Training data acquisition method and related apparatus

Abstract

The present application is applied to the field of artificial intelligence. Provided are a training data acquisition method and a related apparatus. In the technical solution provided in the present application, an input data sequence for training a neural network model is acquired on the basis of a plurality of data sequences, wherein a plurality of pieces of data in the input data sequence are sequentially input into the neural network model. The technical solution provided in the present application enables acquired data to improve the accuracy of neural network models obtained by means of training.

Description

Method and device for obtaining training data

[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on April 3, 2024, with application number 202410407969.4 and application name “Training Data Acquisition Method and Related Device”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present application relates to the field of artificial intelligence, and in particular to methods and devices for acquiring training data. Background Art

[0003] With the development of artificial intelligence research, the application of neural network models continues to expand. Neural network models usually need to be trained to have corresponding functions.

[0004] The training method for a neural network model is as follows: a data acquisition device collects training data to form a data set, and a training device obtains the input data of the neural network model and the label of the input data from the data set; the input data is input into the neural network model to obtain the true output of the neural network model, and the difference between the true output of the neural network model and the label of the input data is compared. If the difference between the true output and the label is greater than a threshold, the model parameters of the neural network model are adjusted until the difference between the true output of the neural network model and the label is less than the threshold. The trained neural network model can learn the correlation between the input data and the label. In this way, the neural network model can process the input data based on this correlation to obtain the corresponding output. However, the accuracy of the output of the neural network model trained by the above method often does not meet the requirements. Summary of the Invention

[0005] The present application provides a method and related apparatus for acquiring training data, and the acquired training data helps to improve the accuracy of the output data of the trained neural network model.

[0006] In a first aspect, the present application provides a method for obtaining training data, the method comprising: obtaining T data sequences, each of the T data sequences containing at least one data, where T is an integer greater than 1; determining a first data sequence based on the T data sequences, the first data sequence containing M data, where M is a positive integer greater than 1, the first data sequence being used as an input data sequence of a neural network model and / or as a label of an input data sequence of a neural network model, wherein the input data sequence is used to be sequentially input into the neural network model.

[0007] In this method, because multiple data in the first data sequence are input into the neural network sequentially, and the label corresponding to the first data sequence is the label of the first data sequence as a whole, the neural network model trained based on the first data sequence and the label corresponding to the first data sequence as a whole can take into account or utilize the impact of the association between multiple data in the input data sequence on the relationship between input and output when obtaining the output data sequence based on the input data sequence. Compared to inputting multiple disordered or random data into the neural network model and training the neural network based on the labels associated with each of these multiple data, because the impact of the association between multiple data in the input data sequence on the relationship between input and output is taken into account or utilized, the trained neural network model can learn or extract more dimensional relationships between output and input, and thus can obtain more accurate output based on the input.

[0008] Some implementations of the method further include: obtaining first information, where the first information indicates P data sequences among the T data sequences used to obtain the first data sequence, where P is a positive integer less than or equal to T. Accordingly, determining the first data sequence based on the T data sequences includes: determining the first data sequence based on the first information and the T data sequences.

[0009] In some implementations, the T data sequences include a first sequence, wherein the first information includes first indication information, and the first indication information indicates: a data sequence in the T data sequences that is spliced ​​with the first sequence.

[0010] In this way, based on the first indication information, a data sequence that is combined with the first sequence to obtain the first data sequence can be determined, thereby determining the first data sequence.

[0011] In some implementations, the T data sequences include a first sequence and a second sequence, wherein the first information includes second indication information, and the second indication information indicates whether the second sequence is spliced ​​with the first sequence.

[0012] In this way, based on the second indication information, it can be determined whether the second sequence can be combined with the first sequence to obtain the first data sequence, thereby determining the first data sequence.

[0013] In some implementations, the second sequence may be a sequence that follows the first sequence in transmission order or storage location.

[0014] In some implementations, the first information includes at least one of the following information: a starting sequence identifier, P, of the first data sequence, or an ending sequence identifier of the first data sequence.

[0015] In this implementation, P data sequences used to obtain the first data sequence are determined from T data sequences based on the starting sequence identifier, P, or the ending sequence identifier of the first data sequence, so that the first data sequence can be determined.

[0016] The first information including P can be understood as the number of data sequences included in the first data sequence included in the first information.

[0017] In some implementations, the first information indicating P data sequences includes: the first information indicating characteristics of the P data sequences. In this way, the P data sequences among the T data sequences that meet the characteristics indicated by the first information can be determined as data sequences used to obtain the first data sequence, thereby determining the first data sequence.

[0018] For example, the first information indicates a first sequence length, and P data sequences with a sequence length of the first sequence length are used to obtain the first data sequence.

[0019] For another example, if the first information indicates an odd number, then P data sequences identified as odd numbers are used to obtain the first data sequence.

[0020] For another example, if the first information indicates an even number, then P data sequences identified as even numbers are used to obtain the first data sequence.

[0021] In some implementations, the first information includes identification information of each data sequence in the P data sequences. The identification information of each data sequence may be an index of the data sequence in the T data sequences.

[0022] In this implementation, the P data sequences indicated by the P identification information included in the first information can be determined as data sequences used to obtain the first data sequence, thereby determining the first data sequence.

[0023] In some implementations, the T data sequences include a first sequence, wherein obtaining the T data sequences includes obtaining the first sequence, and obtaining the first sequence includes: obtaining second information, the second information being used to obtain the first sequence, and obtaining the first sequence based on the second information.

[0024] In some implementations, the second information is used to indicate the format of the first sequence. In some scenarios, the first sequence is generated based on the second information, i.e., the second information is used to generate the first sequence; in some scenarios, the first sequence is identified based on the second information, i.e., the second information is used to identify the first sequence.

[0025] When the second information indicates the format of the first sequence, in some implementations, the second information may include at least one of the following: a representation of the first sequence, a start data identifier of the first sequence, a length of the first sequence, or an end data identifier of the first sequence. In this implementation, the second information may be referred to as auxiliary information associated with the first sequence.

[0026] In some implementations, the second information may include the initial value of the model state corresponding to the first sequence. The initial value of the model state corresponding to the first sequence may also be referred to as auxiliary information associated with the first sequence.

[0027] In some implementations, multiple data sequences in the P data sequences are associated with model state initialization values. In this case, as an example, the model state initialization value corresponding to the first data sequence is the model state initialization value corresponding to the first data sequence in the first data sequence that is input into the neural network model.

[0028] The model state initial value corresponding to the first data sequence refers to the model state initial value used when the neural network model obtains the output data sequence corresponding to the first data sequence.

[0029] In some implementations, the method further includes: obtaining first network state information, where the first network state information indicates: an initial value of a model state used when the neural network model obtains an output data sequence corresponding to the first data sequence.

[0030] In some implementations, the data in the first data sequence includes channel state information.

[0031] In some implementations, a neural network model is used for compression and / or prediction of channel state information.

[0032] In some implementations, the first data sequence is also used as a label corresponding to the first data sequence.

[0033] In some implementations, the T data sequences include a third sequence, wherein obtaining the T data sequences includes obtaining the third sequence, and obtaining the third sequence includes: obtaining third information, the third information is used to obtain the third sequence, and obtaining the third sequence according to the third information.

[0034] In some implementations, the third information is used to indicate a format of the third sequence. In some scenarios, the third sequence is generated based on the third information, i.e., the third information is used to generate the third sequence; in some scenarios, the third sequence is identified based on the third information, i.e., the third information is used to identify the third sequence.

[0035] When the third information indicates the format of the third sequence, in some implementations, the third information may include at least one of the following: a representation method of the third sequence, a start data identifier of the third sequence, a length of the third sequence, or an end data identifier of the third sequence. In this implementation, the third information may be referred to as auxiliary information associated with the third sequence.

[0036] In some implementations, the third information may include the initial value of the model state corresponding to the third sequence. The initial value of the model state corresponding to the third sequence may also be referred to as auxiliary information associated with the third sequence.

[0037] In some implementations, at least part of the auxiliary information corresponding to the first sequence and the auxiliary information corresponding to the third sequence is common auxiliary information.

[0038] When the first data sequence is used only as an input data sequence of the neural network model, in some implementations, the method further includes: obtaining S data sequences, each of the S data sequences contains at least one data, and S is an integer greater than 1; determining a second data sequence based on the S data sequences, the second data sequence contains Q data, Q is a positive integer greater than 1, and the second data sequence is used as a label for the first data sequence, wherein multiple data in the first data sequence are used to be sequentially input into the neural network model.

[0039] In some implementations, each data in each data sequence of the S data sequences has an association relationship with at least one data in the T data sequences. A second data sequence can be generated based on the association relationship and the S data sequences.

[0040] In some implementations, the S data sequences include a fourth sequence, wherein obtaining the S data sequences includes obtaining the fourth sequence, and obtaining the fourth sequence includes: obtaining fourth information, the fourth information being used to obtain the fourth sequence, and obtaining the fourth sequence based on the fourth information.

[0041] In some implementations, the fourth information is used to indicate a format of the fourth sequence. In some scenarios, the fourth sequence is generated based on the fourth information, i.e., the fourth information is used to generate the fourth sequence; in some scenarios, the fourth sequence is identified based on the fourth information, i.e., the fourth information is used to identify the fourth sequence.

[0042] When the fourth information indicates the format of the fourth sequence, in some implementations, the fourth information may include at least one of the following: a representation method of the fourth sequence, a starting data identifier of the fourth sequence, a length of the fourth sequence, or an ending data identifier of the fourth sequence. In this implementation, the fourth information may be referred to as auxiliary information associated with the fourth sequence.

[0043] In some implementations, the fourth information may include the initial value of the model state corresponding to the fourth sequence. The initial value of the model state corresponding to the fourth sequence may also be referred to as auxiliary information associated with the fourth sequence.

[0044] In some implementations, the S data sequences include a fifth sequence. Acquiring the S data sequences includes acquiring the fifth sequence, and acquiring the fifth sequence includes: acquiring fifth information, the fifth information being used to acquire the fifth sequence, and acquiring the fifth sequence based on the fifth information.

[0045] In some implementations, the fifth information is used to indicate the format of the fifth sequence. In some scenarios, the fifth sequence is generated based on the fifth information, i.e., the fifth information is used to generate the fifth sequence; in some scenarios, the fifth sequence is identified based on the fifth information, i.e., the fifth information is used to identify the fifth sequence.

[0046] When the fifth information indicates the format of the fifth sequence, in some implementations, the fifth information may include at least one of the following: a representation method of the fifth sequence, a starting data identifier of the fifth sequence, a length of the fifth sequence, or an ending data identifier of the fifth sequence. In this implementation, the fifth information may be referred to as auxiliary information associated with the fifth sequence.

[0047] In some implementations, the fifth information may include the initial value of the model state corresponding to the fifth sequence. The initial value of the model state corresponding to the fifth sequence may also be referred to as auxiliary information associated with the fifth sequence.

[0048] In some implementations, at least part of the auxiliary information corresponding to the fourth sequence and the auxiliary information corresponding to the fifth sequence is common auxiliary information.

[0049] In some implementations, at least part of the auxiliary information corresponding to the first sequence and the auxiliary information corresponding to the fourth sequence is common auxiliary information.

[0050] In a second aspect, the present application provides a device for acquiring training data. The device may include a module corresponding to each of the methods / operations / steps / actions described in the first aspect, and the module may be implemented as a hardware circuit, software, or a combination of hardware circuit and software.

[0051] In one design, the apparatus may include a processing module and a communication module. The communication module is configured to perform the sending and receiving actions described in the method of the first aspect above, and the processing module is configured to perform the actions involving processing (e.g., obtaining and determining) described in the method of the first aspect above.

[0052] In one design, the device may be a terminal, or a device, module, circuit or chip configured in the terminal, or a device that can be used in conjunction with the terminal.

[0053] In one design, the device can be a network-side device, for example, a network device, or a device, module, circuit or chip configured and set in the network device, or a device that can be used in conjunction with the network device, or a device that communicates with the network device, such as a server.

[0054] According to a third aspect, a device is provided, comprising a processor and a storage medium, wherein the storage medium stores instructions, and when the instructions are executed by the processor, the method according to the first aspect or any possible implementation of the first aspect is implemented.

[0055] In a fourth aspect, a device is provided, comprising a processing circuit, wherein the processing circuit is configured to process data and / or information so that the method according to the first aspect or any possible implementation manner of the first aspect is implemented.

[0056] The processing circuit may include one or more processors, or all or part of the circuitry in one or more processors for processing functions.

[0057] Optionally, the device may further include a memory, the memory being used to store programs or instructions, and the processor being used to run the programs or instructions so that the method in the first aspect or any possible implementation manner of the first aspect is implemented.

[0058] Optionally, the device may further include the transceiver circuit, or an input / output interface.

[0059] In a fifth aspect, a chip is provided, comprising a processing circuit, wherein the processing circuit is configured to run a program or instruction so that the method in the first aspect or any possible implementation manner of the first aspect is implemented.

[0060] Optionally, the chip may further include a memory for storing programs or instructions.

[0061] Optionally, the chip may further include a transceiver circuit, or an input / output interface.

[0062] In a sixth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium includes instructions, and when the instructions are executed by a processor, the method in the first aspect or any possible implementation of the first aspect is implemented.

[0063] In a seventh aspect, a computer program product is provided, comprising computer program code or instructions, which, when executed, enables the method in the first aspect or any possible implementation of the first aspect to be implemented.

[0064] In a fifteenth aspect, a training system for a neural network model is provided, the system comprising an apparatus for executing the first aspect or any possible implementation of the first aspect.

[0065] In some possible implementations, the training system includes a communication system, or the communication system includes the training system. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] FIG1 is a system architecture diagram of an embodiment of the present application;

[0067] FIG2 is a flow chart of a training method according to an embodiment of the present application;

[0068] FIG3 is a flow chart of a training method according to an embodiment of the present application;

[0069] FIG4 is a flow chart of a method for acquiring training data according to an embodiment of the present application;

[0070] FIG5 is an architecture diagram of a communication system according to an embodiment of the present application;

[0071] FIG6 is a flow chart of a communication system according to an embodiment of the present application;

[0072] FIG7 is a flow chart of a communication system according to an embodiment of the present application;

[0073] FIG8 is a flow chart of a communication system according to an embodiment of the present application;

[0074] FIG9 is a structural diagram of a device for acquiring training data according to an embodiment of the present application;

[0075] FIG10 is a structural diagram of a device for acquiring training data according to an embodiment of the present application. DETAILED DESCRIPTION

[0076] The technical solutions in the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application.

[0077] To facilitate the clear description of the technical solutions of the embodiments of the present application, in the embodiments of the present application, the words "first" and "second" are used to distinguish between identical or similar items with substantially the same functions and effects. Those skilled in the art will understand that the words "first" and "second" do not limit the quantity or execution order, and the words "first" and "second" do not necessarily mean different.

[0078] It should be noted that in the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described in this application as "exemplary" or "for example" should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0079] In the embodiments of the present application, "at least one" refers to one or more, and "more" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: the existence of A alone, the existence of A and B at the same time, and the existence of B alone, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, c can be single or multiple.

[0080] In order to better understand the embodiments of the present application, some relevant knowledge about the neural network model is first introduced below.

[0081] The operation of each layer in a neural network model (such as a deep neural network model) can be described by the mathematical expression y = a(W*x+b). The operation of each layer in a neural network model can be understood as transforming the input space (a set of input vectors) to the output space (i.e., from the row space to the column space of a matrix) through five operations. These operations include: 1. Dimensionality increase / decrease; 2. Scaling / reduction; 3. Rotation; 4. Translation; and 5. Warping. Operations 1, 2, and 3 are performed by "W*x," operation 4 is performed by "+b," and operation 5 is performed by "a()." The word "space" is used here because the object being classified is not a single thing, but a class of things, and space refers to the collection of all individuals in that class. W is a weight vector, each value in which represents the weight of a neuron in that layer of the neural network. This vector W determines the spatial transformation from input space to output space described above. That is, the weights W of each layer control how the space is transformed. The purpose of training a neural network model is to ultimately obtain the weight matrices of all layers of the trained neural network model (a weight matrix formed by many layers of vectors W). Therefore, the training process of a neural network model is essentially learning how to control spatial transformations, and more specifically, learning the weight matrices.

[0082] Because we hope that the output of the neural network model is as close as possible to the target value we really want, we can compare the output value of the current neural network model with the target value, and then update the weight vector of each layer of the neural network model according to the difference between the two (of course, there is usually an initialization process before the first update, that is, pre-configuring parameters or pre-configuring initial values ​​for each layer in the neural network model). For example, if the output value of the neural network model is high, the weight vector is adjusted to make the output value of the neural network model lower, and the adjustment is continued until the neural network model can output the target value. Therefore, it is necessary to pre-define "how to compare the difference between the output value and the target value", which is the loss function or objective function, which are important equations for measuring the difference between the output value and the target value. Among them, taking the loss function as an example, the higher the output value (loss) of the loss function, the greater the difference, so the training of the deep neural network becomes a process of minimizing this loss as much as possible. In this application, the target value can be called a label or a true value or an expected value or a target output or an expected output.

[0083] It is understandable that the neural network model in this application can also be replaced by an artificial intelligence (AI) model or a machine learning (ML) model or models.

[0084] 1 , an embodiment of the present application provides a system architecture, wherein the system 100 includes an execution device 110 , a training device 120 , a database 130 , and a data collection device 140 .

[0085] Data acquisition device 140 is used to collect data and store it in database 130. Training device 120 is used to train the neural network model based on the data maintained in database 130 to obtain neural network model 101. Neural network model 101 obtained by training device 120 is applied to execution device 110. Execution device 110 is used to process the input data using neural network model 101 to obtain output.

[0086] In some implementations, database 130 may be located in a training device.

[0087] In some implementations, the data collection device 140 and the training device 120 may be the same device.

[0088] In some implementations, the training device 120 and the execution device 110 may be the same device.

[0089] Figure 2 is a schematic diagram of a model training method according to an embodiment of the present application. Taking the system shown in Figure 1 as an example, the method can be implemented or executed by the training device 120 in Figure 1.

[0090] As shown in Figure 2, input H1 to the neural network model. After H1 is input to the neural network model, the output of the neural network model is recorded as C1. Compare C1 with the label C1' corresponding to H1, and update the weight vector of the neural network model based on the difference between C1 and C1'; input H2 to the neural network model. After H2 is input to the neural network model, the output of the neural network model is recorded as C2. Compare C2 with the label C2' corresponding to H2, and update the neural network model based on the difference between C2 and C2', such as updating the weight vector of the neural network model; and so on. Input H1 to the neural network model. n , H n After inputting the neural network model, the output of the neural network model is recorded as C n , compare C n With H n Corresponding label C n ', based on C n with C n 'The difference between the two updates the weight vector of the neural network model until the difference between the output of the neural network model and the label corresponding to the input is less than a certain threshold.

[0091] However, this training method may have the following problems: the trained neural network model processes inputs H1 and H2 independently, and cannot extract the relationship between H1 and H2 to improve the performance of the neural network.

[0092] To address the above issues, this application proposes a new model training method. Taking the system shown in FIG1 as an example, the method can be implemented or executed by the training device 120 in FIG1 .

[0093] In the model training method proposed in this application, at least one actual output is obtained by inputting data into the neural network model multiple times; these multiple inputs correspond to at least one target output as a whole; this at least one actual output is regarded as a whole, and this at least one target output is regarded as a whole, and the difference between the two wholes is compared, and the weight vector of the neural network model is updated based on this difference.

[0094] As shown in Figure 3, input H1 to the neural network model, the actual output of the neural network model is recorded as C1, and the target output corresponding to H1 is recorded as C1'; next, input H2 to the neural network model, the actual output of the neural network model is recorded as C2, and the target output corresponding to H2 is recorded as C2'; and so on. n , the actual output of the neural network model is recorded as C n , the target output corresponding to Hn is C n '; Compare n times the actual output "C1, C2...C n ” and target output “C1’, C2’…Cn '"; based on this difference, the weight vector of the neural network model is updated. Where n is an integer greater than 1.

[0095] For example, calculate the output sequence "C1, C2...C n ” and the target value sequence “C1’, C2’…C n '" is used as the difference between the two series, and the weight vector of the neural network model is updated based on this difference.

[0096] Among them, H1, H2...H n They are called primary inputs, C1, C2...C n Each of them is recorded as an actual output, C1', C2'...C n 'Recorded as a target output; H1, H2...H n As a whole, it is called an input data sequence, referred to as input sequence; C1, C2...C n As a whole, it is called an actual output data sequence, referred to as the actual output sequence; C1', C2'...C n 'As a whole, it is called a target output data sequence, referred to as a target output sequence. The target output sequence can also be called a label sequence. The target output sequence is the label corresponding to the input sequence or the corresponding label sequence. The input sequence, actual output sequence or target output sequence can be collectively referred to as a data sequence, referred to as a sequence.

[0097] Because these multiple inputs are input into the neural network model based on the agreed sorting, it can be considered that these multiple inputs are ordered, the corresponding at least one output can also be considered ordered, and the labels corresponding to these multiple inputs can also be considered ordered.

[0098] During the training or inference process, the model generates or stores corresponding state information. This state information can affect the output obtained by the model when processing the same input. The state information can be determined by the initial value of the model's state information and the training or inference performed historically. That is, the model shown in Figure 3 has time domain continuity.

[0099] In some possible implementations, the difference between the output sequence and the label sequence may be represented based on the mean square error, squared error, covariance, or correlation between the two sequences.

[0100] In some possible implementations, the label of the input sequence of the neural network model is the input sequence. As an example, multiple data in the first input sequence are sequentially input into the first neural network model to obtain at least one actual output data, which is recorded as the second input sequence; at least one data in the second input sequence is input into the second neural network model to obtain multiple actual output data, which are the output sequence; the difference between the output sequence and the first input sequence is compared, and based on this difference, the weight vectors of the first neural network model and the second neural network model are updated so that the difference between the output of the second neural network model and the input of the first neural network model meets the requirements. In this case, the target neural network model obtained through training can be the first neural network model, or the second neural network model, or both the first neural network model and the second neural network model.

[0101] The function of the second neural network model is exactly the opposite of that of the first neural network model. In other words, when the first data is input into the first neural network model to obtain the second data, and the second data is input into the second neural network model, the expected output of the second neural network model is the first data; when the third data is input into the second neural network model to obtain the fourth data, and the fourth data is input into the first neural network model, the expected output of the first neural network model is the third data.

[0102] As can be seen from the above, in the method newly proposed in this application, because at least one expected output data corresponding to multiple inputs is regarded as a whole to calculate the difference between the at least one actual output corresponding to the multiple inputs and the whole, and based on the difference to update the weight of the neural network model, the conversion relationship from the input space to the output space learned by the trained neural network model can not only include the conversion relationship between one input and one output, but also learn the conversion relationship between other inputs and the current input to the output. In this way, in the application scenario of the neural network model, taking the system shown in Figure 1 as an example, when the execution device 110 uses the trained neural network model 101 to obtain the output based on the input, because it also uses the conversion relationship between the target value corresponding to the other inputs and the current input, it can improve the similarity between the current actual output and the target value, that is, improve the accuracy of the current output.

[0103] In response to the above-mentioned model training method, the present application provides a method for obtaining training data. FIG4 is a schematic flow chart of a method for obtaining training data according to an embodiment of the present application. As shown in FIG4 , the method includes S410 , S420 , and S430 .

[0104] S410: Obtain T data sequences, where each of the T data sequences contains at least one data, and T is an integer greater than 1.

[0105] S420: Obtain first information, where the first information indicates P data sequences among T data sequences used to obtain a first data sequence, where P is a positive integer less than or equal to T.

[0106] S430. Determine a first data sequence based on the first information and T data sequences. The first data sequence includes M data, where M is a positive integer greater than 1. The first data sequence is used as an input data sequence of the neural network model and / or as a label of the input data sequence of the neural network model, wherein the data in the input data sequence is used to be input into the neural network model in sequence.

[0107] In the method of this embodiment, when the first data sequence is used as an input sequence for training a neural network model, multiple data in the first data sequence can be sequentially input into the neural network model. An example of the first data sequence includes H1, H2, ..., H n , where n is an integer greater than 1; an example of a method of sequentially inputting the first data sequence into the neural network model is shown in FIG3 H1, H2 ..., H n How to input the neural network model.

[0108] In some implementations of the present application, P data sequences are arranged in sequence to obtain a first data sequence, where "in sequence" can be understood as the order of the P data sequences in the T data sequences, or the order of the P data sequences indicated by the first information.

[0109] In some implementations of the present application, one understanding of the multiple data in the data sequence being sequentially input into the neural network model is as follows: the multiple data in the data sequence are ordered.

[0110] Optionally, the order in which multiple data in a data sequence are input into the neural network model may be the same as or different from the order of these multiple data in the data sequence. The embodiments of the present application do not impose any restrictions on this, as long as the order in which these multiple data are input into the neural network model can be determined or known based on the order between these multiple data.

[0111] Optionally, the order in which multiple data in the data sequence are input into the neural network model can be predefined, or can be indicated by indication information.

[0112] In the method of this embodiment, the data sequence is used to associate with a tag, which can be understood as: the data sequence can be associated with a tag sequence, or the data sequence can be associated with a tag.

[0113] In the method of this embodiment, the first data sequence is associated with a label sequence or a label, which can be understood as: the label sequence associated with the first data sequence contains at least one label corresponding to multiple data in the first data sequence, and when the multiple data in the first data sequence are sequentially input into the neural network model to obtain at least one actual output corresponding to the multiple data, the at least one label is used to calculate the difference between the at least one label and the at least one actual output, and the difference is used to update the weight vector of the neural network model.

[0114] In some implementations of this embodiment, when the first data sequence is used to train a neural network model, the first data sequence can also be used as a label of an input sequence of the neural network model.

[0115] When a data sequence is used as both an input sequence and a label of the input sequence of a neural network model, an example of training a neural network model based on a data sequence can refer to the aforementioned content about the label of the input sequence of the neural network model being the relevant content of the input sequence, which will not be repeated here.

[0116] In some implementations of this embodiment, obtaining the first data sequence includes: generating the first data sequence. It can be understood that the generation in this embodiment can be understood as a process from nothing to something. For example, generating the first data sequence can be understood as: there are multiple data sequences, and these data sequences can be used to train the neural network model, but whether these data sequences can constitute a new data sequence, or there is no determination of the order in which the data in these multiple data sequences are input into the neural network model, and / or, there is no label sequence that can be associated as a whole, that is, there was originally no concept of the first data sequence, and after the generation operation, the order in which these data are input into the neural network model is determined and / or it is determined that these data as a whole can be associated with a label sequence.

[0117] When acquiring the first data sequence includes generating the first data sequence, in some implementations, the first data sequence is generated by the data acquisition device in FIG1 . For example, the data acquisition device acquires a large amount of data and generates the first data sequence based on the data, or determines which of the data are multiple data included in the first data sequence, or determines which of the data are multiple data that constitute or compose the first data sequence. An example of constituting or composing is splicing.

[0118] When acquiring the first data sequence includes generating the first data sequence, in some implementations, the first data sequence is generated by the database in Figure 1. For example, a data acquisition device generates T data sequences and sends these data sequences to the database, which generates the first data sequence.

[0119] When obtaining the first data sequence includes generating the first data sequence, in some implementations, the first data sequence is generated by the training device in Figure 1. For example, the database obtains T data sequences and sends these data sequences to the training device, which generates the first data sequence.

[0120] The database acquires the T data sequences, which may include: the database generates the T data sequences, or the database receives the T data sequences from a data acquisition device.

[0121] Some implementations of this embodiment further include: sending a first data sequence.

[0122] For example, after the data acquisition device acquires the first data sequence, it may send the first data sequence to the database.

[0123] For another example, after the database acquires the first data sequence, it may send the first data sequence to the training device. The database acquiring the first data sequence here may be the database generating the first data sequence, or the database receiving the first data sequence from the data acquisition device.

[0124] Some implementations of this embodiment further include: using the first data sequence to train a neural network model.

[0125] For example, after the training device acquires the first data sequence, the first data sequence can be used to train the neural network model. Here, the training device acquiring the first data sequence may be the training device generating the first data sequence, or the training device receiving the first data sequence from a database, or, when the database is located on the training device, the training device receiving the first data sequence from a data acquisition device.

[0126] In some implementations of this embodiment, when a training device uses a first data sequence to train a neural network model, the label sequence associated with the first data sequence may be associated with the first data sequence by the training device, wherein the first data sequence may be generated by the training device, a database, or a data acquisition device; or, the label sequence associated with the first data sequence may be associated with the first data sequence by the database, wherein the first data sequence may be generated by the database or a data acquisition device; or, the label sequence associated with the first data sequence may be associated with the first data sequence by the data acquisition device, wherein the first data sequence may be generated by the data acquisition device.

[0127] Some implementations of this embodiment further include: obtaining an initial value of the model state corresponding to the first data sequence.

[0128] The initial value of the model state corresponding to the first data sequence can be understood as: the initial value of the model state of the neural network model when the first data sequence is used as the input of the neural network model and the actual output sequence of the neural network model is expected to be the label sequence associated with the first data sequence.

[0129] In this application, the model state initial value may also be referred to as the model state initial information, or the initial value of the model state information. The model state information may also be referred to as any one or more of the following: model state value, cache information related to the model, storage information related to the model, intermediate information (e.g., intermediate information generated by the model), internal information (e.g., internal information of the device where the model is deployed or internal information of the model), and parameter information generated or updated by the model.

[0130] It can be understood that when the first data sequence is input into the neural network model, the model state value of the neural network model is the initial model state value corresponding to the first data sequence.

[0131] In some implementations, the model state initial value corresponding to the first data sequence may be the model state initial value corresponding to the data sequence that is first input into the neural network model among the P data sequences.

[0132] The model state initial value can be expressed as a scalar, such as 0 or 1. In this case, assuming the network state is actually an N-dimensional tensor, the model state initial value is the scalar assigned to each element in the N-dimensional tensor. Alternatively, the model state initial value can be directly an N-dimensional scalar, where N is a positive integer.

[0133] In some implementations, the first information may include at least one of the following information: a start sequence identifier of the first data sequence, P, or an end sequence identifier of the first data sequence.

[0134] The starting sequence identifier of the first data sequence and the ending sequence identifier of the first data sequence may be the same or may be different.

[0135] When the starting sequence identifier of the first data sequence is the same as the ending sequence identifier of the first data sequence, in some implementations, the starting sequence identifier of the first data sequence and the ending sequence identifier of the first data sequence may be collectively referred to as a sequence identifier.

[0136] In some implementations of this embodiment, at least part of the first information may be predefined or have a default configuration.

[0137] In some implementations of this embodiment, at least part of the first information may be received from other devices, for example, from an OAM device.

[0138] In some implementations, the T data sequences include a first sequence, wherein the first information includes first indication information, and the first indication information indicates: a data sequence in the T data sequences that is spliced ​​with the first sequence.

[0139] In this embodiment, the concatenation of one data sequence with another data sequence can be understood as: the data in the previous data sequence is sequentially arranged before or after the next data sequence. As an example, the sequential arrangement here can be understood as arranging the data in the order in which they appear in the previous data sequence.

[0140] For example, the identifiers of six of the T data sequences are recorded as ID1, ID2, ID3, ID4, ID5, and ID6. Sequence ID2 has an indication containing sequence identifier ID1, indicating that sequence ID2 and sequence ID1 can be combined. Sequence ID4 has an indication containing sequence identifier ID2, indicating that sequence ID4 and sequence ID2 can be combined. Therefore, it can be seen that sequences ID1, ID2, and ID4 are concatenated in order to obtain the first data sequence.

[0141] In some implementations, the T data sequences include a first sequence and a second sequence, wherein the first information includes second indication information, and the second indication information indicates whether the second sequence is spliced ​​with the first sequence.

[0142] For example, the identifiers of six of the T data sequences are recorded as ID1, ID2, ID3, ID4, ID5, and ID6. For sequence ID2, the indication information indicates that sequence ID2 can be combined with the previous sequence ID1. For sequence ID3, the indication information indicates that sequence ID3 cannot be combined with the previous sequence ID2. Therefore, it can be seen that sequences ID1 and ID2 are concatenated in order to obtain the first data sequence.

[0143] In some implementations, the first information includes at least one of the following information: a starting sequence identifier, P, of the first data sequence, or an ending sequence identifier of the first data sequence.

[0144] For example, the identifiers of six of the T data sequences are recorded as ID1, ID2, ID3, ID4, ID5, and ID6. The indication information indicates that the starting sequence is sequence ID2 and P is 3. Therefore, sequences ID2, ID3, and ID4 are sequentially concatenated to obtain the first data sequence.

[0145] The first information including P can be understood as the number of data sequences included in the first data sequence included in the first information.

[0146] For example, the identifiers of six of the T data sequences are recorded as ID1, ID2, ID3, ID4, ID5, and ID6. The indication information indicates that the ending sequence is sequence ID5, and P is 3. Therefore, sequences ID3, ID4, and ID5 are sequentially concatenated to obtain the first data sequence.

[0147] For example, the identifiers of six of the T data sequences are recorded as ID1, ID2, ID3, ID4, ID5, and ID6. The indication information indicates that the starting sequence is sequence ID3 and the ending sequence is sequence ID5. Therefore, sequences ID3, ID4, and ID5 are sequentially concatenated to obtain the first data sequence.

[0148] In some implementations, the first information indicating P data sequences includes: the first information indicating characteristics of the P data sequences. In this way, the P data sequences among the T data sequences that meet the characteristics indicated by the first information can be determined as data sequences used to obtain the first data sequence, thereby determining the first data sequence.

[0149] As an example, the first information indicates a first sequence length, and P data sequences with sequence lengths greater than or equal to the first sequence length are used to obtain the first data sequence.

[0150] For example, if the sequence length indicated by the first information is 100, and the amount of data contained in each of sequences ID3, ID4, and ID5 is greater than or equal to 100, then sequences ID3, ID4, and ID5 can be concatenated in order to obtain a first data sequence.

[0151] As an example, the first information indicates an odd number, and P data sequences identified as odd numbers are used to obtain the first data sequence.

[0152] For example, if the first information indicates that the identifiers of the P sequences are odd numbers, then the sequences ID1, ID3, and ID5 may be concatenated in order to obtain a first data sequence.

[0153] For another example, if the first information indicates an even number, then P data sequences identified as even numbers are used to obtain the first data sequence.

[0154] In some implementations, the first information includes identification information of each data sequence in the P data sequences. The identification information of each data sequence may be an index of the data sequence in the T data sequences.

[0155] For example, the identifiers of 6 sequences in the plurality of T data sequences are recorded as: ID1, ID2, ID3, ID4, ID5, ID6. The indication information indicates: ID1, ID2, ID3, so it can be known that the sequences ID1, ID2, and ID3 are concatenated in order to obtain the first data sequence.

[0156] In some implementations, the order of ID1, ID2, ID3, ID4, ID5, and ID6 may be a transmission order, a storage order, or an order of ID number sizes.

[0157] In some implementations of the present method, when the first data sequence is used only as an input sequence, the present method further includes: obtaining S data sequences, each of the S data sequences containing at least one data, and S being an integer greater than 1; determining a second data sequence based on the S data sequences, the second data sequence containing Q data, Q being a positive integer greater than 1, and the second data sequence being used as a label for the first data sequence, wherein the multiple data in the first data sequence are used to be sequentially input into the neural network model.

[0158] The method for obtaining the Q data sequences can refer to the method for obtaining the P data sequences mentioned above, and will not be repeated here.

[0159] When acquiring T data sequences includes receiving T data sequences, in some implementations, acquiring first information includes receiving first information. In other words, the first information is sent by a device sending the T data sequences to a device receiving the T data sequences.

[0160] As an example, the first information and the T data sequences are included in the same message or the same information and transmitted.

[0161] In some implementations, the first information may be predefined.

[0162] The following takes one of the T data sequences as the first sequence as an example to introduce a method for obtaining T data sequences.

[0163] The first sequence contains multiple data, the first sequence is used as an input data sequence, the multiple data in the first sequence are used to be sequentially input into the neural network model, and the first sequence is used to associate labels.

[0164] In some implementations of this embodiment, obtaining the first sequence includes: generating the first sequence. It can be understood that the generation in this embodiment can be understood as a process from nothing to something. For example, generating the first sequence can be understood as: there are multiple data, and these data can be used to train the neural network model, but these data do not have a fixed order for inputting into the neural network model, and / or, as a whole, cannot be associated with a label sequence, that is, originally there is no concept of a first sequence. After the generation operation, the order in which these data are input into the neural network model and / or it is determined that these data as a whole can be associated with a label sequence is determined.

[0165] When obtaining the first sequence includes generating the first sequence, in some implementations, the first sequence is generated by the data acquisition device in Figure 1. For example, the data acquisition device acquires a large amount of data and generates the first sequence based on this data, or in other words, determines which of these data are multiple data included in the first sequence, or in other words, determines which of these data are multiple data that constitute or compose the first sequence. An example of constituting or composing is splicing. In this implementation, the first data sequence can be obtained by the data acquisition device, database, or training device based on the first sequence.

[0166] When obtaining the first sequence includes generating the first sequence, in some implementations, the first sequence is generated by the database in FIG1 . For example, a data acquisition device collects a large amount of data and sends the data to a database, which then generates the first sequence. In this implementation, the first data sequence can be obtained based on the first sequence by the database or the training device.

[0167] When obtaining the first sequence includes generating the first sequence, in some implementations, the first sequence is generated by the training device in FIG1 . For example, a data acquisition device collects a large amount of data and sends the data to a database, which periodically sends the data to the training device, which generates the first sequence. In this implementation, the training device can obtain the first data sequence based on the first sequence.

[0168] In some implementations of this embodiment, obtaining the first sequence includes: receiving the first sequence.

[0169] When obtaining the first sequence includes receiving the first sequence, in some implementations, the first sequence is received by the database in FIG1 . For example, a data acquisition device generates the first sequence, and the database receives the first sequence from the data acquisition device. In this implementation, the first data sequence can be obtained by the database or the training device based on the first sequence.

[0170] When obtaining the first sequence includes receiving the first sequence, in some implementations, the first sequence is received by the training device in FIG. 1 . For example, after a database generates the first sequence or the database receives the first sequence from a data acquisition device, the training device receives the first sequence from the database. In this implementation, the training device can obtain the first data sequence based on the first sequence.

[0171] Some implementations of this embodiment further include sending a first sequence.

[0172] For example, after the data acquisition device acquires the first sequence, it may send the first sequence to the database. The data acquisition device here may generate the first sequence. In this implementation, the first data sequence may be acquired based on the first sequence by the database or the training device.

[0173] For another example, after the database obtains the first sequence, it may send the first sequence to the training device. The database obtaining the first sequence here may involve the database generating the first sequence, or the database receiving the first sequence from the data acquisition device. In this implementation, the training device may obtain the first data sequence based on the first sequence.

[0174] Some implementations of this embodiment further include: obtaining auxiliary information associated with the first sequence, where the auxiliary information associated with the first sequence includes: an initial value of the model state corresponding to the first sequence, and / or first sequence indication information, where the first sequence indication information is used to indicate the first sequence. The purpose of the indication here can be to generate the first sequence, or to identify the first sequence. Identification can be understood as defining or distinguishing.

[0175] The initial value of the model state corresponding to the first sequence can be understood as: the initial value of the model state of the neural network model when the first sequence is used as the input of the neural network model and the actual output sequence of the neural network model is expected to be the label sequence associated with the first sequence.

[0176] It can be understood that when the first sequence is input into the neural network model, the model state value of the neural network model is the initial model state value corresponding to the first sequence.

[0177] In this implementation, the initial value of the model state corresponding to the first sequence is associated, which can improve the model training performance.

[0178] In this implementation, the auxiliary information associated with the first sequence is obtained. The auxiliary information associated with the first sequence includes the model state initial value corresponding to the first sequence and / or the first sequence indication information. This can be understood as: obtaining the model state initial value corresponding to the first sequence, and / or obtaining the first sequence indication information.

[0179] In some implementations of this embodiment, obtaining the model state initial value corresponding to the first sequence includes: determining the model state initial value corresponding to the first sequence. Determining the model state initial value corresponding to the first sequence can be understood as associating the model state initial value with the first sequence.

[0180] The initial value of the model state may be predefined, randomly generated, or generated by a device that generates the first data sequence.

[0181] For example, when the data acquisition device generates a first sequence, the data acquisition device determines the initial value of the model state corresponding to the first sequence. In this case, the data acquisition device can send not only the first sequence but also the initial value of the model state corresponding to the first sequence.

[0182] For another example, when a database generates a first sequence or a data acquisition device receives the first sequence from a database, the database can determine the initial model state value for the first sequence. In this case, the database can send not only the first sequence but also the initial model state value corresponding to the first sequence.

[0183] For another example, when the training device generates the first sequence or receives the first sequence from a database, the training device may determine an initial value of the model state for the first sequence.

[0184] In some implementations of this embodiment, obtaining the model state initial value corresponding to the first sequence includes: receiving the model state initial value corresponding to the first sequence.

[0185] For example, when the database receives the first sequence from the data acquisition device, the database receives the initial value of the model state from the data acquisition device.

[0186] For another example, when the training device receives the first sequence from a database or a data acquisition device, the training device receives the initial value of the model state from the data acquisition device or the database.

[0187] In some implementations of this embodiment, obtaining the first sequence indication information may include: generating the first sequence indication information. This implementation may be understood as the first sequence indication information being generated along with the generation of the first sequence.

[0188] For example, when a database, data acquisition device, or training device generates a first sequence, the device also generates first sequence indication information. The first sequence indication information generated here can be used by other devices to identify the first sequence. In this case, the method may also include: sending the first sequence indication information.

[0189] In some implementations of this embodiment, obtaining the first sequence indication information may include: receiving the first sequence indication information.

[0190] In some implementations, the received first sequence indication information may be used to generate the first sequence. This implementation may be understood as: generating the first sequence according to the first sequence indication information. In this case, the first sequence indication information is used by the current device to generate the first sequence and by other devices to identify the first sequence.

[0191] For example, a data acquisition device, a database, or a training device receives first sequence indication information and generates a first sequence under the instruction of the first sequence indication information.

[0192] In some implementations, when the database or data acquisition device generates the first sequence indication information, the database or data acquisition device also sends the first sequence indication information so that other devices can identify the first sequence based on the first sequence indication information.

[0193] In some implementations, the received first sequence indication information is used to identify the first sequence. This implementation can be understood as: identifying the first sequence according to the indication of the first sequence indication information.

[0194] For example, when the training device needs to train a neural network model, it can identify the first sequence based on the first sequence indication information.

[0195] In some implementations, the first sequence indication information may be referred to as a pattern of the first sequence.

[0196] In some implementations, the first sequence indication information may include at least one of the following information: a first representation of the first sequence, a start data identifier of the first sequence, a length of the first sequence, or an end data identifier of the first sequence.

[0197] The start data identifier of the first sequence and the end data identifier of the first sequence may be the same or may be different.

[0198] When the start data identifier of the first sequence is the same as the end data identifier of the first sequence, in some implementations, the start data identifier of the first sequence and the end data identifier of the first sequence may be collectively referred to as a data identifier of the sequence.

[0199] In some implementations of this embodiment, at least part of the first sequence indication information may be predefined or have a default configuration. In this implementation, for this part of information, obtaining the first sequence indication information may include: obtaining predefined first sequence indication information.

[0200] In some implementations, the first representation includes: representing each data point in the sequence by a scalar, representing each data point in the sequence by a vector, or representing each data point in the sequence by a basis and a coefficient of each data point relative to the basis. Representing each data point in the sequence by a basis and a coefficient of each data point relative to the basis may be referred to as a basis representation.

[0201] An example of a basis method can refer to a codebook representation method.

[0202] In some implementations, the basis approach can include: representing each datum in a sequence using its basis and its coefficient relative to that basis (known as a unique basis approach); or representing each datum in a sequence using a common basis for all datums in the sequence and its coefficient relative to that common basis (known as a shared basis approach). A unique basis approach means that the basis can be unique to each datum, while a shared basis approach means that the basis is shared by multiple datums.

[0203] In some implementations, the first sequence indication information only indicates that the first representation is a base representation. Whether the base is a specific base or a shared base may be predefined.

[0204] In some implementations, the first sequence indication information directly indicates whether the first representation is a unique basis method or a common basis method.

[0205] In some implementations, the T data sequences include a third sequence, the third sequence includes multiple data, the third sequence serves as an input data sequence, the multiple data in the third sequence are sequentially input into the neural network model, and the third sequence is used to associate labels. Therefore, obtaining the T data sequences may include obtaining the third sequence.

[0206] In this implementation, the relevant content of the third sequence can refer to the relevant content of the first sequence, and will not be repeated here.

[0207] Some implementations of this embodiment further include: acquiring auxiliary information associated with the third sequence.

[0208] In some implementations, the relevant content of the auxiliary information associated with the third sequence may refer to the auxiliary information associated with the first sequence.

[0209] For the sake of simplicity, the auxiliary information associated with the first sequence is referred to as first auxiliary information, and the auxiliary information associated with the third sequence is referred to as second auxiliary information.

[0210] In some implementations, the first auxiliary information and the second auxiliary information are expressed in one of the following forms:

[0211] (1) Public auxiliary information + first data sequence + third data sequence, where "+" represents combination or concatenation. The first auxiliary information and the second auxiliary information are the same, and are referred to as public auxiliary information (hereinafter referred to as public information).

[0212] For example, the public auxiliary information includes at least one of the following information: representation mode, start data identifier, sequence length, or end data identifier;

[0213] (2) public auxiliary information + first difference auxiliary information + first sequence + second difference auxiliary information + third sequence, where "+" represents combination or concatenation. The first difference auxiliary information is auxiliary information specific to the first sequence, and the first auxiliary information includes the first difference auxiliary information and public auxiliary information. The second difference auxiliary information is auxiliary information specific to the third sequence, and the second auxiliary information includes the second difference auxiliary information and public auxiliary information.

[0214] For example, the common auxiliary information includes at least one of the following information: representation method, starting data identifier, sequence length, or ending data identifier; the first difference auxiliary information includes at least one of the following information: representation method, starting data identifier, sequence length, or ending data identifier; the second difference auxiliary information includes at least one of the following information: representation method, starting data identifier, sequence length, or ending data identifier.

[0215] In some implementations, the information included in the common auxiliary information may no longer be included in the difference auxiliary information.

[0216] In some implementations, information included in the common assistance information is also included in the difference assistance information. In this case, if the value in the difference assistance information differs from the value in the common assistance information, the value in the difference assistance information may be prioritized, or the value in the common assistance information may be prioritized. The decision on whether to prioritize the difference assistance information or the common assistance information may be predefined or indicated by a priority or indication information.

[0217] (3) Common auxiliary information + first sequence + second difference auxiliary information + third sequence. The difference between this method and method (2) is that the first auxiliary information is the common auxiliary information, and the second auxiliary information includes the common auxiliary information and the second difference auxiliary information.

[0218] (4) first auxiliary information + first sequence + second auxiliary information + third sequence;

[0219] (5) first sequence + second auxiliary information + third sequence, where the first auxiliary information can be predefined;

[0220] (6) First sequence + third sequence, wherein both the first auxiliary information and the second auxiliary information can be predefined.

[0221] It can be understood that the aforementioned order relationship between the auxiliary information and the data sequence is only an example, and this embodiment does not limit the order of transmission or storage locations between the auxiliary information and the data sequence.

[0222] The following describes a method for obtaining Q data sequences by taking one of the Q data sequences as the fourth sequence as an example.

[0223] The fourth sequence includes at least one data for training the neural network model, wherein the fourth sequence is used as a label for the first sequence.

[0224] In the method of this embodiment, the method for obtaining the fourth sequence can refer to the method for obtaining the first sequence in the previous embodiment, and will not be repeated here.

[0225] Some implementations of this embodiment may further include: acquiring auxiliary information associated with the fourth sequence.

[0226] The method for obtaining the auxiliary information associated with the fourth sequence data may refer to the method for obtaining the auxiliary information associated with the first sequence data in the previous embodiment.

[0227] For the sake of distinction, in this embodiment, the representation of the fourth sequence is recorded as the second representation.

[0228] The fourth sequence is used to train the neural network model. For details about how the first sequence of label sequences is used to train the neural network model in the previous embodiment, details will not be repeated here.

[0229] In some implementations, the auxiliary information associated with the fourth sequence may indicate at least one of the following information: an initial value of the model state corresponding to the fourth sequence, a second representation of the fourth sequence, a sequence length, a start data identifier of the fourth sequence, or an end data identifier of the fourth sequence, wherein the sequence length includes one of the following: the length of the fourth sequence, the minimum value between the length of the first sequence and the length of the fourth sequence, the maximum value between the length of the first sequence and the length of the fourth sequence, or the length of the fourth sequence and the length of the fourth sequence is equal to the length of the first sequence.

[0230] When the sequence length is the minimum value between the length of the first sequence and the length of the fourth sequence, the maximum value between the length of the first sequence and the length of the fourth sequence, or the length of the fourth sequence, and the length of the fourth sequence is equal to the length of the first sequence, the sequence length in the auxiliary information associated with the first sequence is the same as the sequence length in the auxiliary information associated with the fourth sequence.

[0231] An application scenario in which the sequence length is the maximum value is as follows: for the prediction of channel state information, the input sequence contains channel state information at N moments, and the channel state information at each moment is recorded as one data, then the length of the input sequence is N, and the trained neural network model is used to predict the channel state information at the next M moments, that is, the output sequence contains M data, and the length of the output sequence is M. The value of M can be a predefined value, and it is only necessary to indicate the value of N. In this case, the N value can be the maximum value between the length of the input sequence and the length of the output sequence.

[0232] In some implementations, channel information can be used to determine one or more of the resources, modulation and coding scheme (MCS), and precoding configurations of the downlink or uplink data channel of the scheduling terminal device. It can be understood that channel information can reflect channel characteristics, channel quality, etc. Channel information can also be called channel state information (CSI) or channel environment information. It can be understood that the CSI in this application is not limited to traditional CSI, such as one or more of channel quality indication (CQI), precoding matrix indicator (PMI), rank indicator (RI), and channel state information reference signal resource indicator (CSI-RS CRI), but can also be channel response information, such as a channel response matrix, or reference signal receiving power (RSRP), or signal to interference plus noise ratio (SINR).

[0233] An application scenario in which the sequence length is the minimum is as follows: the input sequence contains H1, H2, H3, H4, H5, and H6, with a length of 6. By sliding the window according to the predefined length of 4 (for example), 6 groups of sequences with a length of 4 can be obtained, which are recorded as: H1 to H4, H2 to H5, H3 to H6, H4 to H6 and one 0, H5 to H6 and two 0s, and H6 and three 0s.

[0234] The output sequence is a sequence corresponding to 6 groups of sequences of length 4. These 6 groups of output sequences include: C1_1, C1_2, C1_3, C1_4 (corresponding to H1 to H4), C2_2, C2_3, C2_4, C2_5 (corresponding to H2 to H5), C3_3, C3_4, C3_5, C3_6 (corresponding to H3 to H6), C4_4, C4_5, C4_6, 0 (corresponding to H4 to H6 and 0), C5_5, C5_6, 0, 0 (corresponding to H5 to H6 and two 0s), C6_6, 0, 0, 0 (corresponding to H6 and three 0s). The length of the sequence constituted by these 6 groups of sequences is 24.

[0235] In this example, the sequence length in the auxiliary information may be 6, which is the minimum value between 6 and 24.

[0236] The starting data identifier of the first sequence and the starting data identifier of the fourth sequence may be the same or may be different.

[0237] When the starting data identifier of the first sequence is the same as the starting data identifier of the fourth sequence, in some implementations, the starting data identifier of the first sequence and the starting data identifier of the fourth sequence may be collectively referred to as the starting data identifier of the sequence.

[0238] The end data identifier of the first sequence and the end data identifier of the fourth sequence may be the same or different.

[0239] When the end data identifier of the first sequence is the same as the end data identifier of the fourth sequence, in some implementations, the end data identifier of the first sequence and the end data identifier of the fourth sequence may be collectively referred to as the end data identifier of the sequence.

[0240] In some implementations of this embodiment, the Q sequences further include a fifth sequence, the fifth sequence includes at least one data, and the fifth sequence is used as a tag of the third sequence.

[0241] The relevant content of the fifth sequence is similar to that of the fourth sequence and will not be repeated here.

[0242] Some implementations of this embodiment may further include: obtaining auxiliary information associated with the fifth sequence. The method for obtaining the auxiliary information associated with the fifth sequence data may refer to the method for obtaining the auxiliary information associated with the fourth sequence data, and will not be repeated here.

[0243] In some implementations, the first sequence, the third sequence, and the fourth sequence may be represented by one of the following methods:

[0244] (1) The first sequence + the fourth sequence + the third sequence + the fifth sequence, that is, one training data is represented as a unit, and one training data contains the input sequence and the label sequence of the input sequence;

[0245] (2) The first sequence + the third sequence + the fourth sequence + the fifth sequence, that is, the input data sequence is represented as one unit and the label sequence is represented as one unit.

[0246] Auxiliary information, input sequence, and output sequence can be represented in one of the following ways:

[0247] (1) Common auxiliary information + difference auxiliary information of input sequence + auxiliary information of output sequence + input sequence + output sequence;

[0248] (2) Common auxiliary information + difference auxiliary information of output sequence + input sequence + output sequence;

[0249] (3) Common auxiliary information + difference auxiliary information of input sequence + input sequence + output sequence;

[0250] (4) Auxiliary information of input sequence + auxiliary information of output sequence + input sequence + output sequence;

[0251] (5) Auxiliary information of input sequence + input sequence + output sequence;

[0252] (6) Auxiliary information of output sequence + input sequence + output sequence;

[0253] (7) Public auxiliary information + input sequence + output sequence;

[0254] (8) Common auxiliary information + difference auxiliary information of input sequence + input sequence + difference auxiliary information of output sequence + output sequence;

[0255] (9) Input sequence + auxiliary information of output sequence + output sequence;

[0256] (10) Common auxiliary information + input sequence + difference auxiliary information of output sequence + output sequence;

[0257] (11) Auxiliary information of input sequence + input sequence + auxiliary information of output sequence + output sequence.

[0258] The input sequence includes the first sequence and the third sequence, and the output sequence includes the fourth sequence and the fifth sequence.

[0259] It can be understood that a sequence may not have corresponding difference auxiliary information, but when there is public auxiliary information, the public auxiliary information can be used; when a sequence does not have corresponding difference auxiliary information and has no public auxiliary information, predefined auxiliary information can be used; when there is public auxiliary information and corresponding difference auxiliary information, the public auxiliary information or the difference auxiliary information can be used, and which one to use, the public auxiliary information or the difference auxiliary information, can be determined based on a predefined priority or indicated priority or other indication information.

[0260] The system architecture and related methods and other technical solutions provided above in this application can be applied to communication networks such as the 3rd Generation Partnership Project (3GPP), ZigBee, long-range radio (Lora), Bluetooth (BT), and wireless fidelity (Wi-Fi).

[0261] The technical solutions provided in this application can be applied to various communication systems, such as: fifth generation (5G) or new radio (NR) systems, long term evolution (LTE) systems, LTE frequency division duplex (FDD) systems, LTE time division duplex (TDD) systems, wireless local area networks (WLAN) systems, satellite communication systems, future communication systems, such as sixth generation (6G) mobile communication systems, or a fusion system of multiple systems. The technical solutions provided in this application can also be applied to device to device (D2D) communication, vehicle-to-everything (V2X) communication, machine to machine (M2M) communication, machine type communication (MTC), and Internet of Things (IoT) communication systems or other communication systems.

[0262] A device in a communication system can send signals to or receive signals from another device. The signals may include information, signaling, or data. The term "device" may also be replaced by an entity, a network entity, a communication device, a communication module, a node, a communication node, and the like. This disclosure uses devices as examples for description. For example, a communication system may include at least one terminal device and at least one network device. The network device may send downlink signals to the terminal device, and / or the terminal device may send uplink signals to the network device. It is understood that the terminal device in this disclosure may be replaced by the first device, and the network device may be replaced by the second device, with both devices performing the corresponding communication methods in this disclosure.

[0263] In an embodiment of the present application, the terminal device may also be referred to as user equipment (UE), access terminal, user unit, user station, mobile station, mobile station, remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent or user device.

[0264] The terminal device may be a device that provides voice / data, such as a handheld device or vehicle-mounted device with a wireless connection function. At present, some examples of terminals are: mobile phones, tablet computers, laptop computers, PDAs, mobile internet devices (MIDs), wearable devices, virtual reality (VR) devices, augmented reality (AR) devices, wireless terminals in industrial control, wireless terminals in self-driving, wireless terminals in remote medical surgery, wireless terminals in smart grids, wireless terminals in transportation safety, wireless terminals in smart cities, wireless terminals in smart homes, cellular phones, cordless phones, session initiation protocol (SIP) phones, wireless local loop (WLL) stations, personal digital assistants (PDAs), handheld devices with wireless communication capabilities, computing devices or other processing devices connected to wireless modems, wearable devices, terminal devices in 5G networks or future evolved public land mobile communication networks (PLMNs). The terminal devices in the PLMN network, the devices in the ZigBee network, the devices in the Lora network, the Bluetooth slave (BT slave), the low-power Bluetooth BLE slave, the Wi-Fi station (STA), etc. are not limited in the embodiments of the present application.

[0265] As an example and not a limitation, in the embodiment of the present application, the terminal device may also be a wearable device. Wearable devices may also be called wearable smart devices, which are a general term for wearable devices that are intelligently designed and developed using wearable technology for daily wear, such as glasses, gloves, watches, clothing, and shoes. A wearable device is a portable device that is worn directly on the body or integrated into the user's clothes or accessories. Wearable devices are not only hardware devices, but also achieve powerful functions through software support, data interaction, and cloud interaction. Broadly speaking, wearable smart devices include those that are fully functional, large in size, and can achieve complete or partial functions without relying on smartphones, such as smart watches or smart glasses, as well as those that only focus on a certain type of application function and need to be used in conjunction with other devices such as smartphones, such as various smart bracelets and smart jewelry for vital sign monitoring.

[0266] A terminal device can also be a terminal device in an IoT system, also known as an IoT node. IoT is a crucial component of future information technology development. Its primary technical feature is connecting objects to the network through communication technologies, thereby enabling intelligent networks that interconnect humans and machines, and objects and things. Connections can be achieved through both broadband and narrowband technologies. IoT technology, for example, uses narrowband (NB) technology to achieve massive connections, deep coverage, and power-saving terminals. IoT technologies include reflective communication, spread spectrum, and ultra-wideband (UWB), which are not detailed here.

[0267] In the embodiments of the present application, the device for realizing the function of the terminal device can be a terminal device, or a device capable of supporting the terminal device to realize the function, such as a chip system, which can be installed in the terminal device or used in combination with the terminal device. In the embodiments of the present application, the chip system can be composed of a chip, or it can include a chip and other discrete devices. In the embodiments of the present application, only the terminal device is used as an example for description, and the embodiments of the present application are not limited to the solutions of the embodiments of the present application.

[0268] The network device in the embodiments of the present application may include a device for communicating with a terminal device, and the network device may include an access network device or a radio access network device, such as a base station. The network device in the embodiments of the present application may include a radio access network (RAN) node (or device) that connects the terminal device to a wireless network.

[0269] The radio access network (RAN) device in this application is a device with wireless transceiver functions. The radio access network device can provide wireless communication function services and can connect the terminal device to the wireless network. The network device in the embodiment of the present application can refer to a radio access network (RAN) node (or device) used in a cellular network (or mobile network) to connect the terminal device to the wireless network, and can also be a zigbee base station, a master Bluetooth (BT master), a master low-power Bluetooth (bluetooth low energy master, BLE master), a Lora base station, or a Wi-Fi access point.

[0270] A base station may broadly cover various names as follows, or replace the following names, such as: NodeB, evolved NodeB (eNB), next generation NodeB (gNB), relay station, access point, transmitting and receiving point (TRP), transmitting point (TP), master station, secondary station, multi-standard radio (motor slide retainer, MSR) node, home base station, network controller, access node, wireless node, access point (AP), transmission node, transceiver node, baseband unit (BBU), remote radio unit (RRU), active antenna unit (AAU), remote radio head (RRH), central unit (CU), distributed unit (DU), radio unit (RU), positioning node, etc. A base station may be a macro base station, a micro base station, a relay node, a donor node, or the like, or a combination thereof. A base station may also refer to a communication module, modem, or chip configured within the aforementioned devices or apparatuses. A base station may also be a mobile switching center, a device that performs base station functions in D2D, V2X, or M2M communications, a network-side device in a 6G network, or a device that performs base station functions in future communication systems. A base station may support networks with the same or different access technologies. Optionally, a RAN node may also be a server, a wearable device, a vehicle, or an onboard device. For example, the access network device in vehicle-to-everything (V2X) technology may be a roadside unit (RSU). The embodiments of this application do not limit the specific technology or device form factor employed by the network device. In some deployments, the network device referred to in the embodiments of this application may include a CU, a DU, or both a CU and a DU, or a control plane CU node (central unit-control plane (CU-CP)), a user plane CU node (central unit-user plane (CU-UP)), and a DU node. For example, the network device may include a gNB-CU-CP, a gNB-CU-UP, and a gNB-DU.

[0271] In some deployments, multiple RAN nodes collaborate to assist terminals in achieving wireless access, with different RAN nodes implementing portions of the base station's functionality. For example, a RAN node can be a CU, DU, CU-CP, CU-UP, or RU. The CU and DU can be separate or included in the same network element, such as the BBU. The RU can be included in a radio frequency device or radio unit, such as an RRU, AAU, or RRH.

[0272] The RAN node may support one or more types of fronthaul interfaces, with different fronthaul interfaces corresponding to DUs and RUs with different functions. If the fronthaul interface between the DU and the RU is a common public radio interface (CPRI), the DU is configured to implement one or more baseband functions, and the RU is configured to implement one or more radio frequency functions. If the fronthaul interface between the DU and the RU is another type of interface, relative to the CPRI, some of the downlink and / or uplink baseband functions, such as precoding, digital beamforming (BF), or one or more of inverse fast Fourier transform (IFFT) / cyclic prefix (CP) for downlink, are moved from the DU to the RU for implementation; and for uplink, one or more of digital beamforming (BF), or fast Fourier transform (FFT) / cyclic prefix (CP) removal, are moved from the DU to the RU for implementation. In one possible implementation, the interface may be an enhanced common public radio interface (eCPRI). In the eCPRI architecture, the division between the DU and RU is different, corresponding to different types (category, Cat) of eCPRI, such as eCPRI Cat A, B, C, D, E, and F.

[0273] Taking eCPRI Cat A as an example, for downlink transmission, based on layer mapping, the DU is configured to implement layer mapping and one or more functions preceding it (i.e., one or more of coding, rate matching, scrambling, modulation, and layer mapping). Other functions after layer mapping (e.g., resource element (RE) mapping, digital beamforming (BF), or one or more of inverse fast Fourier transform (IFFT) / cyclic prefix (CP) addition) are moved to the RU for implementation. For uplink transmission, based on RE demapping, the DU is configured to implement demapping and one or more functions preceding it (i.e., one or more of decoding, rate matching, descrambling, demodulation, inverse discrete Fourier transform (IDFT), channel equalization, and RE demapping). Other functions after demapping (e.g., one or more of digital BF or fast Fourier transform (FFT) / CP removal) are moved to the RU for implementation. It is understandable that for the functional description of DU and RU corresponding to various types of eCPRI, reference can be made to the eCPRI protocol, which will not be described in detail here.

[0274] In one possible design, the processing unit for implementing baseband functions in the BBU is called a baseband high layer (BBH) unit, and the processing unit for implementing baseband functions in the RRU / AAU / RRH is called a baseband low layer (BBL) unit.

[0275] In different systems, CU (or CU-CP and CU-UP), DU or RU may also have different names, but those skilled in the art can understand their meanings. For example, in an open radio access network (open RAN, ORAN / O-RAN) system, CU may also be called O-CU (open CU), DU may also be called O-DU, CU-CP may also be called O-CU-CP, CU-UP may also be called O-CU-UP, and RU may also be called O-RU. Any unit in the CU (or CU-CP, CU-UP), DU and RU in this application may be implemented by a software module, a hardware module, or a combination of a software module and a hardware module.

[0276] In the embodiments of the present application, the device for implementing the functions of the network device can be a network device; it can also be a device that can support the network device to implement the functions, such as a chip system, a hardware circuit, a software module, or a hardware circuit and a software module. The device can be installed in the network device or used in conjunction with the network device. In the embodiments of the present application, only the device for implementing the functions of the network device is used as an example to illustrate, and does not constitute a limitation on the solutions of the embodiments of the present application.

[0277] The network device and / or terminal device can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; it can also be deployed on the water surface; it can also be deployed on aircraft, balloons and satellites in the air. The embodiments of this application do not limit the scenarios in which the network device and the terminal device are located. In addition, the terminal device and the network device can be hardware devices, or they can be software functions running on dedicated hardware, software functions running on general-purpose hardware, such as virtualization functions instantiated on a platform (e.g., a cloud platform), or entities including dedicated or general-purpose hardware devices and software functions. This application does not limit the specific forms of the terminal device and the network device.

[0278] To support artificial intelligence (AI) technology in wireless networks, AI nodes can be introduced into the wireless networks. AI nodes can be AI network elements or AI modules.

[0279] Optionally, the AI ​​node can be deployed in one or more of the following locations in the communication system: access network equipment, terminal equipment, or core network equipment. Alternatively, the AI ​​node can be deployed separately, for example, in a location other than any of the above devices, such as a host or cloud server in an over-the-top (OTT) system. The AI ​​node can communicate with other devices in the communication system, such as one or more of the following: network equipment, terminal equipment, or core network elements.

[0280] It is understood that this application does not limit the number of AI nodes. For example, when there are multiple AI nodes, the multiple AI nodes can be divided based on function, such as different AI nodes are responsible for different functions.

[0281] It can also be understood that AI nodes can be independent devices, or they can be integrated into the same device to implement different functions, or they can be network elements in hardware devices, or they can be software functions running on dedicated hardware, or they can be virtualized functions instantiated on a platform (for example, a cloud platform). This application does not limit the specific form of the above-mentioned AI nodes.

[0282] The AI ​​node mentioned here can be used as the data acquisition device, database, training device or execution device in Figure 1.

[0283] Figure 5 is a schematic diagram of a possible application framework in a communication system. As shown in Figure 5, network elements in the communication system are connected through interfaces (such as NG, Xn) or air interfaces. One or more AI modules are provided in these network element nodes, such as core network equipment, access network nodes (RAN nodes), terminals or one or more devices in operation administration and maintenance (OAM) (for clarity, only one is shown in Figure 5). The access network node can be a separate RAN node, or it can include multiple RAN nodes, such as CU and DU, or gNB and / or ng-eNB. The CU and / or DU can also be provided with one or more AI modules. Optionally, the CU can also be split into CU-CP and CU-UP. One or more AI models are provided in the CU-CP and / or CU-UP.

[0284] The AI ​​module is used to implement the corresponding AI function. The AI ​​modules deployed in different network elements may be the same or different. The model of the AI ​​module can implement different functions according to different parameter configurations. The model of the AI ​​module can be configured based on one or more of the following parameters: structural parameters (such as the number of neural network layers, the width of the neural network, the connection relationship between layers, the weight of the neuron, the activation function of the neuron, or at least one of the bias in the activation function), input parameters (such as the type of input parameters and / or the dimension of the input parameters), or output parameters (such as the type of output parameters and / or the dimension of the output parameters). Among them, the bias in the activation function can also be called the bias of the neural network.

[0285] An AI module can have one or more models. A model can infer an output, which includes one or more parameters. The learning, training, or inference processes of different models can be deployed on different nodes or devices, or on the same node or device.

[0286] Figure 6 is a schematic diagram of a possible application framework in a communication system. As shown in Figure 6, the communication system includes a RAN intelligent controller (RIC). For example, the RIC can be the AI ​​module shown in Figure 5, which is used to implement AI-related functions. The RIC includes a near-real-time RIC (near-real time RIC, near-RT RIC) and a non-real-time RIC (non-real time RIC, Non-RT RIC). Among them, the non-real-time RIC mainly processes non-real-time information, such as data that is not sensitive to delay, and the delay of this data can be in the order of seconds. The real-time RIC mainly processes near-real-time information, such as data that is relatively sensitive to delay, and the delay of this data is in the order of tens of milliseconds.

[0287] The near real-time RIC is used for model training and reasoning. For example, it is used to train an AI model and use the AI ​​model for reasoning. The near real-time RIC can obtain network-side and / or terminal-side information from a RAN node (e.g., CU, CU-CP, CU-UP, DU, and / or RU) and / or a terminal. This information can be used as training data or reasoning data. Optionally, the near real-time RIC can deliver the reasoning result to the RAN node and / or the terminal. Optionally, the reasoning result can be exchanged between the CU and the DU, and / or between the DU and the RU. For example, the near real-time RIC delivers the reasoning result to the DU, and the DU sends it to the RU.

[0288] The non-real-time RIC is also used for model training and reasoning. For example, it is used to train an AI model and use the model for reasoning. The non-real-time RIC can obtain network-side and / or terminal-side information from RAN nodes (such as CU, CU-CP, CU-UP, DU and / or RU) and / or terminals. This information can be used as training data or reasoning data, and the reasoning results can be submitted to the RAN node and / or terminal. Optionally, the reasoning results can be exchanged between the CU and the DU, and / or between the DU and the RU. For example, the non-real-time RIC submits the reasoning results to the DU, and the DU sends it to the RU.

[0289] The near real-time RIC and non-real-time RIC may also be separately configured as a network element. Optionally, the near real-time RIC and non-real-time RIC may also be part of other devices. For example, the near real-time RIC is configured in a RAN node (e.g., a CU or DU), while the non-real-time RIC is configured in an OAM, a cloud server, a core network device, or other network device.

[0290] The RIC here can be used as the data acquisition device, database, training device or execution device in Figure 1.

[0291] FIG7 is a schematic diagram of a communication system applicable to the method of an embodiment of the present application. As shown in FIG7 , the communication system 1400 may include at least one network device, such as the network device 1410 shown in FIG7 ; the communication system 1400 may also include at least one terminal device, such as the terminal device 1420 and the terminal device 1430 shown in FIG7 . The network device 1410 and the terminal device (such as the terminal device 1420 and the terminal device 1430) can communicate via a wireless link. The communication devices in the communication system, for example, the network device 1410 and the terminal device 1420, can communicate via a multi-antenna technology.

[0292] The network device in the system can be used as at least one of a data acquisition device, a database, a training device and an execution device, and the terminal can be used as at least one of a data acquisition device, a database and an execution device.

[0293] FIG8 is a schematic diagram of a communication system applicable to the method of an embodiment of the present application. Compared to the communication system 1400 shown in FIG7 , the communication system 1500 shown in FIG8 further includes an AI entity 1440 .

[0294] The AI ​​entity 1440 is used to perform AI-related operations, such as building a training data set or training an AI model.

[0295] In one possible implementation, the network device 1410 may send data related to the training of the AI ​​model to the AI ​​entity 1440, which constructs a training data set and trains the AI ​​model. For example, the data related to the training of the AI ​​model may include data reported by the terminal device. The AI ​​entity 1440 may send the results of the operations related to the AI ​​model to the network device 1410, and forward them to the terminal device through the network device 1410. For example, the results of the operations related to the AI ​​model may include at least one of the following: an AI model that has completed training, an evaluation result or a test result of the model, etc. Exemplarily, a portion of the trained AI model may be deployed on the network device 1410, and another portion may be deployed on the terminal device. Alternatively, the trained AI model may be deployed on the network device 1410. Alternatively, the trained AI model may be deployed on the terminal device.

[0296] It should be understood that FIG8 illustrates only an example of a direct connection between AI entity 1440 and network device 1410. In other scenarios, AI entity 1440 may also be connected to a terminal device. Alternatively, AI entity 1440 may be connected to both network device 1410 and a terminal device simultaneously. Alternatively, AI entity 1440 may be connected to network device 1410 via a third-party network element. This embodiment of the present application does not limit the connection relationship between the AI ​​network element and other network elements.

[0297] The AI ​​entity 1440 can also be set as a module in at least one of the following devices: a network device, a terminal device and a core network element, for example, set in the network device 1410 or the terminal device shown in Figure 8; for example, the AI ​​entity can be located inside the AMF or the location management function (LMF).

[0298] It should be noted that Figures 7 and 8 are simplified schematic diagrams for ease of understanding. For example, the communication system may also include other devices, such as wireless relay devices and / or wireless backhaul devices, and may also include one or more core network elements, which are not shown in Figures 7 and 8. In actual applications, the communication system may include multiple network devices and multiple terminal devices. The embodiments of the present application do not limit the number of network devices and terminal devices included in the communication system.

[0299] The AI ​​entity in the system can be used as at least one of a data acquisition device, a database, a training device, and an execution device.

[0300] The core network element in this application can be used to implement one or more of the following functions: access and mobility management function (AMF), session management function (SMF), policy control function (PCF), user plane function (UPF), network data analytics function (NWDAF), charging function (CHF), and unified data management (UDM) network element.

[0301] AMF is mainly used for terminal attachment, mobility management, and tracking area update processes in mobile networks. The access management network element terminates non-access stratum (NAS) messages, completes registration management, connection management, and reachability management, allocates tracking area lists (TA lists) and mobility management, and transparently routes session management (SM) messages to the session management network element.

[0302] SMF is mainly used for session management in mobile networks, such as session establishment, modification, and release. For example, it can allocate Internet Protocol (IP) addresses to terminals and select user-plane network elements that provide message forwarding functions.

[0303] PCF mainly provides user subscription data management functions, policy control functions, billing policy control functions, and quality of service (QoS) control.

[0304] UPF is mainly responsible for processing user messages, such as forwarding, billing, and legal monitoring.

[0305] NWDAF provides intelligent analysis services, using artificial intelligence and big data analysis to output analysis results in a format defined by specifications. The output analysis results generally come in two forms: one is a statistical analysis of historical data, and the other is a prediction of future data. The service processing network element makes network adjustments based on the NWDAF output to optimize network operation. For example, when the service processing network element finds through the statistical analysis results output by NWDAF that the packet loss rate of the terminal's ongoing service message transmission exceeds the threshold, affecting the service experience, it will take corresponding measures, such as increasing the proportion of retransmitted packets, attempting to improve the service experience, etc.

[0306] CHF mainly collects billing events from various service processing network elements such as SMF, AMF or UPF, and charges users' services according to pre-negotiated rates.

[0307] The UDM network element is used to store and manage user terminal network and service subscription data.

[0308] The communication system of the present application may also include application functions, which may be referred to as third-party applications.

[0309] In this application, the function may also be a network element or an entity, for example, AMF may also be an AMF network element or an AMF entity.

[0310] When the technical solution of the present application is applied to a communication system, in some implementations, the data in the first data sequence is channel state information. For example, H1, H2, ..., Hn in FIG3 are channel state information.

[0311] In some implementations, the trained neural network model is used to compress channel state information. For example, in FIG3 , C1, C2, ..., Cn are the compressed channel state information of H1, H2, ..., Hn, respectively, and C1', C2', ..., Cn' are the true values ​​of the compressed channel state information of H1, H2, ..., Hn, respectively.

[0312] In some implementations, the trained neural network model is used to predict channel state information. For example, in FIG3 , C1, C2, ..., Cm are channel state information predicted based on channel state information H1, H2, ..., Hn, and C1', C2', ..., Cm' are the true values ​​of the channel state information predicted based on channel state information H1, H2, ..., Hn.

[0313] In some implementations, the trained neural network model is used to predict and compress channel state information. For example, in FIG3 , C1, C2, ..., Cm are channel state information predicted and compressed based on channel state information H1, H2, ..., Hn, and C1', C2', ..., Cm' are true values ​​of the channel state information predicted and compressed based on channel state information H1, H2, ..., Hn.

[0314] When the data in the first data sequence is channel state information, in some implementations, the channel state information is denoted as H, and the dimension is (N port ,N band ), where N port Indicates the number of antenna ports, N band Indicates the number of subbands. When H is the original channel information, N port =N T ×N R , where N T Indicates the number of transmit antenna ports, N R Indicates the number of receiving antenna ports. When H is the channel eigenvector, N port =N T .

[0315] The first data sequence can be expressed as H (1:T) (T, N port , N band ), that is, a channel sequence with a sequence length of T.

[0316] In some implementations, the data in the second data sequence is compressed information of the channel state information, that is, compressed channel state information.

[0317] In some implementations, the dimension of the compressed information is N c , indicating that the compressed information is N c A real vector, complex vector, or binary vector of length .

[0318] In some implementations, the second data sequence may be represented as C (1:T) (T, N c ), that is, a compressed information sequence with a sequence length of T.

[0319] When the data in the first data sequence is represented by a basis, in some possible implementations, the i-th data is recorded as H i , H i =V P1 *Coeff i1 *V b1 , where V p1 and V b1 As the base, Coeff i1 is the coefficient of the i-th data on the basis. In this case, the i-th data can be represented by the identity of the basis and the coefficient on the basis.

[0320] When the data in the first data sequence is represented by a common basis, in some possible implementations, the i-th data is determined to be H i , h i =V P3 *Coeff i3 , where V p3 As the base, Coeff i3 is the coefficient of the i-th data on the basis, h i (N port *N band ) is the form of Hi flattened into a long vector.

[0321] When the data in the second data sequence is represented by a basis, in some possible implementations, the i-th data is recorded as C i , Ci=V P2 *Coeff i2 , where V p2 As the base, Coeff i2 is the coefficient of the i-th data on the basis. In this case, the i-th data can be represented by the identity of the basis and the coefficient on the basis.

[0322] In this application, the neural network model used for compression (and prediction) can be called an encoder neural network model, referred to as an encoder; the neural network model used for decompression or recovery can be called a decoder neural network model, referred to as a decoder.

[0323] As an example, the data acquisition device is a network-side device, such as a base station, and the training device is a terminal-side device, such as a terminal or a server communicating with the terminal.

[0324] For example, after the base station has trained its encoder and decoder, it inputs a set of input data sequences into the trained encoder to obtain labels corresponding to the input sequences. The base station then sends the input sequences and labels to the terminal-side device, which then trains its own encoder. This encoder can then be adapted to the decoder on the base station side. The terminal-side device can be a terminal device, a module in the terminal device (such as a chip), software that includes terminal device functions (such as a control subsystem), or other devices that communicate with the terminal device, such as an AI network element, which is a server, such as an OTT device or a cloud server.

[0325] Taking the scenario where a neural network model is applied to the channel state information of a downlink reference signal fed back by a base station to a terminal as an example, in some implementations, the base station obtains N channel state information, and sequentially inputs the N channel state information into a trained encoder to obtain N compressed information, and then sends these N compressed information (compressed information sequence or label sequence) and the corresponding N channel state information (channel state information sequence or input sequence) to the terminal for model training, model monitoring, or model fine-tuning. Alternatively, the base station obtains N channel state information, and sequentially inputs the N channel state information into a trained encoder to obtain N compressed information, and sequentially inputs the N compressed channel information into a trained decoder of the base station to obtain N recovered channel state information, and then feeds back these N compressed information and the corresponding N recovered channel state information to the terminal.

[0326] After receiving these N compressed information and N channel state information, the terminal inputs the N channel state information into the encoder in sequence to obtain N compressed information. By comparing the differences between the output N compressed information and the N compressed information sent by the base station, the encoder is trained so that the trained encoder can match the base station decoder.

[0327] As an example, the data acquisition device is a terminal-side device, such as a terminal or a server communicating with a terminal. The training device is a network-side device, such as a base station or a server communicating with a base station. The network-side device can be a network device, or a module (such as a chip) in a network device, or software containing network device functions (such as a control subsystem), or other devices that communicate with the network device, such as an AI network element, which is a server, such as an OTT device or a cloud server.

[0328] For example, after a terminal has trained its encoder and decoder, it inputs a data sequence into the trained encoder to generate an output sequence. The output sequence is then input into the trained decoder to generate the final output sequence, which is recorded as the encoder output sequence label. The terminal then sends the encoder output sequence and its label to the base station, which then trains its own decoder, allowing the decoder to adapt to the UE's encoder.

[0329] Taking the scenario where a neural network model is applied to the channel state information of a downlink reference signal fed back by a terminal to a base station as an example, in some implementations, the terminal device obtains N channel state information, and the terminal sequentially inputs the N channel state information into the trained neural network model to obtain N compressed information, and feeds back these N compressed information (compressed information sequence or input sequence) and the corresponding N channel state information (channel state information sequence or label sequence) to the base station for model training or model monitoring or model fine-tuning. Alternatively, the terminal device obtains N channel state information, and the terminal sequentially inputs the N channel state information into the trained encoder to obtain N compressed information, and sequentially inputs the N compressed channel information into the decoder trained by the terminal side device to obtain N recovered channel state information, and feeds back these N compressed information and the corresponding N recovered channel state information to the base station.

[0330] After receiving these N compressed information, the base station inputs the N compressed information into the decoder in sequence, thereby obtaining N channel state information. By comparing the differences between the output N state information and the N channel state information sent by the terminal, the decoder is trained so that the trained decoder can match the encoder of the terminal.

[0331] As an example, the data collection device is a third party, and the training device is a terminal-side device, such as a terminal or a server communicating with the terminal; and / or a network-side device, such as a base station or a server communicating with the base station. The third party can be a device, such as a server, that is not provided or operated by the network equipment manufacturer or the terminal manufacturer.

[0332] For example, a third party sends the input sequence and label corresponding to the encoder to the terminal side device for the terminal side device to train its own encoder, or sends the input sequence and label corresponding to the decoder to the network side device for the network side device to train its own decoder. Then, the encoder can be adapted to the decoder on the network side.

[0333] Taking the scenario where a neural network model is applied to the channel state information of a downlink reference signal fed back by a terminal to a base station as an example, in some implementations, a third party obtains N channel state information, and the third party sequentially inputs the N channel state information into a trained encoder to obtain N compressed information, and feeds back these N compressed information (compressed information sequence or label sequence) and the corresponding N channel state information (channel state information sequence or input sequence) to the terminal or base station for model training, model monitoring, or model fine-tuning. Alternatively, the third party obtains N channel state information, and the third party sequentially inputs the N channel state information into a trained encoder to obtain N compressed information, and sequentially inputs the N compressed channel information into a decoder trained by the third party to obtain N recovered channel state information, and feeds back these N compressed information and the corresponding N recovered channel state information to the terminal or base station for model training, model monitoring, or model fine-tuning of the decoder.

[0334] After receiving the N compressed information and the corresponding N channel state information, the terminal uses an encoder to compress the N channel state information, thereby obtaining N compressed information. The encoder is trained by comparing the difference between the N compressed information output by the encoder and the N compressed information received.

[0335] After receiving these N compressed information and the corresponding N channel state information, the base station inputs the N compressed information into the decoder in sequence, thereby obtaining N channel state information. The decoder is trained by comparing the differences between the N channel states output by the decoder and the N channel state information received.

[0336] Ultimately, the decoder trained by the base station can be matched with the encoder trained by the terminal.

[0337] It should be understood that, in this application, indication includes direct indication (also known as explicit indication) and implicit indication. Direct indication of information A refers to including information A; implicit indication of information A refers to indicating information A through the correspondence between information A and information B and the direct indication of information B. The correspondence between information A and information B can be predefined, pre-stored, pre-burned, or pre-configured.

[0338] It should be understood that, in this application, information C is used to determine information D, which includes both information D being determined solely based on information C and information D being determined based on information C and other information. Furthermore, information C can also be used to determine information D indirectly, for example, where information D is determined based on information E, and information E is determined based on information C.

[0339] FIG9 is a schematic diagram of the structure of a device for acquiring training data according to an embodiment of the present application. As shown in FIG9 , the device 1600 may include a processing module 1601 and a communication module 1602 .

[0340] As a first example, the apparatus 1600 can be used in the method for acquiring training data in any of the aforementioned embodiments. For example, the processing module 1601 is used to implement processing-related steps such as acquisition, generation, and determination, and the communication module 1602 is used to implement steps such as sending and / or receiving.

[0341] FIG10 is a schematic diagram of a training data acquisition apparatus according to another embodiment of the present application. As shown in FIG10 , apparatus 1700 includes a processing circuit 1701. The apparatus 1700 may also include a communication circuit 1702. The processing circuit 1701 and the communication circuit 1702 are coupled to each other.

[0342] It can be understood that the processing circuit may be one or more processors, or may be all or part of the circuits of the processing functions in one or more processors.

[0343] It is understandable that the communication circuit 1702 may be a transceiver or an input / output interface.

[0344] Optionally, the apparatus 1700 may further include a memory 1703 for storing instructions executed by the processing circuit 1701 or storing input data required for the processing circuit 1701 to run instructions or storing data generated after the processing circuit 1701 runs instructions.

[0345] It is understandable that the memory 1703 may be located outside the processing circuit 1701 or inside the processing circuit 1701 .

[0346] As an example, the processing circuit 1701 is used to implement the functions of the above-mentioned processing module 1601, and the communication circuit 1702 is used to implement the functions of the above-mentioned communication module 1602.

[0347] As an example, the apparatus 1700 may be a data acquisition device, a database, or a training device, or may be a chip used in a data acquisition device, a database, or a training device.

[0348] As an example, the apparatus 1700 may be a communication device, or a chip used in a communication device.

[0349] When device 1700 is a communication device, the communication circuit can be a transceiver; when device 1700 is a chip, the communication circuit can be an input / output circuit, a bus, a pin or other type of communication interface, wherein the input circuit in the input / output circuit can be used for receiving and the output interface can be used for sending.

[0350] Some embodiments of the present application also provide a computer-readable storage medium, which includes computer instructions. When the computer instructions are executed on a processor, the method implemented by the communication device in any of the above embodiments can be implemented.

[0351] In some embodiments of the present application, a system is also provided that can implement the method implemented by one or more of the data acquisition device, database, training device, and execution device in any of the above embodiments. As an example, the system can be a communication system.

[0352] It is understood that the processor in the embodiments of the present application can be the following devices or all or part of the circuits in the following devices used for processing functions: a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), field programmable gate arrays (FPGA) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. The general-purpose processor can be a microprocessor or any conventional processor.

[0353] The method steps in the embodiments of the present application can be implemented by hardware or by a processor executing software instructions. The software instructions can be composed of corresponding software modules, and the software modules can be stored in random access memory, flash memory, read-only memory, programmable read-only memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, register, hard disk, mobile hard disk, CD-ROM or any other form of storage medium well known in the art. An exemplary storage medium is coupled to the processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be an integral part of the processor. The processor and the storage medium can be located in a chip, such as an application-specific integrated circuit (ASIC), or a chip system, such as a system on a chip (SOC). In addition, the chip or chip system can be located in a network device or a terminal device. Of course, the processor and the storage medium can also be present in a network device or a terminal device as discrete components.

[0354] In the above embodiments, all or part of the embodiments may be implemented using software, hardware, firmware, or any combination thereof. When implemented using software, all or part of the embodiments may be implemented in the form of a computer program product. The computer program product includes one or more computer programs or instructions. When the computer program or instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present application are performed in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user device, or other programmable device. The computer program or instructions may be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another. For example, the computer program or instructions may be transferred from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium, such as a floppy disk, hard disk, or magnetic tape; an optical medium, such as a digital video disk; or a semiconductor medium, such as a solid-state drive.

[0355] In the various embodiments of the present application, unless otherwise specified or there is a logical conflict, the terms and / or descriptions between different embodiments are consistent and can be referenced by each other. The technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationships.

[0356] It is understood that the various numbers used in the embodiments of this application are merely for ease of description and are not intended to limit the scope of the embodiments of this application. The order of the sequence numbers of the above-mentioned processes does not necessarily imply a specific order of execution; the order of execution of the processes should be determined by their functions and inherent logic.

Claims

1. A method for obtaining training data, characterized in that: The method comprises: Obtain T data sequences, each of the T data sequences containing at least one data, where T is an integer greater than 1; Acquire first information, where the first information indicates P data sequences among the T data sequences used to obtain the first data sequence, where P is a positive integer less than or equal to T; A first data sequence is determined based on the first information and the T data sequences, the first data sequence contains M data, M is a positive integer greater than 1, and the first data sequence is used as an input data sequence of the neural network model and / or as a label of the input data sequence of the neural network model, wherein the data in the input data sequence is used to be input into the neural network model in sequence.

2. The method according to claim 1, characterized in that The T data sequences include a first sequence, wherein the first information includes first indication information, and the first indication information indicates: a data sequence among the T data sequences that is spliced ​​with the first sequence.

3. The method according to claim 1, characterized in that The T data sequences include a first sequence and a second sequence, wherein the first information includes second indication information, and the second indication information indicates whether the second sequence is spliced ​​with the first sequence.

4. The method according to claim 1, wherein The first information includes at least one of the following information: a starting sequence identifier of the first data sequence, P, or an ending sequence identifier of the first data sequence.

5. The method according to claim 1, wherein The first information indicates the P data sequences, including: the first information indicates characteristics of the P data sequences.

6. The method according to claim 5, characterized in that The first information includes identification information of each data sequence in the P data sequences.

7. The method according to any one of claims 2 to 6, characterized in that The T data sequences include a first sequence, wherein obtaining the first sequence among the T data sequences includes: Second information is obtained, where the second information indicates at least one of the following information: a start data identifier of the first sequence, a length of the first sequence, or an end data identifier of the first sequence.

8. The method according to claim 7, characterized in that The second information further indicates an initial value of the model state corresponding to the first sequence.

9. The method according to any one of claims 1 to 8, characterized in that The data in the first data sequence includes channel state information.

10. The method according to any one of claims 1 to 9, characterized in that The method further comprises: First network state information is obtained, where the first network state information indicates an initial model state value used by the neural network model when obtaining an output data sequence corresponding to the first data sequence.

11. The method according to any one of claims 1 to 10, characterized in that The neural network model is used for compressing and / or predicting channel state information.

12. The method according to any one of claims 1 to 11, characterized in that The T data sequences include a third sequence, wherein obtaining the third sequence from the T data sequences includes: Acquire third information, where the third information indicates at least one of the following information: a start data identifier of the third sequence, a length of the third sequence, or an end data identifier of the third sequence.

13. The method according to claim 12, characterized in that The third information also indicates the initial state value of the model corresponding to the third sequence.

14. The method according to any one of claims 1 to 13, characterized in that In the case where the first data sequence is used as an input data sequence of the neural network model, the method further comprises: Obtain S data sequences, each of the S data sequences containing at least one data, where S is an integer greater than 1; A second data sequence is determined based on the S data sequences, where the second data sequence contains Q data, where Q is a positive integer greater than 1, and the second data sequence is used as a label for the first data sequence, wherein multiple data in the first data sequence are used to be sequentially input into the neural network model.

15. The method according to claim 14, characterized in that Each data in each data sequence in the S data sequences is associated with at least one data in the T data sequences.

16. The method according to claim 14 or 15, characterized in that The S data sequences include a fourth sequence, wherein obtaining the fourth sequence from the S data sequences includes: Fourth information is obtained, where the fourth information indicates at least one of the following information: a start data identifier of the fourth sequence, a length of the fourth sequence, or an end data identifier of the fourth sequence.

17. The method according to claim 16, characterized in that The fourth information also indicates the initial value of the model state corresponding to the fourth sequence.

18. The method according to claim 16 or 17, characterized in that The S data sequences further include a fifth sequence, wherein obtaining the fifth sequence from the S data sequences includes: Acquire fifth information, where the fifth information indicates at least one of the following information: a start data identifier of the fifth sequence, a length of the fifth sequence, or an end data identifier of the fifth sequence.

19. The method according to claim 18, characterized in that The fifth information also indicates the initial value of the model state corresponding to the fifth sequence.

20. A device for acquiring training data, characterized in that: The method comprises functional modules for implementing the method according to any one of claims 1 to 19.

21. A device for acquiring training data, characterized in that: include: One or more processors, configured to implement the method according to any one of claims 1 to 19.

22. A training system for a neural network model, characterized in that: The system comprises a device for acquiring training data, and the device is used to implement the method according to any one of claims 1 to 19.

23. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 19 when executed by a processor.

24. A computer program product, characterized in that The invention comprises a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 19.

Images