CN116668309A

CN116668309A - Online learning method and device of AI model, communication equipment and readable storage medium

Info

Publication number: CN116668309A
Application number: CN202210157466.7A
Authority: CN
Inventors: 贾承璐; 孙布勒; 王园园
Original assignee: Vivo Mobile Communication Co Ltd
Current assignee: Vivo Mobile Communication Co Ltd
Priority date: 2022-02-21
Filing date: 2022-02-21
Publication date: 2023-08-29
Also published as: WO2023155839A1

Abstract

The application discloses an online learning method and device of an artificial intelligence AI model, communication equipment and a readable storage medium. The online learning method of the artificial intelligence AI model comprises the following steps: the second device configures a first AI model for the first device; the second device configures the first device with online learning information of the first AI model.

Description

Online learning method and device of AI model, communication equipment and readable storage medium

Technical Field

The application belongs to the technical field of electronic information, and particularly relates to an online learning method and device of an AI model, communication equipment and a readable storage medium.

Background

With the wide application of artificial intelligence (Artificial Intelligence, AI) in various fields, the AI is incorporated into a wireless communication network to improve the technical indexes such as network throughput, time delay, user capacity and the like, which becomes an important task of the wireless communication network. Currently, there are a number of implementations of AI modules in wireless communication networks. For example, neural Network (NN) Decision Tree (DT), support vector machine (support vector machine, SVM), genetic algorithm (genetic algorithm, GA), and the like.

In the related art, an AI model is typically trained offline, and then the trained AI model is deployed to a wireless communication system. However, when the wireless communication environment changes, the accuracy of the output result of the AI model is made low. Thus, the calculation accuracy of the AI model is poor.

Disclosure of Invention

The embodiment of the application provides an online learning method, an online learning device, communication equipment and a readable storage medium of an AI model, which can solve the problem of AI model failure caused by dynamic change of a wireless communication environment in an actual scene.

In a first aspect, an online learning method of an AI model is provided, including: the second device configures a first AI model for the first device; the second device configures the first device with online learning information of the first AI model.

In a second aspect, there is provided an online learning apparatus of an AI model, the apparatus including: a configuration module, wherein: the configuration module is used for configuring a first AI model for the first device by the second device; the configuration module is further configured to configure online learning information of the first AI model for the first device by the second device.

In a third aspect, an online learning method of an AI model is provided, including: the device comprises an acquisition module and an execution module, wherein: the acquisition module is used for acquiring a first AI model by the first equipment; the execution module is used for the first equipment to conduct online learning on the first AI model based on online learning information of the first AI model.

In a fourth aspect, there is provided an online learning apparatus of an AI model, the apparatus including: the first equipment acquires a first AI model; the first device performs online learning on the first AI model based on online learning information of the first AI model.

In a fifth aspect, there is provided a communication device 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 as described in the first aspect.

In a sixth aspect, a communication device is provided, including a processor and a communication interface, where the processor is configured to configure a first AI model for a first device; and configuring the first device with online learning information of the first AI model.

In a seventh aspect, there is provided a communication device 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 as described in the first aspect.

In an eighth aspect, a network side device is provided, including a processor and a communication interface, where the processor is configured to obtain a first AI model, and perform online learning on the first AI model based on online learning information of the first AI model.

In a ninth 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 a tenth 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 an eleventh 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 online learning method of AI models as in the first aspect, or as in the third aspect.

In the embodiment of the application, a first device acquires a first AI model and performs online learning on the first AI model based on online learning information of the first AI model. According to the method, the first AI model is deployed at the first equipment side, and parameters required by online learning are configured for the first model, so that the first AI model can be continuously adjusted online at the first equipment side, the prediction performance of the first AI model is maintained, and the service quality of the first equipment is further ensured.

Drawings

Fig. 1 is a block diagram of a wireless communication system provided by an embodiment of the present application;

FIG. 2 is a flowchart of an online learning method of an AI model according to an embodiment of the application;

FIG. 3 is a second flowchart of an online learning method of an AI model according to an embodiment of the application;

FIG. 4 is a schematic structural diagram of an online learning device of an AI model according to an embodiment of the application;

FIG. 5 is a second schematic diagram of an online learning device of an AI model according to an embodiment of the application;

fig. 6 is a schematic structural diagram of a communication device according to an embodiment of the present application;

fig. 7 is a schematic diagram of a hardware structure of a terminal according to an embodiment of the present application;

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

fig. 9 is a second schematic hardware structure 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 New air interface (NR) system for purposes of example and uses NR terminology in much of the description that follows, but these techniques are also applicable to applications other than NR 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.

Some terms involved in the embodiments of the present invention are explained below:

(1) Artificial intelligence (Artificial Intelligence, AI): artificial intelligence is a very broad science that consists of different fields such as machine learning, computer vision, etc.

(2) Machine Learning (ML): machine learning is an important branch of artificial intelligence, mainly studying how to enable computers to learn themselves. The algorithms of machine learning include Neural Network (NN) Decision Tree (DT), support vector machine (support vector machine, SVM), genetic algorithm (genetic algorithm, GA), and the like.

(3) Neural network: neural networks are composed of a large number of nodes, which are called neurons. The composition information of the neuron includes: the weight/multiplicative coefficient (w) is input (a 1, a2, … aK), which is the bias/additive coefficient (b), and the function (σ ()). Common activation functions include Sigmoid, tanh, reLU (Rectified Linear Unit), linear rectification functions, modified linear units), and the like.

Further, parameters of the neural network may be optimized by a gradient optimization algorithm. Gradient optimization algorithms are a class of algorithms that minimize or maximize an objective function (sometimes called a loss function), which is often a mathematical combination of model parameters and data. For example, given data X and its corresponding label Y, after a neural network model f (), a predicted output f (X) can be obtained from the input X, and the difference (f (X) -Y), i.e., the loss function, between the predicted value and the actual value can be calculated. If a proper W is found, b minimizes the value of the loss function, and the smaller the loss value is, the closer the model is to the real situation.

Illustratively, the optimization algorithms that are currently common are generally based on the BP (error Back Propagation ) algorithm. The basic idea of the BP algorithm is that the learning process consists of two processes, forward propagation of the signal and backward propagation of the error. In forward propagation, an input sample is transmitted from an input layer, is processed layer by each hidden layer, and is transmitted to an output layer. If the actual output of the output layer does not match the desired output, the back propagation phase of the error is shifted. The error back transmission is to make the output error pass through hidden layer to input layer in a certain form and to distribute the error to all units of each layer, so as to obtain the error signal of each layer unit, which is used as the basis for correcting the weight of each unit. The process of adjusting the weights of the layers of forward propagation and error back propagation of the signal is performed repeatedly. The constant weight adjustment process is the learning training process of the network. This process is continued until the error in the network output is reduced to an acceptable level or until a preset number of learnings is performed.

For example, common optimization algorithms are Gradient Descent (Gradient Descepting), random Gradient Descent (Stochastic Gradient Descent, SGD), small lot Gradient Descent (mini-batch Gradient Descent), momentum method (Momentum), random Gradient Descent with Momentum (Nesterov), adaptive Gradient Descent (ADAptive GRADient Descent, adagrad), adadelta, root mean square error Descent (root mean square prop, RMSprop), adaptive Momentum estimation (Adaptive Moment Estimation, adam), and the like.

For example, the above optimization algorithm derives the gradient by applying the influence of learning rate, previous gradient/derivative/bias to the derivative/bias of the current neuron according to the error/loss obtained by the loss function during error back propagation, and transmits the gradient to the previous layer.

The online learning method provided by the embodiment of the application is described in detail below through some embodiments and application scenes thereof with reference to the accompanying drawings.

The current AI research in the wireless communication field mainly focuses on offline learning and deployment, and as the wireless environment is constantly changed, a fixed AI model obtained through offline training gradually fails in a dynamic environment, and how to improve the adaptability of the model in a new changing environment becomes a problem to be solved.

The application solves the problems by way of example, and provides an online learning method of an AI model. Further, the following difficulties are encountered in achieving online learning: 1) Limited by the storage capacity and data collection capacity of the device (e.g., the time cost and hardware cost of collecting data is relatively high), it is often difficult to acquire a sufficiently large number of data sets for online training; 2) Limited by the computational power of the device and the limited data set, multiple rounds of model fine tuning may not be performed or may result in overfitting after multiple rounds of model fine tuning; 3) For wireless communication, there are also problems of communication delay limitation and communication continuity, which puts demands on the time of data acquisition of the first device and the time of online learning.

Fig. 2 shows a flowchart of an online learning method of an AI model according to an embodiment of the present application. As shown in fig. 2, the online learning method of an AI model provided by the embodiment of the present application may include the following steps 201 and 202:

step 201: the second device configures a first AI model for the first device.

In an embodiment of the present application, the second device may include at least one of: core network equipment, access network equipment and terminals; the first device may include at least one of: core network equipment, access network equipment and terminals.

The second device is illustratively a core network device, and the first device may be an access network device or a terminal, respectively.

The second device is an access network device, and the first device may be a core network device or a terminal, respectively.

The second device is illustratively a terminal, and the first device may be a core network device or an access network device, respectively.

In the embodiment of the present application, the first AI model is an AI model obtained by offline training on the second device side.

Optionally, in an embodiment of the present application, the algorithm of the first AI model may include at least one of: neural networks, decision trees, support vector machines, bayesian classifiers.

Alternatively, in the embodiment of the present application, the first AI model may be an AI model for terminal positioning, network optimization, large input data set processing, and network recommendation for a user.

Optionally, in an embodiment of the present application, the second device may train the AI model based on a preset learning framework to obtain the first AI model. Illustratively, the first AI model is taken as an example of a neural network model. The second device may train the neural network model based on a preset learning framework to obtain the first neural network model.

Optionally, in the embodiment of the present application, the second device may send the first AI model obtained by training to the first device, and deploy the first AI model on the first device side.

Step 202: the second device configures online learning information of the first AI model for the first device.

In the embodiment of the application, the second device can send the information required by the online learning of the first AI model to the first device, and the first device can perform online learning on the first AI model based on the online learning information under the condition that the online learning information sent by the second device is received.

In the embodiment of the application, the online learning information is configured by the network device or is autonomously determined by the second device.

Optionally, in the embodiment of the present application, the first device may be a terminal, and the second device may be a core network device.

The core network device transmits the first AI model to the terminal, and transmits online learning information of the first AI model to the terminal, which receives the first AI model transmitted by the core network device, and the online learning information of the first AI model.

In the online learning method provided by the embodiment of the application, the second device configures a first AI model for the first device, and configures online learning information of the first AI model for the first device. According to the method, the first AI model is deployed at the first equipment side, and parameters required by online learning are configured for the first model, so that the first AI model can be continuously adjusted online at the first equipment side, the prediction performance of the first AI model is maintained, and the service quality of the first equipment is further ensured.

Optionally, in an embodiment of the present application, the online learning information includes at least one of:

a triggering mode of online learning;

an abort condition for online learning;

parameter configuration information for online learning;

an online learned data set.

The triggering method is illustratively related to the state information of the first device and the channel information of the first device-related channel. For example, in a case where the moving speed of the first device is fast, or in a case where the channel environment of the operating channel of the first device changes, online learning of the first AI model may be triggered, so that the first AI model continuously adapts to the changing environment.

For example, the first device may suspend online learning of the first AI model if the number of online learning of the first AI model is greater than a preset number of iterations; or under the condition that the first AI model reaches the preset precision, stopping online learning of the first AI model; alternatively, if the error information of the output result of the first AI model is small, the online learning of the first AI model is suspended. Because the online learning frequency of the first AI model is larger than the preset iteration frequency, or the precision of the first AI model reaches the preset precision, and the current first AI model is characterized as effective, the online learning of the first AI model can be terminated, so that the power consumption is saved.

Optionally, in an embodiment of the present application, the triggering condition corresponding to the triggering manner includes at least one of the following:

the state information of the first equipment meets a first preset condition;

the data volume acquired by the first equipment is larger than a first threshold value;

the measurement information of the first equipment meets a second preset condition;

error information of the output result of the first AI model is larger than a second threshold value;

the statistical information of the first information corresponding to the first AI model meets a third preset condition;

The statistical information of the measurement information of the first device meets a fourth preset condition;

optionally, the status information includes at least one of: mobile speed, beam switching information, cell switching information.

Alternatively, the data amount collected by the first device may be the online data collection amount of the first device, for example, the channel information collected by the first device in real time.

Illustratively, in the case of a fast movement of the first device, triggering the first device to learn the first AI model online; or under the condition that the data volume acquired by the first equipment is large, triggering the first equipment to learn the first AI model online; or triggering the first equipment to perform online learning on the first AI model under the condition that the measurement information of the first equipment indicates that the channel environment of the current channel changes; or triggering the first equipment to perform online learning on the first AI model under the condition that the error of the output result of the first AI model is large or the accuracy of the first AI model is low. The first AI model may be disabled due to a rapid movement of the terminal or a change in the channel environment. Therefore, the first equipment can learn the first AI model online under the condition that the first AI model fails based on the information such as the moving speed of the first equipment, the channel environment of the related channel, the accuracy value of the AI model and the like, so that the prediction accuracy of the first AI model is improved.

Illustratively, the channel information may include at least one of: angle of emission information of a signal, angle of arrival information of a signal, delay information of a signal in a channel, signal quality in a channel, and the like.

Alternatively, the first threshold may be 3000, 5000, 7000, or the like.

Optionally, the measurement information includes at least one of: and the first measurement information of the reference signal received by the first equipment and the second measurement information acquired by the sensor of the first equipment. Illustratively, the first measurement information includes at least one of: instantaneous measurement information of the reference signal, statistical measurement information of the reference signal. Illustratively, the instantaneous measurement information of the reference signal may be: measurement information of a reference signal at a specific time; the statistical measurement information of the reference signal may be: measurement information of reference signals over a period of time.

Illustratively, the reference signal includes at least one of: synchronization signal block SSB, CSI reference signal CSI-RS, sounding reference signal SRS, positioning reference signal PRS. Optionally, the sensor may include at least one of: vision sensors, radar sensors, position sensors, etc.

Optionally, the first information includes at least one of: input information of the first AI model, and output information of the first AI model.

For example, taking the first device as an example of the terminal, in the case that the first information includes input information of the first AI model, the first information may be channel information of an operating channel or a peripheral channel of the terminal, and in the case that the first information includes output information of the first AI model, the first information may be location information of the terminal.

Optionally, in an embodiment of the present application, the statistical information of the first information includes at least one of the following:

the method comprises the steps of first statistics of the first information in a first time window, second statistics corresponding to the first information in at least two continuous second time windows, statistics of the first information of at least two terminals in a first cell at a first moment, and correlation information of the first information.

Wherein the second statistic is calculated based on the statistics in each second time window.

Alternatively, the statistical information may include at least one of the following; mean, variance, etc.

Alternatively, the statistical information may include statistical information in time and statistical information in space. For example, the statistical information over time may be: the statistical information of the channels of the same terminal in a continuous period of time, and the statistical information in space can be the statistical information of the channels of a plurality of different terminals in a cell.

Illustratively, the statistical information of the first information is used to characterize whether a wireless network environment within the region of action of the first AI model changes.

Illustratively, the first information is taken as channel information as an example. When the average value of the channel information in a certain continuous time window is smaller than a certain threshold value, or the correlation index of the channel information in two time windows is smaller than a certain threshold value, or the correlation index of the front and rear data in the current time window is smaller than a certain threshold value, the channel environment of the relevant channel of the first equipment is represented to change, the first equipment can be triggered to conduct online learning on the first AI model so as to adapt to the current channel environment, and therefore the prediction accuracy of the first AI model is improved.

The statistical information of the first information will be explained below by taking the first information as channel information as an example.

For example, in a case where the statistical information of the first information includes a first statistics of the first information within a first time window, the statistical information of the first information may be: the mean or variance of channel information for the same terminal over a continuous time window.

For example, in a case where the statistical information of the first information includes at least two second statistics corresponding to the first information in the consecutive second time window, the statistical information of the first information may be: and determining the average value based on the average value of the channel information in each continuous time window of the plurality of continuous time windows of the terminal, wherein if the average value of the channel information in the time window 1 is a, the average value of the channel information in the time window 2 is b, the average value of the channel information in the time window 3 is c, and the statistical information of the channel information is the average value of a, b and c.

For example, in a case where the statistical information of the first information includes statistical information of the first information of at least two terminals in the first cell at the first time, the statistical information of the first information may be: and the average value of the channel information of a plurality of different terminals in one cell at a certain moment, for example, the average value of the channel information of a terminal A, a terminal B and a terminal C in the same cell at a moment 1 is d, e and f respectively, and the statistical information of the channel information is the average value of d, e and f.

For example, in the case where the statistical information of the first information includes correlation information of the first information, the statistical information of the first information may be: the correlation index of the channel information of the two time windows is lower than a certain threshold, for example, the distance between the front and rear data in the current time window is smaller than a certain threshold.

Optionally, in an embodiment of the present application, the statistical information of the measurement information of the first device includes at least one of the following:

and third statistics of the measurement information in a third time window, fourth statistics of the measurement information corresponding to at least two continuous fourth time windows, and correlation information of the measurement information.

Wherein the fourth statistic is calculated based on the statistics in each fourth time window.

Optionally, the above-mentioned correlation information includes at least one of: distance between data, covariance, correlation coefficient.

Illustratively, the measurement information is channel measurement information of a reference signal received by the first device. The statistical information of the measurement information may be: the variance of the channel measurement information in a certain continuous time window, or the average of the channel measurement information in two continuous time windows, or the distance between the data of the channel measurement information in the current time window.

Illustratively, the statistical information of the measurement information is used to characterize whether the wireless network environment within the region of action of the first AI model has changed.

Illustratively, the measurement information is taken as channel information. When the average value of measurement information in a certain continuous time window is smaller than a certain threshold value, or the correlation index of the measurement information in two time windows is smaller than a certain threshold value, or the correlation index of front and rear data in the current time window is smaller than a certain threshold value, the wireless environment representing the action of the first AI model changes, the first equipment can be triggered to learn the first AI model online so as to adapt to the current wireless environment, and accordingly the prediction accuracy of the first AI model is improved.

Further optionally, in an embodiment of the present application, the triggering condition includes: the state information of the first equipment meets a first preset condition; the meeting the first preset condition includes at least one of the following:

the moving speed of the first device is greater than a third threshold value;

the beam switching information indicates that the first device performs beam switching, and the beam switching frequency is greater than a fourth threshold;

the cell switch information indicates that the first device performs cell switch.

Illustratively, the third threshold may be 60km/h,80km/h,100 km/h, and so on.

Further optionally, in an embodiment of the present application, the triggering condition includes: the measurement information of the first equipment meets a second preset condition; optionally, the meeting the second preset condition includes: the measurement information of the first device indicates that the channel environment of the associated channel of the first device has changed.

Illustratively, the measurement information is a reference signal received by the first device. The second preset condition may be: the first device estimates the downlink channel based on the CSI-RS measurements, detects a change in channel environment, e.g., from a line of sight (LOS) environment to a non-line of sight (not line of sight, NLOS) environment, e.g., the signal-to-noise ratio SINR is below a threshold, etc.

Illustratively, the measurement information is measurement information collected by a sensor of the first device. The second preset condition may be: the measurement information obtained by the visual sensor indicates that the first device is in an LOS environment.

Further optionally, in an embodiment of the present application, the triggering condition includes: the statistical information of the first information corresponding to the first AI model meets a third preset condition; the meeting the third preset condition includes at least one of:

the first statistics are greater than a maximum value of a first threshold interval;

the second statistic is greater than the maximum value of the second threshold interval;

the correlation information of the first information acquired in at least two time windows meets a first condition;

correlation information between different first information acquired within the current time window satisfies a second condition.

Illustratively, taking the first statistics as an example of the average value of channel information of the terminal in a certain continuous time window. The third preset condition may be: the first device detects that the statistics of channel information within a certain continuous time window exceeds a maximum value of a certain threshold interval.

Illustratively, taking the second statistic as an example of an average of the channel information of the terminal over a plurality of consecutive time windows. The third preset condition may be: the first device detects that the average of the channel information over a plurality of consecutive time windows exceeds a maximum of a certain threshold interval.

Illustratively, the first information is taken as channel information as an example. The third preset condition may be: the correlation index of the channel information in the two time windows is lower than a certain threshold value, or the correlation of the front and rear data in the current time window is lower than a certain threshold value.

Further optionally, in an embodiment of the present application, the triggering condition includes: the statistical information of the measurement information of the first device meets a fourth preset condition; the satisfaction of the fourth preset condition includes at least one of:

the third statistic is greater than the maximum value of the third threshold interval;

the fourth statistic is greater than the maximum value of the fourth threshold interval;

the correlation information of the measurement information collected in at least two time windows meets a third condition;

the correlation information among different measurement information acquired in the current time window meets a fourth condition;

the difference between the distribution of the measurement information and the reference distribution, which is information configured by the second device for the first device, is greater than a fifth threshold.

For example, in the case where the third statistic is a mean value of the channel measurement information in a certain continuous time window, the fourth preset condition may be: the average value of the channel measurement information in a certain continuous time window exceeds the maximum value of a certain threshold interval.

For example, in the case where the fourth statistic is a mean value calculated based on a mean value of channel measurement information in each fourth time window, the fourth preset condition may be: the average value calculated based on the average value of the channel measurement information in each fourth time window exceeds the maximum value of a certain threshold interval.

Illustratively, the third condition may be: the covariance of the data of the measurement information respectively acquired in at least two time windows is less than a certain threshold.

Illustratively, the fourth condition may be: the distance between different data of the measurement information acquired in the current time window is smaller than a certain threshold value.

The distribution of the measurement information is illustratively a statistical distribution of the measurement information. Illustratively, the above reference distribution is a statistical distribution of the first AI model. It can be understood that the training set during offline training of the first AI model follows the reference distribution, and the first AI model performs best when the measurement information follows the reference distribution.

Illustratively, the index describing the difference of the distribution of the above measurement information from the reference distribution may include at least one of: wasserstein distance; kullback-Leibler divergence; hellinger distance, etc.

Optionally, in an embodiment of the present application, the triggering condition of the triggering manner includes at least one of the following:

the second device instructs the first device to perform online learning;

the output accuracy of the first AI model is less than or equal to the sixth threshold.

Further optionally, in an embodiment of the present application, the second device instructs the first device to perform online learning, including at least one of:

the second device instructs the first device to periodically perform online learning;

the second device instructs the first device to perform online learning in a half-cycle manner;

the second device instructs the first device to perform online learning aperiodically.

The period adopted when the first equipment periodically or semi-periodically performs online learning is as follows: the second device may be preconfigured with a period, or the first device may be autonomously configured with a period.

Further optionally, the second device instructs the first device to perform online learning periodically, including:

the second device instructs the first device to perform online learning in a half-cycle manner through first signaling, wherein the first signaling comprises at least one of the following steps: a media access control-control unit MAC-CE, downlink control information DCI.

Optionally, in an embodiment of the present application, the suspension condition includes at least one of:

The online learning frequency of the first AI model is larger than the preset iteration frequency;

the first AI model achieves a preset accuracy;

error information of the output result of the first AI model is less than a seventh threshold;

the second device suddenly instructs the first device to end the current online learning process;

a target task associated with the first AI model is aborted;

the difference information of the measured information distribution of the first device and the reference distribution is smaller than an eighth threshold.

For example, the target task associated with the first AI model may be a task currently performed by the first AI model, such as locating a terminal, making a network recommendation for a user, and so on.

Optionally, in an embodiment of the present application, the parameter configuration information includes at least one of the following:

an online learning mode of the first AI model;

the size of the sample batch of the first AI model;

the state of the optimizer of the first AI model;

a partitioning manner of a first data set of the first AI model;

composition information of a first data set of a first AI model;

contribution weights of the first data set of the first AI model to the first AI model update;

an AI model identification associated with the first information;

a reference distribution of the first AI model;

wherein said first data set is at least one of: the original data set used by the first AI model, and the data set newly acquired by the first AI model; the parameter information of the reference distribution includes at least one of: variance, mean, standard deviation; the size of the sample lot refers to the size of the number of samples included in one sample lot (Batch).

Optionally, the above-mentioned original data set is: in the data set used for offline training of the first AI model (i.e., the old data set), the new acquired data set is: and after the first AI model is deployed on line at the first equipment side, namely the first AI model is configured for the first equipment, the data set (namely a new data set) acquired by the first equipment in a new environment.

Optionally, the online learning mode includes any one of the following: transient training mode (i.e., one-shot mode), continuous learning mode. Illustratively, in One-shot mode, the first device performs online learning when the acquired data reaches a specified number; in the continuous learning mode, the first device continuously performs online learning as the number of acquired data increases.

Illustratively, the Batch (Batch) size is N, and N is a positive integer.

Illustratively, the states of the optimizer of the first AI model can include a loss function, a learning rate, and the like.

Illustratively, the first data set partitioning manner may include a training set, a validation set, a test set partitioning ratio, and the like.

Illustratively, the composition information of the data sets described above includes a ratio between the number of original data sets and the number of newly acquired data sets. It should be noted that, in the embodiment of the present application, the original data set during training can effectively prevent the overfitting of the newly acquired data in the online learning process, so as to effectively improve the performance of the AI model.

Illustratively, the contribution weights of the first data set to the first AI model update may be: the contribution weights of the old data set and the new data set to the update of the first AI model may be smaller weights for the original data set and larger weights for the new data set, for example, when the first AI model performs online learning.

Further optionally, in an embodiment of the present application, the parameter configuration information includes an online learning mode of the first AI model, and the online learning mode is an instantaneous training mode, and the parameter configuration information further includes at least one of the following: and the data quantity acquired by the first equipment is the acquisition time length of the data quantity.

Further optionally, in an embodiment of the present application, the parameter configuration information includes an online learning mode of the first AI model, and the online learning mode is a continuous learning mode, and the parameter configuration information further includes at least one of: time interval of two adjacent online learning, data volume interval of two adjacent online learning.

For example, the interval between the data amounts of the two adjacent online learning may be 100, such as performing online training once every 100 sets of data are collected.

Fig. 3 shows a flowchart of an online learning method of an AI model according to an embodiment of the present application. As shown in fig. 3, the online learning method of an AI model provided by the embodiment of the present application may include the following steps 301 and 302:

step 301: the first device obtains a first AI model.

Step 302: the first device performs online learning on the first AI model based on online learning information of the first AI model.

In the embodiment of the application, the first device can perform online learning on the first AI model obtained through offline training based on the online learning information of the first AI model, and obtain the first AI model after parameter adjustment.

In the embodiment of the present application, the online learning information is information determined by the first device.

In the online learning method of the AI model provided by the embodiment of the application, a first device acquires a first AI model and carries out online learning on the first AI model based on online learning information of the first AI model. According to the method, the first AI model is deployed at the first equipment side, and parameters required by online learning are configured for the first model, so that the first AI model can be continuously adjusted online at the first equipment side, the prediction performance of the first AI model is maintained, and the service quality of the first equipment is further ensured.

a triggering mode of online learning;

an abort condition for online learning;

parameter configuration information for online learning;

An online learned data set.

Optionally, in an embodiment of the present application, the step 301 may include the following step 301a:

step 301a: the first device receives a first AI model of a second device configuration.

Optionally, in an embodiment of the present application, the second device includes at least one of: core network equipment, access network equipment and terminals; the first device includes at least one of: core network equipment, access network equipment, and terminals.

Alternatively, the second device may send the first AI model obtained by offline training to the first device, and the first device may receive the first AI model sent by the second device.

Optionally, in the embodiment of the present application, before the step 302, the online learning method provided by the embodiment of the present application further includes the following step A1:

step A1: the first device obtains online learning information of the first AI model from the second device.

Alternatively, in a case where the second device can transmit information required for the first AI model to the first device for online learning, the first device may receive online learning information transmitted by the second device, and online learning may be performed on the first AI model based on the online learning information.

Optionally, the first device may be a terminal, and the second device may be a core network device.

In this way, the second device configures the first AI model for the first device, and configures parameters required by online learning of the first AI model, so that continuous online adjustment can be performed on the first AI model at the first device side, thereby maintaining the prediction performance of the first AI model, and further ensuring the service quality of the first device.

Optionally, in the embodiment of the present application, the online learning method provided in the embodiment of the present application further includes the following step 303:

step 303: the first device configures online learning information of the first AI model for a third device.

Optionally, the third device comprises at least one of: core network equipment, access network equipment, and terminals.

Optionally, the first device may be a core network device, and the third device may be a terminal.

Alternatively, the second device may autonomously perform the online learning method of the AI model, or the second device may deploy the AI model to the first device and configure information required for the online learning of the AI model for the first device, and the first device may perform the online learning of the AI model, or the first device may deploy the AI model to the third device and configure information required for the online learning of the AI model for the first device, and the third device may perform the online learning of the AI model.

the state information of the first equipment meets a first preset condition;

the data amount acquired by the first equipment is larger than a first threshold value;

error information of the output result of the first AI model is larger than a second threshold;

the statistical information of the measurement information of the first equipment meets a fourth preset condition;

wherein the status information includes at least one of: mobile speed, beam switching information, cell switching information.

Illustratively, the channel information may include at least one of: the angle of emission of the signal, the delay information of the signal in the channel, the signal quality in the channel, etc.

Alternatively, the first threshold may be 3000, 5000, 7000, or the like.

For example, the first device may acquire the state information in real time or periodically, and perform online learning on the first AI model if the state information satisfies the first preset condition.

For example, in the calculation process of the first AI model, the first device may acquire, in real time or periodically, error information of an output result of the first AI model, or prediction accuracy of the first AI model, and perform online learning on the first AI model if the error information is greater than the second threshold.

For example, the first device may count input information and output information of the first AI model, and perform online learning on the first AI model if the statistical information of the input information or the output information satisfies a third preset condition.

For example, the first device may detect the reference signal or measurement information collected by the sensor, and perform online learning on the first AI model if the measurement information satisfies the second preset condition and/or if the statistical information of the measurement information satisfies the fourth preset condition.

Optionally, in an embodiment of the present application, the statistical information of the first information includes at least one of the following: and the first statistics of the first information in the first time window, the second statistics corresponding to the first information in at least two continuous second time windows, the statistics of the first information of at least two terminals in the first cell at the first moment, and the correlation information of the first information.

The statistical information of the first information is explained below using the first information as channel information.

For example, in a case where the statistical information of the first information includes at least two second statistics corresponding to the first information in the consecutive second time window, the statistical information of the first information may be: and the average value of the channel information of the terminal in a plurality of continuous time windows is, for example, the average value of the channel information in a time window 1 is a, the average value of the channel information in a time window 2 is b, the average value of the channel information in a time window 3 is c, and the statistical information of the channel information is the average value of a, b and c.

Optionally, in an embodiment of the present application, the statistical information of the measurement information of the first device includes at least one of the following: and third statistics of the measurement information in a third time window, fourth statistics of the measurement information corresponding to at least two continuous fourth time windows, and correlation information of the measurement information.

the moving speed of the first device is greater than a third threshold value;

Illustratively, the third threshold may be 60km/h,80km/h,100 km/h, and so on.

Further optionally, in an embodiment of the present application, the triggering condition includes: the measurement information of the first equipment meets a second preset condition; the meeting the second preset condition includes: the measurement information of the first device indicates that the channel environment of the associated channel of the first device has changed.

Wherein the measurement information includes at least one of: first measurement information of a reference signal received by first equipment, and second measurement information acquired by a sensor of the first equipment; optionally, the first measurement information includes at least one of: instantaneous measurement information of the reference signal, statistical measurement information of the reference signal;

wherein the reference signal includes at least one of: synchronization signal block SSB, CSI reference signal CSI-RS, sounding reference signal SRS, positioning reference signal PRS.

Further optionally, in an embodiment of the present application, the triggering condition includes: the statistical information of the second information corresponding to the first AI model meets a third preset condition; the meeting the third preset condition includes at least one of:

the second device instructs the first device to perform online learning;

The period adopted when the first equipment periodically or semi-periodically performs online learning is as follows: the second device pre-configures the period, or the first device autonomously configures the period.

the second device instructs the first device to learn online periodically through first signaling, wherein the first signaling comprises at least one of the following: a media access control-control unit MAC-CE, downlink control information DCI.

the first AI model achieves a preset accuracy;

a target task associated with the first AI model is aborted;

For example, in the process of online learning the first AI model, the first device may detect whether the suspension condition is satisfied in real time or periodically, and in the case where the suspension condition is satisfied, suspend online learning of the first AI model.

In one example, the first device may suspend online learning of the first AI model if it detects that the number of iterations of the first AI model is greater than a preset number of iterations (e.g., 10000). In another example, the second device may suspend online learning of the first AI model if it detects that the accuracy value of the first AI model reaches a preset accuracy. In yet another example, the second device may suspend online learning of the first AI model if an error of the output result of the first AI model is detected to be less than a preset error value. In this way, the first device can suspend online learning of the AI model based on the above-described number of online learning of the AI model, the achieved accuracy, the error of the output result, and the difference from the reference distribution satisfying the conditions, thereby saving power consumption while improving the model prediction accuracy.

Note that, suspending the online learning of the first AI model may be to temporarily stop the online learning of the first AI model, or to end the online learning of the first AI model.

an online learning mode of the first AI model;

the size of the sample batch of the first AI model;

the state of the optimizer of the first AI model;

a partitioning manner of a first data set of the first AI model;

composition information of a first data set of a first AI model;

contribution weights of a first data set of a first AI model to the AI model update;

AI model identification associated with the online learning information;

a reference distribution of the first AI model;

wherein said first data set is at least one of: the original data set used by the first AI model, the data set newly acquired by the first AI model. The parameter information of the reference distribution includes at least one of: variance, mean, standard deviation.

Optionally, the above-mentioned original data set is: in the data set used for offline training of the first AI model (i.e., the old data set), the new acquired data set is: after the first device side is deployed online, that is, the first AI model is configured for the first device, a data set collected by the first device in a new environment (that is, a new data set).

Illustratively, the Batch (Batch) size is N, and N is a positive integer.

For example, the first device may obtain parameter configuration information of the first AI model, and learn the first AI model online based on the parameter configuration information, thereby ensuring prediction accuracy of the AI model in a changing environment.

Further optionally, in an embodiment of the present application, the parameter configuration information includes an online learning mode of the first AI model, and the online learning mode is an instantaneous training mode, and the parameter configuration information further includes at least one of the following: the data volume collected by the first device, and the collection time length of the data volume.

According to the online learning method of the AI model, provided by the embodiment of the application, the execution subject can be an online learning device of the AI model. In the embodiment of the application, an online device of an AI model is taken as an example to execute an online method of the AI model, and the online learning device of the AI model provided by the embodiment of the application is described.

An embodiment of the present application provides an online learning apparatus 400 for an AI model, as shown in fig. 4, the online learning apparatus 400 for an AI model includes: a configuration module 401, wherein: the configuration module 401 is configured to deploy, by a second device, a first AI model for a first device; the configuration 401 is further configured to configure, by the second device, online learning information of the first AI model for the first device.

a triggering mode of online learning;

an abort condition for online learning;

parameter configuration information for online learning;

an online learned data set.

Alternatively, in an embodiment of the present application,

the triggering conditions corresponding to the triggering modes comprise at least one of the following:

the state information of the first equipment meets a first preset condition;

error information of an output result of the first AI model is larger than a second threshold;

Wherein the status information includes at least one of: moving speed, beam switching information and cell switching information;

the first information includes at least one of: input information of the first AI model and output information of the first AI model.

Optionally, in an embodiment of the present application, the statistical information of the first information includes at least one of: a first statistic of the first information in a first time window, a second statistic corresponding to the first information in at least two continuous second time windows, and statistic information of the first information at a first moment of at least two terminals in a first cell, wherein the correlation information of the first information;

the second statistic is calculated based on the statistics within each of the second time windows;

the statistics of the measurement information of the first device include at least one of: a third statistic for the measurement information in a third time window, wherein the measurement information corresponds to a fourth statistic in at least two continuous fourth time windows, and the correlation information of the measurement information;

the fourth statistic is calculated based on the statistics within each of the fourth time windows;

The correlation information includes at least one of: distance between data, covariance, correlation coefficient.

Optionally, in an embodiment of the present application, the triggering condition includes: the state information of the first equipment meets a first preset condition;

the meeting the first preset condition includes at least one of:

the speed of movement of the first device is greater than a third threshold;

the beam switching information indicates that the first device performs beam switching and the beam switching frequency is greater than a fourth threshold;

the cell switch information indicates that the first device is cell switched.

Optionally, in an embodiment of the present application, the triggering condition includes: the measurement information of the reference signal received by the first equipment meets a second preset condition;

the meeting the second preset condition includes: the measurement information of the reference signal indicates that the channel environment of the relevant channel of the first device changes;

wherein the measurement information includes at least one of: first measurement information of a reference signal received by the first equipment, and second measurement information acquired by a sensor of the first equipment;

the first measurement information includes at least one of: instantaneous measurement information of the reference signal, statistical measurement information of the reference signal;

The reference signal includes at least one of: synchronization signal block SSB, CSI reference signal CSI-RS, sounding reference signal SRS, positioning reference signal PRS.

Optionally, in an embodiment of the present application, the triggering condition includes: the statistical information of the first information corresponding to the first AI model meets a third preset condition;

the meeting the third preset condition includes at least one of:

the second statistic is greater than a maximum value of a second threshold interval;

correlation information between different first information acquired in the current time window meets a second condition.

Optionally, in an embodiment of the present application, the triggering condition includes: the statistical information of the measurement information of the first equipment meets a fourth preset condition;

the meeting the fourth preset condition includes at least one of:

the third statistic is greater than a maximum value of a third threshold interval;

the fourth statistic is greater than a maximum value of a fourth threshold interval;

the correlation information of the measurement information acquired in at least two time windows meets a third condition;

the difference between the distribution of the measurement information and a reference distribution, which is information configured by the second device for the first device, is greater than a fifth threshold.

Alternatively, in an embodiment of the present application,

the triggering conditions of the triggering mode comprise at least one of the following:

the second device instructs the first device to learn online;

the output accuracy of the first AI model is less than or equal to a sixth threshold.

Optionally, in an embodiment of the present application, the second device instructs the first device to perform online learning, including at least one of:

the second device instructs the first device to perform online learning periodically;

The period adopted when the first equipment periodically or semi-periodically performs online learning is as follows: the second device may be pre-configured for a period, or the first device may be autonomously configured for a period.

Optionally, in an embodiment of the present application, the second device instructs the first device to perform online learning periodically, including:

The second device instructs the first device to learn online in half-cycle through first signaling, the first signaling including at least one of: a media access control-control unit MAC-CE, downlink control information DCI.

the first AI model achieves a preset precision;

error information of an output result of the first AI model is less than a seventh threshold;

a target task associated with the first AI model is aborted;

the difference information of the measurement information distribution of the first device and the reference distribution is smaller than an eighth threshold.

Optionally, in an embodiment of the present application, the parameter configuration information includes at least one of:

an online learning mode of the AI model;

the size of the sample lot of the AI model;

the state of the optimizer of the AI model;

a partitioning manner of a first data set of the AI model;

composition information of a first dataset of the AI model;

a contribution weight of the first data set of the AI model to the AI model update;

An AI model identification associated with the first information;

a reference distribution of the first AI model;

wherein the first data set is at least one of: the original data set used by the first AI model, the data set newly acquired by the first AI model;

the parameter information of the reference distribution includes at least one of: variance, mean, standard deviation.

Optionally, in an embodiment of the present application, the parameter configuration information includes an online learning mode of the first AI model, and the online learning mode is an instantaneous training mode, and the parameter configuration information further includes at least one of: and the data volume is acquired by the first equipment, and the acquisition time length of the data volume is longer.

Optionally, in an embodiment of the present application, the parameter configuration information includes an online learning mode of the first AI model, and the online learning mode is a continuous learning mode, and the parameter configuration information further includes at least one of: time interval of two adjacent online learning, data volume interval of two adjacent online learning.

Optionally, in an embodiment of the present application, the second device includes at least one of: core network equipment, access network equipment and terminals; the first device comprises at least one of: core network equipment, access network equipment and terminals.

In the online learning device of an AI model provided in the embodiment of the present application, a second device configures a first AI model for a first device, and configures online learning information of the first AI model for the first device. According to the method, the first AI model is deployed at the first equipment side, and parameters required by online learning are configured for the first model, so that the first AI model can be continuously adjusted online at the first equipment side, the prediction performance of the first AI model is maintained, and the service quality of the first equipment is further ensured.

An embodiment of the present application provides an online learning apparatus 500 for an AI model, as shown in fig. 5, the online learning apparatus 500 for an AI model includes: an acquisition module 501 and an execution module 502, wherein: the acquiring module 501 is configured to acquire a first AI model by the first device; the execution module 502 is configured to perform online learning on the first AI model by using the first device based on online learning information of the first AI model.

Optionally, in an embodiment of the present application, the first device acquires a first AI model, including:

the first device receives a first AI model of a second device configuration.

Optionally, in an embodiment of the present application, the acquiring module is specifically configured to

Online learning information of the first AI model is acquired from the second device.

Optionally, in an embodiment of the present application, the apparatus further includes: and the configuration module is used for configuring online learning information of the first AI model for the third equipment.

a triggering mode of online learning;

an abort condition for online learning;

parameter configuration information for online learning;

an online learned data set.

the state information of the first equipment meets a first preset condition;

the meeting the first preset condition includes at least one of:

the speed of movement of the first device is greater than a third threshold;

the cell switch information indicates that the first device is cell switched.

Optionally, in an embodiment of the present application, the triggering condition includes: the statistical information of the second information corresponding to the first AI model meets a third preset condition;

the meeting the third preset condition includes at least one of:

the meeting the fourth preset condition includes at least one of:

the second device instructs the first device to learn online;

the first AI model achieves a preset precision;

a target task associated with the first AI model is aborted;

an online learning mode of the AI model;

the size of the sample lot of the AI model;

the state of the optimizer of the AI model;

a partitioning manner of a first data set of the AI model;

composition information of a first dataset of the AI model;

An AI model identification associated with the first information;

a reference distribution of the first AI model;

Optionally, in an embodiment of the present application, the second device, the first device, and the third device include at least one of: core network equipment, access network equipment, and terminals.

In the online learning device for an AI model provided by the embodiment of the present application, a first device acquires a first AI model, and performs online learning on the first AI model based on online learning information of the first AI model. According to the method, the first AI model is deployed at the first equipment side, and parameters required by online learning are configured for the first model, so that the first AI model can be continuously adjusted online at the first equipment side, the prediction performance of the first AI model is maintained, and the service quality of the first equipment is further ensured.

The online learning device of the AI model in the embodiment of the application may be an electronic device, for example, an electronic device with an operating system, or may be a component in the 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 online learning device for the AI model provided by the embodiment of the present application can implement each process implemented by the embodiments of the methods of fig. 1 to 3, and achieve the same technical effects, and for avoiding repetition, a detailed description is omitted here.

Optionally, as shown in fig. 6, the embodiment of the present application further provides a communication device 600, including a processor 601 and a memory 602, where the memory 602 stores a program or an instruction that can be executed on the processor 601, for example, when the communication device 600 is a terminal, the program or the instruction is executed by the processor 601 to implement each step of the above-mentioned online learning method embodiment of the AI model, and the same technical effects can be achieved. When the communication device 600 is a network side device, the program or the instruction, when executed by the processor 601, implements the steps of the above-mentioned embodiment of the online learning method of the AI model, and the same technical effects can be achieved, so that repetition is avoided, and no further description is given here.

Taking a first device as an example.

The embodiment of the application also provides a terminal which comprises a processor and a communication interface, wherein the processor is used for acquiring the first AI model and carrying out online learning on the first AI model based on online learning information of the first AI model. The terminal embodiment corresponds to the terminal-side method embodiment, and each implementation process and implementation manner of the method embodiment can be applied to the terminal embodiment, and the same technical effects can be achieved. Specifically, fig. 7 is a schematic diagram of a hardware structure of a terminal for implementing an embodiment of the present application.

The terminal 700 includes, but is not limited to: at least some of the components of the radio frequency unit 701, the network module 702, the audio output unit 703, the input unit 704, the sensor 705, the display unit 706, the user input unit 707, the interface unit 708, the memory 709, and the processor 710.

Those skilled in the art will appreciate that the terminal 700 may further include a power source (e.g., a battery) for powering the various components, and that the power source may be logically coupled to the processor 710 via a power management system so as to perform functions such as managing charging, discharging, and power consumption via the power management system. The terminal structure shown in fig. 7 does not constitute a limitation of the terminal, and the terminal may include more or less components than shown, or may combine certain components, or may be arranged in different components, which will not be described in detail herein.

It should be appreciated that in embodiments of the present application, the input unit 704 may include a graphics processing unit (Graphics Processing Unit, GPU) 7041 and a microphone 7042, with the graphics processor 7041 processing image data of still pictures or video obtained by an image capturing device (e.g., a camera) in a video capturing mode or an image capturing mode. The display unit 706 may include a display panel 7061, and the display panel 7061 may be configured in the form of a liquid crystal display, an organic light emitting diode, or the like. The user input unit 707 includes at least one of a touch panel 7071 and other input devices 7072. The touch panel 7071 is also referred to as a touch screen. The touch panel 7071 may include two parts, a touch detection device and a touch controller. Other input devices 7072 may include, but are not limited to, a physical keyboard, function keys (e.g., volume control keys, switch keys, etc.), a trackball, a mouse, a joystick, and so forth, which are not described in detail herein.

In the embodiment of the present application, after receiving downlink data from a network side device, the radio frequency unit 701 may transmit the downlink data to the processor 710 for processing; in addition, the radio frequency unit 701 may send uplink data to the network side device. Typically, the radio unit 701 includes, but is not limited to, an antenna, an amplifier, a transceiver, a coupler, a low noise amplifier, a duplexer, and the like.

The memory 709 may be used to store software programs or instructions and various data. The memory 709 may mainly include a first storage area storing programs or instructions and a second storage area storing data, wherein the first storage area may store an operating system, application programs or instructions (such as a sound playing function, an image playing function, etc.) required for at least one function, and the like. Further, the memory 709 may include volatile memory or nonvolatile memory, or the memory 709 may include both volatile and nonvolatile memory. The nonvolatile Memory may be a Read-Only Memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an Electrically Erasable EPROM (EEPROM), or a flash Memory. The volatile memory may be random access memory (Random Access Memory, RAM), static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (ddr SDRAM), enhanced SDRAM (Enhanced SDRAM), synchronous DRAM (SLDRAM), and Direct RAM (DRRAM). Memory 709 in embodiments of the application includes, but is not limited to, these and any other suitable types of memory.

Processor 710 may include one or more processing units; optionally, processor 710 integrates an application processor that primarily processes operations involving an operating system, user interface, application programs, and the like, and a modem processor that primarily processes wireless communication signals, such as a baseband processor. It will be appreciated that the modem processor described above may not be integrated into the processor 710.

The processor 710 is configured to obtain a first AI model by the first device, where the processor 710 is further configured to perform online learning on the first AI model by the first device based on online learning information of the first AI model.

Optionally, in an embodiment of the present application, the radio frequency unit 701 is configured to receive a first AI model configured by a second device.

Optionally, in an embodiment of the present application, the processor 710 is specifically configured to obtain online learning information of the first AI model from the second device.

Optionally, in an embodiment of the present application, the processor 710 is further configured to configure online learning information of the first AI model for a third device.

A triggering mode of online learning;

an abort condition for online learning;

parameter configuration information for online learning;

an online learned data set.

the state information of the first equipment meets a first preset condition;

the meeting the first preset condition includes at least one of:

the speed of movement of the first device is greater than a third threshold;

the cell switch information indicates that the first device is cell switched.

the meeting the third preset condition includes at least one of:

the meeting the fourth preset condition includes at least one of:

the second device instructs the first device to learn online;

the first AI model achieves a preset precision;

A target task associated with the first AI model is aborted;

an online learning mode of the AI model;

the size of the sample lot of the AI model;

the state of the optimizer of the AI model;

a partitioning manner of a first data set of the AI model;

composition information of a first dataset of the AI model;

an AI model identification associated with the first information;

a reference distribution of the first AI model;

In the terminal provided by the embodiment of the application, the terminal acquires the first AI model and carries out online learning on the first AI model based on the online learning information of the first AI model. According to the method, the first AI model is deployed at the terminal side, and parameters required by online learning are configured for the first model, so that continuous online adjustment can be performed on the first AI model at the terminal side, the prediction performance of the first AI model is maintained, and the service quality of the terminal is further ensured.

Take the second device as a network side device as an example.

The embodiment of the application also provides network side equipment, which comprises a processor and a communication interface, wherein the processor is used for configuring a first AI model for the first equipment; and configuring the first device with online learning information of the first AI model. 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: an antenna 81, a radio frequency device 82, a baseband device 83, a processor 84 and a memory 85. The antenna 81 is connected to a radio frequency device 82. In the uplink direction, the radio frequency device 82 receives information via the antenna 81, and transmits the received information to the baseband device 83 for processing. In the downlink direction, the baseband device 83 processes information to be transmitted, and transmits the processed information to the radio frequency device 82, and the radio frequency device 82 processes the received information and transmits the processed information through the antenna 81.

The method performed by the network side device in the above embodiment may be implemented in the baseband apparatus 83, and the baseband apparatus 83 includes a baseband processor.

The baseband device 83 may, for example, include at least one baseband board, where a plurality of chips are disposed, as shown in fig. 8, where one chip, for example, a baseband processor, is connected to the memory 85 through a bus interface, so as to call a program in the memory 85 to perform the network device operation shown in the above method embodiment.

The network-side device may also include a network interface 86, such as 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 85 and executable on the processor 84, the processor 84 invokes the instructions or programs in the memory 85 to perform the method performed by the modules shown in fig. 4 and achieve the same technical effects, and are not repeated here.

Specifically, the embodiment of the application also provides network side equipment. As shown in fig. 9, the network side device 900 includes: a processor 901, a network interface 902, and a memory 903. The network interface 902 is, for example, a common public radio interface (common public radio interface, CPRI).

Specifically, the network side device 900 of the embodiment of the present application further includes: instructions or programs stored in the memory 903 and executable on the processor 901, the processor 901 invokes the instructions or programs in the memory 903 to execute the method executed by each module shown in fig. 4, and achieve the same technical effects, so that repetition is avoided and thus a description thereof is omitted.

The embodiment of the application also provides a readable storage medium, wherein the readable storage medium stores a program or an instruction, and the program or the instruction realizes each process of the online learning method embodiment of the AI model when being executed by a processor, and can achieve the same technical effect, so that repetition is avoided and no description is repeated 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 a program or instructions, the processes of the online learning method embodiment of the above AI model are 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 processes of the above-described embodiments of the online learning method of the AI model, and achieve the same technical effects, so that repetition is avoided, and details are not repeated herein.

The embodiment of the application also provides a communication system, which comprises: the terminal can be used for executing the steps of the online learning method of the AI model, and the network side equipment can be used for executing the steps of the online learning method of the AI model.

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. An on-line learning method of an artificial intelligence AI model, the method comprising:

the first equipment acquires a first AI model;

the first device performs online learning on the first AI model based on online learning information of the first AI model.

2. The method of claim 1, wherein the first device obtains a first AI model, comprising:

the first device receives a first AI model of a second device configuration.

3. The method of claim 1 or 2, wherein the first device, prior to online learning the first AI model based on the online learning information of the first AI model, further comprises:

the first device obtains online learning information of the first AI model from the second device.

4. The method according to claim 1, wherein the method further comprises:

the first device configures online learning information of the first AI model for a third device.

5. The method of claim 1, wherein the online learning information comprises at least one of:

a triggering mode of online learning;

an abort condition for online learning;

parameter configuration information for online learning;

an online learned data set.

6. The method of claim 5, wherein the step of determining the position of the probe is performed,

the state information of the first equipment meets a first preset condition;

7. The method of claim 6, wherein the step of providing the first layer comprises,

the statistical information of the first information includes at least one of: a first statistic of the first information in a first time window, a second statistic corresponding to the first information in at least two continuous second time windows, and statistic information of the first information at a first moment of at least two terminals in a first cell, wherein the correlation information of the first information;

the second statistic is calculated based on a statistics within each of the second time windows.

8. The method of claim 7, wherein the third preset condition comprises at least one of:

9. The method of claim 6, wherein the step of providing the first layer comprises,

The statistics of the measurement information of the first device include at least one of: third statistics of the measurement information in a third time window, fourth statistics of the measurement information corresponding to at least two continuous fourth time windows, and correlation information of the measurement information;

10. The method of claim 9, wherein the meeting a fourth preset condition comprises at least one of:

11. The method of claim 6, wherein the trigger condition comprises: the state information of the first equipment meets a first preset condition;

the meeting the first preset condition includes at least one of:

the speed of movement of the first device is greater than a third threshold;

the cell switch information indicates that the first device is cell switched.

12. The method of claim 6, wherein the trigger condition comprises: the measurement information of the first equipment meets a second preset condition;

the meeting the second preset condition includes: the measurement information indicates that a channel environment of a relevant channel of the first device changes;

13. The method of claim 5, wherein the step of determining the position of the probe is performed,

the second device instructs the first device to learn online;

14. The method of claim 13, wherein the step of determining the position of the probe is performed,

the second device instructs the first device to learn online, including at least one of:

the second device instructs the first device to perform online learning aperiodically;

15. The method of claim 5, wherein the abort condition comprises at least one of:

the first AI model achieves a preset precision;

a target task associated with the first AI model is aborted;

16. The method of claim 5, wherein the parameter configuration information comprises at least one of:

an online learning mode of the first AI model;

the size of the sample batch of the first AI model;

a state of an optimizer of the first AI model;

a partitioning manner of a first data set of the first AI model;

composition information of a first data set of the first AI model;

a contribution weight of a first dataset of the first AI model to the first AI model update;

an AI model identification associated with the online learning information;

a reference distribution of the first AI model;

17. The method of claim 16, wherein the parameter configuration information includes an online learning mode of the first AI model, and the online learning mode is a transient training mode, the parameter configuration information further including at least one of: and the data volume is acquired by the first equipment, and the acquisition time length of the data volume is longer.

18. The method of claim 16, wherein the parameter configuration information includes an online learning mode of the first AI model and the online learning mode is a continuous learning mode, the parameter configuration information further including at least one of: time interval of two adjacent online learning, data volume interval of two adjacent online learning.

19. The method of claim 1, wherein the second device, first device, and third device comprise at least one of: core network equipment, access network equipment, and terminals.

20. An on-line learning method of an artificial intelligence AI model, the method comprising:

the second device configures a first AI model for the first device;

The second device configures online learning information of the first AI model for the first device.

21. The method of claim 20, wherein the online learning information comprises at least one of:

a triggering mode of online learning;

an abort condition for online learning;

parameter configuration information for online learning;

an online learned data set.

22. The method of claim 21, wherein the step of determining the position of the probe is performed,

the state information of the first equipment meets a first preset condition;

23. The method of claim 22, wherein the step of determining the position of the probe is performed,

24. The method of claim 23, wherein the third preset condition comprises at least one of:

25. The method of claim 22, wherein the step of determining the position of the probe is performed,

26. The method of claim 25, wherein the meeting a fourth predetermined condition comprises at least one of:

27. The method of claim 22, wherein the trigger condition comprises: the state information of the first equipment meets a first preset condition;

the meeting the first preset condition includes at least one of:

the speed of movement of the first device is greater than a third threshold;

the cell switch information indicates that the first device is cell switched.

28. The method of claim 22, wherein the trigger condition comprises: the measurement information of the first equipment meets a second preset condition;

29. The method of claim 21, wherein the step of determining the position of the probe is performed,

The second device instructs the first device to learn online;

30. The method of claim 29, wherein the step of providing the first information comprises,

31. The method of claim 21, wherein the abort condition comprises at least one of:

the first AI model achieves a preset precision;

A target task associated with the first AI model is aborted;

32. The method of claim 21, wherein the parameter configuration information comprises at least one of:

an online learning mode of the first AI model;

the size of the sample batch of the first AI model;

a state of an optimizer of the first AI model;

a partitioning manner of a first data set of the first AI model;

composition information of a first data set of the first AI model;

a contribution weight of a first dataset of the first AI model to the AI model update;

an AI model identification associated with the online learning information;

a reference distribution of the first AI model;

33. The method of claim 32, wherein the parameter configuration information includes an online learning mode of the first AI model, and the online learning mode is a transient training mode, the parameter configuration information further including at least one of: and the data volume is acquired by the first equipment, and the acquisition time length of the data volume is longer.

34. The method of claim 33, wherein the parameter configuration information includes an online learning mode of the first AI model and the online learning mode is a continuous learning mode, the parameter configuration information further including at least one of: time interval of two adjacent online learning, data volume interval of two adjacent online learning.

35. The method of claim 20, wherein the second device comprises at least one of: core network equipment, access network equipment and terminals; the first device comprises at least one of: core network equipment, access network equipment and terminals.

36. An on-line learning device for AI models, the device comprising: a configuration module, wherein:

the configuration module is used for configuring a first AI model for the first device by the second device;

the configuration module is further configured to configure online learning information of the first AI model for the first device by the second device.

37. An on-line learning device for an artificial intelligence AI model, the device comprising: the device comprises an acquisition module and an execution module, wherein:

the acquisition module is used for acquiring a first AI model by the first equipment;

The execution module is used for the first equipment to conduct online learning on the first AI model based on online learning information of the first AI model.

38. A communication device 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 online learning method of the AI model of any of claims 1-19.

39. A communication device 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 online learning method of the AI model of any of claims 20-35.

40. A readable storage medium, wherein a program or instructions is stored on the readable storage medium, which when executed by a processor, implements the method of online learning of AI models of any of claims 1-19, or the steps of the method of online learning of AI models of any of claims 20-35.