CN116776988A

CN116776988A - Data processing method in communication network and network side equipment

Info

Publication number: CN116776988A
Application number: CN202210950629.7A
Authority: CN
Inventors: 程思涵; 崇卫微; 吴晓波
Original assignee: Vivo Mobile Communication Co Ltd
Current assignee: Vivo Mobile Communication Co Ltd
Priority date: 2022-03-07
Filing date: 2022-08-09
Publication date: 2023-09-19

Abstract

The application discloses a data processing method in a communication network and network side equipment, belonging to the technical field of communication, and the data processing method in the communication network in the embodiment of the application comprises the following steps: the first network element determines a first accuracy of a first model, wherein the first accuracy is used for indicating the accuracy degree of the first model for actual reasoning; and under the condition that the first accuracy meets the preset condition, the first network element retrains the first model or reselects a second model.

Description

Data processing method in communication network and network side equipment

Technical Field

The application belongs to the technical field of communication, and particularly relates to a data processing method in a communication network and network side equipment.

Background

In a communication network, a plurality of network elements can be generally introduced to conduct intelligent model training, and an inference task is executed based on a model obtained by training to obtain an inference result, wherein the inference result can assist in policy decision-making of devices inside and outside the network so as to improve the intelligent degree of the policy decision-making of the devices.

However, in practical applications, the accuracy achieved by the model through the training phase does not represent the accuracy of reasoning that the model can achieve in practical reasoning use, that is, when the reasoning task is performed based on the model, the accuracy of the reasoning results may be lower than the accuracy of the model training phase, in which case, if the reasoning results are provided to the in-network and out-of-network devices, wrong policy decisions may result or improper operations may be performed.

Disclosure of Invention

The embodiment of the application provides a data processing method in a communication network and network side equipment, which can solve the problem that policy decision is affected when the accuracy of model actual reasoning is low.

In a first aspect, a method for processing data in a communication network is provided, the method comprising:

the first network element determines a first accuracy of a first model, wherein the first accuracy is used for indicating the accuracy degree of the first model for actual reasoning;

and under the condition that the first accuracy meets the preset condition, the first network element retrains the first model or reselects a second model.

In a second aspect, there is provided a data processing apparatus in a communication network, the apparatus comprising:

a determination module for determining a first accuracy of a first model, the first accuracy being indicative of how accurate the first model is for actual reasoning;

and the model training module is used for retraining the first model or reselecting the second model under the condition that the first accuracy meets the preset condition.

In a third aspect, a method of data processing in a communication network is provided, the method comprising:

the second network element executes an reasoning task based on a first model, wherein the first model is obtained by training the first network element, and the first network element comprises a model training functional network element;

And transmitting at least one of the use information and the first data of the first model to the first network element, and/or transmitting first indication information to a seventh network element, wherein the first indication information is used for indicating the seventh network element to store the first data of the reasoning task, and the seventh network element comprises a data storage function network element.

In a fourth aspect, there is provided a data processing apparatus in a communication network, the apparatus comprising:

the task execution module is used for executing an reasoning task based on a first model, wherein the first model is obtained by training a first network element, and the first network element comprises a model training function network element;

the sending module is configured to send at least one of usage information of the first model and first data to the first network element, and/or send first indication information to a seventh network element, where the first indication information is used to indicate the seventh network element to store the first data of the reasoning task, and the seventh network element includes a data storage function network element.

In a fifth aspect, a network side device is provided, the terminal comprising a processor and a memory storing a program or instructions executable on the processor, which program or instructions, when executed by the processor, implement the steps of the method as described in the first aspect.

In a sixth aspect, a network side device is provided, including a processor and a communication interface, where the processor is configured to determine a first accuracy of a first model, where the first accuracy is used to indicate an accuracy of the first model for actual reasoning; retraining the first model or reselecting the second model if the first accuracy meets a preset condition.

In a seventh aspect, a network side device is provided, comprising a processor and a memory storing a program or instructions executable on the processor, which when executed by the processor, implement the steps of the method according to the third aspect.

An eighth aspect provides a network side device, including a processor and a communication interface, where the processor is configured to perform an inference task based on a first model, where the first model is obtained by training a first network element, and the first network element includes a model training functional network element; the communication interface is configured to send at least one of usage information of the first model and first data to the first network element, and/or send first indication information to a seventh network element, where the first indication information is used to indicate the seventh network element to store the first data of the reasoning task, and the seventh network element includes a data storage function network element.

A ninth aspect provides a data processing system in a communication network, comprising: a first network side device operable to perform the steps of the method as described in the first aspect and a second network side device operable to perform the steps of the method as described in the third aspect.

In a tenth aspect, there is provided a readable storage medium having stored thereon a program or instructions which when executed by a processor, performs the steps of the method according to the first aspect, or performs the steps of the method according to the third aspect.

In an eleventh aspect, there is provided a chip comprising a processor and a communication interface, the communication interface and the processor being coupled, the processor being for running a program or instructions to implement the method according to the first aspect or to implement the method according to the third aspect.

In a twelfth aspect, there is provided a computer program/program product stored in a storage medium, the computer program/program product being executed by at least one processor to implement the steps of the method as described in the first aspect, or to implement the steps of the method as described in the third aspect.

In the embodiment of the application, since the first accuracy of the first model used for actual reasoning can be determined, and the first model is retrained under the condition that the first accuracy does not meet the preset condition, when the accuracy of the first model used for actual reasoning is reduced, the accuracy of the first model can be adjusted by retrained or reselected to improve the accuracy of the first model used for actual reasoning, thereby better assisting the network internal and external equipment to make correct strategy decisions or combined behavior operation.

Drawings

Fig. 1 is a schematic diagram of a wireless communication system according to an embodiment of the present application;

FIG. 2 is a schematic flow chart of a data processing method in a communication network according to an embodiment of the application;

FIG. 3 is a schematic flow chart of a data processing method in a communication network according to an embodiment of the application;

FIG. 4 is a schematic flow chart of a data processing method in a communication network according to an embodiment of the application;

fig. 5 is a schematic diagram of a data processing apparatus in a communication network according to an embodiment of the present application;

fig. 6 is a schematic diagram of a data processing apparatus in a communication network according to an embodiment of the present application;

Fig. 7 is a schematic structural view of a communication device according to an embodiment of the present application;

fig. 8 is a schematic structural diagram of a network side device according to an embodiment of the present application.

Detailed Description

The technical solutions of the embodiments of the present application will be clearly described below with reference to the drawings in the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which are derived by a person skilled in the art based on the embodiments of the application, fall within the scope of protection of the application.

The terms first, second and the like in the description and in the claims, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the terms so used are interchangeable under appropriate circumstances such that the embodiments of the application are capable of operation in sequences other than those illustrated or otherwise described herein, and that the "first" and "second" distinguishing between objects generally are not limited in number to the extent that the first object may, for example, be one or more. Furthermore, in the description and claims, "and/or" means at least one of the connected objects, and the character "/" generally means a relationship in which the associated object is an "or" before and after.

It should be noted that the techniques described in the embodiments of the present application are not limited to long term evolution (Long Term Evolution, LTE)/LTE evolution (LTE-Advanced, LTE-a) systems, but may also be used in other wireless communication systems, such as code division multiple access (Code Division Multiple Access, CDMA), time division multiple access (Time Division Multiple Access, TDMA), frequency division multiple access (Frequency Division Multiple Access, FDMA), orthogonal frequency division multiple access (Orthogonal Frequency Division Multiple Access, OFDMA), single carrier frequency division multiple access (Single-carrier Frequency Division Multiple Access, SC-FDMA), and other systems. The terms "system" and "network" in embodiments of the application are often used interchangeably, and the techniques described may be used for both the above-mentioned systems and radio technologies, as well as other systems and radio technologies. The following description describes a 5G system for purposes of example and uses 5G terminology in much of the description that follows, but the techniques are also applicable to applications other than 5G system applications, such as generation 6 (6) ^th Generation, 6G) communication system.

Fig. 1 shows a block diagram of a wireless communication system to which an embodiment of the present application is applicable. The wireless communication system includes a terminal 11 and a network device 12. The terminal 11 may be a mobile phone, a tablet (Tablet Personal Computer), a Laptop (Laptop Computer) or a terminal-side Device called a notebook, a personal digital assistant (Personal Digital Assistant, PDA), a palm top, a netbook, an ultra-mobile personal Computer (ultra-mobile personal Computer, UMPC), a mobile internet appliance (Mobile Internet Device, MID), an augmented reality (augmented reality, AR)/Virtual Reality (VR) Device, a robot, a Wearable Device (weather Device), a vehicle-mounted Device (VUE), a pedestrian terminal (PUE), a smart home (home Device with a wireless communication function, such as a refrigerator, a television, a washing machine, or a furniture), a game machine, a personal Computer (personal Computer, PC), a teller machine, or a self-service machine, and the Wearable Device includes: intelligent wrist-watch, intelligent bracelet, intelligent earphone, intelligent glasses, intelligent ornament (intelligent bracelet, intelligent ring, intelligent necklace, intelligent anklet, intelligent foot chain etc.), intelligent wrist strap, intelligent clothing etc.. It should be noted that the specific type of the terminal 11 is not limited in the embodiment of the present application. The network-side device 12 may comprise an access network device or a core network device, wherein the access network device 12 may also be referred to as a radio access network device, a radio access network (Radio Access Network, RAN), a radio access network function or a radio access network element. Access network device 12 may include a base station, a WLAN access point, a WiFi node, or the like, which may be referred to as a node B, an evolved node B (eNB), an access point, a base transceiver station (Base Transceiver Station, BTS), a radio base station, a radio transceiver, a basic service set (Basic Service Set, BSS), an extended service set (Extended Service Set, ESS), a home node B, a home evolved node B, a transmission and reception point (Transmitting Receiving Point, TRP), or some other suitable terminology in the art, and the base station is not limited to a particular technical vocabulary so long as the same technical effect is achieved, and it should be noted that in the embodiment of the present application, only a base station in the NR system is described as an example, and the specific type of the base station is not limited. The core network device may include, but is not limited to, at least one of: core network nodes, core network functions, mobility management entities (Mobility Management Entity, MME), access mobility management functions (Access and Mobility Management Function, AMF), session management functions (Session Management Function, SMF), user plane functions (User Plane Function, UPF), policy control functions (Policy Control Function, PCF), policy and charging rules function units (Policy and Charging Rules Function, PCRF), edge application service discovery functions (Edge Application Server Discovery Function, EASDF), unified data management (Unified Data Management, UDM), unified data repository (Unified Data Repository, UDR), home subscriber server (Home Subscriber Server, HSS), centralized network configuration (Centralized network configuration, CNC), network storage functions (Network Repository Function, NRF), network opening functions (Network Exposure Function, NEF), local NEF (or L-NEF), binding support functions (Binding Support Function, BSF), application functions (Application Function, AF), and the like. It should be noted that, in the embodiment of the present application, only the core network device in the NR system is described as an example, and the specific type of the core network device is not limited.

In the communication network, some network elements can be introduced to conduct intelligent data analysis and generate data analysis results of some tasks, the data analysis results can assist the devices inside and outside the network to conduct policy decision, and the purpose is to improve the intelligent degree of the policy decision of the devices by using an AI method.

For example, the network data analysis function (Network Data Analytics Function, NWDAF) may perform artificial intelligence/Machine learning (AI/ML) model training based on the training data to obtain a model corresponding to an AI task. The NWDAF performs model reasoning based on the AI/ML model and the reasoning input data to obtain a data analysis result (or referred to as a reasoning result, analysis) corresponding to a specific AI reasoning task. The PCF in the network executes intelligent policy control and charging (Policy control and Charging, PCC) policies based on certain reasoning result analysis, such as formulating intelligent user residence policies according to user business behavior analysis results, and improving user business experience; or, AMF performs intelligent mobility management operation based on certain reasoning result analysis, such as intelligent paging of user according to user movement track analysis result, and improves paging accessibility.

The in-network and out-network equipment makes a correct and optimized strategy decision according to the AI data analysis result, provided that the correct data analysis result is needed. If the accuracy of the data analysis result is low, it is provided as error information to the in-network and out-of-network devices for reference, which eventually results in erroneous policy decisions or improper operations, so that it is necessary to ensure the accuracy of the data analysis result. However, in practical application, due to different data distribution, insufficient model generalization capability and the like, the model accuracy achieved by the model in the training stage does not represent the reasoning accuracy achieved by the model in the practical reasoning use (the accuracy of the practical reasoning of the general model is lower than that of the model in the training stage), so that the problem of wrong strategy decision or unsuitable operation execution is easy to occur.

In order to solve the above technical problems, embodiments of the present application provide a data processing method in a communication network and a network side device, where when accuracy of a first model in use for actual reasoning is reduced, accuracy of the first model may be adjusted by retraining the first model or reselecting a second model, so as to improve accuracy of the first model in use for actual reasoning, thereby better assisting in making correct policy decisions or co-setting behavior operations by devices inside and outside the network.

The data processing method and the network side device in the communication network provided by the embodiments of the present application are described in detail below with reference to the accompanying drawings through some embodiments and application scenarios thereof.

As shown in fig. 2, an embodiment of the present application provides a data processing method 200 in a communication network, which may be performed by a first network element, which may be a network side device in the embodiment shown in fig. 1, in other words, the method may be performed by software or hardware installed in the first network element or the network side device, and the method includes the following steps.

S202: the first network element determines a first accuracy of the first model, the first accuracy being indicative of how accurate the first model is for actual reasoning.

The first network element may be a model training function network element in the communication network, where the model training function network element has an AI/ML model training function and may be used for AI/ML model training based on training data. Alternatively, the first network element may be a model training logical network element (Model Training Logical Function, MTLF).

The first model may be trained by the first network element. Specifically, the first network element may acquire training data from other network elements (such as a network element that may provide training data), and perform AI/ML training based on the training data to obtain a first model, where the training data includes input data and tag data, and the tag data corresponds to the input data, and may be a real data value (group score), that is, actually occurring facts and data.

The first network element, in case of training to obtain the first model, may determine a first Accuracy of the first model, which may be an Accuracy of reasoning (Accuracy in Use) of the first model. In this embodiment, the first accuracy may be used to indicate the accuracy of the first model for actual reasoning, and in particular, the first accuracy may be used to indicate the correctness and/or the error degree of the reasoning result in the actual reasoning.

The expression form of the first accuracy may be various, may be a specific percentage value, for example, 90%, may be a classified expression form, for example, high, medium, low, etc., or may be normalized data, for example, 0.9, and the expression form of the first accuracy is not particularly limited herein. The first accuracy may indicate from the front or back a degree of accuracy or a degree of error of the first model in the task's reasoning results. For example, the degree of inference accuracy of the first model may be indicated from the opposite side by calculating an inference error or an inference error rate for the models. The calculation method of the inference error or the inference error rate is various, and may be, for example, mean absolute error (Mean Absolute Error, MAE), mean square error (Mean Square Error, MSE), or the like.

Optionally, as an embodiment, the determining, by the first network element, the first accuracy of the first model may include:

the first network element acquires first data;

the first network element determines a first accuracy based on the first data.

Wherein the first data comprises at least one of:

reasoning input data;

inference result data corresponding to the inference input data;

tag data corresponding to the inferential input data.

The inference input data may be model input data when the first model performs an inference task, the inference result data may be a model output result obtained by inferring the inference input data based on the first model, and the tag data may be actual result data corresponding to the inference input data.

In this embodiment, the first network element may acquire the first data in various manners. Optionally, in a first implementation, the first network element obtains the first data, and the method may include the following steps:

receiving use information of the first model sent by the second network element;

determining source information of the first data according to the use information;

and acquiring the first data according to the source information.

In a second implementation, the first network element obtains the first data, which may include the following steps:

And receiving the first data sent by the second network element.

The second network element may perform an inference task based on the first model. The second network element can be a model reasoning function network element in the communication network, the model reasoning function network element has a model reasoning function, and reasoning input data corresponding to the reasoning task can be deduced based on the first model, so that reasoning result data can be obtained. Alternatively, the second network element may be an analysis logical network element (Analytics Logical Function, anLF). Specifically, the second network element may be a network element that requests the first network element to obtain the first model before, or a network element that requests the first network element to obtain the model information of the first model before, and when receiving the first data sent by the second network element, the first network element may receive the first data from the network element that requests the first network element to obtain the first model or the model information of the first model before.

In a third implementation manner, the first network element acquires the first data, and may include the following steps:

the first data sent by a seventh network element is received, the seventh network element comprises a data storage function network element, and the data storage function network element can be a data analysis repository network element (Analytics Data Repository Function, ADRF).

In the first implementation manner, before receiving the usage information of the first model sent by the second network element, the first network element may train to obtain the first model, and then send the model information of the first model to the second network element. In particular, the second network element may send a model request message to the first network element when performing the reasoning task using the first model, the model request message being used to request acquisition of the first model. After the first network element receives the model request message, if the first network element is trained to obtain a first model, the model information of the first model can be sent to the second network element, if the first network element is not trained on the first model, the first model can be trained, and after the first model is obtained by training, the model information of the first model is sent to the second network element. Or after the first network element is trained and acquired to obtain the first model, model information of the first model can be actively sent to the second network element, and when the second network element needs to execute the reasoning task, the first model can be used for executing the reasoning task without sending a model request message to the first network element. When the first network element trains the first model, training data can be acquired from other network elements (network elements capable of providing training data), and then the first model is obtained based on training data. The first network element may transmit the model information of the first model to the second network element through nnwdaf_mlmodelprovision_notify or nnwdaf_mlmodelinfo_response.

The model information of the first model may include at least one of:

a first model;

and a second accuracy of the first model, the second accuracy being used to indicate how accurate the model output results presented by the first model during the training phase or the testing phase.

The second accuracy may be a training accuracy (accuracy in training, aiT) of the first model, may be used to indicate a degree of accuracy of the model training phase, and in particular, the second accuracy may be used to indicate a degree of correctness and/or a degree of error of the training results of the model training phase. Alternatively, the second accuracy may be equal to the number of times the model decision result is correct divided by the total number of decisions, i.e. the second accuracy = number of times the decision is correct/total number of decisions. The decision result may be correct, where the decision result is consistent with the tag data, and/or the difference between the decision result and the tag data is within an allowable range. Optionally, the first network element may set a verification data set for evaluating the second accuracy of the model, where the verification data set includes data for model input and real tag data, the first network element may input the verification input data into the trained model to obtain output data, compare the output data with the real tag data to determine whether the decision result is correct, and finally obtain the second accuracy of the first model through a calculation formula of the second accuracy.

It should be noted that, the expression form of the second accuracy may be various, may be a specific percentage value, for example, 90%, may be a classified expression form, for example, high, medium, low, etc., or may be normalized data, for example, 0.9, and the expression form of the second accuracy is not particularly limited herein. The second accuracy may indicate from the front or back a degree of accuracy or a degree of error of the first model in the model training phase. For example, the accuracy of the first model during the training phase may be indicated from the reverse side by calculating the training error or training error rate of the model. The training error or training error rate may be calculated by various methods, such as MAE and MSE.

After the first network element sends the model information of the first model to the second network element, the second network element may perform an inference task based on the first model. The specific implementation of the second network element to perform the reasoning task may be referred to the specific implementation of the corresponding steps in the embodiment shown in fig. 3, which will not be described in detail here. After the second network element executes the reasoning task, the second network element can send the use information of the second network element on the first model to the first network element, and the first network element can receive the use information of the second network element on the first model. Optionally, before receiving the usage information of the first model sent by the second network element, the first network element may send a first request message to the second network element, where the first request message is used to request to obtain the usage information of the first model by the second network element. The second network element may send information about use of the first model by the second network element to the first network element when the first request message is received. That is, the usage information of the first model by the second network element may be actively sent to the first network element by the second network element, or may be sent to the first network element by the second network element when the first request message of the first network element is received, which is not limited herein. Wherein the usage information of the first model by the second network element may include at least one of:

Model identification information of the first model;

task identification information of an inference task executed based on the first model;

condition definition information of reasoning tasks;

the object information of the task is inferred.

Model identification information (Model ID) may be used to indicate which Model the second network element uses, and Model identification information of the first Model may be used to indicate the first Model. Task identification information (analytical ID) may identify a type of inference task that may be used to determine a corresponding model, and task identification information of an inference task performed based on the first model may be used to determine the first model. The inference task condition definition information (analytics filter information) can be used to define the scope of execution of the inference task, such as time scope, area scope, etc. The inference task object information (analysis targets) may be used to indicate an object for which the inference task is directed, such as Target UE (Target UE, i.e. when the task Target is a certain UE), or may be a certain NF instance.

Optionally, the usage information of the first model by the second network element may also include at least one of reasoning input data, reasoning result data corresponding to the reasoning input data, and tag data corresponding to the reasoning input data when the second network element performs the reasoning task based on the first model. The inference input data may be collected by the second network element, for example, the second network element may collect the inference input data to perform inference when receiving task requests of other network elements (such as consumer network elements) for an inference task, or the inference input data may be actively collected by the second network element. The inference result data may be derived by the second network element based on the first model and the inference input data when performing the inference task. The tag data may be obtained by the second network element from other network elements (e.g. source devices of the tag data), for example, the second network element may send a data obtaining request to the other network element to request obtaining the tag data.

After receiving the usage information of the second network element on the first model, the first network element may determine source information of the first data according to the usage information. Wherein the source information of the first data includes at least one of:

the third network element is used for providing reasoning input data corresponding to the reasoning task;

and the fourth network element is used for providing label data corresponding to the reasoning task.

The third network element may comprise a source device for reasoning about the input data. The third network element may be determined by the first network element, specifically, the first network element may determine, according to information such as condition definition information (analytics filter information) of an inference task and object information (analysis target) of the inference task in the usage information, an object and a scope related to the inference task, and according to the object and scope and metadata (metadata) information, determine which network elements specifically obtain inference input data corresponding to the inference task, where the determined network elements are the third network element.

The fourth network element may comprise a source device of the tag data. The fourth network element may be determined by the first network element, specifically, the first network element may determine, according to an output data type (data type) of the first model, a corresponding network element device type (NF type) that may provide the data type, and then determine, according to an object of an inference task, limitation information, etc., a network element instance corresponding to the network element device type, and use the network element instance as the fourth network element. For example, the first model is UE mobility model, the output data type of the first model is UE location, based on the output data type, it may be determined that AMF type is UE location data information, and according to constraint information such as reasoning task object UE1 and AOI, the corresponding AMF instance is AMF1 queried from UDM or NRF, where the AMF1 is the fourth network element, and the first network element may use AMF1 as a source of tag data, and obtain a real UE location tag data value from AMF 1.

Optionally, the source information may further include at least one of the following:

a first model corresponding to the reasoning task;

an input data type of the first model;

the output data type of the first model.

The first model may be determined by the first network element based on model identification information and/or task identification information in the usage information. For example, the usage information includes task identification information (analysis ID), and a mapping relationship is provided between the task identification information and model identification information (model ID), and for a certain analysis ID (for example, analysis id=ue mobility, used for predicting a user movement track), the corresponding model identification information may be determined to be model 1 based on the mapping relationship, and the model corresponding to the model 1 is the first model corresponding to the reasoning task.

The input data type (which may be referred to as metadata information) and the output data type of the first model are related to the inference task or result data for prediction to which the first model is specifically applied. For example, if the first model is used for predicting the movement track of the user, the input data type of the first model may include UE ID, time, current service status of the UE, and the output data type may include UE location (e.g. TA/cell).

After determining the source information of the first data, the first network element may acquire the first data according to the source information. The first network element obtains first data according to the source information, and may include at least one of the following:

an input data acquisition request message is sent to a third network element, and the input data acquisition request message is used for requesting acquisition of reasoning input data;

and sending a tag data acquisition request message to the fourth network element, wherein the tag data acquisition request message is used for requesting to acquire tag data.

The input data acquisition request message may be used by the third network element to determine which inference input data to feed back to the first network element, and may include at least one of:

reasoning type information of input data;

reasoning object information corresponding to input data;

and reasoning time information corresponding to the input data.

The type information of the reasoning input data, the object information corresponding to the reasoning input data and the time information corresponding to the reasoning input data are respectively determined by the first network element according to the input data type of the reasoning process, the reasoning object and the time aimed by the reasoning, namely, the first network element determines the type of the reasoning input data required to be acquired according to the input data type of the reasoning process, determines the object of the reasoning input data required to be acquired according to the reasoning object, and determines the time information (time stamp, time period and the like) of the reasoning input data according to the time aimed by the reasoning. Where the inference process is a statistical calculation made for some past time or a prediction made for some future time, the time information may be the past time or the future time.

The tag data obtaining request message may be used by the fourth network element to determine which tag data to feed back to the first network element, where the tag data obtaining request message may include at least one of:

type information of the tag data;

object information corresponding to the tag data;

time information corresponding to the tag data.

The type information of the tag data, the object information corresponding to the tag data and the time information corresponding to the tag data are respectively determined by the first network element according to the output data type of the reasoning process, the reasoning object and the time aimed by the reasoning, namely, the first network element determines the type of the tag data to be acquired according to the output data type of the reasoning process, determines the object of the tag data to be acquired according to the reasoning object and determines the time information (time stamp, time period and the like) of the tag data according to the time aimed by the reasoning. Where the inference process is a statistical calculation made for some past time or a prediction made for some future time, the time information may be the past time or the future time.

For example, the MTLF (first network element) sends a tag data acquisition request message to the AMF or the LMF (third network element), where the request message carries that the tag data type is UE location, the object information is UE ID1, and the time information is a specific time period, and the request message is used to request the AMF/LMF to feed back the value of UE location of UE ID1 in a specific time period.

After receiving the input data acquisition request message, the third network element may send the corresponding inferred input data to the first network element. In this way, the first network element can obtain the inferential input data from the third network element. Likewise, after receiving the tag data obtaining request message, the fourth network element may send the corresponding tag data to the first network element, and the first network element may obtain the tag data from the fourth network element. It should be noted that, if the second network element performs one or more reasoning processes to obtain multiple reasoning output results when performing the reasoning task, the first network element needs to correspondingly obtain multiple tag data values corresponding to the multiple reasoning output results from the fourth network element.

It should be further noted that, how the first network element obtains the inference input data and/or the tag data is described above, for the inference result data, the first network element may be obtained from the second network element when obtaining the inference result data, and the inference result data obtained from the second network element may be obtained after the second network element performs the inference task based on the first model. In addition, because the inference result data can be obtained by inference based on the inference input data and the first model, when the first network element obtains the inference result data, the first network element can also obtain the inference result data by inference based on the inference input data and the first model after obtaining the inference input data, so that the first network element can not need to obtain the inference result data from other network elements, thereby simplifying the data obtaining step.

In the second implementation manner, that is, in the case that the first network element obtains the first data by receiving the first data sent by the second network element, the first data obtained by the first network element may include at least one of inference input data, inference result data, and tag data. The inference input data may be collected by the second network element, for example, the second network element may collect the inference input data to perform inference when receiving task requests of other network elements (such as consumer network elements) for an inference task, or the inference input data may be actively collected by the second network element. The inference result data may be derived by the second network element based on the first model and the inference input data when performing the inference task. The tag data may be obtained by the second network element from other network elements (such as a source device of the tag data), for example, the second network element may obtain the tag data by sending a data obtaining request to the other network element.

Optionally, before receiving the first data sent by the second network element, the first network element may further include the following steps:

the first network element sends a second request message to the second network element, where the second request message is used to request to acquire the first data collected by the second network element.

That is, the first network element may send a second request message to the second network element when acquiring the first data, where the second request message is used to request to acquire the first data collected by the second network element. The second network element may send the first data to the first network element upon receiving the second request message.

Alternatively, the second request message may be a subscription message. Optionally, the second request information includes at least one of:

the identification information of the reasoning task;

the limiting condition information of the reasoning task;

reasoning object information of the task;

identification information of the first model;

input data type information of the first model;

output data type information of the first model.

Optionally, the second request information may further include a request reason, where the request reason may be, for example, that the first model needs to be retrained, or that the accuracy of the first model does not meet the accuracy requirement or decreases, etc.

In the third implementation manner, that is, in the case that the first network element obtains the first data by receiving the first data sent by the seventh network element, the first data obtained by the first network element may include at least one of inference input data, inference result data, and tag data. The first data may be stored by the second network element into the seventh network element, and specifically, after the second network element performs the reasoning task, the second network element may send first indication information to the seventh network element, where the first indication information is used to indicate the seventh network element to store the first data of the reasoning task. Optionally, the first indication information includes at least one of:

The identification information of the reasoning task;

the limiting condition information of the reasoning task;

reasoning object information of the task;

reasoning input data corresponding to the reasoning task;

reasoning result data corresponding to the reasoning task;

label data corresponding to the reasoning task;

the reason information may be, for example, that the second network element completes the reasoning task, the seventh network element needs to store the first data periodically, the accuracy of the first model used when the second network element performs the reasoning task does not meet the accuracy requirement or is degraded, and so on.

The first data stored by the seventh network element may be sent by the second network element to the seventh network element. The manner in which the second network element obtains the first data may be: the inference input data is collected by the second network element, for example, the second network element may collect the inference input data to perform inference when receiving task requests of other network elements (such as consumer network elements) for the inference task, or the inference input data may be actively collected by the second network element. The inference result data may be derived by the second network element based on the first model and the inference input data when performing the inference task. The tag data may be obtained by the second network element from other network elements (such as a source device of the tag data), for example, the second network element may obtain the tag data by sending a data obtaining request to the other network element.

Optionally, before receiving the first data sent by the seventh network element, the first network element may further include the following steps:

the first network element sends a third request message to the seventh network element, where the third request message is used to request to acquire the first data.

That is, when the first network element acquires the first data, the first network element may send a third request message to the seventh network element, where the third request message is used to request to acquire the first data stored in the seventh network element. The seventh network element may send the first data to the first network element in case the third request message is received.

Alternatively, the third request message may be a subscription message. Optionally, the third request information includes at least one of:

the identification information of the reasoning task;

the limiting condition information of the reasoning task;

reasoning object information of the task;

identification information of the first model;

input data type information of the first model;

output data type information of the first model.

Optionally, the third request information may further include a request reason, where the request reason may be, for example, that the first model needs to be retrained, or that the accuracy of the first model does not meet the accuracy requirement or decreases, etc.

It should be noted that, in practical application, the first network element may acquire the first data in any one or more of the three manners, that is, the first network element may acquire the first data from the second network element, and/or determine source information of the first data according to usage information of the second network element on the first model, and acquire the first data according to the source information, and/or acquire the first data from the seventh network element.

After the first network element obtains the first data, a first accuracy of the first model may be determined according to the first data.

When determining the first accuracy according to the first data, optionally, if the first data includes the inference input data and the tag data and does not include the inference result data, the first network element may first input the inference input data into the first model, determine the inference result data corresponding to the inference input data, and then determine the first accuracy according to the inference result data and the tag data. If the first data includes the inference result data and the tag data, the first network element may directly determine the first accuracy according to the inference result data and the tag data.

When determining the first accuracy according to the inference result data and the tag data, specifically, the inference result data and the tag data may be compared, and a ratio of the number of times the inference result is correct to the total number of times of inference may be determined, where the ratio is the first accuracy of the first model. The reasoning result is correct, namely the reasoning result data is consistent with the label data, or the difference value between the reasoning result data and the label data is within an allowable range. The first accuracy may be expressed in a percentage (e.g., 90%), a classification expression (e.g., high, medium, low), or a normalized value (e.g., 0.9), which is not particularly limited herein.

S204: the first network element retrains or reselects the first model if the first accuracy meets a preset condition.

After determining the first accuracy of the first model, the first network element may determine whether the first accuracy meets a preset condition to determine whether the first model needs to be retrained or the second model needs to be reselected.

Alternatively, as an embodiment, the first accuracy meeting the preset condition may include at least one of:

the first accuracy is smaller than the second accuracy, and the second accuracy is used for indicating the accuracy degree of a model output result presented by the first model in a training stage or a testing stage;

the first accuracy is smaller than the second accuracy, and the difference between the first accuracy and the second accuracy is larger than a preset value;

the first accuracy is smaller than a preset accuracy, which may be set according to actual requirements, and is not particularly limited herein.

In the case that the first accuracy meets the preset condition, it may be stated that the accuracy of the first model actually used for reasoning does not meet the actual requirement, and at this time, the first network element may retrain the first model or reselect the second model. The first model may be retrained by correcting the first model without changing the model structure of the first model, or by training the first model with a new model structure while changing the model structure of the first model. Alternatively, the retraining of the model may be accomplished in two ways, one by retraining the model from scratch based on the training data and the other by fine-tuning the first model based on the training data, which may converge faster and save resources. The re-selection of the second model may be the re-selection of a new, already existing, other model, wherein the re-selected second model may be a model with an accuracy above the first threshold or meeting the performance requirements of the model.

Optionally, as an embodiment, the first network element may include the following steps when retraining the first model:

acquiring target training data, wherein the target training data comprises target input data and target label data corresponding to the target input data;

the first model is retrained based on the target training data.

The target training data is different from training data used when the first model was previously trained. When the first model is retrained, new training data can be used for model training so as to adjust the accuracy of the first model and improve the accuracy of the retrained first model.

Optionally, as an embodiment, the first network element acquires the target training data, which may include at least one of the following:

acquiring first training data used in training a first model;

determining a fifth network element; acquiring second training data from a fifth network element, wherein the fifth network element is used for providing the training data;

determining a sixth network element; the reasoning data is obtained from a sixth network element, which is arranged to provide the reasoning data.

The first training data, the second training data, and the reasoning data each include input data and label data, respectively.

The fifth network element may comprise a source device of training data. Optionally, as an embodiment, the determining, by the first network element, the fifth network element may include:

and determining a fifth network element according to the second information.

The second information includes task identification information of an inference task executed based on the first model and/or condition definition information of the inference task. Wherein task identification information (analytical ID) may identify a type of inference task that may be used to determine a corresponding model. The inference task condition definition information (analytics filter information) can be used to define the scope of execution of the inference task, such as time scope, area scope, etc.

The sixth network element may comprise a source device for the inferred data. Optionally, as an embodiment, the determining, by the first network element, the sixth network element may include:

and determining a sixth network element according to the second information.

The first network element may obtain the second model after retraining the first model or reselecting the second model. After the second model is obtained, optionally, as an embodiment, at least one of the following may be further included:

transmitting the re-trained second model or the re-selected model information of the second model to a second network element, and executing an reasoning task by the second network element based on the second model;

And transmitting the model information of the second model obtained by retraining or the reselected second model to a seventh network element, and storing the model information of the second model by the seventh network element.

The second network element may be the second network element in the current task, that is, the second network element that sends its usage information of the first model to the first network element after performing the reasoning task based on the first model. After the first network element sends the model information of the second model to the second network element, the second network element can re-execute the previous reasoning task or execute the new reasoning task based on the second model, and the accuracy of the reasoning result data is higher because the second model is a re-trained model. Optionally, the second network element may also be other network elements that perform the reasoning task based on the second model, and after the first network element sends the model information of the second model to other second network elements, other second network elements may perform the reasoning task based on the second model, and because the second model is a retrained model, the accuracy of the reasoning result data is higher.

The seventh network element comprises a data storage function network element, i.e. a network element storing model information of the second model. Optionally, the seventh network element may be a data analysis repository network element (Analytics Data Repository Function, ADRF). Storing the model information of the second model in the seventh network element may facilitate other network elements to find the model or data.

The model information of the second model may include at least one of:

model identification information of the second model;

task identification information of the reasoning task executed based on the second model;

application range information of the second model;

a third accuracy of the second model, the third accuracy being used to indicate a degree of accuracy of the model output result presented by the second model in the training phase or the testing phase;

training data of the second model;

and a second model.

Model identification information (Model ID) of the second Model is used to indicate the second Model. Task identification information (analytical ID) of the inference task performed based on the second model may be used to determine the corresponding second model. The scope of applicability information of the second model may be used to define the scope of the execution of the inference task, such as time scope, area scope, etc. The third accuracy of the second model may be referred to as training accuracy (accuracy in training, aiT) of the second model for describing the accuracy of the recognition or decision that the model can reach after training, and in particular, the third accuracy may be used to indicate the accuracy and/or the error level of the model output results presented by the second model during the training phase or the testing phase. The third accuracy may be determined in the same manner as the second accuracy of the first model, and the third accuracy may be expressed in the same manner as the second accuracy, and the description thereof will not be repeated. The training data of the second model is training data used when training the second model, and may include input data and tag data, and specifically may be target training data obtained when retraining the first model. The second model includes, but is not limited to, description information and/or model files for the second model, which may include elements such as complete network structure and parameter information for generating the second model.

It should be noted that, the model information sent by the first network element to the second network element and the seventh network element may be the same or different. For example, the first network element may send the second model to the second network element, and may send the model identification information of the second model, task identification information based on an inference task performed by the second model, application scope information of the second model, third accuracy of the second model, training data of the second model, and the second model to the seventh network element.

As shown in fig. 3, an embodiment of the present application provides a data processing method 300 in a communication network, where the method may be performed by a second network element, which may be a network side device in the embodiment shown in fig. 1, in other words, the method may be performed by software or hardware installed in the second network element or the network side device. It should be noted that, the second network element may be deployed as a different network element device independently from the first network element in the embodiment shown in fig. 2, or may be deployed in the same network element device, for example, in an NWDAF, where the NWDAF may provide both a model training function and a model reasoning function. The method shown in fig. 3 includes the following steps.

S302: the second network element performs an inference task based on a first model, the first model being trained by the first network element, the first network element comprising a model training function network element.

The first model may be trained by the first network element based on training data that may be obtained by the first network element from other network elements (network elements that may provide the training data). In this step, the second network element may perform an inference task based on the first model trained by the first network element.

Optionally, as an embodiment, the second network element may further acquire the first model from the first network element before performing the reasoning task based on the first model, and specifically may include the following steps:

sending a model request message to a first network element, wherein the model request message is used for requesting to acquire a first model;

and receiving the model information of the first model sent by the first network element.

After the second network element sends the model request message to the first network element, if the first network element is trained to obtain the first model, the second network element can send the model information of the first model to the second network element, and at this time, the second network element can receive the model information of the first model sent by the first network element. If the first network element does not train the first model, training data can be obtained from other network elements (network elements capable of providing training data), then the first model is obtained based on training data, and then model information of the first model is sent to the second network element, and at this time, the second network element can receive the model information of the first model sent by the first network element. The second network element may receive the model information of the first model through nnwdaf_mlmodelprovision_notify or nnwdaf_mlmodelinfo_response.

The model information of the first model includes at least one of:

a first model;

The second accuracy of the first model may be a training accuracy (accuracy in training, aiT) of the first model, may be used to indicate a degree of accuracy of the model training phase, and in particular, the second accuracy may be used to indicate a degree of correctness and/or a degree of error of the training results of the model training phase. Alternatively, the second accuracy may be equal to the number of times the model decision result is correct divided by the total number of decisions, i.e. the second accuracy = number of times the decision is correct/total number of decisions. The decision result may be correct, where the decision result is consistent with the tag data, and/or the difference between the decision result and the tag data is within an allowable range. The second accuracy may be expressed in a percentage (e.g., 90%), a classification expression (e.g., high, medium, low), or a normalized value (e.g., 0.9), which is not particularly limited herein. The second accuracy may indicate from the front or back a degree of accuracy or a degree of error of the first model in the model training phase. For example, the accuracy of the first model during the training phase may be indicated from the reverse side by calculating the training error or training error rate of the model. The training error or training error rate may be calculated by various methods, such as MAE and MSE.

The second network element may perform an inference task based on the first model after receiving the model information of the first model.

Optionally, as an embodiment, the second network element performs the inference task based on the first model, including at least one of:

receiving a task request aiming at an reasoning task, which is sent by an eighth network element; performing an inference task based on the first model;

acquiring a task request of an reasoning task triggered by the second network element simulation; an inference task is performed based on the first model.

That is, the second network element performs the reasoning task based on the first model, and may be triggered when the reasoning task sent by the eighth network element is received, and/or the second network element sets a verification test stage, in which the second network element itself triggers the reasoning task in a simulation manner, to measure and calculate the accuracy of model reasoning. The eighth network element includes a consumer network element, specifically may be a consumer NF, and the consumer NF may be a 5G network element or an AF terminal. The eighth network element may send the reasoning task to the second network element through nnwdaf_analytics subscription_subscriber or nnwdaf_analytics info_request.

Under the condition that the reasoning task is sent to the second network element by the eighth network element, if the reasoning task carries the reasoning input data, the reasoning input data can be input into the first model, and the reasoning result data can be obtained. If the reasoning task does not carry the reasoning input data, the second network element may include:

An input data acquisition request message is sent to a third network element, and the input data acquisition request message is used for requesting to acquire reasoning input data corresponding to a reasoning task;

receiving reasoning input data sent by a third network element;

and inputting the reasoning input data into the first model to obtain the reasoning result data.

The third network element may be the third network element in the embodiment shown in fig. 2, and the third network element may provide inference input data corresponding to the inference task. The input data acquisition request message may be used by the third network element to determine which inferential input data to feed back to the second network element, and may include at least one of: reasoning type information of input data; reasoning object information corresponding to input data; and reasoning time information corresponding to the input data. Reference may be made in particular to the corresponding content of the embodiment shown in fig. 2, and the description will not be repeated here. The second network element may send the input data acquisition Request message to the third network element through nnwdaf_analytics description_subscnribe or nnwdaf_analytics info_request.

After receiving the input data acquisition request message, the third network element may send the corresponding inferred input data to the second network element. After receiving the reasoning input data sent by the third network element, the second network element can input the reasoning input data into the first model, and corresponding reasoning result data can be obtained.

For example, the second network element performs inference calculation by using the value of the model1 corresponding to the analysis id=ue mobility (for identifying an inference task type, such as analysis id=ue mobility, for predicting a user movement track), and the value of the input data (such as UE ID, time, and current service state of the UE) corresponding to the model1, so as to obtain an output value of the inference result UE location.

When the reasoning task is triggered by the second network element, the second network element can generate or determine the reasoning input data by the second network element when the reasoning task is executed based on the first model, and then input the reasoning input data into the first model, so that the reasoning result data can be obtained.

It should be noted that, when the second network element performs the reasoning task, the value of a plurality of output results may be obtained by performing the reasoning calculation process once; alternatively, performing multiple inferences can obtain values of the results of the multiple inferences output.

Optionally, as an embodiment, if the reasoning task performed by the second network element is sent by the eighth network element, after the second network element performs the reasoning task by using the first model, the second network element may further send the obtained reasoning result data to the eighth network element, so as to assist the eighth network element in making a policy decision.

S304: and transmitting at least one of the use information of the first model and the first data to a first network element, and/or transmitting first indication information to a seventh network element, wherein the first indication information is used for indicating the seventh network element to store the first data of the reasoning task, and the seventh network element comprises a data storage function network element.

After performing the reasoning task, the second network element may send at least one of its usage information for the first model and the first data to the first network element. The first network element, after receiving at least one of the usage information and the first data, may determine whether to retrain the first model or reselect the second model, and the specific implementation may refer to the corresponding steps of the embodiment shown in fig. 2, which is not repeated here.

The information of the second network element on the use of the first model comprises at least one of the following:

model identification information of the first model;

condition definition information of reasoning tasks;

the object information of the task is inferred.

The Model identification information (Model ID) of the first Model may be used to indicate that the Model used by the second network element is the first Model. Task identification information (analytical ID) may identify a type of inference task, and task identification information of an inference task performed based on the first model may be used to determine the first model. The inference task condition defining information (analytics filter information) may be used to define an execution scope, such as a time scope, a region scope, etc., in which the second network element performs the inference task. The inference task object information (analysis Target) may be used to indicate an object for which the inference task performed by the second network element is directed, such as a Target UE (i.e. when the task Target is a certain UE), or may be a certain NF instance. Optionally, the usage information of the first model by the second network element may also include reasoning input data and/or reasoning result data when the second network element performs the reasoning task based on the first model.

Optionally, as an embodiment, before sending the usage information of the first model to the first network element, the second network element further includes at least one of the following:

the second network element receives a first request message sent by the first network element, where the first request message is used to request to obtain usage information of the first model by the second network element.

That is, the second network element may send the usage information of the first model to the first network element upon receiving the first request message of the first network element.

The first data may include at least one of:

reasoning input data;

inference result data corresponding to the inference input data;

tag data corresponding to the inferential input data.

Optionally, as an embodiment, before the second network element sends the first data to the first network element, the second network element further includes at least one of the following:

the second network element collects reasoning input data;

The second network element collects tag data corresponding to the inferential input data.

That is, in case the first data comprises inferential input data, the inferential input data may be collected by the second network element. In case that the first data includes tag data corresponding to the inferential input data, the tag data may also be collected by the second network element. Alternatively, in case that the first data includes the inference result data, the inference result data may be obtained after the second network element performs the inference based on the first model and the inference input data.

Optionally, the second network element collects the inferential input data, which may include at least one of:

the second network element collects reasoning input data under the condition that a task request of an eighth network element aiming at a reasoning task is received, wherein the eighth network element comprises a consumer network element;

the second network element actively collects the inferential input data.

That is, the second network element may collect the inference input data to perform the inference when receiving the task request of the inference task, or the second network element may actively collect the inference input data. The task request of the reasoning task may be sent by the eighth network element, where the eighth network element includes a consumer network element, and specifically may be a consumer NF, where the consumer NF may be a 5G network element or an AF terminal, etc.

Optionally, the second network element collects tag data corresponding to the reasoning input data, which may include:

the second network element sends a data acquisition request to the fourth network element, wherein the data acquisition request is used for requesting to acquire tag data corresponding to the reasoning input data.

The fourth network element may be a source device of the tag data. When the second network element collects the tag data, the second network element can send a data request to the fourth network element, and the fourth network element can send the corresponding tag data to the second network element under the condition that the fourth network element receives the data acquisition request.

Optionally, before sending the first data to the first network element, the second network element may further include the following steps:

and receiving a second request message sent by the first network element, wherein the second request message is used for requesting to acquire the first data collected by the second network element.

That is, the second network element may send the first data to the first network element upon receiving the second request message sent by the first network element.

the identification information of the reasoning task;

the limiting condition information of the reasoning task;

reasoning object information of the task;

Identification information of the first model;

input data type information of the first model;

output data type information of the first model.

In this embodiment, after the second network element performs the reasoning task, the second network element may also send first indication information to the seventh network element, where the first indication information is used to indicate the seventh network element to store the first data of the reasoning task, and the seventh network element includes a network element with a data storage function, and may specifically be ADRF.

Optionally, the first indication information includes at least one of:

the identification information of the reasoning task;

the limiting condition information of the reasoning task;

reasoning object information of the task;

reasoning input data corresponding to the reasoning task;

reasoning result data corresponding to the reasoning task;

label data corresponding to the reasoning task;

the storage of the cause information, for example, the second network element completes task reasoning, the seventh network element needs to store the first data periodically, the accuracy of the first model used when the second network element performs the reasoning task does not meet the accuracy requirement or is reduced, and the like.

After the second network element sends the first indication information to the seventh network element, the seventh network element may store the first data according to the first indication information. The first data may be determined by the second network element by the method described above, and will not be repeated here. After the seventh network element stores the first data, when the first network element needs to acquire the first data, the first network element may acquire the first data from the seventh network element, and a specific implementation manner may refer to corresponding content in the embodiment shown in fig. 2, which is not repeated herein.

Optionally, as an embodiment, the second network element may further include:

and receiving model information of a second model sent by the first network element, wherein the second model is obtained after the first network element retrains the first model or is a model reselected by the first network element.

Specifically, after receiving the usage information and/or the first data of the second network element on the first model, the first network element may retrain the first model or reselect the second model if it is determined that the first accuracy of the first model meets the preset condition. After the first network element retrains to obtain the second model or reselects to obtain the second model, the second network element may send the model information of the second model to the second network element, where the second network element may receive the model information of the second model sent by the first network element. The second network element may then re-perform the previously performed reasoning tasks using the second model or perform new reasoning tasks. Because the second model is a retrained model, the accuracy of the inference result data is higher.

The model information of the second model may include at least one of:

model identification information of the second model;

task identification information of an inference task executed based on the second model;

application range information of the second model;

training data of the second model;

and a second model.

Model identification information (Model ID) of the second Model is used to indicate the second Model. Task identification information (analytical ID) of the inference task performed based on the second model may be used to determine the corresponding second model. The scope of applicability information of the second model may be used to define the scope of the execution of the inference task, such as time scope, area scope, etc. The third accuracy of the second model may be referred to as training accuracy (accuracy in training, aiT) of the second model for describing the accuracy of the recognition or decision that the model can reach after training, and in particular, the third accuracy may be used to indicate the accuracy and/or the error level of the model output results presented by the second model during the training phase or the testing phase. The third accuracy may be determined in the same manner as the second accuracy of the first model in the embodiment shown in fig. 2, and the third accuracy may be expressed in the same manner as the second accuracy, which will not be repeated here. The training data of the second model is training data used when training the second model, and may include input data and tag data, and specifically may be target training data obtained when retraining the first model. The second model includes, but is not limited to, description information and/or model files for the second model, which may include elements such as complete network structure and parameter information for generating the second model.

In the embodiment of the application, since the first accuracy of the first model used for actual reasoning can be determined, and the first model is retrained under the condition that the first accuracy does not meet the preset condition, when the accuracy of the first model used for actual reasoning is reduced, the accuracy of the first model can be adjusted in a retrained mode to improve the accuracy of the first model used for actual reasoning, thereby better assisting the network internal and external equipment to make correct strategy decisions or combined behavior operations.

In a possible application scenario, the method for processing data in a communication network according to the embodiment of the present application may be as shown in fig. 4. The data processing method in the communication network shown in fig. 4 may include the following steps.

Step 1: the first network element obtains training data from the fifth network element.

Step 2: the first network element trains the first model based on the training data.

Step 3: the second network element sends a model acquisition request for the first model to the first network element.

Step 4: the first network element sends the model information of the first model to the second network element.

It should be noted that the sequence of steps 1 to 4 may be step 1, step 2, step 3, and step 4, or may be step 3, step 1, step 2, and step 4.

Step 5: the second network element receives the reasoning task of the eighth network element.

It should be noted that the sequence of steps 1-4 and 5 may be steps 1-4 and 5, or may be steps 5 and 1-4.

Step 6: the second network element obtains the inference input data from the third network element and performs an inference task based on the inference input data and the first model.

Step 7: the second network element obtains the tag data from the fourth network element.

Step 8: the second network element sends the reasoning result data to the eighth network element.

It should be noted that the second network element may also be provided with a verification test stage, in which the second network element may itself simulate and trigger the reasoning task, for measuring and calculating the accuracy of model reasoning. Specifically, the second network element may simulate triggering the reasoning task and perform the reasoning task based on the first model after performing step 4, where step 5-7 is replaced by a step of the second network element simulating triggering the reasoning task and performing the reasoning task based on the first model, and fig. 3 only illustrates that step 5-7 is performed by the second network element.

Step 9: the first network element sends a first request message to the second network element, wherein the first request message is used for requesting to acquire the use information of the second network element on the first model and/or the first data collected by the second network element.

Step 10: the second network element sends the usage information and/or the first data of the first model to the first network element.

The usage information of the first model includes at least one of:

model identification information of the first model;

condition definition information of reasoning tasks;

the object information of the task is inferred.

The first data includes at least one of:

reasoning input data;

inference result data corresponding to the inference input data;

tag data corresponding to the inferential input data.

Step 11: the first network element determines source information of the first data according to the use information.

The first data includes at least one of:

reasoning input data;

inference result data corresponding to the inference input data;

tag data corresponding to the inferential input data.

The source information of the first data includes at least one of:

Step 12a: the first network element sends an input data acquisition request message to the third network element.

The input data acquisition request message is used to request acquisition of inferred input data. The input data acquisition request message includes at least one of the following:

Reasoning type information of input data;

reasoning object information corresponding to input data;

and reasoning time information corresponding to the input data.

Step 12b: the first network element sends a tag data acquisition request message to the fourth network element.

The tag data acquisition request message is used for requesting acquisition of tag data. The tag data acquisition request message includes at least one of the following:

type information of the tag data;

object information corresponding to the tag data;

time information corresponding to the tag data.

It should be noted that the first network element may perform at least one of the steps 12a and 12b.

It should be further noted that, the first network element may optionally perform step 11, step 12a and step 12b. For example, if the second network element does not send the usage information of the first model and only sends the first data, and the first data includes the inference input data, the inference result data, and the tag data, the first network element may not execute step 11, step 12a, and step 12b; if the second network element transmits both the usage information for the first model and the first data, the first network element may perform steps 11 and 12a in the case where only the tag data is included in the first data, and the first network element may perform steps 11 and 12b in the case where only the inference input data is included in the first data. The specific steps executed by the first network element may be determined according to actual situations, so long as the first network element is guaranteed to obtain the inference input data, the inference result data and the tag data, which are not limited herein.

In addition, when the first network element acquires the first data, the first data may be acquired from the seventh network element, and fig. 4 only illustrates an example of acquiring the usage information and/or the first data of the first model from the second network element.

Step 13: the first network element determines a first accuracy of the first model.

The first accuracy is used to indicate the accuracy of the first model for actual reasoning, which may be a degree of correctness or a degree of error.

Step 14: the first network element determines whether the first accuracy meets a preset condition.

The first accuracy meets a preset condition including at least one of:

the first accuracy is less than a preset accuracy.

At least one of the steps 15a-15c may be performed in case the first accuracy fulfils a preset condition. In case the first accuracy does not meet the preset condition, it may not be necessary to perform subsequent steps, here illustrated by way of example of performing at least one of the steps 15a-15 c.

Step 15a: the first network element obtains first training data for use in training the first model.

Step 15b: the first network element obtains second training data from the fifth network element.

Step 15c: the first network element obtains the reasoning data from the sixth network element.

Step 16: the first network element re-selects the first model or re-trains the first model based on the target training data to obtain a second model.

The target training data here includes data obtained in at least one of the above steps 15a-15 c. I.e. the target training data comprises at least one of the following:

first training data for use in training the first model.

And second training data acquired from the fifth network element.

And (3) reasoning data acquired from the sixth network element.

Step 17: the first network element sends the model information of the second model to the second network element.

The model information of the second model includes at least one of:

model identification information of the second model;

application range information of the second model;

Training data of the second model;

and a second model.

Step 18: the first network element sends the model information of the second model to other second network elements.

Step 19: the first network element sends the model information of the second model to the seventh network element.

The first network element may or may not perform the above steps 17-19 in case the first model is retrained or reselected to obtain the second model, and may perform at least one of the steps 17-19 in case it is performed.

The specific implementation of the steps shown in fig. 4 may refer to the specific implementation of the corresponding steps in fig. 2 and 3, and the description thereof will not be repeated here.

According to the data processing method in the communication network provided by the embodiment of the application, the execution main body can be a data processing device in the communication network. In the embodiment of the present application, a data processing device in a communication network is described by taking a data processing method in the communication network as an example.

Fig. 5 is a schematic structural diagram of a data processing apparatus in a communication network according to an embodiment of the present application, which may correspond to the first network element in other embodiments. As shown in fig. 5, the apparatus 500 includes the following modules.

A determining module 501 for determining a first accuracy of a first model, the first accuracy being indicative of how accurate the first model is for actual reasoning;

the model training module 502 is configured to retrain the first model or reselect the second model if the first accuracy meets a preset condition.

Optionally, as an embodiment, the first accuracy is used to indicate at least one of:

the first model is used for the accuracy degree of the reasoning result in actual reasoning;

the first model is used for the error degree of the reasoning result in the actual reasoning.

Optionally, as an embodiment, the determining module 501 is configured to:

acquiring first data; determining the first accuracy based on the first data;

the first data includes at least one of:

reasoning input data;

reasoning result data corresponding to the reasoning input data;

tag data corresponding to the inferential input data.

Optionally, as an embodiment, the determining module 501 is configured to:

receiving usage information of the first model sent by a second network element;

and acquiring the first data according to the source information.

Optionally, as an embodiment, the determining module 501 is configured to:

and receiving the first data sent by a second network element, wherein the second network element comprises a model reasoning function network element.

Optionally, as an embodiment, the determining module 501 is configured to: :

and receiving the first data sent by a seventh network element, wherein the seventh network element comprises a data storage function network element.

Optionally, as an embodiment, the determining module 501 is further configured to:

and sending a first request message to the second network element, wherein the first request message is used for requesting to acquire the use information of the second network element on the first model.

and sending a third request message to the seventh network element, wherein the third request message is used for acquiring the first data.

Optionally, as an embodiment, the model training module 502 is further configured to:

training to obtain the first model; transmitting the model information of the first model to the second network element;

wherein the model information of the first model includes at least one of:

the first model;

and a second accuracy of the first model, wherein the second accuracy is used for indicating the accuracy degree of a model output result presented by the first model in a training stage or a testing stage.

Optionally, as an embodiment, the usage information includes at least one of:

model identification information of the first model;

Condition limiting information of the reasoning task;

and the object information of the reasoning task.

Optionally, as an embodiment, the source information includes at least one of:

and the fourth network element is used for providing the label data corresponding to the reasoning task.

Optionally, as an embodiment, the determining module 501 is configured to at least one of:

an input data acquisition request message is sent to the third network element, and the input data acquisition request message is used for requesting to acquire the reasoning input data;

a tag data acquisition request message is sent to the fourth network element, and the tag data acquisition request message is used for requesting to acquire the tag data;

wherein the input data acquisition request message includes at least one of the following:

the type information of the reasoning input data;

the object information corresponding to the reasoning input data;

the time information corresponding to the reasoning input data;

the tag data acquisition request message includes at least one of the following:

type information of the tag data;

Object information corresponding to the tag data;

and the time information corresponding to the tag data.

Optionally, as an embodiment, the determining module 501 is configured to:

inputting the reasoning input data into the first model to determine reasoning result data;

and determining the first accuracy according to the reasoning result data and the label data.

Optionally, as an embodiment, the first accuracy meets a preset condition, including at least one of:

the first accuracy is less than the second accuracy, and a difference between the first accuracy and the second accuracy is greater than a preset value;

the first accuracy is less than a preset accuracy.

Optionally, as an embodiment, the second accuracy is used to indicate at least one of:

the accuracy degree of the model output result presented by the first model in the training stage or the testing stage;

the first model outputs the error degree of the result in the model presented in the training stage or the testing stage.

Optionally, as an embodiment, the model training module 502 is configured to:

retraining the first model based on the target training data.

Optionally, as an embodiment, the model training module 502 is configured to at least one of:

acquiring first training data used in training the first model;

determining a fifth network element; acquiring second training data from the fifth network element, wherein the fifth network element is used for providing training data;

determining a sixth network element; and obtaining the reasoning data from the sixth network element, wherein the sixth network element is used for providing the reasoning data.

Optionally, as an embodiment, the model training module 502 is configured to:

determining the fifth network element according to the second information;

the second information comprises task identification information of an reasoning task executed based on the first model and/or condition limiting information of the reasoning task.

Optionally, as an embodiment, the model training module 502 is configured to:

and determining the sixth network element according to the second information.

Optionally, as an embodiment, the model training module 502 is further configured to at least one of:

transmitting the re-trained second model or the re-selected model information of the second model to a seventh network element, and storing the model information of the second model by the seventh network element;

wherein the model information of the second model includes at least one of:

model identification information of the second model;

application range information of the second model;

a third accuracy of the second model, the third accuracy being used to indicate a degree of accuracy of a model output result presented by the second model in a training phase or a testing phase;

training data of the second model;

the second model.

Optionally, as an embodiment, the third accuracy is used to indicate at least one of:

the accuracy degree of the model output result presented by the second model in the training stage or the testing stage;

The second model outputs the error degree of the result in the model presented in the training stage or the testing stage.

Optionally, as an embodiment, the first network element includes a model training function network element;

the second network element comprises a model reasoning function network element;

the third network element comprises source equipment for reasoning input data;

the fourth network element comprises source equipment of tag data;

the fifth network element comprises source equipment of training data;

the sixth network element comprises source equipment for reasoning data;

the seventh network element comprises a data storage function network element.

The apparatus 500 according to the embodiment of the present application may refer to the flow of the method 200 corresponding to the embodiment of the present application, and each unit/module in the apparatus 500 and the other operations and/or functions described above are respectively for implementing the corresponding flow in the method 200, and may achieve the same or equivalent technical effects, which are not described herein for brevity.

Fig. 6 is a schematic structural diagram of a data processing apparatus in a communication network according to an embodiment of the present application, which may correspond to the second network element in other embodiments. As shown in fig. 6, the apparatus 600 includes the following modules.

The task execution module 601 is configured to execute an inference task based on a first model, where the first model is obtained by training a first network element, and the first network element includes a model training functional network element;

A sending module 602, configured to send at least one of usage information of the first model and first data to the first network element, and/or send first indication information to a seventh network element, where the first indication information is used to indicate the seventh network element to store the first data of the reasoning task, and the seventh network element includes a data storage function network element.

Optionally, as an embodiment, the first data includes at least one of:

reasoning input data;

reasoning result data corresponding to the reasoning input data;

tag data corresponding to the inferential input data.

Optionally, as an embodiment, the apparatus 600 further includes a first receiving module 603, where the first receiving module 603 is configured to:

and receiving a first request message sent by the first network element, wherein the first request message is used for requesting to acquire the use information of the second network element on the first model.

Optionally, as an embodiment, the apparatus 600 further includes a second receiving module 604, where the second receiving module 604 is configured to:

Optionally, as an embodiment, the apparatus 600 further includes a collecting module 604, where the collecting module 604 is configured to at least one of:

the second network element collects the reasoning input data;

and the second network element collects label data corresponding to the reasoning input data.

Optionally, as an embodiment, the collecting module 604 is further configured to at least one of:

collecting the reasoning input data under the condition that a task request of an eighth network element aiming at the reasoning task is received, wherein the eighth network element comprises a consumer network element;

and the second network element actively collects the reasoning input data.

Optionally, as an embodiment, the collecting module 604 is further configured to:

the second network element sends a data acquisition request to a fourth network element, wherein the data acquisition request is used for requesting to acquire tag data corresponding to the reasoning input data.

Optionally, as an embodiment, the task execution module 601 is further configured to:

receiving model information of a second model sent by the first network element, wherein the second model is obtained by retraining the first model by the first network element and is a model reselected by the first network element;

The model information of the second model includes at least one of:

model identification information of the second model;

application range information of the second model;

training data of the second model;

the second model.

sending a model request message to the first network element, wherein the model request message is used for requesting to acquire the first model;

receiving model information of the first model sent by the first network element;

wherein the model information of the first model includes at least one of:

The first model;

Optionally, as an embodiment, the task execution module 601 is configured to at least one of:

receiving a task request aiming at the reasoning task, which is sent by an eighth network element; performing the reasoning task based on the first model, the eighth network element comprising a consumer network element;

acquiring a task request of the reasoning task triggered by the second network element simulation; and executing the reasoning task based on the first model.

Optionally, as an embodiment, the task execution module 601 is configured to:

an input data acquisition request message is sent to a third network element, and the input data acquisition request message is used for requesting to acquire reasoning input data corresponding to the reasoning task;

Receiving the reasoning input data sent by the third network element;

and inputting the reasoning input data into the first model to obtain reasoning result data.

and sending the reasoning result data to the eighth network element.

Optionally, as an embodiment, the usage information includes at least one of:

model identification information of the first model;

condition limiting information of the reasoning task;

and the object information of the reasoning task.

The apparatus 600 according to the embodiment of the present application may refer to the flow of the method 300 corresponding to the embodiment of the present application, and each unit/module in the apparatus 600 and the other operations and/or functions described above are respectively for implementing the corresponding flow in the method 300, and may achieve the same or equivalent technical effects, which are not described herein for brevity.

The data processing apparatus in the communication network in the embodiment of the present application may be an electronic device, for example, an electronic device with an operating system, or may be a component in an electronic device, for example, an integrated circuit or a chip. The electronic device may be a terminal, or may be other devices than a terminal. By way of example, terminals may include, but are not limited to, the types of terminals 11 listed above, other devices may be servers, network attached storage (Network Attached Storage, NAS), etc., and embodiments of the application are not specifically limited.

The data processing device in the communication network provided by the embodiment of the present application can implement each process implemented by the embodiments of the methods of fig. 2 to fig. 4, and achieve the same technical effects, and in order to avoid repetition, a detailed description is omitted here.

Optionally, as shown in fig. 7, the embodiment of the present application further provides a communication device 700, including a processor 701 and a memory 702, where the memory 702 stores a program or an instruction that can be executed on the processor 701, and when the communication device 700 is a network side device, the program or the instruction is executed by the processor 701 to implement each step of the data processing method embodiment in the communication network, and the same technical effects can be achieved, so that repetition is avoided and no further description is given here.

The embodiment of the application also provides network side equipment, which comprises a processor and a communication interface, wherein the processor is used for determining the first accuracy of a first model, and the first accuracy is used for indicating the accuracy of the first model for actual reasoning; retraining the first model or reselecting the second model if the first accuracy meets a preset condition. Or the processor is used for executing an reasoning task based on a first model, the first model is obtained by training a first network element, and the first network element comprises a model training function network element; the communication interface is configured to send at least one of usage information of the first model and first data to the first network element, and/or send first indication information to a seventh network element, where the first indication information is used to indicate the seventh network element to store the first data of the reasoning task, and the seventh network element includes a data storage function network element. The network side device embodiment corresponds to the network side device method embodiment, and each implementation process and implementation manner of the method embodiment can be applied to the network side device embodiment, and the same technical effects can be achieved.

Specifically, the embodiment of the application also provides network side equipment. As shown in fig. 8, the network side device 800 includes: a processor 801, a network interface 802, and a memory 803. The network interface 802 is, for example, a common public radio interface (common public radio interface, CPRI).

Specifically, the network side device 800 of the embodiment of the present application further includes: instructions or programs stored in the memory 803 and capable of being executed on the processor 801, the processor 801 calls the instructions or programs in the memory 803 to execute the method executed by each module shown in fig. 5 or fig. 6, and achieve the same technical effect, so that repetition is avoided and therefore no description is given here.

The embodiment of the application also provides a readable storage medium, on which a program or an instruction is stored, which when executed by a processor, implements each process of the data processing method embodiment in the communication network, and can achieve the same technical effects, so that repetition is avoided, and no further description is given here.

Wherein the processor is a processor in the terminal described in the above embodiment. The readable storage medium includes computer readable storage medium such as computer readable memory ROM, random access memory RAM, magnetic or optical disk, etc.

The embodiment of the application further provides a chip, the chip comprises a processor and a communication interface, the communication interface is coupled with the processor, the processor is used for running programs or instructions, the processes of the data processing method embodiment in the communication network can be realized, the same technical effects can be achieved, and the repetition is avoided, and the description is omitted here.

It should be understood that the chips referred to in the embodiments of the present application may also be referred to as system-on-chip chips, or the like.

The embodiments of the present application further provide a computer program/program product stored in a storage medium, where the computer program/program product is executed by at least one processor to implement the respective processes of the data processing method embodiments in the communication network, and achieve the same technical effects, and are not repeated herein.

The embodiment of the application also provides a data processing system in the communication network, which comprises: the first network side device may be configured to perform the steps of the data processing method in the communication network as described in fig. 2, and the second network side device may be configured to perform the steps of the data processing method in the communication network as described in fig. 3.

It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element. Furthermore, it should be noted that the scope of the methods and apparatus in the embodiments of the present application is not limited to performing the functions in the order shown or discussed, but may also include performing the functions in a substantially simultaneous manner or in an opposite order depending on the functions involved, e.g., the described methods may be performed in an order different from that described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art in the form of a computer software product stored in a storage medium (e.g. ROM/RAM, magnetic disk, optical disk) comprising instructions for causing a terminal (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the method according to the embodiments of the present application.

The embodiments of the present application have been described above with reference to the accompanying drawings, but the present application is not limited to the above-described embodiments, which are merely illustrative and not restrictive, and many forms may be made by those having ordinary skill in the art without departing from the spirit of the present application and the scope of the claims, which are to be protected by the present application.

Claims

1. A method of data processing in a communication network, comprising:

2. The method of claim 1, wherein the first accuracy is used to indicate at least one of:

3. The method of claim 1, wherein the first network element determining a first accuracy of a first model comprises:

The first network element acquires first data; the first network element determining the first accuracy based on the first data;

the first data includes at least one of:

reasoning input data;

reasoning result data corresponding to the reasoning input data;

tag data corresponding to the inferential input data.

4. A method according to claim 3, wherein the first network element obtaining the first data comprises:

and acquiring the first data according to the source information.

5. A method according to claim 3, wherein the first network element obtaining the first data comprises:

6. A method according to claim 3, wherein the first network element obtaining the first data comprises:

7. The method of claim 4, wherein prior to receiving usage information for the first model sent by the second network element, the method further comprises:

The first network element sends a first request message to the second network element, where the first request message is used to request to obtain usage information of the second network element on the first model.

8. The method of claim 5, wherein prior to receiving the first data sent by the second network element, the method further comprises:

9. The method of claim 6, wherein prior to receiving the first data sent by the seventh network element, the method further comprises:

the first network element sends a third request message to the seventh network element, wherein the third request message is used for acquiring the first data.

10. The method of claim 4, wherein prior to receiving usage information for the first model sent by a second network element, the method further comprises:

wherein the model information of the first model includes at least one of:

The first model;

11. The method of claim 4, wherein the usage information comprises at least one of:

model identification information of the first model;

condition limiting information of the reasoning task;

and the object information of the reasoning task.

12. The method of claim 11, wherein the source information comprises at least one of:

13. The method of claim 12, wherein the obtaining the first data from the source information comprises at least one of:

the type information of the reasoning input data;

the object information corresponding to the reasoning input data;

the time information corresponding to the reasoning input data;

type information of the tag data;

object information corresponding to the tag data;

and the time information corresponding to the tag data.

14. A method according to claim 3, wherein the first network element determining a first accuracy of the first model comprises:

15. The method of claim 1, wherein the first accuracy meets a preset condition, comprising at least one of:

the first accuracy is less than a preset accuracy.

16. The method of claim 15, wherein the second accuracy is used to indicate at least one of:

17. The method of claim 1, wherein the first network element retrains the first model, comprising:

retraining the first model based on the target training data.

18. The method of claim 17, wherein the acquiring the target training data comprises at least one of:

acquiring first training data used in training the first model;

19. The method of claim 18, wherein the determining a fifth network element comprises:

determining the fifth network element according to the second information;

20. The method of claim 19, wherein the determining a sixth network element comprises:

and determining the sixth network element according to the second information.

21. The method of claim 1, wherein the first network element, after retraining the first model or reselecting a second model, further comprises at least one of:

Wherein the model information of the second model includes at least one of:

model identification information of the second model;

application range information of the second model;

training data of the second model;

the second model.

22. The method of claim 21, wherein the third accuracy is used to indicate at least one of:

23. The method of claim 1, wherein the step of determining the position of the substrate comprises,

the first network element comprises a model training function network element;

the third network element comprises source equipment for reasoning input data;

the fourth network element comprises source equipment of tag data;

the fifth network element comprises source equipment of training data;

The sixth network element comprises source equipment for reasoning data;

the seventh network element comprises a data storage function network element.

24. A method of data processing in a communication network, comprising:

25. The method of claim 24, wherein the first data comprises at least one of:

reasoning input data;

reasoning result data corresponding to the reasoning input data;

tag data corresponding to the inferential input data.

26. The method of claim 24, wherein prior to said sending the usage information of the first model to the first network element, the method further comprises:

27. The method of claim 24, wherein prior to said transmitting said first data to said first network element, said method further comprises:

the second network element receives a second request message sent by the first network element, where the second request message is used to request to acquire the first data collected by the second network element.

28. The method of claim 24, wherein prior to said transmitting the first data to the first network element, the method further comprises at least one of:

the second network element collects the reasoning input data;

29. The method of claim 28, wherein the second network element gathers the inferential input data comprising at least one of:

the second network element collects the reasoning input data under the condition of receiving a task request of an eighth network element aiming at the reasoning task, wherein the eighth network element comprises a consumer network element;

and the second network element actively collects the reasoning input data.

30. The method of claim 28, wherein the second network element gathers tag data corresponding to the inferential input data, comprising:

31. The method of claim 24, wherein the method further comprises:

receiving model information of a second model sent by the first network element, wherein the second model is obtained after the first network element retrains the first model or is a model reselected by the first network element;

the model information of the second model includes at least one of:

model identification information of the second model;

application range information of the second model;

training data of the second model;

the second model.

32. The method of claim 31, wherein the third accuracy is used to indicate at least one of:

33. The method of claim 24, wherein the second network element, prior to performing the inference task based on the first model, further comprises:

wherein the model information of the first model includes at least one of:

the first model;

34. The method of claim 33, wherein the second accuracy is used to indicate at least one of:

35. The method of claim 24, wherein the second network element performs the inference task based on the first model, comprising at least one of:

36. The method of claim 34, wherein the performing the inference task based on the first model comprises:

receiving the reasoning input data sent by the third network element;

37. The method of claim 35, wherein after receiving a task request for the inference task sent by an eighth network element and performing the inference task using the first model, the method further comprises:

and sending the reasoning result data to the eighth network element.

38. The method of claim 24, wherein the usage information comprises at least one of:

Model identification information of the first model;

condition limiting information of the reasoning task;

and the object information of the reasoning task.

39. A data processing apparatus in a communication network, comprising:

40. The apparatus of claim 39, wherein the first accuracy is used to indicate at least one of:

41. The apparatus of claim 39, wherein the means for determining is configured to:

acquiring first data; determining the first accuracy based on the first data;

the first data includes at least one of:

Reasoning input data;

reasoning result data corresponding to the reasoning input data;

tag data corresponding to the inferential input data.

42. The apparatus of claim 41, wherein the means for determining is configured to:

acquiring the first data according to the source information;

wherein the usage information includes at least one of:

model identification information of the first model;

condition limiting information of the reasoning task;

object information of the reasoning task;

the source information includes at least one of:

43. The apparatus of claim 41, wherein the means for determining is configured to:

44. The apparatus of claim 41, wherein the means for determining is configured to:

45. The apparatus of claim 42, wherein the means for determining is further configured to:

46. The apparatus of claim 43, wherein the determining module is further configured to:

and sending a second request message to the second network element, wherein the second request message is used for requesting to acquire the first data collected by the second network element.

47. The apparatus of claim 44, wherein the means for determining is further for:

48. The apparatus of claim 42, wherein the model training module is further configured to:

Wherein the model information of the first model includes at least one of:

the first model;

49. The apparatus of claim 42, wherein the means for determining is configured to at least one of:

the type information of the reasoning input data;

the object information corresponding to the reasoning input data;

the time information corresponding to the reasoning input data;

type information of the tag data;

object information corresponding to the tag data;

and the time information corresponding to the tag data.

50. The apparatus of claim 41, wherein the means for determining is configured to:

51. The apparatus of claim 39, wherein the model training module is configured to:

retraining the first model based on the target training data.

52. The apparatus of claim 51, wherein the model training module is configured to at least one of:

acquiring first training data used in training the first model;

53. The apparatus of claim 52, wherein the model training module is further configured to:

Determining the fifth network element according to the second information;

54. The apparatus of claim 53, wherein the model training module is further configured to:

and determining the sixth network element according to the second information.

55. The apparatus of claim 39, wherein the model training module is further configured to at least one of:

wherein the model information of the second model includes at least one of:

model identification information of the second model;

application range information of the second model;

training data of the second model;

the second model.

56. A data processing apparatus in a communication network, comprising:

57. The apparatus of claim 56, further comprising a first receiving module, said first receiving module configured to:

58. The apparatus of claim 56, further comprising a second receiving module for:

59. The apparatus of claim 56, further comprising a collection module for at least one of:

collecting the inference input data;

and collecting label data corresponding to the reasoning input data.

60. The apparatus of claim 59, wherein the collection module is further configured to at least one of:

and actively collecting the reasoning input data.

61. The apparatus of claim 59, wherein the collection module is further configured to:

62. An apparatus as defined in claim 56, wherein the task execution module is further to:

receiving model information of a second model sent by the first network element, wherein the second model is obtained after the first model is retrained by the first network element;

the model information of the second model includes at least one of:

model identification information of the second model;

application range information of the second model;

training data of the second model;

the second model.

63. An apparatus as defined in claim 56, wherein the task execution module is further to:

wherein the model information of the first model includes at least one of:

The first model;

64. A network side device comprising a processor and a memory storing a program or instructions executable on the processor, which when executed by the processor, performs the steps of the method of any of claims 1-23, or performs the steps of the method of any of claims 24-38.

65. A readable storage medium, characterized in that the readable storage medium has stored thereon a program or instructions which, when executed by a processor, implement the steps of the method according to any of claims 1-23 or the steps of the method according to any of claims 24-38.