, if the amount of pre-stored sample data is less than a preset data amount threshold, the first network device acquires the collected sample data. When the total amount of the collected sample data and the pre-stored sample data reaches the data amount threshold, the first network device uses the collected sample data and the pre-stored sample data to form training sample data, and executes the step of training the artificial intelligence model based on the training sample data and the model configuration information.

[0122] Step 402: If the amount of the pre-stored sample data is greater than or equal to a preset data amount threshold, the pre-stored sample data is used as training sample data.

[0123] Through this solution, the data volume of the pre-stored sample data is compared with the preset data volume threshold. When the data volume of the pre-stored sample data is insufficient, the collected sample data and the pre-stored sample data constitute the training sample data; when the data volume of the pre-stored sample data is sufficient, the pre-stored sample data is directly used as the training sample data, which can ensure that the data volume used to train the artificial intelligence model is sufficient, thereby enabling the first network device to perform the artificial intelligence model training and / or data collection process, thereby enabling the deployment of the artificial intelligence model on the first network device, that is, the wireless access network side. In addition, this solution can achieve the decoupling of the training sample data and the artificial intelligence model.

[0124] In one embodiment, acquiring training sample data may include: after data collection is completed, using the collected sample data as training sample data.

[0125] Through this solution, the sample data collected this time is directly used as training sample data, which not only ensures that the amount of data used to train the artificial intelligence model is sufficient, thereby enabling the first network device to execute the artificial intelligence model training and / or data collection process, and then realize the deployment of the artificial intelligence model on the first network device, that is, the wireless access network side, but also can improve the real-time nature of the sample training data, thereby improving the accuracy of the artificial intelligence model.

[0126] In one embodiment, as shown in FIG. 5 , based on the execution result of the artificial intelligence model training and / or data collection process, sending a model configuration response message to the second network device may include at least one of the following:

[0127] Step 501: If the artificial intelligence model training or data collection process fails to start, a startup failure response message is sent to the second network device.

[0128] Optionally, the startup failure response message may include failure reason information. The failure reason information may be an error code. The failure reason information may include at least one of insufficient computing power, insufficient storage space, busy base station, and invalid configuration.

[0129] Optionally, the first network device may execute a processing procedure for startup failure of artificial intelligence model training or data collection process according to preset startup failure processing rules.

[0130] Step 502: If the artificial intelligence model training is started successfully, a model training start response message is sent to the second network device.

[0131] In an embodiment of the present application, the model training start response message indicates that the artificial intelligence model training is successfully started.

[0132] Step 503: If the artificial intelligence model training is completed, a model training completion message is sent to the second network device.

[0133] In an embodiment of the present application, the model training completion message indicates that the artificial intelligence model training is completed.

[0134] Through this solution, a startup failure response message is sent to the second network device, so that the processing process of the failure to start the artificial intelligence model training or data acquisition process can be executed in time, ensuring that the artificial intelligence model is deployed on the first network device, that is, the wireless access network side, and improving the efficiency of deploying the artificial intelligence model on the first network device, that is, the wireless access network side; sending a model training startup response message and a model training completion message to the second network device can ensure the stable progress of the model training and ensure that the artificial intelligence model is deployed on the first network device, that is, the wireless access network side.

[0135] In one embodiment, as shown in FIG6 , after sending the model training completion message to the second network device, the method may further include at least one of the following:

[0136] Step 601, execute the model deployment process for the trained artificial intelligence model.

[0137] In an embodiment of the present application, after sending a model training completion message to the second network device, the first network device executes a model deployment process on the trained artificial intelligence model.

[0138] Step 602: Receive a model deployment message sent by the second network device, and execute a model deployment process on the trained artificial intelligence model.

[0139] In an embodiment of the present application, the model deployment message (also called the model online message) instructs the first network device to execute the model deployment process on the trained artificial intelligence model.

[0140] Specifically, after sending the model training completion message to the second network device, the first network device waits for the model deployment message sent by the second network device. Then, the first network device receives the model deployment message sent by the second network device and executes the model deployment process on the trained artificial intelligence model.

[0141] Through this solution, after sending the model training completion message to the second network device, the model deployment process is directly executed on the trained artificial intelligence model, which can improve the efficiency of deploying the artificial intelligence model on the first network device, that is, the wireless access network side; after receiving the model deployment message sent by the second network device, the model deployment process is executed on the trained artificial intelligence model, which can improve the flexibility of deploying the artificial intelligence model on the first network device, that is, the wireless access network side.

[0142] In one embodiment, the method may further include: when the artificial intelligence model training fails, determining the cumulative number of model training times of the artificial intelligence model.

[0143] Optionally, if the artificial intelligence model does not converge after a preset number of training rounds of model training, the first network device determines that the artificial intelligence model training has failed.

[0144] Optionally, when the artificial intelligence model training fails, the first network device obtains the cumulative model training times of the artificial intelligence model.

[0145] In one embodiment, as shown in FIG7 , the method may further include at least one of the following:

[0146] Step 701: If the cumulative number of model training times is less than the maximum number of model training times, the artificial intelligence model is retrained and the cumulative number of model training times is updated.

[0147] In an embodiment of the present application, the first network device compares the cumulative model training times with the maximum model training times included in the model configuration information. If the cumulative model training times are less than the maximum model training times, the first network device retrains the artificial intelligence model and adds 1 to the cumulative model training times.

[0148] Step 702: If the cumulative number of model training times is greater than or equal to the maximum number of model training times, the model training of the artificial intelligence model is stopped, the model training timer is started, and the artificial intelligence model is retrained when the timing length of the model training timer is greater than or equal to the training interval.

[0149] In an embodiment of the present application, if the cumulative number of model training times is greater than or equal to the maximum number of model training times, the first network device stops the model training of the artificial intelligence model, starts the model training timer, or resets the model training timer. When the timing duration of the model training timer is greater than or equal to the training interval, the first network device retrains the artificial intelligence model.

[0150] Through this solution, when the artificial intelligence model training fails, the cumulative model training times of the artificial intelligence model are determined, and when the cumulative model training times are greater than or equal to the maximum model training times, the model training of the artificial intelligence model is stopped, and the artificial intelligence model is retrained after waiting for the training interval, which can avoid the impact of the artificial intelligence model training caused by the current unsuitable network conditions and other reasons, save computing power, and reduce energy consumption. In addition, this solution proposes a variety of response mechanisms when the model training is successful or failed, supports the sending of model training related configurations, and supports multiple attempts to perform model training under the configuration of timer and / or maximum training times when model training fails, which increases the probability of successful model construction and can be implemented in the existing network.

[0151] In one embodiment, before retraining the artificial intelligence model, the method may further include: updating the artificial intelligence model according to a preset model update strategy.

[0152] In an embodiment of the present application, the model update strategy may include at least one of a model structure update strategy, an impulse function update strategy, a loss function update strategy, and a model type update strategy.

[0153] Through this solution, the artificial intelligence model is updated before retraining the artificial intelligence model, which improves the success probability of model training and the success probability of deploying the artificial intelligence model on the first network device, that is, the wireless access network side, which can be achieved in the existing network.

[0154] In one embodiment, as shown in FIG8 , updating the artificial intelligence model according to a preset model update strategy may include at least one of the following:

[0155] Step 801, changing the model structure of the artificial intelligence model according to the preset model update strategy.

[0156] In an embodiment of the present application, changing the model structure of the artificial intelligence model may include adding convolutional layers, reducing convolutional layers, adding self-attention modules, reducing self-attention modules, adding residual structures, reducing residual structures, adding fully connected layers, and reducing at least one of the fully connected layers.

[0157] Specifically, the first network device determines the target model structure and the target number of layers corresponding to the target model structure according to a preset model update strategy. Then, the first network device updates the number of layers of the target model structure to the target number of layers. The target model structure may include at least one of a convolutional layer, a self-attention module, a residual structure, and a fully connected layer.

[0158] Optionally, when the model update strategy includes a model structure change sequence and a mapping relationship between a preset model structure, the number of model structure updates, and the number of layers, the first network device determines the target model structure of this change according to the model structure change sequence. Then, the first network device queries the target number of layers corresponding to the target model structure of this model structure update in the mapping relationship between the preset model structure, the number of model structure updates, and the number of layers. The model structure change sequence represents the order in which the model structure is changed.

[0159] Optionally, the first network device randomly determines the target model structure of this change in a preset model structure. Then, the first network device randomly determines the target number of layers of the target model structure of this change in a preset number of layers.

[0160] Step 802, changing the impulse function or loss function of the artificial intelligence model according to the preset model update strategy.

[0161] In an embodiment of the present application, the first network device determines the target impulse function or target loss function of the artificial intelligence model according to a preset model update strategy. Then, the first network device changes the impulse function or loss function of the artificial intelligence model to the target impulse function or target loss function.

[0162] Optionally, the first network device determines the target impulse function or target loss function of this change according to a preset impulse function or loss function complexity sequence, wherein the impulse function or loss function complexity sequence represents the complexity ranking of each impulse function or loss function.

[0163] Optionally, the first network device randomly determines a target impulse function or a target loss function of the artificial intelligence model in a preset impulse function or loss function.

[0164] Step 803: Change the model type of the artificial intelligence model according to the preset model update strategy.

[0165] In an embodiment of the present application, the first network device determines the target model type of the artificial intelligence model according to a preset model update strategy, and then changes the model type of the artificial intelligence model to the target model type.

[0166] Optionally, the first network device determines the target impulse function or target model type for this change according to a preset model type complexity sequence, wherein the model type complexity sequence represents the complexity ranking of each model type.

[0167] Optionally, the first network device randomly determines a target model type of the artificial intelligence model from preset model types.

[0168] Optionally, the first network device determines the target model update type of this artificial intelligence model update according to the artificial intelligence model update complexity sequence. Then, the first network device updates the artificial intelligence model according to the target model update type. Among them, the artificial intelligence model update complexity sequence represents the order of the complexity of the model update types of various artificial intelligence models from small to large. The model update type may include at least one of changing the model structure of the artificial intelligence model, changing the impulse function or loss function of the artificial intelligence model, and transforming the model type of the artificial intelligence model. For example, the artificial intelligence model update complexity sequence may be changing the model structure of the artificial intelligence model, changing the impulse function or loss function of the artificial intelligence model, and transforming the model type of the artificial intelligence model.

[0169] Through this solution, before retraining the artificial intelligence model, the artificial intelligence model is updated by changing at least one of the model structure of the artificial intelligence model, changing the impulse function or loss function of the artificial intelligence model, and transforming the model type of the artificial intelligence model. A variety of model enhancement mechanisms are proposed before retraining the model after model training fails, so as to optimize the model and enhance the model training performance, thereby further improving the success probability of model training and the success probability of deploying the artificial intelligence model on the first network device, i.e., the wireless access network side.

[0170] In one embodiment, before retraining the artificial intelligence model, the method may further include: changing the training sample data according to a preset training sample data update strategy.

[0171] Optionally, the first network device uses data sets with different numbers as training sample data.

[0172] Optionally, the first network device uses a plurality of numbered data sets to merge and construct training sample data.

[0173] Through this solution, the training sample data is changed before the artificial intelligence model is retrained, which improves the success probability of model training and the success probability of deploying the artificial intelligence model on the first network device, that is, the wireless access network side, which can be achieved in the existing network.

[0174] In one embodiment, as shown in FIG. 9 , a model lifecycle management method is provided, which is described by taking the method applied to the second network device 200 in FIG. 1 as an example, and includes the following steps:

[0175] Step 901: Send signaling to a first network device.

[0176] The signaling includes at least model configuration information, and the model configuration information instructs the first network device to execute artificial intelligence model training and / or data collection process.

[0177] Step 902: Receive a model configuration response message sent by the first network device.

[0178] Among them, the model configuration response message is generated based on the execution results of the artificial intelligence model training and / or data collection process.

[0179] Through this solution, the second network device sends a signaling containing at least model configuration information to the first network device, instructing the first network device to perform artificial intelligence model training and / or data collection process, and receives the model configuration response message generated by the first network device based on the execution result of the artificial intelligence model training and / or data collection process. Through the second network device, the artificial intelligence model can be deployed on the first network device, that is, the wireless access network side, which can improve spectrum efficiency, user experience rate and positioning accuracy, reduce latency, enhance security and privacy protection, and optimize AI-related indicators. In addition, this solution can implement signaling configuration for AI / ML model deployment between the wireless network and the base station on the RAN side, realize the network's manageability and controllability of AI model generation and training, and enable AI to improve the performance of the wireless air interface system. In addition, this solution defines a specific configuration scheme for wireless access network model construction and model training, and defines a specific design scheme for related signaling interaction and configuration information for configuring model construction and model training for RAN-side network elements such as base stations, gateways or core network elements through base stations, gateways or core network elements. It can support base stations, gateways or core network elements to build models on various network element devices such as base stations, intelligent computing boards, main control boards or new network elements on the RAN side through model configuration information, thereby achieving strong flexibility.

[0180] In one embodiment, as shown in FIG. 10 , receiving a model configuration response message sent by the first network device may include at least one of the following:

[0181] Step 1001: Receive a startup failure response message sent by a first network device.

[0182] Step 1002: Receive a model training start response message sent by the first network device.

[0183] Step 1003: Receive a model training completion message sent by the first network device.

[0184] Through this solution, by receiving the startup failure response message sent by the first network device, the processing process of the failure to start the artificial intelligence model training or the data acquisition process can be executed in time, ensuring the deployment of the artificial intelligence model on the first network device, that is, the wireless access network side, and improving the efficiency of deploying the artificial intelligence model on the first network device, that is, the wireless access network side; receiving the model training startup response message and the model training completion message sent by the first network device can ensure the stable progress of the model training, and ensure that the artificial intelligence model is deployed on the first network device, that is, the wireless access network side, through the second network device.

[0185] In one embodiment, after receiving the model training completion message sent by the first network device, the method may further include: sending a model deployment message to the first network device.

[0186] Among them, the model deployment message is used to instruct the first network device to execute the model deployment process for the trained artificial intelligence model.

[0187] Through this solution, after receiving the model training completion message sent by the first network device, a model deployment message is sent to the first network device, instructing the first network device to execute the model deployment process on the trained artificial intelligence model, thereby improving the flexibility of deploying the artificial intelligence model on the first network device, i.e., the wireless access network side.

[0188] In one embodiment, a signaling interaction flow chart of a model lifecycle management method is provided in Figure 11. As shown in Figure 11, the method is applied to a model lifecycle management system, the model lifecycle management system includes at least a first network device and a second network device, and the method includes the following steps.

[0189] Step 1101: The second network device sends a signaling to the first network device, wherein the signaling includes at least model configuration information, and the model configuration information instructs the first network device to perform artificial intelligence model training and / or data collection process.

[0190] Step 1102: The first network device executes artificial intelligence model training and / or data collection process according to the model configuration information.

[0191] Step 1103: The first network device determines whether the artificial intelligence model training or data collection process fails to start.

[0192] If the artificial intelligence model training or data collection process fails to start, execute step 1104; if the artificial intelligence model training is successfully started, execute step 1105.

[0193] Step 1104: The first network device sends a startup failure response message to the second network device.

[0194] Step 1105: The first network device sends a model training start response message to the second network device.

[0195] Step 1106: The first network device determines whether the artificial intelligence model has been trained.

[0196] If the artificial intelligence model training is completed, execute step 1107; otherwise, no processing is performed.

[0197] Step 1107: The first network device sends a model training completion message to the second network device.

[0198] Step 1108: The second network device sends a model deployment message to the first network device.

[0199] Step 1109: The first network device executes a model deployment process on the trained artificial intelligence model.

[0200] It should be understood that, although the various steps in the flowchart of Figure 2-11 are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in Figure 2-11 may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0201] In one embodiment, as shown in FIG. 12 , a model lifecycle management device 1200 is provided. The device is applied to a first network device and includes: a first receiving module 1210, a first executing module 1220 and a first sending module 1230, wherein:

[0202] The first receiving module 1210 is used to receive a signaling sent by a second network device; wherein the signaling at least includes model configuration information;

[0203] A first execution module 1220, configured to execute artificial intelligence model training and / or data collection process according to the model configuration information;

[0204] The first sending module 1230 is used to send a model configuration response message to the second network device based on the execution results of the artificial intelligence model training and / or data collection process.

[0205] Optionally, the model configuration information includes at least one of indication information and training configuration information; the indication information is used to indicate whether to perform artificial intelligence model training; the training configuration information includes at least one of model construction geographical scope, storage space information, computing power information, model structure configuration information, training parameter information and data collection configuration information.

[0206] Optionally, the storage space information represents the storage space required for the artificial intelligence model training and / or data collection process; the computing power information represents the computing power required for the artificial intelligence model training and / or data collection process.

[0207] Optionally, the model structure configuration information includes at least one of pre-training indication information, model type and model structure parameters; the pre-training indication information indicates whether it is a pre-trained model.

[0208] Optionally, the training parameter information includes at least one of a data set size, a loss function type, a training cycle, deployment conditions, a maximum number of model training times, and a training interval.

[0209] Optionally, the data collection configuration information includes at least one of a data set number, a data set type, a data set function, a data collection pilot signal configuration, a data collection period, a data collection sample type, a data collection accuracy and a data collection amount.

[0210] Optionally, when the model configuration information includes data collection configuration information, the first execution module 1220 is specifically configured to:

[0211] Execute a data collection process based on the data collection configuration information; and,

[0212] Acquire training sample data, and train the artificial intelligence model based on the training sample data and the model configuration information; the training sample data includes at least one of the collected sample data and / or pre-stored sample data.

[0213] Optionally, the first execution module 1220 is specifically configured to:

[0214] If the data volume of the pre-stored sample data is less than a preset data volume threshold, the collected sample data is obtained, and when the total data volume of the collected sample data and the pre-stored sample data reaches the data volume threshold, the step of training the artificial intelligence model based on the training sample data and the model configuration information is executed;

[0215] If the data volume of the pre-stored sample data is greater than or equal to the preset data volume threshold, the pre-stored sample data is used as the training sample data.

[0216] Optionally, the first execution module 1220 is specifically configured to:

[0217] After the data collection is completed, the collected sample data is used as training sample data.

[0218] Optionally, the first sending module 1230 is specifically configured to:

[0219] If the artificial intelligence model training or the data collection process fails to start, sending a startup failure response message to the second network device;

[0220] If the artificial intelligence model training is successfully started, sending a model training start response message to the second network device;

[0221] If the artificial intelligence model training is completed, a model training completion message is sent to the second network device.

[0222] Optionally, the device 1200 further includes at least one of the following:

[0223] A second execution module, used to execute a model deployment process on the trained artificial intelligence model;

[0224] The third execution module is used to receive the model deployment message sent by the second network device and execute the model deployment process on the trained artificial intelligence model.

[0225] Optionally, the device 1200 further includes:

[0226] A determination module is used to determine the cumulative number of model training times of the artificial intelligence model when the artificial intelligence model training fails.

[0227] Optionally, the device 1200 further includes at least one of the following:

[0228] A first retraining module, configured to retrain the artificial intelligence model and update the cumulative model training times if the cumulative model training times are less than the maximum model training times;

[0229] The second retraining module is used to stop the model training of the artificial intelligence model if the cumulative model training times are greater than or equal to the maximum model training times, start the model training timer, and retrain the artificial intelligence model when the timing length of the model training timer is greater than or equal to the training interval.

[0230] Optionally, the device 1200 further includes:

[0231] The first updating module is used to update the artificial intelligence model according to a preset model updating strategy.

[0232] Optionally, the first updating module is specifically used to:

[0233] According to a preset model update strategy, the model structure of the artificial intelligence model is changed;

[0234] According to a preset model update strategy, changing the impulse function or loss function of the artificial intelligence model;

[0235] According to a preset model update strategy, the model type of the artificial intelligence model is changed.

[0236] Optionally, the device 1200 further includes:

[0237] The second updating module is used to change the training sample data according to a preset training sample data updating strategy.

[0238] Optionally, the first network device includes at least one of a terminal, a base station, a main control board, an intelligent computing board, a computing power board, a baseband board and a wireless access network element.

[0239] Optionally, the signaling includes at least one of user plane signaling, intelligent plane signaling, data plane signaling, computing plane signaling, wireless resource control signaling, media access control-control element, data center interconnection, management plane signaling, control plane signaling, system message, non-access layer signaling, access layer signaling and dedicated configuration signaling.

[0240] In one embodiment, as shown in FIG. 13 , a model lifecycle management device 1300 is provided. The device is applied to a second network device and includes: a second sending module 1310 and a second receiving module 1320, wherein:

[0241] The second sending module 1310 is used to send a signaling to the first network device; wherein the signaling includes at least model configuration information, and the model configuration information instructs the first network device to perform artificial intelligence model training and / or data collection process;

[0242] The second receiving module 1320 is used to receive a model configuration response message sent by the first network device; the model configuration response message is generated based on the execution results of the artificial intelligence model training and / or data collection process.

[0243] Optionally, the second receiving module 1320 is specifically configured to:

[0244] Receiving a startup failure response message sent by the first network device;

[0245] Receiving a model training start response message sent by the first network device;

[0246] Receive a model training completion message sent by the first network device.

[0247] Optionally, the device 1300 further includes:

[0248] The third sending module is used to send a model deployment message to the first network device; the model deployment message is used to instruct the first network device to execute a model deployment process on the trained artificial intelligence model.

[0249] Optionally, the second network device includes at least one of a base station, a core network element, and network management operation and maintenance management.

[0250] For the specific definition of the model lifecycle management device, please refer to the definition of the model lifecycle management method above, which will not be repeated here. Each module in the above-mentioned model lifecycle management device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0251] FIG14 is a schematic diagram of the structure of an access network device provided in an embodiment of the present application. The access network device may include a receiver 1401, a memory 1402, a processor 1403, at least one communication bus 1404 and a transmitter 1405. The communication bus 1404 is used to realize the communication connection between the components. The memory 1402 may include a high-speed RAM memory, and may also include a non-volatile storage NVM, such as at least one disk storage. Various programs may be stored in the memory 1402 to complete various processing functions and implement the method steps of this embodiment. In this embodiment, the transmitter 1405 may be a radio frequency processing module or a baseband processing module in the access network device, and the receiver 1401 may also be a radio frequency processing module or a baseband processing module in the access network device. The transmitter 1405 and the receiver 1401 can be integrated together to realize a transceiver, and the transmitter 1405 and the receiver 1401 can be coupled to the processor 1403, which can realize the action of receiving or sending under the instruction or control of the processor 1403.

[0252] In this embodiment, the receiver 1401 is used to receive the signaling sent by the second network device; wherein the signaling at least includes the model configuration information;

[0253] Processor 1403, used to perform artificial intelligence model training and / or data collection process according to model configuration information;

[0254] Transmitter 1405 is used to send a model configuration response message to the second network device based on the execution results of the artificial intelligence model training and / or data collection process.

[0255] In one embodiment, the model configuration information includes at least one of indication information and training configuration information; the indication information is used to indicate whether to perform artificial intelligence model training; the training configuration information includes at least one of the model construction geographical scope, storage space information, computing power information, model structure configuration information, training parameter information and data collection configuration information.

[0256] In one embodiment, the storage space information represents the storage space required for the artificial intelligence model training and / or data collection process; the computing power information represents the computing power required for the artificial intelligence model training and / or data collection process.

[0257] In one embodiment, the model structure configuration information includes at least one of pre-training indication information, model type and model structure parameters; the pre-training indication information indicates whether it is a pre-training model.

[0258] In one embodiment, the training parameter information includes at least one of a data set size, a loss function type, a training cycle, a deployment condition, a maximum number of model training times, and a training interval.

[0259] In one embodiment, the data collection configuration information includes at least one of a data set number, a data set type, a data set function, a data collection pilot signal configuration, a data collection period, a data collection sample type, a data collection accuracy, and a data collection amount.

[0260] In one embodiment, when the model configuration information includes data acquisition configuration information, the processor 1403 is specifically configured to:

[0261] Execute a data collection process based on the data collection configuration information; and,

[0262] Acquire training sample data, and train the artificial intelligence model based on the training sample data and model configuration information; the training sample data includes at least one of the collected sample data and / or pre-stored sample data.

[0263] In one embodiment, the processor 1403 is specifically configured to:

[0264] If the amount of pre-stored sample data is less than a preset data amount threshold, the collected sample data is obtained, and when the total amount of the collected sample data and the pre-stored sample data reaches the data amount threshold, the step of training the artificial intelligence model based on the training sample data and the model configuration information is executed;

[0265] If the data volume of the pre-stored sample data is greater than or equal to a preset data volume threshold, the pre-stored sample data is used as training sample data.

[0266] In one embodiment, the processor 1403 is specifically configured to:

[0267] After the data collection is completed, the collected sample data is used as training sample data.

[0268] In one embodiment, the transmitter 1405 is specifically configured to:

[0269] If the artificial intelligence model training or data collection process fails to start, a startup failure response message is sent to the second network device;

[0270] If the artificial intelligence model training is successfully started, a model training start response message is sent to the second network device;

[0271] If the artificial intelligence model training is completed, a model training completion message is sent to the second network device.

[0272] In one embodiment, the processor 1403 is further configured to:

[0273] Execute the model deployment process for the trained AI model;

[0274] Receive a model deployment message sent by the second network device, and execute a model deployment process on the trained artificial intelligence model.

[0275] In one embodiment, the processor 1403 is further configured to:

[0276] When artificial intelligence model training fails, the cumulative model training times of the artificial intelligence model are determined.

[0277] In one embodiment, the processor 1403 is further configured to:

[0278] If the cumulative number of model training times is less than the maximum number of model training times, the artificial intelligence model is retrained and the cumulative number of model training times is updated;

[0279] If the cumulative number of model training times is greater than or equal to the maximum number of model training times, the model training of the artificial intelligence model is stopped, the model training timer is started, and the artificial intelligence model is retrained when the timing length of the model training timer is greater than or equal to the training interval.

[0280] In one embodiment, the processor 1403 is further configured to:

[0281] Update the artificial intelligence model according to the preset model update strategy.

[0282] In one embodiment, the processor 1403 is further configured to:

[0283] Change the model structure of the artificial intelligence model according to the preset model update strategy;

[0284] Change the impulse function or loss function of the artificial intelligence model according to the preset model update strategy;

[0285] Change the model type of the artificial intelligence model according to the preset model update strategy.

[0286] In one embodiment, the processor 1403 is further configured to:

[0287] The training sample data is changed according to the preset training sample data update strategy.

[0288] In one embodiment, the first network device includes at least one of a terminal, a base station, a main control board, an intelligent computing board, a computing power board, a baseband board and a wireless access network element.

[0289] In one embodiment, the signaling includes at least one of user plane signaling, intelligence plane signaling, data plane signaling, computing plane signaling, radio resource control signaling, media access control-control element, data center interconnect, management plane signaling, control plane signaling, system message, non-access layer signaling, access layer signaling and dedicated configuration signaling.

[0290] FIG15 is a schematic diagram of the structure of an access network device provided in an embodiment of the present application. The access network device may include a receiver 1501, a memory 1502, a processor 1503, at least one communication bus 1504 and a transmitter 1505. The communication bus 1504 is used to realize the communication connection between the components. The memory 1502 may include a high-speed RAM memory, and may also include a non-volatile storage NVM, such as at least one disk storage. Various programs may be stored in the memory 1502 to complete various processing functions and implement the method steps of this embodiment. In this embodiment, the transmitter 1505 may be a radio frequency processing module or a baseband processing module in the access network device, and the receiver 1501 may also be a radio frequency processing module or a baseband processing module in the access network device. The transmitter 1505 and the receiver 1501 can be integrated together to realize a transceiver. The transmitter 1505 and the receiver 1501 can be coupled to the processor 1503, which can realize the action of receiving or sending under the instruction or control of the processor 1503.

[0291] In this embodiment, the transmitter 1505 is used to send a signaling to the first network device; wherein the signaling includes at least model configuration information, and the model configuration information instructs the first network device to perform artificial intelligence model training and / or data collection process;

[0292] Receiver 1501 is used to receive a model configuration response message sent by the first network device; the model configuration response message is generated based on the execution results of the artificial intelligence model training and / or data collection process.

[0293] In one embodiment, the receiver 1501 is specifically configured to:

[0294] Receiving a startup failure response message sent by the first network device;

[0295] Receiving a model training start response message sent by the first network device;

[0296] Receive a model training completion message sent by the first network device.

[0297] In one embodiment, the transmitter 1505 is further configured to:

[0298] Send a model deployment message to the first network device; the model deployment message is used to instruct the first network device to execute a model deployment process for the trained artificial intelligence model.

[0299] In one embodiment, the second network device includes at least one of a base station, a core network element, and a network management operation and maintenance management.

[0300] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0301] The embodiment of the present application also provides a computer program product including instructions, which, when executed on a computer, enables the computer to execute the steps of the above method.

[0302] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0303] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0304] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A model lifecycle management method, characterized in that: The method is applied to a first network device, and the method includes: Receiving a signaling sent by a second network device; wherein the signaling at least includes model configuration information; Execute artificial intelligence model training and / or data collection process according to the model configuration information; Based on the execution results of the artificial intelligence model training and / or data collection process, a model configuration response message is sent to the second network device.

2. The method according to claim 1, characterized in that The model configuration information includes at least one of indication information and training configuration information; the indication information is used to indicate whether to perform artificial intelligence model training; the training configuration information includes at least one of the model construction geographical scope, storage space information, computing power information, model structure configuration information, training parameter information and data collection configuration information.

3. The method according to claim 2, characterized in that The storage space information represents the storage space required for the artificial intelligence model training and / or data collection process; the computing power information represents the computing power required for the artificial intelligence model training and / or data collection process.

4. The method according to claim 2, characterized in that: The model structure configuration information includes at least one of pre-training indication information, model type and model structure parameters; the pre-training indication information indicates whether it is a pre-training model.

5. The method according to claim 2, characterized in that: The training parameter information includes at least one of a data set size, a loss function type, a training cycle, a deployment condition, a maximum number of model training times, and a training interval.

6. The method according to claim 2, characterized in that The data collection configuration information includes at least one of a data set number, a data set type, a data set function, a data collection pilot signal configuration, a data collection period, a data collection sample type, a data collection accuracy, and a data collection amount.

7. The method according to claim 1 or 2, characterized in that: In the case where the model configuration information includes data collection configuration information, performing artificial intelligence model training and / or data collection process according to the model configuration information includes at least one of the following: Execute a data collection process based on the data collection configuration information; as well as, Acquire training sample data, and train the artificial intelligence model based on the training sample data and the model configuration information; the training sample data includes at least one of the collected sample data and / or pre-stored sample data.

8. The method according to claim 7, characterized in that The obtaining of training sample data includes at least one of the following: If the data volume of the pre-stored sample data is less than a preset data volume threshold, the collected sample data is obtained, and when the total data volume of the collected sample data and the pre-stored sample data reaches the data volume threshold, the step of training the artificial intelligence model based on the training sample data and the model configuration information is executed; If the data volume of the pre-stored sample data is greater than or equal to the preset data volume threshold, the pre-stored sample data is used as the training sample data.

9. The method according to claim 7, characterized in that: The obtaining of training sample data comprises: After the data collection is completed, the collected sample data is used as training sample data.

10. The method according to claim 1, characterized in that The sending of a model configuration response message to the second network device based on the execution result of the artificial intelligence model training and / or data collection process includes at least one of the following: If the artificial intelligence model training or the data collection process fails to start, sending a startup failure response message to the second network device; If the artificial intelligence model training is successfully started, sending a model training start response message to the second network device; If the artificial intelligence model training is completed, a model training completion message is sent to the second network device.

11. The method according to claim 10, characterized in that After sending the model training completion message to the second network device, the method further includes at least one of the following: Execute a model deployment process on the trained artificial intelligence model; Receive the model deployment message sent by the second network device, and execute the model deployment process on the trained artificial intelligence model.

12. The method according to claim 1, characterized in that The method further comprises: When the artificial intelligence model training fails, the cumulative model training times of the artificial intelligence model are determined.

13. The method according to claim 12, characterized in that The method further comprises at least one of the following: If the cumulative model training times are less than the maximum model training times, the artificial intelligence model is retrained and the cumulative model training times are updated; If the cumulative number of model training times is greater than or equal to the maximum number of model training times, the model training of the artificial intelligence model is stopped, the model training timer is started, and the artificial intelligence model is retrained when the timing length of the model training timer is greater than or equal to the training interval.

14. The method according to claim 13, characterized in that Before retraining the artificial intelligence model, the method further includes: Update the artificial intelligence model according to a preset model update strategy.

15. The method according to claim 14, characterized in that The updating of the artificial intelligence model according to the preset model updating strategy includes at least one of the following: According to a preset model update strategy, the model structure of the artificial intelligence model is changed; According to a preset model update strategy, changing the impulse function or loss function of the artificial intelligence model; According to a preset model update strategy, the model type of the artificial intelligence model is changed.

16. The method according to claim 13, characterized in that Before retraining the artificial intelligence model, the method further includes: The training sample data is changed according to a preset training sample data update strategy.

17. The method according to claim 1, characterized in that The first network device includes at least one of a terminal, a base station, a main control board, an intelligent computing board, a computing power board, a baseband board and a wireless access network element.

18. The method according to claim 1, characterized in that The signaling includes at least one of user plane signaling, intelligent plane signaling, data plane signaling, computing plane signaling, radio resource control signaling, media access control control element, data center interconnection, management plane signaling, control plane signaling, system message, non-access layer signaling, access layer signaling and dedicated configuration signaling.

19. A model lifecycle management method, characterized in that: The method is applied to a second network device, and the method includes: Sending a signaling to a first network device; wherein the signaling includes at least model configuration information, and the model configuration information instructs the first network device to perform an artificial intelligence model training and / or data collection process; Receive a model configuration response message sent by the first network device; the model configuration response message is generated based on the execution results of the artificial intelligence model training and / or data collection process.

20. The method according to claim 19, characterized in that The receiving a model configuration response message sent by the first network device includes at least one of the following: Receiving a startup failure response message sent by the first network device; Receiving a model training start response message sent by the first network device; Receive a model training completion message sent by the first network device.

21. The method according to claim 20, characterized in that After receiving the model training completion message sent by the first network device, the method further includes: Send a model deployment message to the first network device; the model deployment message is used to instruct the first network device to perform a model deployment process on the trained artificial intelligence model.

22. The method according to claim 19, characterized in that The second network equipment includes at least one of a base station, a core network element, and network management operation and maintenance management.

23. A model lifecycle management device, characterized in that: The device is applied to a first network device, and the device includes: A first receiving module, configured to receive a signaling sent by a second network device; wherein the signaling at least includes model configuration information; A first execution module, used to execute artificial intelligence model training and / or data collection process according to the model configuration information; The first sending module is used to send a model configuration response message to the second network device based on the execution results of the artificial intelligence model training and / or data collection process.

24. A model lifecycle management device, characterized in that: The device is applied to a second network device, and the device includes: A second sending module, configured to send a signaling to the first network device; wherein the signaling includes at least model configuration information, and the model configuration information instructs the first network device to perform an artificial intelligence model training and / or data collection process; The second receiving module is used to receive a model configuration response message sent by the first network device; the model configuration response message is generated based on the execution results of the artificial intelligence model training and / or data collection process.

25. A communication device, characterized in that: include: Receivers, processors, and transmitters; The receiver is used to receive signaling sent by the second network device; wherein the signaling at least includes model configuration information; The processor is used to perform artificial intelligence model training and / or data collection process according to the model configuration information; The transmitter is used to send a model configuration response message to the second network device based on the execution results of the artificial intelligence model training and / or data collection process.

26. A communication device, characterized in that: include: Transmitters and receivers; The transmitter is used to send a signaling to the first network device; wherein the signaling includes at least model configuration information, and the model configuration information instructs the first network device to perform artificial intelligence model training and / or data collection process; The receiver is used to receive a model configuration response message sent by the first network device; the model configuration response message is generated based on the execution results of the artificial intelligence model training and / or data collection process.

27. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 22 are implemented.

28. A chip, characterized in that: The chip comprises a programmable logic circuit and / or program instructions, and the steps of the method according to any one of claims 1 to 22 are implemented when the chip is executed.

29. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 22 are implemented.

Images