Search

CN-121980222-A - Label verification method, device, equipment, medium and product

CN121980222ACN 121980222 ACN121980222 ACN 121980222ACN-121980222-A

Abstract

The embodiment of the invention provides a tag verification method, a device, equipment, a medium and a product, and relates to the technical field of multimedia intelligent processing. The method comprises the steps of obtaining a label to be verified of target content, obtaining a description text of the target content as the description text to be utilized, processing the label to be verified, the description text to be utilized and a prompt word by utilizing a large language model to obtain a verification result indicating whether the label to be verified has errors, wherein the prompt word is used for indicating the large language model to judge whether the label to be verified can be used as a label of the content represented by the description text to be utilized, and obtaining the verification result. The invention can improve the accuracy of the label.

Inventors

  • WEN XU
  • CHENG QIJIAN
  • LI DAWEI

Assignees

  • 北京奇艺世纪科技有限公司

Dates

Publication Date
20260505
Application Date
20260123

Claims (11)

  1. 1. A method of tag verification, the method comprising: Acquiring a label to be verified of target content, and acquiring a description text of the target content as the description text to be utilized; And processing the label to be verified, the description text to be utilized and the prompt word by using a large language model to obtain a verification result indicating whether the label to be verified has errors, wherein the prompt word is used for indicating the large language model to judge whether the label to be verified can be used as the label of the content represented by the description text to be utilized to obtain the verification result.
  2. 2. The method of claim 1, wherein the hint word comprises a plurality of preset tags; the processing of the label to be verified, the description text to be used and the prompt word by using the large language model to obtain a verification result indicating whether the label to be verified has errors or not comprises the following steps: the following steps are performed using the large language model: Detecting whether the tags to be verified exist in the preset tags or not; If the label to be verified does not exist, determining that the label to be verified is a label with a first type of error, and obtaining a verification result, wherein the first type of error indicates that the label to be verified does not belong to the plurality of preset labels; If the label to be verified exists, judging whether the label to be verified can be used as the label of the content represented by the description text to be utilized, and obtaining a verification result.
  3. 3. The method of claim 2, wherein the prompt word further includes a relationship description text for describing a parent-child relationship among the plurality of preset labels, and a child label corresponding to a parent label represents a child category under a category represented by the parent label; if the label to be verified exists, judging whether the label to be verified can be used as the label of the content represented by the description text to be utilized, and obtaining a verification result, wherein the method comprises the following steps: Judging whether sub-class labels corresponding to the to-be-verified labels exist in all the to-be-verified labels existing in the plurality of preset labels based on the father-son relationship aiming at each to-be-verified label existing in the plurality of preset labels under the condition that the plurality of to-be-verified labels exist in the plurality of preset labels; If the sub-class label corresponding to the label to be verified exists, judging whether the label to be verified can be used as the label of the content represented by the description text to be utilized; If the label to be verified cannot be used as the label of the content represented by the description text to be utilized, determining the label to be verified and the corresponding sub-class label as the label with the first class error, and obtaining a verification result.
  4. 4. A method according to claim 2 or 3, wherein the step of the large language model performing further comprises: Under the condition that a plurality of tags to be verified exist, determining the text proportion of the content represented by the tags to be verified in the description text to be utilized aiming at each tag to be verified without first-class errors; And determining the label to be verified with the text proportion smaller than the preset proportion as a label with a second type error to obtain a verification result, wherein the second type error indicates that the label to be verified is not matched with the target content.
  5. 5. The method according to claim 4, wherein the method further comprises: calculating the proportion of the tags to be verified, which have the first type of errors and the second type of errors, in all the tags to be verified, so as to obtain the accuracy of all the tags to be verified; And if the accuracy rate does not reach the preset accuracy rate, outputting alarm information to prompt that other operations for distributing labels to the target content are required to be executed.
  6. 6. The method according to claim 5, wherein calculating the proportion of the tags to be verified having the first type of errors and the second type of errors in all the tags to be verified, to obtain the accuracy of all the tags to be verified, includes: calculating a weighted sum of the number of tags to be verified with the first type of errors and the number of tags to be verified with the second type of errors, wherein the weight of the number of tags to be verified with the first type of errors is larger than the weight of the number of tags to be verified with the second type of errors; and calculating the ratio of the obtained weighted sum to the total number of all the tags to be verified to obtain the accuracy of all the tags to be verified.
  7. 7. The method according to any one of claims 1-3, wherein the prompt word comprises a positive example text and a negative example text, wherein the positive example text comprises a positive sample description text, a positive sample label and a reason why the positive sample label can be used as a label of the content represented by the positive sample description text; The prompt word is used for indicating a large language model to judge whether the label to be verified can be used as the label of the content represented by the description text to be utilized based on the positive example text and the negative example text, so as to obtain a verification result, and a reason for obtaining the verification result is generated.
  8. 8. A tag verification apparatus, the apparatus comprising: The acquisition module is used for acquiring a label to be verified of target content and acquiring a description text of the target content as the description text to be utilized; The utilization module is used for processing the label to be verified, the description text to be utilized and the prompt word by utilizing a large language model to obtain a verification result which indicates whether the label to be verified has errors, wherein the prompt word is used for indicating the large language model to judge whether the label to be verified can be used as the label of the content represented by the description text to be utilized to obtain the verification result.
  9. 9. The electronic equipment is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory are communicated with each other through the communication bus; a memory for storing a computer program; a processor for implementing the method of any of claims 1-7 when executing a program stored on a memory.
  10. 10. A computer readable storage medium, characterized in that the computer readable storage medium has stored therein a computer program which, when executed by a processor, implements the method of any of claims 1-7.
  11. 11. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any of claims 1-7.

Description

Label verification method, device, equipment, medium and product Technical Field The invention relates to the technical field of multimedia intelligent processing, in particular to a tag verification method, a device, equipment, a medium and a product. Background The tag is a text describing contents such as news, blogs and videos, and enterprises providing the contents for users can conveniently conduct classified management on the contents through the tag and help the users to accurately and quickly find the required contents. By way of example, labels for content may be generally determined in a manual fashion. In the prior art, the type of each content can be manually analyzed, and the accuracy of the label of the content can be verified. However, the verification of the accuracy of the tag by the human effort consumes a lot of labor cost and time cost, and the effect of the manual verification is also subjectively affected by the staff, resulting in lower accuracy of the verification result. Disclosure of Invention The embodiment of the invention aims to provide a label verification method, device, equipment, medium and product, so as to avoid manual label verification, reduce labor cost and time cost and improve the accuracy of verification results. The specific technical scheme is as follows: In a first aspect, an embodiment of the present invention provides a tag verification method, where the method includes: The method comprises the steps of obtaining a label to be verified of target content, obtaining a description text of the target content as the description text to be utilized, processing the label to be verified, the description text to be utilized and a prompt word by utilizing a large language model to obtain a verification result indicating whether the label to be verified has errors, wherein the prompt word is used for indicating the large language model to judge whether the label to be verified can be used as a label of the content represented by the description text to be utilized, and obtaining the verification result. Optionally, the prompt word includes a plurality of preset labels; the processing of the label to be verified, the description text to be used and the prompt word by using the large language model to obtain a verification result indicating whether the label to be verified has errors or not comprises the following steps: The large language model is used for executing the following steps of detecting whether the label to be verified exists in the plurality of preset labels, determining that the label to be verified is a label with a first type of errors to obtain a verification result if the label to be verified does not exist, judging whether the label to be verified can be used as the label of the content represented by the description text to be verified if the label to be verified exists, and obtaining the verification result. Optionally, the prompt word further comprises a relationship description text, wherein the relationship description text is used for describing father-son relationships among the preset labels, and a sub-class label corresponding to a father-class label represents sub-classes under the category represented by the father-class label; if the label to be verified exists, judging whether the label to be verified can be used as the label of the content represented by the description text to be utilized, and obtaining a verification result, wherein the method comprises the following steps: Judging whether sub-class labels corresponding to the to-be-verified labels exist in all the to-be-verified labels existing in the plurality of preset labels based on the father-son relationship aiming at each to-be-verified label existing in the plurality of preset labels under the condition that the plurality of to-be-verified labels exist in the plurality of preset labels; If the sub-class label corresponding to the label to be verified exists, judging whether the label to be verified can be used as the label of the content represented by the description text to be utilized; If the label to be verified cannot be used as the label of the content represented by the description text to be utilized, determining the label to be verified and the corresponding sub-class label as the label with the first class error, and obtaining a verification result. Optionally, the large language model further comprises the steps of determining the text proportion of the content represented by the to-be-verified label in the description text aiming at each to-be-verified label without first type errors when the to-be-verified labels are multiple, determining the to-be-verified label with the text proportion smaller than the preset proportion as the label with second type errors, and obtaining a verification result, wherein the second type errors represent that the to-be-verified label is not matched with the target content. Optionally, the method further comprises the steps of calcula