Model updating method and communication device

Abstract

The application discloses a model updating method and communication equipment, and belongs to the technical field of communication. The model updating method comprises the following steps: a first communication equipment acquires model updating information; the first communication equipment updates a time sequence prediction model used by the first communication equipment according to the model updating information, and the time sequence prediction model is used for predicting the execution of a task; wherein the model updating information comprises at least one of the following: updating information of a kernel function; updating information of model hyperparameters; updating information of a prediction mode; and updating information of a calculation mode.

Description

Technical Field

[0001] This application belongs to the field of communication technology, specifically relating to a model update method and a communication device. Background Technology

[0002] With the increasing mobility of communication scenarios, such as non-terrestrial networks (NTN) and high-speed rail, the rapid changes in the scattering environment around communication equipment severely shorten the coherence time of the wireless channel. This means that the detection results of the wireless channel by the communication equipment will quickly become invalid, so more frequent wireless channel detection is necessary to maintain good communication performance.

[0003] Time series forecasting is a method used by communication devices to predict future communication measurements based on observations of past communication measurements. Therefore, time series forecasting models are widely used in predicting key time-related parameters in wireless communication networks. Common time series forecasting models include the Gaussian process (GP), which achieves good forecasting performance with low model storage, computation, and parameter interaction overhead.

[0004] However, in certain scenarios (such as when the prediction task and / or prediction requirements change), the model parameters of the time series prediction model may need to be updated in order to ensure the prediction performance of the time series prediction model. Therefore, how to update the model parameters of the time series prediction model has become an urgent problem to be solved. Summary of the Invention

[0005] This application provides a model update method and communication device that can solve the problem of how to update the model parameters of a time series prediction model.

[0006] In a first aspect, a model update method is provided, comprising: a first communication device acquiring model update information; the first communication device updating a time series prediction model used by the first communication device according to the model update information, the time series prediction model being used to predict the execution of a task; wherein the model update information includes at least one of the following: kernel function update information; model hyperparameter update information; prediction mode update information; and computation mode update information.

[0007] Secondly, a model update method is provided, comprising: a second communication device sending model update information to a first communication device, the model update information being used to update a time series prediction model used by the first communication device, the time series prediction model being used to predict the execution of a task; wherein the model update information includes at least one of the following: kernel function update information; model hyperparameter update information; prediction mode update information; and computation mode update information.

[0008] Thirdly, a model updating device is provided, comprising an acquisition module and an update module; the acquisition module is used to acquire model update information; the update module is used to update the time series prediction model used by the first communication device according to the model update information, the time series prediction model being used to predict the execution of a task; wherein, the model update information includes at least one of the following: kernel function update information; model hyperparameter update information; prediction mode update information; and computation mode update information.

[0009] Fourthly, a model update apparatus is provided, comprising a sending module; the sending module is configured to send model update information to a first communication device, the model update information being used to update a time series prediction model used by the first communication device, the time series prediction model being used to predict the execution of a task; wherein the model update information includes at least one of the following: kernel function update information; model hyperparameter update information; prediction mode update information; and computation mode update information.

[0010] Fifthly, a communication device is provided, the terminal including a processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the method as described in the first or second aspect.

[0011] In a sixth aspect, a communication device is provided, including a processor and a communication interface. When the communication device is a first communication device, the communication interface is used to acquire model update information; the processor updates the time series prediction model used by the first communication device according to the model update information; or, when the communication device is a second communication device, the communication interface is used to send model update information to the first communication device. The model update information is used to update the time series prediction model used by the first communication device, and the time series prediction model is used to predict the execution of a task. The model update information includes at least one of the following: kernel function update information; model hyperparameter update information; prediction mode update information; and computation mode update information.

[0012] A seventh aspect provides a communication system comprising: a terminal and a network-side device, wherein the terminal is configured to perform the steps of the model update method as described in the first aspect, and the network-side device is configured to perform the steps of the model update method as described in the second aspect; or the network-side device is configured to perform the steps of the model update method as described in the first aspect, and the terminal is configured to perform the steps of the model update method as described in the second aspect.

[0013] Eighthly, a readable storage medium is provided, on which a program or instructions are stored, which, when executed by a processor, implement the steps of the model update method as described in the first aspect, or implement the steps of the model update method as described in the second aspect.

[0014] In a ninth aspect, a chip is provided, the chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the model update method as described in the first aspect, or to implement the model update method as described in the second aspect.

[0015] In a tenth aspect, a computer program / program product is provided, the computer program / program product being stored in a storage medium, the computer program / program product being executed by at least one processor to implement the steps of the model update method as described in the first or second aspect.

[0016] In this embodiment, the first communication device acquires model update information; based on the model update information, the first communication device updates the time series prediction model used by the first communication device, which is used to execute the prediction task; wherein, the model update information includes at least one of the following: kernel function update information; model hyperparameter update information; prediction mode update information; and computation mode update information. Through this scheme, since the kernel function, model hyperparameters, prediction mode, and computation mode are all parameter information of the time series prediction model, after the first communication device acquires the model update information, it can update the time series prediction model used by the first communication device according to the model update information, thereby enabling the time series prediction model to execute the prediction task more accurately, and thus enabling the first communication device to obtain good communication performance. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of a wireless communication system provided in an embodiment of this application;

[0018] Figure 2 This is a schematic diagram of a neural network provided in an embodiment of this application;

[0019] Figure 3 This is a schematic diagram of a neuron provided in an embodiment of this application;

[0020] Figure 4 This is a schematic diagram illustrating the variation of normalized mean square error with prediction time unit for different training window sizes, provided in an embodiment of this application.

[0021] Figure 5 This is a schematic diagram of the smooth-turn motion trajectory of a terminal according to an embodiment of this application;

[0022] Figure 6 This is a schematic diagram illustrating the average error of prediction at different sampling intervals provided in an embodiment of this application;

[0023] Figure 7 This is a schematic flowchart of a model update method provided in an embodiment of this application;

[0024] Figure 8 This is a schematic diagram of the structure of a model update device provided in an embodiment of this application;

[0025] Figure 9 This is a schematic diagram of another model update device provided in an embodiment of this application;

[0026] Figure 10 This is a schematic diagram of the structure of the communication device provided in the embodiments of this application;

[0027] Figure 11 This is a hardware schematic diagram of the communication device provided in the embodiments of this application. Detailed Implementation

[0028] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0029] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first" and "second" are generally of the same class, not limited in number; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0030] It is worth noting that the technologies described in this application are not limited to Long Term Evolution (LTE) / LTE-Advanced (LTE-A) systems, but can also be used in other wireless communication systems, such as Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Orthogonal Frequency Division Multiple Access (OFDMA), Single-carrier Frequency Division Multiple Access (SC-FDMA), and other systems. The terms "system" and "network" in this application are often used interchangeably, and the described technologies can be used in the systems and radio technologies mentioned above, as well as in other systems and radio technologies. The following description describes New Radio (NR) systems for illustrative purposes, and NR terminology is used in most of the following description; however, these technologies can also be applied to applications beyond NR systems, such as 6th Generation (6G) communication systems.

[0031] Figure 1This diagram illustrates a block diagram of a wireless communication system applicable to embodiments of this application. The wireless communication system includes a terminal 11 and a network-side device 12. Terminal 11 can be a mobile phone, tablet computer, laptop computer, personal digital assistant (PDA), handheld computer, netbook, ultra-mobile personal computer (UMPC), mobile internet device (MID), augmented reality (AR) / virtual reality (VR) device, robot, wearable device, vehicle-mounted device (VUE), pedestrian terminal (PUE), smart home (home devices with wireless communication capabilities, such as refrigerators, televisions, washing machines, or furniture), game console, personal computer (PC), ATM, or self-service machine, etc. Wearable devices include: smartwatches, smart bracelets, smart headphones, smart glasses, smart jewelry (smart bracelets, smart chains, smart rings, smart necklaces, smart anklets, smart anklets, etc.), smart wristbands, smart clothing, etc. It should be noted that the specific type of terminal 11 is not limited in this embodiment. Network-side equipment 12 may include access network equipment or core network equipment. Access network equipment 12 may also be referred to as radio access network equipment, radio access network (RAN), radio access network function, or radio access network unit. Access network equipment 12 may include base stations, WLAN access points, or WiFi nodes, etc. Base stations may be referred to as Node B, evolved Node B (eNB), access point, base transceiver station (BTS), radio base station, radio transceiver, Basic Service Set (BSS), Extended Service Set (ESS), home B node, home evolved B node, Transmitting Receiving Point (TRP), or any other suitable term in the field, as long as the same technical effect is achieved. The base station is not limited to specific technical terms. It should be noted that in this application embodiment, only a base station in an NR system is used as an example for description, and the specific type of base station is not limited.Core network equipment may include, but is not limited to, at least one of the following: core network node, core network function, Mobility Management Entity (MME), Access and Mobility Management Function (AMF), Session Management Function (SMF), User Plane Function (UPF), Policy Control Function (PCF), Policy and Charging Rules Function (PCRF), Edge Application Server Discovery Function (EASDF), Unified Data Management (UDM), Unified Data Repository (UDR), Home Subscriber Server (HSS), Centralized network configuration (CNC), Network Repository Function (NRF), Network Exposure Function (NEF), Local NEF (or L-NEF), Binding Support Function (BSF), and Application Function. Function (AF), etc. It should be noted that the embodiments of this application only use the core network equipment in the NR system as an example for introduction, and do not limit the specific type of core network equipment.

[0032] Due to the explosive growth of mobile devices and mobile data traffic, as well as the emergence of numerous application scenarios, future wireless communication networks will require high speed, low latency, and enhanced mobility. Especially for enhanced mobility scenarios such as NTN and high-speed rail, the rapid changes in the surrounding scattering environment severely shorten the coherence time of the wireless channel. This causes communication devices (such as wireless communication equipment like terminals and base stations) to quickly lose their detection results on the wireless channel. Consequently, communication devices must conduct more frequent wireless channel probing to maintain good communication performance, posing a significant challenge to network load and energy consumption. Time series forecasting is a method for communication devices to predict future communication measurements based on observations of past communication measurements. Essentially, it uses historical measurement data to mine the temporal correlation between the future and the past.

[0033] Currently, artificial intelligence (AI) has been widely applied in various fields. Integrating AI into wireless communication networks to significantly improve technical indicators such as throughput, latency, and user capacity is an important task for future wireless communication networks. AI modules can be implemented in various ways, such as neural networks, decision trees, support vector machines, and Bayesian classifiers. This application uses neural networks as an example for illustration, but it does not limit the specific type of AI module. Figure 2 The diagram shown is a schematic of a neural network. The neural network is composed of neurons, and a schematic diagram of a neuron is shown below. Figure 3 As shown in the diagram. Here, a1, a2, ..., aK are the input values, w is the weight (multiplicative coefficient), b is the bias (additive coefficient), and σ(.) is the activation function. Common activation functions include Sigmoid, tanh, rectified linear unit (ReLU), etc.

[0034] The parameters of a neural network can be optimized using gradient optimization algorithms. Gradient optimization algorithms are a class of algorithms that minimize or maximize an objective function (also known as 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 building a neural network model f(), based on the input x, we can obtain the predicted output f(x), and calculate the difference between the predicted value and the true value (f(x) - Y), which is the loss function. When building the model, it is necessary to determine appropriate W and b to minimize the value of the loss function. The smaller the value of the loss function, the closer the model is to the reality, that is, the higher the accuracy of the model.

[0035] Most optimization algorithms are based on the error back propagation (BP) 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. During forward propagation, the input sample is introduced from the input layer, processed layer by layer by the hidden layers, and then propagated to the output layer. If the actual output of the output layer does not match the expected output, the process transitions to the error back propagation stage. Error back propagation involves propagating the output error back to the input layer layer by layer through the hidden layers in a certain form, distributing the error to all units in each layer, thereby obtaining the error signal of each unit. This error signal serves as the basis for adjusting the weights of each unit. This process of adjusting the weights of each layer through forward and backward propagation is repeated continuously. This continuous adjustment of weights is the learning and training process of the neural network. This process continues until the error of the network output is reduced to an acceptable level, or until the predetermined number of learning iterations is reached.

[0036] For example, optimization algorithms may include gradient descent, stochastic gradient descent (SGD), mini-batch gradient descent, momentum, Nesterov with momentum, adaptive gradient descent (Adagrad), Adadelta, root mean square propagation (RMSprop), and adaptive moment estimation (Adam). During error backpropagation, these optimization algorithms calculate the gradient by taking the derivative / partial derivative of the error / loss obtained from the loss function with respect to the current neuron, adding the learning rate, previous gradients / derivatives / partial derivatives, etc., and then propagating this gradient to the previous layer.

[0037] Generalized regression (GP) is a commonly used time series forecasting model. GP is a machine learning method developed based on statistical learning theory and Bayesian theory. It is suitable for handling complex regression and classification problems with high dimensionality, small sample sizes, and nonlinearity, and has strong generalization ability. Compared with methods such as neural networks, GP has advantages such as ease of implementation, small sample training, adaptive hyperparameter acquisition, flexible nonparametric inference, interpretable output, and probabilistic meaning, making it a potential solution to complex time series forecasting problems in future wireless communication systems.

[0038] Simulation results show that GP has good predictive performance for variables with strong time correlation, such as CSI and location. For CSI prediction, Figure 4The paper shows the variation of normalized mean squared error (NMSE) with prediction time units for different training window sizes at a speed of 30 km / h. Even with a training window size of only 4 time units, relatively reliable CSI predictions were achieved for the initial 2 time units, although prediction accuracy gradually decreased over time. For position prediction, in... Figure 5 Under the smooth-turn trajectory shown with a terminal speed of 15 km / h, Figure 6 The diagram shows the prediction of the position for the next 5 time units (e.g., 200, 210, 220, 230, 240) based on the position of the previous 20 time units (e.g., time units 0, 10, 20, 30, ..., 190) with a position sampling interval of 10 time units. The average position error of the prediction varies with the position estimation error (the position of the previous 20 time units is estimated by the position reference signal PRS). Even when the estimation error of the historical position information is 3m, the prediction error is still less than 0.5m.

[0039] The model update method provided in this application will be described in detail below with reference to the accompanying drawings, through some embodiments and application scenarios.

[0040] like Figure 7 As shown in the embodiments of this application, a model update method is provided, which can be applied to, for example... Figure 1 The wireless communication system shown may include steps 201 and 202 as described below.

[0041] Step 201: The first communication device obtains model update information.

[0042] Step 202: The first communication device updates the time series prediction model used by the first communication device according to the model update information.

[0043] The aforementioned time series prediction model can be used to predict the execution of tasks, and the model update information may include at least one of the following:

[0044] Kernel function update information;

[0045] Update information for model hyperparameters;

[0046] Update information for the prediction model;

[0047] Information on the update of the computing model.

[0048] It should be noted that, in the embodiments of this application, the above-mentioned model update information may also include any other possible update information, which can be determined according to actual usage requirements, and is not limited in the embodiments of this application.

[0049] In this embodiment, after the first communication device obtains model update information (e.g., at least one of kernel function update information, model hyperparameter update information, prediction mode update information, or computation mode update information), the first communication device can update the time series prediction model used by the first communication device according to the model update information, thereby enabling the time series prediction model to perform the prediction task more accurately. This allows the first communication device to achieve good communication performance.

[0050] Optionally, in this embodiment of the application, the prediction mode may include the structure of the time window for the prediction task, and the structure of the time window may include at least one of the following:

[0051] The length of the time unit in the training window of a time series prediction model;

[0052] The length of the time unit of the prediction window in a time series forecasting model;

[0053] The interval between adjacent time units in the training window of a time series prediction model;

[0054] The interval between adjacent time units in the prediction window of a time series forecasting model.

[0055] Optionally, in embodiments of this application, the prediction task may include at least one of predicting location and predicting channel (e.g., predicting channel state information (CSI)).

[0056] Of course, in actual implementation, the prediction task involved in the embodiments of this application may also include any other possible tasks, which can be determined according to actual usage requirements. The embodiments of this application do not limit this.

[0057] Optionally, in the embodiments of this application, the time unit may include, but is not limited to: reference signal period, prediction period, time slot, half time slot, symbol (e.g., orthogonal frequency division multiplexing (OFDM) symbol), subframe, radio frame, millisecond, second, which can be determined according to actual usage requirements.

[0058] In this embodiment, the length of the time unit of the training window of the aforementioned time series prediction model refers to the number of consecutive time units contained within a training window. Correspondingly, the length of the time unit of the prediction window of the aforementioned time series prediction model refers to the number of consecutive time units contained within a prediction window. Each time unit contains one data input or output; that is, one time unit corresponds to one data point.

[0059] For example, assuming the above prediction task is CSI prediction, that is, predicting the CSI of the next M time units based on the CSI of the past N time units, then the training window of the time series prediction model has a time unit length of N time units, and the prediction window of the time series prediction model has a time unit length of M. Furthermore, the interval between adjacent time units in the training window of the time series prediction model is 2, and the interval between adjacent time units in the prediction window of the time series prediction model is also 2. Therefore, based on the CSI of 0, 2, ..., 2N time units, the CSI of 2N+2, 2N+4, ..., 2N+2M ​​time units can be predicted.

[0060] Optionally, in the embodiments of this application, the model update information may further include at least one of the following:

[0061] The identifier of the prediction task to be updated (also known as the target task);

[0062] The identifier of the model to be updated (also known as the target model, such as the time series prediction model in the embodiments of this application);

[0063] The timestamp information of the model to be updated.

[0064] It should be noted that in the embodiments of this application, the first communication device can be a terminal or a network-side device, which can be determined according to actual usage requirements, and is not limited in the embodiments of this application.

[0065] Optionally, the network-side device mentioned in the embodiments of this application can be:

[0066] a) Core network nodes, including network data analytics function (NWDAF), location management function (LMF), or neural network processing nodes;

[0067] b) Access network nodes, including base stations or newly defined neural network processing nodes;

[0068] c) A combination of the above nodes.

[0069] In some embodiments, the first communication device may directly determine model update information based on information such as the prediction task and / or prediction requirements.

[0070] In other embodiments, the first communication device may receive model update information from the second communication device.

[0071] In some other embodiments described above, step 202 can be implemented by steps 203 and 202a as described below.

[0072] Step 203: The second communication device sends model update information to the first communication device.

[0073] Step 202a: The first communication device receives model update information from the second communication device.

[0074] Optionally, in this embodiment, the second communication device can be a network-side device or a terminal. For example, if the first communication device is a terminal, the second communication device can be a network-side device; if the first communication device is a network-side device, the second communication device can be a terminal. The specific choice can be determined based on actual usage requirements, and this embodiment does not impose any limitations.

[0075] Optionally, in this embodiment, the first communication device may store the same list of prediction models as the second communication device, and the list of prediction models may include identifiers of prediction modes. Thus, when the model update information includes update information of prediction modes, the first communication device can determine the prediction mode indicated by the update information of the prediction mode from the list of prediction models.

[0076] Correspondingly, the second communication device may store the same list of prediction models as the first communication device. This list of prediction models may include identifiers of prediction modes, thereby enabling the second communication device to configure a suitable prediction mode for the first communication device based on the list of prediction models.

[0077] In this embodiment, the second communication device can periodically send the aforementioned model update information to the first communication device, or it can send the model update information to the first communication device aperiodically. It is understood that the first communication device can periodically update the time series prediction model, or it can update the time series prediction model aperiodically.

[0078] Optionally, in this embodiment of the application, the above step 203 can be implemented by the following step 203a, based on the second communication device periodically sending model update information to the first communication device.

[0079] Step 203a: The second communication device sends model update information to the first communication device based on the fourth information.

[0080] The aforementioned fourth piece of information may include at least one of the following:

[0081] Prediction error of time series forecasting models;

[0082] Statistical information on the prediction error of time series forecasting models;

[0083] Mobility information of the first communication device;

[0084] Statistical information on the prediction results of time series forecasting models;

[0085] Update information for the prediction task of the time series forecasting model;

[0086] Updated information on the forecasting performance requirements of time series forecasting models;

[0087] Environmental perception information.

[0088] It is understood that in the embodiments of this application, when the prediction task and / or prediction performance requirements of the time series prediction model used by the first communication device are updated, the second communication device may send model update information to the first communication device.

[0089] Optionally, in this embodiment of the application, the mobility information of the first communication device may include at least one of the following: the moving speed of the first communication device, the beam switching information of the first communication device, and the cell handover information of the first communication device.

[0090] For example, when the moving speed of the first communication device is greater than a first preset threshold, the second communication device may send model update information to the first communication device; and / or, when the beam switching frequency of the first communication device is greater than a second preset threshold, the second communication device may send model update information to the first communication device; and / or, when the first communication device completes cell handover, the second communication device may send model update information to the first communication device.

[0091] Optionally, in the embodiments of this application, the above prediction task may include at least one of prediction location and prediction channel.

[0092] Optionally, in the embodiments of this application, the statistical information of the prediction results of the above-mentioned time series prediction model may include the mean and / or the variance of the prediction results.

[0093] For example, assuming the above prediction task is to predict location, and the statistical information of the prediction result is the variance of the predicted location, then if the variance of the predicted location by the first communication device is greater than a third preset threshold, the second communication device can send model update information to the first communication device.

[0094] Optionally, in embodiments of this application, the aforementioned prediction performance requirements may include at least one of prediction accuracy and computation latency. The aforementioned prediction task may include at least one of prediction location and prediction channel.

[0095] Of course, in actual implementation, the prediction performance requirements and prediction tasks involved in the embodiments of this application may also include any other possible performance requirements and tasks, which can be determined according to actual usage requirements. The embodiments of this application do not limit them.

[0096] Optionally, in the embodiments of this application, the above-mentioned environmental perception information may include, but is not limited to: line of sight (OS) environment, non-line of sight (NLOS) environment, dense urban environment, and rural environment, which can be determined according to actual usage needs.

[0097] Optionally, in this embodiment, the first communication device may send a model update request to the second communication device to obtain model update information. Based on this, before step 203 above, the model update method provided in this embodiment may further include steps 204 and 205 as described below.

[0098] Step 204: The first communication device sends a model update request to the second communication device.

[0099] Step 205: The second communication device receives model update request information from the first communication device.

[0100] The aforementioned model update request information can be used to request the first update information; the aforementioned model update information may include at least one of the following:

[0101] First update information;

[0102] Second update information in addition to the first update information.

[0103] It is understood that after the second communication device receives the above-mentioned model update request information, the second communication device may send the model update request information to the first communication device to request the first update information, and / or send the second update information other than the first update information to the first communication device.

[0104] For example, assuming that the first update information indicated by the model update request information is the update information of the model hyperparameters, then the second communication device can send the update information of the model hyperparameters to the first communication device, that is, the above-mentioned model update information includes the first update information; and / or, the second communication device sends at least one of the update information of the kernel function, the update information of the prediction mode, and the update information of the calculation mode to the first communication device, that is, the above-mentioned model update information includes the second update information other than the first update information.

[0105] Optionally, in the embodiments of this application, the above-mentioned model update request information may also indicate the conditions or events that trigger the time series prediction model update.

[0106] Optionally, in this embodiment of the application, the first communication device may periodically or non-periodically send model update requests to the second communication device to obtain model update information.

[0107] In some embodiments, step 204 can be implemented by step 204a as described below.

[0108] Step 204a: The first communication device sends a model update request to the second communication device based on the first information.

[0109] The aforementioned first information may include at least one of the following:

[0110] Prediction error of time series forecasting models;

[0111] Statistical information on the prediction error of time series forecasting models;

[0112] Mobility information of the first communication device;

[0113] Statistical information on the prediction results of time series forecasting models;

[0114] Environmental perception information.

[0115] It should be noted that, in the embodiments of this application, the explanations and descriptions of the mobility information of the first communication device, the prediction task, the statistical information of the prediction results, and the environmental perception information can be found in the detailed descriptions in the above embodiments. To avoid repetition, they will not be repeated here.

[0116] For example, when the moving speed of the first communication device is greater than a first preset threshold, the first communication device may send a model update request to the second communication device; and / or, when the beam switching frequency of the first communication device is greater than a second preset threshold, the first communication device may send a model update request to the second communication device; and / or, when the first communication device completes cell handover, the first communication device may send a model update request to the second communication device.

[0117] For example, assuming the above prediction task is to predict location, and the statistical information of the prediction result is the variance of the predicted location, then if the variance of the predicted location by the first communication device is greater than the third preset threshold, the first communication device can send a model update request to the second communication device.

[0118] It is understandable that step 204a above is an update of the conditionally triggered time series prediction model.

[0119] In some other embodiments, step 204 described above can be implemented by step 204b described below.

[0120] Step 204b: If the first condition is met, the first communication device sends a model update request to the second communication device.

[0121] The first condition mentioned above may include at least one of the following:

[0122] The first communication device receives the model update instruction information sent by the second communication device;

[0123] Update the forecast task of the time series forecasting model;

[0124] The forecasting performance requirements of time series forecasting models need to be updated.

[0125] It should be noted that, in the embodiments of this application, the above-mentioned model update indication information can be sent by the second communication device to the first communication device non-periodically (burstingly). In this way, after the first communication device receives the model update indication information, the first communication device can send model update request information to the second communication device, thereby realizing non-periodic updates of the time series prediction model.

[0126] It is understandable that step 204b above is an update of the event-triggered time series prediction model.

[0127] Optionally, the model update method provided in this application embodiment may further include the following steps 206 and 207.

[0128] Step 206: The first communication device sends the first capability information to the second communication device.

[0129] Step 207: The second communication device receives the first capability information from the first communication device.

[0130] The aforementioned first capability information can be used to indicate at least one of the following:

[0131] The kernel functions supported by the first communication device;

[0132] Does the first communication device have a prediction task?

[0133] The model optimizer configuration supported by the first communication device;

[0134] Model training configurations supported by the first communication device;

[0135] The first communication device supports the model prediction configuration.

[0136] It is understood that the aforementioned first capability information can be used to indicate the first communication device's ability to update the configuration of its time series prediction model.

[0137] In this embodiment, the aforementioned model optimizer configuration refers to the configuration of the optimizer used to update the time series prediction model. Optionally, the optimizer configuration may include at least one of the optimizer type and optimizer state. Of course, in actual implementation, the optimizer configuration may also include other configuration information of the optimizer, which can be determined according to actual usage requirements, and this embodiment does not limit it.

[0138] In this embodiment, the model training configuration may include the storage capacity and computing power of the first communication device, and the second communication device may configure corresponding model update information for the first communication device according to the model training configuration. For example, assuming the first communication device has weak storage and / or computing power, it may not support a long training window size.

[0139] Optionally, in one embodiment, the second communication device may configure a time series prediction model to the first communication device. After the first communication device receives the configuration information of the time series prediction model, the first communication device may provide feedback to the second communication device on whether the first communication device supports the time series prediction model.

[0140] Based on this, the model update method provided in the embodiments of this application may further include the following steps 208-211.

[0141] Step 208: The second communication device sends the first model configuration to the first communication device.

[0142] Step 209: The first communication device receives the first model configuration from the second communication device.

[0143] Step 210: The first communication device sends the first feedback information to the second communication device.

[0144] Step 211: The second communication device receives the first feedback information from the first communication device.

[0145] The aforementioned first feedback information can be used to indicate whether the first communication device supports the time series prediction model indicated by the first model configuration.

[0146] For example, assuming the first communication device is a terminal and the second communication device is a network-side device, the network-side device can first configure the time series prediction model to the terminal, and then the terminal can report to the network-side device whether it supports the time series prediction model configured by the network-side device.

[0147] Alternatively, in another implementation, the first communication device may first report its computing power to the second communication device, and then the second communication device may configure a time series prediction model for the first communication device based on the computing power.

[0148] Based on this, the model update method provided in the embodiments of this application may further include the following steps 212-215.

[0149] Step 212: The first communication device sends computing capability information to the second communication device.

[0150] The computing power information can be used to indicate the computing power of the first communication device.

[0151] Step 213: The second communication device receives computing capability information from the first communication device.

[0152] Step 214: The second communication device sends the second model configuration to the first communication device.

[0153] Step 215: The first communication device receives the second model configuration from the second communication device.

[0154] The time series prediction model indicated by the second model configuration can be determined based on at least one of the computing power, processing latency, and prediction performance requirements of the first communication device.

[0155] For example, assuming the first communication device is a terminal and the second communication device is a network-side device, the terminal can first report its computing power to the network-side device. After receiving the computing power information, the network-side device can determine how to configure the time series prediction model for the terminal based on at least one of the terminal's computing power, processing latency, and the prediction performance requirements of the terminal's time series prediction model. Thus, the network-side device can send the second model configuration to the terminal to configure the time series prediction model for the terminal.

[0156] Optionally, in this embodiment, the processing latency may include, but is not limited to: model loading latency, data reading latency, model calculation latency, etc., which can be determined according to actual usage requirements.

[0157] Optionally, the model update method provided in this application embodiment may further include the following steps 216 and 217.

[0158] Step 216: The first communication device sends the second information or the updated recommendation information of the time series prediction model to the second communication device.

[0159] Step 217: The second communication device receives the second information or the updated recommendation information of the time series prediction model from the first communication device.

[0160] The model update information mentioned above is determined based on the second piece of information mentioned above.

[0161] Optionally, in this embodiment, the aforementioned update recommendation information may include at least one of the following: kernel function recommendation information, model hyperparameter recommendation information, prediction mode recommendation information, and computation mode recommendation information. Thus, after the second communication device receives the update recommendation information, it can send corresponding model update information to the first communication device based on the update recommendation information.

[0162] Of course, in actual implementation, the above-mentioned updated recommendation information can also include any other possible recommendation information, which can be determined according to actual usage needs.

[0163] Optionally, in embodiments of this application, the second information may include at least one of the following:

[0164] Statistical information on the prediction results of time series forecasting models;

[0165] Estimation error of time series forecasting models;

[0166] Statistical information on the estimation error of the time series prediction model;

[0167] The prediction error of a time series forecasting model;

[0168] Statistical information on the prediction error of the time series prediction model;

[0169] Mobility information of the first communication device;

[0170] Noise statistics;

[0171] Predictive performance requirements of time series forecasting models.

[0172] In this embodiment, the statistical information of the estimation error of the time series prediction model can be the mean and / or variance of the estimation error of the time series prediction model. Since GP predicts unknown information based on known information, the estimation error of the time series prediction model can be the estimation error of known information (historical information), such as channel state information (CSI) prediction. The CSI of the past N time units is estimated based on pilot (sounding reference signal (SRS)) estimation, and the estimation has errors.

[0173] In this embodiment of the application, the statistical information of the model prediction error can be at least one of the mean of the model prediction error and the variance of the model prediction error, which can be determined according to the actual use requirements.

[0174] Optionally, in this embodiment, the statistical information of the noise may include the signal-to-noise ratio and the signal-to-dryness ratio. The specific details can be determined based on actual usage requirements, and this embodiment does not impose any limitations.

[0175] Optionally, in this embodiment of the application, if the model update information includes the update information of the calculation mode, before step 203 above, the model update method provided in this embodiment of the application may further include the following steps 218 and 219.

[0176] Step 218: The first communication device sends the second capability information to the second communication device.

[0177] Step 219: The second communication device receives the second capability information from the first communication device.

[0178] The aforementioned second capability information may include at least one of the following:

[0179] The computing power information of the first communication device;

[0180] Storage capacity information of the first communication device;

[0181] Configuration information of the computing unit of the first communication device.

[0182] It is understood that the aforementioned second capability information can be used to indicate the hardware capabilities of the first communication device.

[0183] Optionally, in this embodiment of the application, the second capability information may further include at least one of the following:

[0184] The first communication device supports the maximum amount of data to be cached, for example, the maximum number of time units of CSI that can be cached;

[0185] The maximum amount of computation that the first communication device can support, such as how many time units of data are contained in the training window of a time series prediction model;

[0186] The first communication device supports the maximum number of parallel computing threads, for example, how many sub-models in a time series prediction model can be trained simultaneously.

[0187] In this embodiment of the application, the calculation mode may include the following calculation modes:

[0188] a) Serial computing mode, that is, the GP model (i.e., time series prediction model) is trained in the same computing unit, such as in a certain core or thread of the CPU.

[0189] b) Parallel computing mode, that is, the training of GP model (i.e., time series prediction model) is carried out simultaneously in multiple computing units, such as multiple CPU cores or multiple CPU threads.

[0190] Optionally, when the above model update information indicates that the computation mode of the time series prediction model should be updated to a parallel computation mode, the model update method provided in this application embodiment may further include the following steps 220 and 221.

[0191] Step 220: The second communication device sends the third information to the first communication device.

[0192] Step 221: The first communication device receives the third information from the second communication device.

[0193] The aforementioned third information may include at least one of the following:

[0194] The dataset can be partitioned in a way that, for example, divides the dataset into N parts and trains N sub-models of the time series prediction model, with the training window length of each sub-model becoming 1 / N of the original. This reduces computational complexity; the computational complexity of GP is O(n^2). 3 (where n is the number of samples in the training window. Therefore, when the number of samples in the dataset is large, the computational complexity of GP is extremely high. By dividing the dataset into multiple sub-models for simultaneous training through parallel computing, the computational complexity can be significantly reduced.)

[0195] The choice of optimization objective, such as a weighted sum of the marginal functions of multiple models;

[0196] Selection of model optimizer;

[0197] The initial state of the model optimizer;

[0198] The fusion method of prediction results from multiple sub-models, such as the weighted sum of prediction results from multiple sub-models.

[0199] It should be noted that, in this embodiment of the application, in order to reduce the amount of computation, the dataset can be divided into N parts to train N sub-models of the time series prediction model. During the training process, these N sub-models are trained together and share a common optimization objective or loss function. It can be understood that the process of optimizing the optimization objective is the process of training these N models simultaneously.

[0200] The model update method provided in this application embodiment uses kernel function, model hyperparameters, prediction mode, and calculation mode as parameter information of time series prediction model. Therefore, after the first communication device obtains the model update information, the first communication device can update the time series prediction model used by the first communication device according to the model update information, so that the time series prediction model can perform prediction tasks more accurately, thereby enabling the first communication device to obtain good communication performance.

[0201] The model update method provided in this application can be executed by a model update device. This application uses an example of a model update device executing the model update method to illustrate the model update device provided in this application.

[0202] like Figure 8As shown, this application embodiment provides a model update device 300, which may include an acquisition module 301 and an update module 302. The acquisition module 301 is used to acquire model update information; the update module 302 is used to update the time series prediction model used by the first communication device according to the model update information. The time series prediction model is used to predict the execution of a task. The model update information includes at least one of the following:

[0203] Kernel function update information;

[0204] Update information for model hyperparameters;

[0205] Update information for the prediction model;

[0206] Information on the update of the computing model.

[0207] Optionally, the acquisition module may include a receiving submodule, which can be used to receive model update information from the second communication device.

[0208] Optionally, the model update device may further include a sending module. The sending module can be used to send model update request information to the second communication device, the model update request information being used to request the acquisition of first update information; wherein the model update information includes at least one of the following:

[0209] First update information;

[0210] Second update information in addition to the first update information.

[0211] Optionally, the sending module can be specifically used to send model update request information to the second communication device based on the first information; wherein the first information includes at least one of the following:

[0212] Prediction error of time series forecasting models;

[0213] Statistical information on the prediction error of time series forecasting models;

[0214] Mobility information of the first communication device;

[0215] Statistical information on the prediction results of time series forecasting models;

[0216] Environmental perception information.

[0217] Optionally, the sending module can be specifically used to send model update request information to the second communication device when the first condition is met;

[0218] The first condition includes at least one of the following:

[0219] The first communication device receives the model update instruction information sent by the second communication device;

[0220] Update the forecast task of the time series forecasting model;

[0221] The prediction performance requirements of the time series prediction model configured for the first communication device are updated.

[0222] Optionally, the prediction performance requirements include at least one of prediction accuracy and computation latency.

[0223] Optionally, the model update apparatus may further include a transmitting module. This transmitting module can be used to transmit first capability information to a second communication device, the first capability information indicating at least one of the following:

[0224] The kernel functions supported by the first communication device;

[0225] Does the first communication device have a prediction task?

[0226] The model optimizer configuration supported by the first communication device;

[0227] Model training configurations supported by the first communication device;

[0228] The first communication device supports the model prediction configuration.

[0229] Optionally, the model update apparatus may further include a receiving module and a sending module. The receiving module may be used to receive a first model configuration from a second communication device; the sending module may be used to send first feedback information to the second communication device, the first feedback information being used to indicate whether the first communication device supports the time series prediction model indicated by the first model configuration.

[0230] Optionally, the model update apparatus may further include a sending module and a receiving module. The sending module may be used to send computing power information to the second communication device, the computing power information indicating the computing power of the first communication device; the receiving module may be used to receive a second model configuration from the second communication device, the time series prediction model indicated by the second model configuration being determined based on at least one of computing power, processing latency, and the prediction performance requirements of the time series prediction model.

[0231] Optionally, the model update device may further include a sending module. This sending module can be used to send second information or updated recommendation information of the time series prediction model to the second communication device; wherein the model update information is determined based on the second information.

[0232] Optionally, the second information includes at least one of the following:

[0233] Statistical information on the prediction results of time series forecasting models;

[0234] Estimation error of time series forecasting models;

[0235] Statistical information on the estimation error of time series forecasting models;

[0236] The prediction error of a time series forecasting model;

[0237] Statistical information on the prediction error of time series forecasting models;

[0238] Mobility information of the first communication device;

[0239] Noise statistics;

[0240] Predictive performance requirements of time series forecasting models.

[0241] Optionally, the prediction model includes the structure of the time window for the prediction task, and the structure of the time window includes at least one of the following:

[0242] The length of the time unit in the training window of a time series prediction model;

[0243] The length of the time unit of the prediction window in a time series forecasting model;

[0244] The interval between adjacent time units in the training window of a time series prediction model;

[0245] The interval between adjacent time units in the prediction window of a time series forecasting model.

[0246] Optionally, the first communication device stores the same list of prediction models as the second communication device, and the list of prediction models includes identifiers of prediction modes.

[0247] Optionally, the model update information includes update information of the computation mode; the model update device may further include a sending module. This sending module can be used to send second capability information to the second communication device before the receiving submodule receives the model update information from the second communication device. The second capability information includes at least one of the following:

[0248] The computing power information of the first communication device;

[0249] Storage capacity information of the first communication device;

[0250] Configuration information of the computing unit of the first communication device.

[0251] Optionally, the second capability information may also include at least one of the following:

[0252] The maximum amount of data that the first communication device can buffer;

[0253] The maximum computational load supported by the first communication device;

[0254] The first communication device supports the maximum number of parallel computing threads.

[0255] Optionally, the computation mode update information indicates that the computation mode of the time series forecasting model should be updated to a parallel computation mode. Optionally, the model update device may further include a receiving module. This receiving module can be used to receive third information from the second communication device, the third information including at least one of the following:

[0256] The method of dividing the dataset;

[0257] Optimize the selection of objectives;

[0258] Selection of model optimizer;

[0259] The initial state of the model optimizer;

[0260] The method of fusing prediction results from multiple sub-models.

[0261] Optionally, the model update information may also include at least one of the following:

[0262] Identifier of prediction tasks pending update;

[0263] Identifier of the model to be updated;

[0264] The timestamp information of the model to be updated.

[0265] The model update device provided in this application embodiment uses kernel function, model hyperparameters, prediction mode and calculation mode as parameter information of time series prediction model. Therefore, after the model update device obtains the model update information, it can update the time series prediction model used by the first communication device according to the model update information, so that the time series prediction model can perform prediction tasks more accurately, thereby enabling the first communication device to obtain good communication performance.

[0266] like Figure 9 As shown, this application embodiment provides a model update device 400, which may include a sending module 401. The sending module 401 can be used to send model update information to a first communication device. The model update information is used to update the time series prediction model used by the first communication device, and the time series prediction model is used to predict the execution of a task. The model update information includes at least one of the following:

[0267] Kernel function update information;

[0268] Update information for model hyperparameters;

[0269] Update information for the prediction model;

[0270] Information on the update of the computing model.

[0271] Optionally, the sending module can be specifically used to send model update information to the first communication device based on the fourth information; wherein the fourth information includes at least one of the following:

[0272] Prediction error of time series forecasting models;

[0273] Statistical information on the prediction error of time series forecasting models;

[0274] Mobility information of the first communication device;

[0275] Statistical information on the prediction results of time series forecasting models;

[0276] Update information for the prediction task of the time series forecasting model;

[0277] Updated information on the forecasting performance requirements of time series forecasting models;

[0278] Environmental perception information.

[0279] Optionally, the model update information may further include a receiving module. This receiving module can be used to receive model update request information from the first communication device, the model update request information being used to request the acquisition of first update information; wherein the model update information includes at least one of the following:

[0280] First update information;

[0281] Second update information in addition to the first update information.

[0282] Optionally, the model update information may further include a receiving module. This receiving module can be used to receive first capability information from a first communication device, the first capability information indicating at least one of the following:

[0283] The kernel functions supported by the first communication device;

[0284] Does the first communication device have a prediction task?

[0285] The model optimizer configuration supported by the first communication device;

[0286] Model training configurations supported by the first communication device;

[0287] The first communication device supports the model prediction configuration.

[0288] Optionally, the model update information may further include a receiving module. The sending module may also be used to send a first model configuration to the first communication device; the receiving module may be used to receive first feedback information from the first communication device, the first feedback information being used to indicate whether the first communication device supports the time series prediction model indicated by the first model configuration.

[0289] Optionally, the model update information may further include a receiving module. This receiving module may be used to receive computing power information from the first communication device, the computing power information indicating the computing power of the first communication device; the sending module may also be used to send a second model configuration to the first communication device, the second model configuration indicating a time series prediction model determined based on at least one of computing power, processing latency, and the prediction performance requirements of the time series prediction model.

[0290] Optionally, the model update information may further include a receiving module. This receiving module can be used to receive second information or updated recommendation information of the time series prediction model from the first communication device; wherein the model update information is determined based on the second information.

[0291] Optionally, the second information may include at least one of the following:

[0292] Statistical information on the prediction results of time series forecasting models;

[0293] Estimation error of time series forecasting models;

[0294] Statistical information on the estimation error of time series forecasting models;

[0295] The prediction error of a time series forecasting model;

[0296] Statistical information on the prediction error of time series forecasting models;

[0297] Mobility information of the first communication device;

[0298] Noise statistics;

[0299] Predictive performance requirements of time series forecasting models.

[0300] Optionally, the prediction model includes the structure of the time window for the prediction task, and the structure of the time window includes at least one of the following:

[0301] The length of the time unit in the training window of a time series prediction model;

[0302] The length of the time unit of the prediction window in a time series forecasting model;

[0303] The interval between adjacent time units in the training window of a time series prediction model;

[0304] The interval between adjacent time units in the prediction window of a time series forecasting model.

[0305] Optionally, the second communication device stores the same list of prediction models as the first communication device, and the list of prediction models includes identifiers of prediction modes.

[0306] Optionally, the model update information includes update information of the computation mode; the model update information may also include a receiving module. This receiving module can be used to receive second capability information from the first communication device before the sending module sends the model update information to the first communication device. The second capability information includes at least one of the following:

[0307] The computing power information of the first communication device;

[0308] Storage capacity information of the first communication device;

[0309] Configuration information of the computing unit of the first communication device.

[0310] Optionally, the second capability information may also include at least one of the following:

[0311] The maximum amount of data that the first communication device can buffer;

[0312] The maximum computational load supported by the first communication device;

[0313] The first communication device supports the maximum number of parallel computing threads.

[0314] Optionally, the computation mode update information indicates that the computation mode of the time series prediction model should be updated to a parallel computation mode. The sending module is also used to send third information to the first communication device, the third information including at least one of the following:

[0315] The method of dividing the dataset;

[0316] Optimize the selection of objectives;

[0317] Selection of model optimizer;

[0318] The initial state of the model optimizer;

[0319] The method of fusing prediction results from multiple sub-models.

[0320] Optionally, the model update information may also include at least one of the following:

[0321] Identifier of prediction tasks pending update;

[0322] Identifier of the model to be updated;

[0323] The timestamp information of the model to be updated.

[0324] In this embodiment, since the kernel function, model hyperparameters, prediction mode, and calculation mode are all parameter information of the time series prediction model, the model update device can send model update information to the first communication device, so that the first communication device can update the time series prediction model used by the first communication device according to the model update information, thereby enabling the time series prediction model to perform prediction tasks more accurately, and thus enabling the first communication device to obtain good communication performance.

[0325] In this embodiment, the model update device can be a communication device, such as a communication device with an operating system, or a component within the communication device, such as an integrated circuit or a chip. The communication device can be a terminal, or other devices besides a terminal, such as network-side devices. Exemplarily, the terminal can include, but is not limited to, the types of terminals 11 listed above. Other devices can be core network devices or access network devices, servers, network attached storage (NAS), etc., and this embodiment does not impose specific limitations.

[0326] The model update apparatus provided in this application embodiment can implement all the processes implemented in the above-described model update method embodiments and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0327] Optionally, such as Figure 10 As shown in the illustration, this application also provides a communication device 500, including a processor 501 and a memory 502. The memory 502 stores programs or instructions that can run on the processor 501. For example, when the communication device 500 is a terminal, the program or instructions executed by the processor 501 implement the various steps of the above-described model update method embodiment and achieve the same technical effect. When the communication device 500 is a network-side device, the program or instructions executed by the processor 501 implement the various steps of the above-described model update method embodiment and achieve the same technical effect. To avoid repetition, further details are omitted here.

[0328] This application also provides a communication device, including a processor and a communication interface. When the communication device is a first communication device, the communication interface is used to acquire model update information; the processor updates the time series prediction model used by the first communication device according to the model update information; or, when the communication device is a second communication device, the communication interface is used to send model update information to the first communication device. The model update information is used to update the time series prediction model used by the first communication device. The time series prediction model is used for the execution of prediction tasks. The model update information includes at least one of the following: kernel function update information; model hyperparameter update information; prediction mode update information; and computation mode update information. This communication device embodiment corresponds to the above method embodiment. All implementation processes and methods of the above method embodiments can be applied to this communication device embodiment and can achieve the same technical effects.

[0329] This application also provides a communication device. For example... Figure 11 As shown, the communication device 600 includes: an antenna 61, a radio frequency (RF) device 62, a baseband device 63, a processor 64, and a memory 65. The antenna 61 is connected to the RF device 62. In the uplink direction, the RF device 62 receives information through the antenna 61 and transmits the received information to the baseband device 63 for processing. In the downlink direction, the baseband device 63 processes the information to be transmitted and sends it to the RF device 62. The RF device 62 processes the received information and transmits it through the antenna 61.

[0330] The method executed by the communication device (including the first communication device and the second communication device) in the above embodiments can be implemented in the baseband device 63, which includes a baseband processor.

[0331] Baseband device 63 may include, for example, at least one baseband board on which multiple chips are disposed, such as Figure 11 As shown, one of the chips is, for example, a baseband processor, which is connected to the memory 65 via a bus interface to call the program in the memory 65 and execute the network-side device operations shown in the above method embodiment.

[0332] The communication device may also include a network interface 66, such as a common public radio interface (CPRI).

[0333] Specifically, the communication device 600 of this embodiment further includes: instructions or programs stored in a memory 65 and executable on a processor 64, wherein the processor 64 calls the instructions or programs in the memory 65 to execute. Figure 8 or Figure 9 The methods executed by each module shown achieve the same technical effect, and to avoid repetition, they will not be described in detail here.

[0334] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described model update method embodiments and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0335] The processor is the processor in the terminal described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0336] This application also provides a chip, which includes a processor and a communication interface. The communication interface and the processor are coupled. The processor is used to run programs or instructions to implement the various processes of the above-described model update method embodiments and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0337] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0338] This application also provides a computer program / program product, which is stored in a storage medium and executed by at least one processor to implement the various processes of the above-described model update method embodiments, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0339] This application also provides a communication system, including a terminal and a network-side device. The terminal can be used to execute the steps of the above-described model update method, and the network-side device can be used to execute the steps of the above-described model update method.

[0340] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0341] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network-side device, etc.) to execute the methods of the various embodiments of this application.

[0342] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A model update method, characterized in that, include: The first communication device sends a model update request to the second communication device; The first communication device obtains model update information from the second communication device; The first communication device updates the time series prediction model used by the first communication device according to the model update information. The time series prediction model is used to predict the execution of the task. The model update information includes at least one of the following: Update information for the prediction model; Identifier of the model to be updated; The timestamp information of the model to be updated; The prediction mode includes a time window structure for the prediction task, and the time window structure includes at least one of the following: The length of the time unit of the training window for the time series prediction model; The length of the time unit of the prediction window in the time series prediction model; The interval between adjacent time units in the training window of the time series prediction model; The interval between adjacent time units of the prediction window of the time series prediction model.

2. The method according to claim 1, characterized in that, The model update information also includes at least one of the following: Kernel function update information; Update information for model hyperparameters; Information on the update of the computing model.

3. The method according to claim 1, characterized in that, The first communication device obtains model update information from the second communication device, including: The first communication device receives the model update information from the second communication device.

4. The method according to claim 1, characterized in that, The model update request information is used to request the first update information; The model update information includes at least one of the following: First update information; A second update in addition to the first update information.

5. The method according to claim 1, characterized in that, The first communication device sends a model update request to the second communication device, including: The first communication device sends the model update request information to the second communication device based on the first information; The first information includes at least one of the following: The prediction error of the time series prediction model; Statistical information on the prediction error of the time series prediction model; Mobility information of the first communication device; Statistical information of the prediction results of the time series prediction model; Environmental perception information.

6. The method according to claim 1, characterized in that, The first communication device sends a model update request to the second communication device, including: If the first condition is met, the first communication device sends the model update request information to the second communication device; The first condition includes at least one of the following: The first communication device receives model update indication information sent by the second communication device; The prediction task of the time series prediction model is updated; The prediction performance of the time series prediction model needs to be updated.

7. The method according to claim 6, characterized in that, The prediction performance requirements include at least one of prediction accuracy and computation latency.

8. The method according to claim 3, characterized in that, The method further includes: The first communication device sends first capability information to the second communication device, the first capability information indicating at least one of the following: The kernel functions supported by the first communication device; Does the first communication device have a prediction task? The model optimizer configurations supported by the first communication device; The model training configurations supported by the first communication device; The model prediction configuration supported by the first communication device.

9. The method according to claim 3, characterized in that, The method further includes: The first communication device receives the first model configuration from the second communication device; The first communication device sends a first feedback message to the second communication device, the first feedback message being used to indicate whether the first communication device supports the time series prediction model indicated by the first model configuration.

10. The method according to claim 3, characterized in that, The method further includes: The first communication device sends computing power information to the second communication device, the computing power information being used to indicate the computing power of the first communication device; The first communication device receives a second model configuration from the second communication device, wherein the time series prediction model indicated by the second model configuration is determined based on at least one of the computing power, processing latency, and prediction performance requirements of the time series prediction model.

11. The method according to claim 3, characterized in that, The method further includes: The first communication device sends second information or updated recommendation information of the time series prediction model to the second communication device; The model update information is determined based on the second information.

12. The method according to claim 11, characterized in that, The second information includes at least one of the following: Statistical information of the prediction results of the time series prediction model; The estimation error of the time series prediction model; Statistical information on the estimation error of the time series prediction model; The prediction error of the time series prediction model; Statistical information on the prediction error of the time series prediction model; Mobility information of the first communication device; Noise statistics; The prediction performance requirements of the time series prediction model.

13. The method according to claim 3, characterized in that, The first communication device stores the same list of prediction models as the second communication device, and the list of prediction models includes identifiers of prediction modes.

14. The method according to claim 3, characterized in that, The model update information includes update information for the calculation mode; Before the first communication device receives the model update information from the second communication device, the method further includes: The first communication device sends second capability information to the second communication device, the second capability information including at least one of the following: The computing power information of the first communication device; Storage capacity information of the first communication device; The computing unit configuration information of the first communication device.

15. The method according to claim 14, characterized in that, The second capability information also includes at least one of the following: The first communication device supports a maximum amount of data to be cached; The first communication device supports the maximum amount of computation. The first communication device supports a maximum number of parallel computing threads.

16. The method according to claim 2, characterized in that, The update information of the computation mode indicates that the computation mode of the time series prediction model should be updated to a parallel computation mode, and the method further includes: The first communication device receives third information from the second communication device, the third information including at least one of the following: The method of dividing the dataset; Optimize the selection of objectives; Selection of model optimizer; The initial state of the model optimizer; The method of fusing prediction results from multiple sub-models.

17. The method according to claim 1, characterized in that, The model update information also includes: the identifier of the prediction task to be updated.

18. A model update method, characterized in that, include: The second communication device receives model update request information from the first communication device; The second communication device sends model update information to the first communication device. The model update information is used to update the time series prediction model used by the first communication device. The time series prediction model is used to predict the execution of the task. The model update information includes at least one of the following: Update information for the prediction model; Identifier of the model to be updated; The timestamp information of the model to be updated; The prediction mode includes a time window structure for the prediction task, and the time window structure includes at least one of the following: The length of the time unit of the training window for the time series prediction model; The length of the time unit of the prediction window in the time series prediction model; The interval between adjacent time units in the training window of the time series prediction model; The interval between adjacent time units of the prediction window of the time series prediction model.

19. The method according to claim 18, characterized in that, The model update information also includes at least one of the following: Kernel function update information; Update information for model hyperparameters; Information on the update of the computing model.

20. The method according to claim 18, characterized in that, The second communication device sends model update information to the first communication device, including: The second communication device sends the model update information to the first communication device according to the fourth information; The fourth piece of information includes at least one of the following: The prediction error of the time series prediction model; Statistical information on the prediction error of the time series prediction model; Mobility information of the first communication device; Statistical information of the prediction results of the time series prediction model; Update information for the prediction task of the time series prediction model; Updated information on the prediction performance requirements of the time series prediction model; Environmental perception information.

21. The method according to claim 18, characterized in that, The model update request information is used to request the first update information; The model update information includes at least one of the following: First update information; A second update in addition to the first update information.

22. The method according to claim 18, characterized in that, The method further includes: The second communication device receives first capability information from the first communication device, the first capability information being used to indicate at least one of the following: The kernel functions supported by the first communication device; Does the first communication device have a prediction task? The model optimizer configurations supported by the first communication device; The model training configurations supported by the first communication device; The model prediction configuration supported by the first communication device.

23. The method according to claim 18, characterized in that, The method further includes: The second communication device sends the first model configuration to the first communication device; The second communication device receives first feedback information from the first communication device, the first feedback information being used to indicate whether the first communication device supports the time series prediction model indicated by the first model configuration.

24. The method according to claim 18, characterized in that, The method further includes: The second communication device receives computing power information from the first communication device, the computing power information being used to indicate the computing power of the first communication device; The second communication device sends a second model configuration to the first communication device. The time series prediction model indicated by the second model configuration is determined based on at least one of computing power, processing latency, and prediction performance requirements of the time series prediction model.

25. The method according to claim 18, characterized in that, The method further includes: The second communication device receives the second information or the updated recommendation information of the time series prediction model from the first communication device; The model update information is determined based on the second information.

26. The method according to claim 25, characterized in that, The second information includes at least one of the following: Statistical information of the prediction results of the time series prediction model; The estimation error of the time series prediction model; Statistical information on the estimation error of the time series prediction model; The prediction error of the time series prediction model; Statistical information on the prediction error of the time series prediction model; Mobility information of the first communication device; Noise statistics; The prediction performance requirements of the time series prediction model.

27. The method according to claim 18, characterized in that, The second communication device stores the same list of prediction models as the first communication device, and the list of prediction models includes identifiers of prediction modes.

28. The method according to claim 18, characterized in that, The model update information includes update information for the calculation mode; Before the second communication device sends model update information to the first communication device, the method further includes: The second communication device receives second capability information from the first communication device, the second capability information including at least one of the following: The computing power information of the first communication device; Storage capacity information of the first communication device; The computing unit configuration information of the first communication device.

29. The method according to claim 28, characterized in that, The second capability information also includes at least one of the following: The first communication device supports a maximum amount of data to be cached; The first communication device supports the maximum amount of computation. The first communication device supports a maximum number of parallel computing threads.

30. The method according to claim 19, characterized in that, The update information of the computation mode indicates that the computation mode of the time series prediction model should be updated to a parallel computation mode, and the method further includes: The second communication device sends third information to the first communication device, the third information including at least one of the following: The method of dividing the dataset; Optimize the selection of objectives; Selection of model optimizer; The initial state of the model optimizer; The method of fusing prediction results from multiple sub-models.

31. The method according to claim 18, characterized in that, The model update information also includes: the identifier of the prediction task to be updated.

32. A model update device, characterized in that, The model update device includes: a sending module, an acquisition module, and an update module; The sending module is used to send model update request information to the second communication device; The acquisition module is used to acquire model update information from the second communication device; The update module is used to update the time series prediction model used by the first communication device according to the model update information. The time series prediction model is used to predict the execution of the task. The model update information includes at least one of the following: Update information for the prediction model; Identifier of the model to be updated; The timestamp information of the model to be updated; The prediction mode includes a time window structure for the prediction task, and the time window structure includes at least one of the following: The length of the time unit of the training window for the time series prediction model; The length of the time unit of the prediction window in the time series prediction model; The interval between adjacent time units in the training window of the time series prediction model; The interval between adjacent time units of the prediction window of the time series prediction model.

33. A model update device, characterized in that, The model update device includes: a receiving module and a sending module; The receiving module is used to receive model update request information from the first communication device; The sending module is used to send model update information to the first communication device. The model update information is used to update the time series prediction model used by the first communication device. The time series prediction model is used to predict the execution of the task. The model update information includes at least one of the following: Update information for the prediction model; Identifier of the model to be updated; The timestamp information of the model to be updated; The prediction mode includes a time window structure for the prediction task, and the time window structure includes at least one of the following: The length of the time unit of the training window for the time series prediction model; The length of the time unit of the prediction window in the time series prediction model; The interval between adjacent time units in the training window of the time series prediction model; The interval between adjacent time units of the prediction window of the time series prediction model.

34. A communication device, characterized in that, It includes a processor and a memory, the memory storing a program or instructions that can run on the processor, the program or instructions being executed by the processor to implement the steps of the model update method as described in any one of claims 1 to 31.

35. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the model update method as described in any one of claims 1-31.

Images