Text content tag generation method and apparatus, and medium and electronic device

Abstract

The present invention relates to the technical field of data processing. Specifically provided are a text content tag generation method and apparatus, and a medium and an electronic device, which aims to solve the problem of cold start of existing tags and tag data. For this purpose, the text content tag generation method of the present invention comprises: acquiring a target text to be tagged; on the basis of the target text, generating recommended tags by means of a preset tagging model and displaying the recommended tags; determining whether the recommended tags have been modified; if the recommended tags have been modified, performing quality inspection on the modified tags; and if the modified tags pass the quality inspection, using the modified tags as tags of the target text, and storing in a database the correspondence between the modified tags and the target text. The correspondence between the modified tags and the target text can be stored in the database, thereby facilitating tagging when similar target texts appear and also avoiding the problem of cold start of tag data caused by the lack of suitable tags.

Description

Text content tag generation method, device, medium and electronic device Priority claim

[0001] This application claims priority to Chinese patent application CN202410161023.4, filed on February 4, 2024, entitled “Text content label generation method, device, medium and electronic device”. The entire contents of the above Chinese patent application are incorporated into this application by reference. Technical Field

[0002] The present invention relates to the field of data processing technology, and specifically provides a method, device, medium and electronic device for generating text content tags. Background Art

[0003] In the internet age, users generate content on platforms, which then understand and distribute it to other users in a personalized manner. Content tagging is a primary method for understanding content. Its advantages are structured and interpretable, and it can generally build multiple levels of tags to cover a wide range of content. Content tagging systems are platforms that tag content, demonstrating their significant value.

[0004] Traditional content tagging systems follow a strict synchronous process: users create content on the platform, and the platform then tags it based on its understanding of the content. Depending on whether users can perceive the platform's tagging results, tagging can be categorized as back-end or front-end. Back-end tagging is a completely automated process performed in the background of the system, leaving users unaware of the tagging process and the results. Front-end tagging involves the system providing tagging results after users have finished editing content and are preparing to publish it. These results are displayed on the content editing page, allowing users to perceive the system's tagging results. However, both back-end and front-end tagging are performed independently by the tagging system, resulting in a cold start problem for both labels and tagged data. This requires defining the labels and manually labeling samples for model training. If new labels are not discovered promptly, a significant amount of content will not be covered by existing labels. Even if new labels are discovered, manual collection of sufficient samples is required to train the model.

[0005] Accordingly, this field requires a new technical solution to solve the above problems. Summary of the Invention

[0006] The present invention aims to solve the above technical problems and resolve the cold start problem of existing tags and tag data.

[0007] The present invention provides a method for generating a text content label, the method comprising:

[0008] Get the target text to be marked;

[0009] Based on the target text, generate recommended tags through a preset tagging model and display the recommended tags;

[0010] Determining whether the recommended tag has been modified;

[0011] If the recommended label is modified, then quality inspection is performed on the modified label;

[0012] If the quality inspection is passed, the modified label is used as the label of the target text, and the corresponding relationship between the modified label and the target text is stored in a database.

[0013] When using the above technical solution, the tagging model can generate and display recommended tags. Users can select the recommended tags as the tags for the target text, avoiding a cold start of the tag and label data. Furthermore, if the user modifies the recommended tags, the modified tags are quality-checked. By using the modified tags as the tags for the target text, the target text can be annotated with the new tags that passed the quality check. The correspondence between the modified tags and the target text is stored in the database, facilitating the tagging of similar target texts when they appear, while also avoiding the problem of cold starts of the label data caused by a lack of suitable tags.

[0014] In a specific embodiment of the above-mentioned method for generating text content tags, the step of "generating recommended tags based on the target text using a preset tagging model and displaying the recommended tags" further includes:

[0015] Based on the target text, one or more first candidate tags are selected from a preset tag library;

[0016] One or more of the first candidate tags are displayed to the user as the recommended tags.

[0017] In a specific embodiment of the above-mentioned method for generating text content labels, the step of "if the recommended label is modified, then performing quality inspection on the modified label" further includes:

[0018] Determine whether the modified tag belongs to the tag library;

[0019] If the modified label belongs to the label library, the modified label is quality-checked using a preset quality-check model.

[0020] When the above technical solution is adopted, a quality inspection model is used to inspect the modified labels to ensure that the modified labels correspond to the target text.

[0021] In a specific embodiment of the above-mentioned method for generating text content labels, the step of "if the recommended label is modified, then performing quality inspection on the modified label" further includes: if the modified label does not belong to the label library, then recommending a second candidate label based on the modified label by a preset machine recommendation model;

[0022] Determining whether the second candidate label is adopted;

[0023] Based on the judgment results, quality inspection is carried out again.

[0024] When adopting the above technical solution, if the modified label does not belong to the label library, other labels can be recommended through the machine recommendation model, so that the label can accurately represent the text content and comply with the label specifications.

[0025] In a specific embodiment of the above-mentioned method for generating text content labels, the step of "performing quality inspection again based on the judgment result" further includes:

[0026] If the second candidate label is adopted, performing quality inspection on the second candidate label using the quality inspection model;

[0027] If the second candidate label is not adopted, manual quality inspection is performed on the modified label.

[0028] In a specific embodiment of the above-mentioned method for generating text content labels, the step of “performing quality inspection on the second candidate label by the quality inspection model” further includes:

[0029] The target text is compared with the second candidate label by using the quality inspection model. If the target text and the second candidate label correspond to each other, the quality inspection passes.

[0030] In a specific implementation of the above text content tag generation method, the method further includes:

[0031] If the quality inspection is passed, the second candidate tag is added to the tag library.

[0032] When the above technical solution is adopted, the second candidate label that passes the quality inspection can be added to the label library, so that the labels in the label library can cover more content and can be more accurate when annotating text content.

[0033] In a specific embodiment of the above-mentioned method for generating text content tags, after the step of “determining whether the recommended tag has been modified”, the following steps are further included:

[0034] If the recommended tag is not modified, the recommended tag is used as the tag of the target text.

[0035] In a specific embodiment of the above-mentioned method for generating text content tags, the step of "generating recommended tags based on the target text using a preset tagging model and displaying the recommended tags" further includes:

[0036] During the target text editing process, the first candidate tag is synchronously selected on the backend through the tagging model;

[0037] After the target text is edited, the first alternative label is displayed.

[0038] When the above technical solution is adopted, the tagging model tags the target text being edited synchronously, the tagging speed is faster, and the recommended tags can be quickly displayed when the editing is completed.

[0039] The present invention also discloses a text content label generation device, comprising:

[0040] an acquisition module, configured to acquire a target text to be marked;

[0041] A recommendation module, configured to generate a recommended tag based on the target text using a preset tagging model and display the recommended tag;

[0042] a determination module configured to determine whether the recommendation tag has been modified;

[0043] a quality inspection module, configured to perform a quality inspection on the modified label if the recommended label is modified;

[0044] The tagging module is configured to use the modified label as a label of the target text if the quality inspection is passed.

[0045] When using the above technical solution, the tagging model can generate and display recommended tags. Users can select the recommended tags as the tags for the target text, avoiding a cold start of the tag and label data. Furthermore, if the user modifies the recommended tags, the modified tags are quality-checked. By using the modified tags as the tags for the target text, the target text can be annotated with the new tags that have passed quality inspection, thus avoiding the problem of cold starting the label data due to the lack of suitable tags.

[0046] The present invention also discloses an electronic device, comprising:

[0047] processor;

[0048] Memory, used to store programs,

[0049] When the program is executed by the processor, the processor implements the above-mentioned method for generating text content tags.

[0050] The present invention also discloses a computer-readable medium on which a computer program is stored. When the computer program is executed by a processor, the method for generating text content tags is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The preferred embodiments of the present invention are described below with reference to the accompanying drawings, in which:

[0052] FIG1 is a flow chart of the main steps of the text content label generation method;

[0053] FIG. 2 is a flow chart of the expanded steps of step S104 in the text content tag generation method.

[0054] FIG3 is a flow chart of a possible implementation of a method for generating text content tags;

[0055] FIG4 is a schematic diagram of the main modules of the label generating device.

[0056] List of reference numerals: 300 - label generating device; 301 - acquisition module; 302 - recommendation module; 303 - judgment module; 304 - quality inspection module; 305 - labeling module. DETAILED DESCRIPTION

[0057] The preferred embodiments of the present application are described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely intended to illustrate the technical principles of the present application and are not intended to limit the scope of protection of the present application. Those skilled in the art may adjust these embodiments as needed to suit specific applications.

[0058] It should be noted that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance. In addition, "plurality" in this application means at least two.

[0059] First, referring to FIG. 1 , the text content tag generation method of the present application is introduced.

[0060] As shown in FIG1 , in order to solve the cold start problem of existing labels and label data, the present invention provides a method for generating text content labels, which includes:

[0061] S101, obtaining target text to be marked; for example, a user enters text on a smart terminal device, and the text is obtained while the user is typing. In another example, the text can be obtained after the user completes the input. The smart terminal device can be a mobile phone, tablet computer, laptop computer, desktop computer, or other terminal device with input function.

[0062] S102, based on the target text, generate recommended tags through a preset tagging model and display the recommended tags; in this application, the tagging model is used to characterize the correspondence between the recommended tags and the target text. For example, the preset tagging model is preferably a trained large language model. The target text is input into the tagging model. The tagging model can generate recommended tags that fit the target text and display the recommended tags on the user's smart terminal device. Of course, the preset tagging model can also adopt a machine learning model. The trained large language model can be Chat-GPT, Wenxin Yiyan or Pangu, etc.

[0063] S103, determine whether the recommended tag has been modified; for example, during the interaction process, the user can modify the displayed preset tag, or directly select the recommended tag to use. After the modification or selection is completed, determine whether the recommended tag has been modified to facilitate subsequent operations.

[0064] S104: If the recommended tag is modified, perform a quality check on the modified tag; for example, if the recommended tag is modified, perform a quality check on whether the target text and the modified tag match, and output the quality check result.

[0065] S105: If the quality inspection passes, the modified label is used as the label of the target text, and the corresponding relationship between the modified label and the target text is stored in the database. For example, if the target text and the modified label match each other after the quality inspection, the modified label is used as the label of the target text, and the corresponding relationship between the modified label and the target text is stored in the database, further ensuring that the target text can use the label that best matches the target text.

[0066] In this way, the tagging model can generate and display recommended tags. Users can select the recommended tags as the tags for the target text, avoiding a cold start in the tag and label data. Furthermore, if the user modifies the recommended tags, the modified tags are quality-checked. By using the modified tags as the tags for the target text, the target text can be annotated with the new tags that passed quality inspection. The correspondence between the modified tags and the target text is stored in the database, making it easier to annotate similar target text when it appears. This also avoids the problem of cold starts in the label data caused by a lack of suitable tags.

[0067] In a preferred embodiment, the step of “generating recommended tags based on the target text using a preset tagging model and displaying the recommended tags” includes:

[0068] Based on the target text, one or more first candidate tags are selected from the preset tag library. For example, the target text input by the user is "I made a braised pork dish in the kitchen today. It tastes very good. I miss the delicious food from my childhood." The tags in the preset tag library include: food, health, life, home, finance, sports, music, emotion, entertainment, film and television, fashion, and games. Through the labeling model, "food" and "life" are selected as the first candidate tags for the target text in the tag library.

[0069] One or more first candidate tags are displayed to the user as recommended tags. For example, after determining "food" and "life" as the first candidate tags, the "food" and "life" tags are displayed on the user's smart terminal device and can be selected by the user.

[0070] After the recommended tags are displayed to the user, it is determined whether the recommended tags are adopted. For example, the user can adopt one or two of the recommended tags. Of course, the user can also choose not to adopt the recommended tags and choose to create a new tag or modify the existing tag.

[0071] If a user selects one or more recommended tags, the target text will be tagged. If the user does not select a recommended tag and does not create a new tag, the target text will not be tagged. This way, tagged target texts can be recommended to people interested in the tagged content, increasing their exposure. Untagged target texts will be randomly recommended to other users, reducing their exposure.

[0072] It should be noted that the above steps are preferred and those skilled in the art may optimize the above steps according to specific application scenarios. For example, the number of first candidate labels may be one, two, three, or more. For another example, the user may select one, two, three, or more first candidate labels as labels for the target text.

[0073] Furthermore, the step of "generating recommended tags based on the target text using a preset tagging model and displaying the recommended tags" also includes:

[0074] During the target text editing process, the tagging model synchronously selects the first candidate tag on the backend. For example, if a user pauses for more than one second while entering text, the tagging model will tag the text once. Another example is if a user enters punctuation or hits Enter, the tagging model will tag the text once.

[0075] After the target text is edited, the tagging model generates a recommended tag and displays the first candidate tag. For example, after the user completes editing and clicks the "Done" button, the tagging model generates a recommended tag that matches the target text and displays the recommended tag. For another example, after the user completes editing and clicks the "Tag" icon, the tagging model generates a recommended tag that matches the target text and displays the recommended tag.

[0076] As shown in FIG2 , in a preferred embodiment, the step of “if the recommended label is modified, performing quality inspection on the modified label” further includes:

[0077] S1041: Determine whether the modified tag belongs to the tag library. For example, if a user modifies a preset tag from "food" to "sports," and the "sports" tag is in the tag library, then the tag belongs to the tag library. For another example, if a user modifies a preset tag from "food" to "psychology," and the "psychology" tag is not in the tag library, then the tag does not belong to the tag library.

[0078] If the changed label belongs to the label library, the modified changed label is quality checked by a preset quality inspection model. In this application, the quality inspection model is used to characterize the matching relationship between the target text and the label. For example, the quality inspection model is preferably a trained large language model. If "food" is changed to "sports", the "sports" label and the target text "I made a braised pork dish in the kitchen today. The taste is very good. I miss the food from my childhood" are quality checked. Specifically, the target text and the changed label are input into the quality inspection model. The quality inspection model determines whether the changed label matches the target text. If they match, the quality inspection passes. If they do not match, the quality inspection fails.

[0079] S1042: If the modified label does not belong to the label library, a second candidate label is recommended based on the modified label through a preset machine recommendation model. In the present application, the machine recommendation model is used to characterize the correspondence between the modified label and the second candidate label. For example, the machine recommendation model is preferably a trained large language model. If "delicious food" is changed to "braised pork in red sauce", and the modified label "braised pork in red sauce" does not belong to the label library, the machine recommendation model recommends a second candidate label based on the modified label, such as "cooking skills".

[0080] It should be noted that the above is a preferred implementation method of this step, and those skilled in the art can adjust the above steps according to specific application scenarios. For example, the steps can also be: if the changed label does not meet the label specification, then the preset machine recommendation model recommends a second alternative label before quality inspection. The label specification can be two characters or one word, etc. For another example, the steps can also be: if the changed label does not belong to the label library and does not meet the label specification, then the preset machine recommendation model recommends a second alternative label before quality inspection; if the changed label does not belong to the label library, but meets the label specification, then quality inspection is directly carried out. For another example, changing labels is not limited to changing the labels that appear based on the recommended labels selected by the user, but also includes new labels created by the user.

[0081] It should be further explained that the step of "if the recommended label is modified, then quality inspection is performed on the modified label" can also be performed directly by manual quality inspection instead of using a quality inspection model to quality inspect the modified label.

[0082] S1043, determining whether the second candidate tag is adopted; for example, obtaining whether the user selects the second candidate tag on the smart terminal device; if the second candidate tag is selected, the second candidate tag is adopted; if the second candidate tag is not selected and the tag is re-edited, returning to S1041.

[0083] S1044, based on the judgment result, perform quality inspection again.

[0084] Specifically, if the second candidate label is adopted, the quality inspection model will be used to perform a quality inspection on the second candidate label. Specifically, the quality inspection model will compare the target text with the second candidate label. If the target text and the second candidate label match, the quality inspection passes. For example, if the second candidate label "cooking skills" is adopted, the quality inspection model will be used to determine whether the second candidate label "cooking skills" matches the target text. If the target text and the second candidate label match, the quality inspection passes.

[0085] If the second candidate label is not adopted, the modified label is manually inspected. For example, if the user does not adopt "cooking skills" as a label and insists on selecting "braised pork in red sauce" as the label for the target text, the modified label "braised pork in red sauce" is manually inspected.

[0086] Furthermore, if the quality inspection passes, the second alternative label is added to the label library. And the correspondence between the modified label and the target text is stored in the database. For example, if the second alternative label "cooking" is adopted, and the quality inspection model determines that the second alternative label "cooking" matches the target text, the second alternative label "cooking" is added to the label library. At the same time, the correspondence between the modified label and the target text is stored in the database. For another example, if the modified label "braised pork in red sauce" is adopted, and the quality inspection model determines that the modified label "braised pork in red sauce" matches the target text, the modified label "braised pork in red sauce" is added to the label library. In this way, the second alternative label that passes the quality inspection can be added to the label library, so that the labels in the label library can cover more content and can be more accurate when annotating the text content.

[0087] It should be noted that the step of "if the recommended label is modified, then quality inspection is performed on the modified label" can also be performed directly by manual quality inspection instead of using a quality inspection model to quality inspect the modified label.

[0088] It should be further explained that the modified label, the second candidate label and the new label in this embodiment all belong to the modified label in claim 1.

[0089] A possible implementation process of the present application is described below with reference to FIG3 .

[0090] As shown in FIG3 , the process of the text content tag generation method according to one embodiment of the present invention includes the following steps:

[0091] S201, obtaining the target text to be marked;

[0092] S202, the labeling model selects two first candidate labels from a preset label library;

[0093] S203, displaying the two first candidate tags as recommended tags to the user;

[0094] S204, determine whether the recommended tag is adopted; if yes, execute S205, if not, execute S206;

[0095] S205, determine whether the recommended tag has been modified; if yes, execute S206, if not, execute S213;

[0096] S206, determine whether the modified tag belongs to the tag library; if yes, execute S209, if not, execute S207;

[0097] S207, the machine recommendation model recommends a second candidate tag;

[0098] S208, determine whether the second candidate label is adopted; if yes, execute S209, otherwise, execute S210;

[0099] S209, performing quality inspection on the label using a quality inspection model;

[0100] S210, perform manual quality inspection on the modified labels;

[0101] S211, determine whether the quality inspection has passed; if passed, execute 212; if not, execute 214;

[0102] S212, adding the tag to the tag library;

[0103] S213, using the label as a label of the target text;

[0104] S214, target text does not use labels.

[0105] It should be noted that although the above describes the detailed steps of the method of the present application in detail, those skilled in the art may combine, split, and change the order of the above steps without departing from the basic principles of the present application. The modified technical solution does not change the basic concept of the present application and therefore falls within the scope of protection of the present application. For example, the order of S212 and S213 can be interchanged. For another example, S205 can be omitted. If the recommended tag in S204 is adopted, S213 is executed. If the recommended tag is not adopted in S204, but the tag is modified or newly created, S206 is executed.

[0106] Based on the same inventive concept, as shown in FIG4 , the present invention further discloses a text content tag generation device 300 , comprising:

[0107] An acquisition module 301 is configured to acquire a target text to be marked;

[0108] A recommendation module 302 is configured to generate recommended tags based on the target text using a preset tagging model and display the recommended tags;

[0109] A determination module 303 is configured to determine whether the recommendation tag has been modified;

[0110] A quality inspection module 304 is configured to perform a quality inspection on the modified recommended label if the recommended label is modified;

[0111] The tagging module 305 is configured to use the modified label as the label of the target text if the quality inspection is passed.

[0112] When using the above technical solution, the tagging model can generate and display recommended tags. Users can select the recommended tags as the tags for the target text, avoiding a cold start of the tag and label data. Furthermore, if the user modifies the recommended tags, the modified tags are quality-checked. By using the modified tags as the tags for the target text, the target text can be annotated with the new tags that have passed quality inspection, thus avoiding the problem of cold starting the label data due to the lack of suitable tags.

[0113] It should be noted that the text content label generation device provided in the above embodiment is only illustrated by the division of the above-mentioned functional modules (such as acquisition module 301, recommendation module 302, judgment module 303, quality inspection module 304, annotation module 305, etc.). In actual applications, the above-mentioned functional modules can be completed by different functional units as needed, that is, the functional modules in this embodiment can be decomposed or combined again. For example, the functional modules of the above embodiment can be combined into one functional module, or further divided into multiple sub-modules to complete all or part of the functions described above. The names of the functional modules involved in this embodiment are only for distinction and are not regarded as improper limitations on this application.

[0114] Those skilled in the art will appreciate that all or part of the process steps in the method of the above embodiment of the present invention can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, it can implement the steps of each of the above method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file, or some intermediate form. The computer-readable storage medium can include: any entity or device, medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory, random access memory, electric carrier signal, telecommunication signal, and software distribution medium that can carry computer program code. It should be noted that the content contained in the computer-readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable storage media do not include electric carrier signals and telecommunication signals.

[0115] The various component embodiments of the present invention can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It will be appreciated by those skilled in the art that a microprocessor or digital signal processor (DSP) can be used in practice to implement some or all of the functions of some or all of the components in the server or client according to an embodiment of the present invention. The present invention can also be implemented as a device or apparatus program (e.g., a PC program and a PC program product) for executing a part or all of the methods described herein. Such a program implementing the present invention can be stored on a PC readable medium, or can have the form of one or more signals. Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0116] Based on the same inventive concept, the present invention also discloses an electronic device comprising: a processor; and a memory for storing a program. When the program is executed by the processor, the processor implements the above-described method for generating text content tags. For ease of illustration, only the portions relevant to the embodiments of the present invention are shown. For specific technical details not disclosed, please refer to the method section of the embodiments of the present invention. The electronic device may include various devices capable of implementing the above-described method.

[0117] Based on the same inventive concept, the present invention also discloses a computer-readable medium having a computer program stored thereon, which implements the above-mentioned method when the program is executed by a processor. For ease of explanation, only the parts related to the embodiment of the present invention are shown. For specific technical details not disclosed, please refer to the method section of the embodiment of the present invention. The computer-readable storage medium can be a storage device formed by various electronic devices. Optionally, the computer-readable storage medium in the embodiment of the present invention is a non-transitory computer-readable storage medium.

[0118] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

Claims

1. A method for generating text content labels, characterized in that: The method includes: Get the target text to be marked; Based on the target text, generate recommended tags through a preset tagging model and display the recommended tags; Determining whether the recommended tag has been modified; If the recommended label is modified, then quality inspection is performed on the modified label; If the quality inspection is passed, the modified label is used as the label of the target text, and the corresponding relationship between the modified label and the target text is stored in a database.

2. The text content label generation method according to claim 1, characterized in that: The step of “generating recommended tags based on the target text using a preset tagging model and displaying the recommended tags” further includes: Based on the target text, one or more first candidate tags are selected from a preset tag library; One or more of the first candidate tags are displayed to the user as the recommended tags.

3. The method for generating text content labels according to claim 2, wherein: The step of "if the recommended label is modified, performing quality inspection on the modified label" further includes: Determine whether the modified tag belongs to the tag library; If the modified label belongs to the label library, the modified label is quality-checked using a preset quality-check model.

4. The method for generating text content labels according to claim 3, wherein: The step of "if the recommended label is modified, performing quality inspection on the modified label" further includes: If the modified tag does not belong to the tag library, recommending a second candidate tag based on the modified tag through a preset machine recommendation model; Determining whether the second candidate label is adopted; Based on the judgment results, quality inspection is carried out again.

5. The method for generating text content labels according to claim 4, wherein: The step of "conducting quality inspection again based on the judgment result" further includes: If the second candidate label is adopted, performing quality inspection on the second candidate label using the quality inspection model; If the second candidate label is not adopted, manual quality inspection is performed on the modified label.

6. The method for generating text content labels according to claim 5, wherein: The step of “performing quality inspection on the second candidate label by the quality inspection model” further includes: The target text is compared with the second candidate label by using the quality inspection model. If the target text and the second candidate label correspond to each other, the quality inspection passes.

7. The method for generating text content labels according to claim 6, wherein: The method further comprises: If the quality inspection is passed, the second candidate tag is added to the tag library.

8. The method for generating text content labels according to claim 2, wherein: After the step of "determining whether the recommendation tag has been modified", the following steps are also included: If the recommended tag is not modified, the recommended tag is used as the tag of the target text.

9. The method for generating text content labels according to claim 2, wherein: The step of “generating recommended tags based on the target text using a preset tagging model and displaying the recommended tags” further includes: During the target text editing process, the first candidate tag is synchronously selected on the backend through the tagging model; After the target text is edited, the first alternative label is displayed.

10. A text content label generation device, characterized in that: include: an acquisition module, configured to acquire a target text to be marked; A recommendation module, configured to generate a recommended tag based on the target text using a preset tagging model and display the recommended tag; a determination module configured to determine whether the recommendation tag has been modified; a quality inspection module, configured to perform a quality inspection on the modified label if the recommended label is modified; The tagging module is configured to use the modified label as a label of the target text if the quality inspection is passed.

11. An electronic device, characterized in that: include: processor; Memory, used to store programs, When the program is executed by the processor, the processor implements the method for generating text content tags according to any one of claims 1 to 9.

12. A computer-readable medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for generating text content tags according to any one of claims 1 to 9 is implemented.

Images