A data annotation method, device and equipment in an AI model

Abstract

The present invention discloses a data annotation method, device and equipment in an AI model, the method comprising: obtaining and recording chat records between a user and a customer service staff, generating multiple paired text data samples, recorded as text data sample pairs; performing a word segmentation operation on the text data sample pairs to generate word data sample pairs; performing reliability identification on the word data sample pairs; obtaining text data pairs corresponding to reliable word data samples for annotation, and using the annotated text data pairs for AI model training. The embodiments of the present invention can achieve the online acquisition of a large number of reliable text data sample pairs, which can enhance the manual collection and annotation of data samples, so that the number of training samples is further increased, and the generalization ability of the training model is improved.

Description

Technical Field

[0001] The present invention relates to the field of data annotation technology, and in particular to a data annotation method, device and equipment in an AI model. Background Art

[0002] The current text data enhancement and annotation work is widely used in AI customer service, question answering and other systems, mainly for the training of AI text models. However, the training of AI models usually requires a large number of data samples, which mainly come from the paired annotation of data by specialized annotators, which consumes a lot of manpower costs. Summary of the invention

[0003] The purpose of the present invention is to provide a data labeling method, device and equipment in an AI model, aiming to solve the technical problem in the prior art that the training of AI models usually requires a large number of data samples, such data mainly comes from specialized labeling personnel to perform data pairing labeling, which requires a lot of manpower costs.

[0004] The technical solution of the present invention is as follows:

[0005] A data annotation method in an AI model, the method comprising:

[0006] Obtain and record the chat records between the user and the customer service staff to generate multiple paired text data samples, which are recorded as text data sample pairs;

[0007] Performing a word segmentation operation on the text data sample pairs to generate word data sample pairs;

[0008] Conduct reliability identification on word data sample pairs;

[0009] Obtain text data pairs corresponding to reliable word data samples for annotation, and use the annotated text data pairs for AI model training.

[0010] Furthermore, after the reliability identification of the word data sample pair is performed, the method further includes:

[0011] The text data pairs judged to be unreliable are deleted.

[0012] Further preferably, the acquiring and recording of the chat records between the user and the customer service personnel to generate a plurality of paired text data samples, recorded as text data sample pairs, includes:

[0013] Obtain and record the chat history between the user and the customer service staff;

[0014] The user's chat record is recorded as the first data, and the customer service staff's chat record is recorded as the second data;

[0015] A large number of paired text data samples are generated according to the first data and the second data, and are recorded as text data sample pairs.

[0016] Further preferably, the generating word data sample pairs after performing word segmentation operation on the text data sample pairs comprises:

[0017] Perform keyword splitting operation on the text in the text data sample pair to generate corresponding word data sample pairs.

[0018] Preferably, performing a keyword splitting operation on the text in the text data sample pair to generate corresponding word data sample pairs includes:

[0019] Performing a keyword splitting operation on a first data sample corresponding to the first data in the text data sample pair to generate a first word data sample;

[0020] Performing a keyword splitting operation on a second data sample corresponding to the second data in the text data sample to generate a second word data sample;

[0021] According to the first word data sample and the second word data sample, a corresponding word data sample pair is generated.

[0022] Furthermore, the reliability identification of the word data sample pair includes:

[0023] Performing a clustering operation on the first word data sample to obtain the number of categories of the first word data sample corresponding to the second word data sample;

[0024] sorting the proportion of each category in the second word data sample to the second word data sample;

[0025] When it is detected that the proportion of a certain category in the second word data sample meets a preset condition, the word data sample pair formed by the first word data sample and the current category in the second word data sample is determined to be a reliable word data sample pair.

[0026] Furthermore, when it is detected that the proportion of a certain category in the second word data sample is greater than a preset threshold and greater than the proportion of any other category, the word data sample pair formed by the first word data sample and the current category in the second word data sample is determined to be a reliable word data sample pair.

[0027] Another embodiment of the present invention provides a data annotation device in an AI model, the device comprising:

[0028] A text data pair generation module is used to obtain and record the chat records between the user and the customer service staff, and generate multiple paired text data samples, which are recorded as text data sample pairs;

[0029] A word segmentation operation module, used to generate word data sample pairs after performing word segmentation operation on the text data sample pairs;

[0030] A reliability identification module, used for performing reliability identification on word data sample pairs;

[0031] The data annotation module is used to obtain text data pairs corresponding to reliable word data samples for annotation, and use the annotated text data pairs for AI model training.

[0032] Another embodiment of the present invention provides a data annotation device in an AI model, the device comprising at least one processor; and

[0033] a memory communicatively connected to the at least one processor; wherein,

[0034] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the data labeling method in the above-mentioned AI model.

[0035] Another embodiment of the present invention further provides a non-volatile computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are executed by one or more processors, the one or more processors can execute the data labeling method in the above-mentioned AI model.

[0036] Beneficial effects: The embodiments of the present invention can realize the online acquisition of a large number of reliable text data sample pairs, which can enhance the manual collection and labeling of data samples, so that the number of training samples is further increased, and the generalization ability of the training model is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:

[0038] Figure 1 A flowchart of a preferred embodiment of a data labeling method in an AI model of the present invention;

[0039] Figure 2 A schematic diagram of a sample collection process in a data labeling method in an AI model of the present invention;

[0040] Figure 3 This is a schematic diagram of the sample screening process in a data labeling method in an AI model of the present invention;

[0041] Figure 4 A schematic diagram of functional modules of a preferred embodiment of a data labeling device in an AI model of the present invention;

[0042] Figure 5 A schematic diagram of the hardware structure of a preferred embodiment of a data labeling device in an AI model of the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solution and effect of the present invention clearer and more specific, the present invention is further described in detail below. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. The embodiments of the present invention are described below in conjunction with the accompanying drawings.

[0044] The embodiment of the present invention provides a data annotation method in an AI model. Figure 1 , Figure 1 Flow chart of a preferred embodiment of a data labeling method in an AI model of the present invention. Figure 1 As shown, it includes the steps of:

[0045] Step S100: Obtain and record the chat records between the user and the customer service staff to generate a plurality of paired text data samples, which are recorded as text data sample pairs;

[0046] Step S200, performing a word segmentation operation on the text data sample pair to generate a word data sample pair;

[0047] Step S300, performing reliability identification on word data sample pairs;

[0048] Step S400: obtain text data pairs corresponding to reliable word data samples for annotation, and use the annotated text data pairs for AI model training.

[0049] In specific implementation, taking AI customer service and question-and-answer systems as an example, customer service personnel can be the system's sales, front desk customer service, class teachers and other back-end personnel. In the actual online chat and question-and-answer system, the chat records of both parties are recorded in real time. Since the records come from the actual online chat and question-and-answer process, the data can be considered to be authentic and reliable. This is equivalent to completing the data labeling process invisibly during the chat and question-and-answer process.

[0050] In the actual implementation process, the interactive chat information between customer service personnel (sales, customer service, class teachers, etc.) and users is recorded to form a large number of paired text data samples. The paired text data samples are segmented by keywords, and some meaningless modal particles and other words are removed to generate a large number of paired word data sample pairs. The word data sample pairs are identified for reliability based on the clustering algorithm, and the results are annotated as reliable data for subsequent AI model training.

[0051] A large number of reliable text data sample pairs are available online. These data can enhance the manual collection and labeling of data samples, further increase the number of training samples, and improve the generalization ability of the training model.

[0052] Furthermore, after reliability identification of the word data sample pair, the following is also included:

[0053] The text data pairs judged to be unreliable are deleted.

[0054] In specific implementation, text data pairs that are judged to be unreliable in reliability identification are deleted, thereby reducing the number of samples and reducing the burden on the server.

[0055] Furthermore, the chat records between the user and the customer service staff are obtained and recorded to generate multiple paired text data samples, which are recorded as text data sample pairs, including:

[0056] Obtain and record the chat history between the user and the customer service staff;

[0057] The user's chat record is recorded as the first data, and the customer service staff's chat record is recorded as the second data;

[0058] A large number of paired text data samples are generated according to the first data and the second data, and are recorded as text data sample pairs.

[0059] In specific implementation, when a user chats with a customer service staff, the chat record is recorded in real time, the chat message sent by the user is recorded as the first data, and the chat message sent by the customer service staff who is talking to the user is recorded as the second data, thereby obtaining a text data sample pair corresponding to the user and the customer service staff.

[0060] Furthermore, performing a word segmentation operation on the text data sample pairs to generate word data sample pairs includes:

[0061] Perform keyword splitting operation on the text in the text data sample pair to generate corresponding word data sample pairs.

[0062] In the specific implementation, the collected text data is segmented to obtain text data samples to word data sample pairs. By splitting a text of the text data sample pair into several keywords, it is convenient to process.

[0063] Furthermore, a keyword splitting operation is performed on the text in the text data sample pair to generate corresponding word data sample pairs, including:

[0064] Performing a keyword splitting operation on a first data sample corresponding to the first data in the text data sample pair to generate a first word data sample;

[0065] Performing a keyword splitting operation on a second data sample corresponding to the second data in the text data sample to generate a second word data sample;

[0066] According to the first word data sample and the second word data sample, a corresponding word data sample pair is generated.

[0067] When implementing it, Figure 2 As shown in , by recording the interactive chat information between customer service personnel (sales, customer service, class teachers, etc.) and users, a large number of paired text data samples can be formed. Figure 2 The text A and text B parts in the chat can be simply understood as a question and answer in the chat process, which is an AB data pair.

[0068] Furthermore, reliability identification is performed on the word data sample pairs, including:

[0069] Performing a clustering operation on the first word data sample to obtain the number of categories of the first word data sample corresponding to the second word data sample;

[0070] sorting the proportion of each category in the second word data sample to the second word data sample;

[0071] When it is detected that the proportion of a certain category in the second word data sample meets a preset condition, the word data sample pair formed by the first word data sample and the current category in the second word data sample is determined to be a reliable word data sample pair.

[0072] In the specific implementation, clustering is performed according to the A part of the sample pair to obtain the situation of all the B parts corresponding to the A part ( Figure 3 B1, B2, B3...BN, N is the number of all categories of the corresponding B part when the A part is the same, and the categories are classified according to keywords), and the proportion of each B situation to all B situations is ranked (the first is B1, the second is B2, and the last is BN). If the proportion of B1 meets the preset conditions, the A-B1 data sample pair in the data is determined to be a reliable data sample pair and can be used for model training, otherwise it is unreliable.

[0073] Further, when it is detected that the proportion of a certain category in the second word data sample meets a preset condition, it is determined that the word data sample pair formed by the first word data sample and the current category in the second word data sample is a reliable word data sample pair, including:

[0074] When it is detected that the proportion of a certain category in the second word data sample is greater than a preset threshold and greater than the proportion of any other category, the word data sample pair formed by the first word data sample and the current category in the second word data sample is determined to be a reliable word data sample pair.

[0075] In the specific implementation, taking the second word data sample as case B1 as an example, if the proportion of B1 is greater than 50% and the proportion of B1 is greater than 2 times the proportion of B2, and B2 is the category with the largest proportion except B1, then the word data sample pair formed by the first word data sample and the current category in the second word data sample is determined to be a reliable word data sample pair, which can be used for model training.

[0076] It can be seen from the above method embodiments that the embodiments of the present invention can realize the real-time recording of the chat records of both parties in the actual online chat and question-answering system, taking the AI ​​customer service and question-answering system as an example. Since the records come from the actual online chat and question-answering process, the data can be considered to be authentic and reliable. This is equivalent to completing the data labeling process invisibly during the chat and question-answering process, and using these labeled data in the training of the actual customer service and question-answering models, which plays a role in data enhancement. The entire process does not require manual intervention to construct data, and there is no labeling step, which saves labor costs.

[0077] It should be noted that there is not necessarily a certain order between the above-mentioned steps. A person skilled in the art can understand, based on the description of the embodiments of the present invention, that in different embodiments, the above-mentioned steps may have different execution orders, that is, they may be executed in parallel, may be executed interchangeably, and so on.

[0078] Another embodiment of the present invention provides a data annotation device in an AI model, such as Figure 4 As shown, the device 1 comprises:

[0079] The text data pair generation module 11 is used to obtain and record the chat records between the user and the customer service personnel, and generate a plurality of paired text data samples, which are recorded as text data sample pairs;

[0080] A word segmentation operation module 12, used to generate word data sample pairs after performing word segmentation operation on the text data sample pairs;

[0081] A reliability identification module 13, used for performing reliability identification on word data sample pairs;

[0082] The data annotation module 14 is used to obtain text data pairs corresponding to reliable word data samples for annotation, and use the annotated text data pairs for AI model training.

[0083] The specific implementation method is shown in the method embodiment, which will not be described in detail here.

[0084] Furthermore, the device is also provided with a data deletion module for deleting text data pairs determined to be unreliable.

[0085] The specific implementation method is shown in the method embodiment, which will not be described in detail here.

[0086] Furthermore, the text data pair generation module 11 is also used to obtain and record the chat records between the user and the customer service personnel;

[0087] The user's chat record is recorded as the first data, and the customer service staff's chat record is recorded as the second data;

[0088] A large number of paired text data samples are generated according to the first data and the second data, and are recorded as text data sample pairs.

[0089] The specific implementation method is shown in the method embodiment, which will not be described in detail here.

[0090] Furthermore, the word segmentation operation module 12 is also used to perform a keyword segmentation operation on the text in the text data sample pair to generate a corresponding word data sample pair.

[0091] The specific implementation method is shown in the method embodiment, which will not be described in detail here.

[0092] Furthermore, the word segmentation operation module 12 is further used to perform a keyword splitting operation on the first data sample corresponding to the first data in the text data sample pair to generate a first word data sample;

[0093] Performing a keyword splitting operation on a second data sample corresponding to the second data in the text data sample to generate a second word data sample;

[0094] According to the first word data sample and the second word data sample, a corresponding word data sample pair is generated.

[0095] The specific implementation method is shown in the method embodiment, which will not be described in detail here.

[0096] Furthermore, the reliability identification module 13 is also used to perform a clustering operation on the first word data sample to obtain the number of categories of the first word data sample corresponding to the second word data sample;

[0097] sorting the proportion of each category in the second word data sample to the second word data sample;

[0098] When it is detected that the proportion of a certain category in the second word data sample meets a preset condition, the word data sample pair formed by the first word data sample and the current category in the second word data sample is determined to be a reliable word data sample pair.

[0099] The specific implementation method is shown in the method embodiment, which will not be described in detail here.

[0100] Furthermore, the reliability identification module 13 is also used to determine that the word data sample pair formed by the first word data sample and the current category in the second word data sample is a reliable word data sample pair when it is detected that the proportion of a certain category in the second word data sample is greater than a preset threshold and greater than the proportion of any other category.

[0101] The specific implementation method is shown in the method embodiment, which will not be described in detail here.

[0102] Another embodiment of the present invention provides a data annotation device in an AI model, such as Figure 5 As shown, the device 10 includes:

[0103] One or more processors 110 and memory 120, Figure 5 A processor 110 is used as an example for description. The processor 110 and the memory 120 may be connected via a bus or other means. Figure 5 The example of connecting through bus is taken in the following.

[0104] The processor 110 is used to complete various control logics of the device 10, which can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a single-chip microcomputer, an ARM (Acorn RISCMachine) or other programmable logic device, a discrete gate or transistor logic, a discrete hardware component, or any combination of these components. In addition, the processor 110 can also be any traditional processor, microprocessor or state machine. The processor 110 can also be implemented as a combination of computing devices, for example, a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors combined with a DSP core, or any other such configuration.

[0105] The memory 120, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions corresponding to the data labeling method in the AI ​​model in the embodiment of the present invention. The processor 110 executes various functional applications and data processing of the device 10 by running the non-volatile software programs, instructions and units stored in the memory 120, that is, implements the data labeling method in the AI ​​model in the above method embodiment.

[0106] The memory 120 may include a program storage area and a data storage area, wherein the program storage area may store applications required for operating the device and at least one function; the data storage area may store data created according to the use of the device 10, etc. In addition, the memory 120 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 120 may optionally include a memory remotely arranged relative to the processor 110, and these remote memories may be connected to the device 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0107] One or more units are stored in the memory 120, and when executed by one or more processors 110, the data labeling method in the AI ​​model in any of the above method embodiments is executed, for example, the data labeling method in the AI ​​model described above is executed. Figure 1 The method comprises steps S100 to S400.

[0108] An embodiment of the present invention provides a non-volatile computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are executed by one or more processors, for example, to execute the above-described Figure 1 The method comprises steps S100 to S400.

[0109] As an example, the non-volatile storage medium can include a read-only memory (ROM), a programmable ROM (PROM), an electrically programmable ROM (EPROM), an electrically erasable ROM (EEPROM), or a flash memory. Volatile memory can include a random access memory (RAM) as an external cache memory. By way of illustration and not limitation, RAM can be obtained in many forms such as synchronous RAM (SRAM), dynamic RAM, (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), and direct Rambus RAM (DRRAM). The disclosed memory components or memories of the operating environment described herein are intended to include one or more of these and / or any other suitable types of memory.

[0110] Another embodiment of the present invention provides a computer program product, the computer program product includes a computer program stored on a non-volatile computer-readable storage medium, the computer program includes program instructions, when the program instructions are executed by a processor, the processor executes the data labeling method in the AI ​​model of the above method embodiment. For example, executing the above described Figure 1 The method comprises steps S100 to S400.

[0111] The embodiments described above are merely illustrative, in which the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0112] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a general hardware platform, and of course can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can exist in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or some parts of the embodiment.

[0113] Conditional language such as "can," "can," "might," or "may," among other things, unless specifically stated otherwise or otherwise understood within the context as used, is generally intended to convey that particular embodiments can include (while other embodiments do not) particular features, elements, and / or operations. Thus, such conditional language is also generally intended to imply that features, elements, and / or operations are anyway required for one or more embodiments or that one or more embodiments must include logic for determining, with or without input or prompting, whether such features, elements, and / or operations are included or will be performed in any particular embodiment.

[0114] What has been described herein in this specification and the accompanying drawings includes examples of data annotation methods and devices in AI models. Of course, it is not possible to describe every conceivable combination of elements and / or methods for the purpose of describing the various features of the present disclosure, but it is recognized that many other combinations and permutations of the disclosed features are possible. Therefore, it is obvious that various modifications can be made to the present disclosure without departing from the scope or spirit of the present disclosure. In addition, or in an alternative, other embodiments of the present disclosure may be apparent from consideration of the present specification and the accompanying drawings and from the practice of the present disclosure as presented herein. It is intended that the examples presented in this specification and the accompanying drawings are considered to be illustrative and not restrictive in all respects. Although specific terms are employed herein, they are used in a general and descriptive sense and are not used for limiting purposes.

Claims

1. A data annotation method in an AI model, characterized in that , the method comprises: Obtain and record the chat records between the user and the customer service staff to generate multiple paired text data samples, which are recorded as text data sample pairs; Performing a word segmentation operation on the text data sample pairs to generate word data sample pairs; Conduct reliability identification on word data sample pairs; Obtain text data pairs corresponding to reliable word data samples for annotation, and use the annotated text data pairs for AI model training; The step of generating word data sample pairs after performing word segmentation operation on the text data sample pairs comprises: Perform keyword splitting operation on the text in the text data sample pair to generate corresponding word data sample pairs; The step of performing a keyword splitting operation on the text in the text data sample pair to generate a corresponding word data sample pair includes: Performing a keyword splitting operation on a first data sample corresponding to the first data in the text data sample pair to generate a first word data sample; Performing a keyword splitting operation on a second data sample corresponding to the second data in the text data sample to generate a second word data sample; generating a corresponding word data sample pair according to the first word data sample and the second word data sample; The reliability identification of the word data sample pair includes: Performing a clustering operation on the first word data sample to obtain the number of categories of the first word data sample corresponding to the second word data sample; sorting the proportion of each category in the second word data sample to the second word data sample; When it is detected that the proportion of a certain category in the second word data sample meets a preset condition, the word data sample pair formed by the first word data sample and the current category in the second word data sample is determined to be a reliable word data sample pair.

2. The data annotation method in the AI ​​model according to claim 1, characterized in that: After the reliability identification of the word data sample pair is performed, the method further includes: The text data pairs judged to be unreliable are deleted.

3. The data annotation method in the AI ​​model according to claim 2, characterized in that: The chat records between the user and the customer service personnel are obtained and recorded to generate multiple paired text data samples, which are recorded as text data sample pairs, including: Obtain and record the chat history between the user and the customer service staff; The user's chat record is recorded as the first data, and the customer service staff's chat record is recorded as the second data; A large number of paired text data samples are generated according to the first data and the second data, and are recorded as text data sample pairs.

4. The data annotation method in the AI ​​model according to claim 1, characterized in that: When it is detected that the proportion of a certain category in the second word data sample meets a preset condition, determining that the word data sample pair formed by the first word data sample and the current category in the second word data sample is a reliable word data sample pair includes: When it is detected that the proportion of a certain category in the second word data sample is greater than a preset threshold and greater than the proportion of any other category, the word data sample pair formed by the first word data sample and the current category in the second word data sample is determined to be a reliable word data sample pair.

5. A data annotation device in an AI model, characterized in that: The device comprises: A text data pair generation module is used to obtain and record the chat records between the user and the customer service staff, and generate multiple paired text data samples, which are recorded as text data sample pairs; A word segmentation operation module, used to generate word data sample pairs after performing word segmentation operation on the text data sample pairs; A reliability identification module, used for performing reliability identification on word data sample pairs; The data annotation module is used to obtain text data pairs corresponding to reliable word data samples for annotation, and use the annotated text data pairs for AI model training; The step of generating word data sample pairs after performing word segmentation operation on the text data sample pairs comprises: Perform keyword splitting operation on the text in the text data sample pair to generate corresponding word data sample pairs; The step of performing a keyword splitting operation on the text in the text data sample pair to generate a corresponding word data sample pair includes: Performing a keyword splitting operation on a first data sample corresponding to the first data in the text data sample pair to generate a first word data sample; Performing a keyword splitting operation on a second data sample corresponding to the second data in the text data sample to generate a second word data sample; generating a corresponding word data sample pair according to the first word data sample and the second word data sample; The reliability identification of the word data sample pair includes: Performing a clustering operation on the first word data sample to obtain the number of categories of the first word data sample corresponding to the second word data sample; sorting the proportion of each category in the second word data sample to the second word data sample; When it is detected that the proportion of a certain category in the second word data sample meets a preset condition, the word data sample pair formed by the first word data sample and the current category in the second word data sample is determined to be a reliable word data sample pair.

6. A data annotation device in an AI model, characterized in that: The device comprises at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the data labeling method in the AI ​​model described in any one of claims 1-4.

7. A non-volatile computer-readable storage medium, characterized in that: The non-volatile computer-readable storage medium stores computer-executable instructions, which, when executed by one or more processors, enable the one or more processors to execute the data labeling method in the AI ​​model described in any one of claims 1-4.

Images