Search

CN-122019909-A - Product architecture information generation method and device based on large model and electronic equipment

CN122019909ACN 122019909 ACN122019909 ACN 122019909ACN-122019909-A

Abstract

The disclosure provides a method and a device for generating product architecture information based on a large model and electronic equipment, relates to the technical field of computers, and particularly relates to the technical fields of artificial intelligence, large models, information retrieval, data mining and the like. The method comprises the steps of obtaining a plurality of product identifiers, grouping the product identifiers to obtain an identifier group under each grouping granularity in N grouping granularities, wherein N is more than or equal to 2, N is an integer, the N grouping granularities have different thickness degrees, performing identifier integration aiming at a first identifier group by using a first large model to obtain a first identifier integration result, wherein the first identifier group is an identifier group under the ith grouping granularity in the N grouping granularities, i is more than or equal to 1 and less than or equal to N, and i is an integer, and obtaining product architecture information related to the product identifiers based on the first identifier integration result. By adopting the method and the device, the accuracy and the acquisition efficiency of the product architecture information can be improved.

Inventors

  • LIU JIAQI
  • YUAN MINGCHEN
  • YU FEI

Assignees

  • 北京百度网讯科技有限公司

Dates

Publication Date
20260512
Application Date
20260129

Claims (17)

  1. 1. A product architecture information generation method based on a large model comprises the following steps: acquiring a plurality of product identifiers; Grouping the product identifiers to obtain an identifier group under each grouping granularity in N grouping granularities, wherein N is more than or equal to 2 and N is an integer; performing identification integration aiming at a first identification group by using a first large model to obtain a first identification integration result, wherein the first identification group is an identification group with the ith grouping granularity in the N grouping granularities, i is more than or equal to 1 and less than or equal to N, and i is an integer; and obtaining product architecture information related to the plurality of product identifiers based on the first identifier integration result.
  2. 2. The method of claim 1, wherein grouping the plurality of product identifiers to obtain an identifier group at each of N grouping granularities comprises: The method comprises the steps of obtaining a plurality of code groups corresponding to a plurality of product identifiers one by one, wherein for each product identifier in the plurality of product identifiers, the code group corresponding to the product identifier comprises N+1 code results, and the N+1 code results are used for quantizing the product identifier from thick to thin; And grouping the product identifiers based on the code groups to obtain identifier groups under each grouping granularity in the N grouping granularities.
  3. 3. The method of claim 2, wherein the obtaining a plurality of code sets that are in one-to-one correspondence with the plurality of product identifiers comprises: Vectorizing the product identifiers aiming at each product identifier in the plurality of product identifiers to obtain an identifier vector; and utilizing a target encoder to encode the identification vector to obtain an encoding group corresponding to the product identification.
  4. 4. The method of claim 2, wherein grouping the plurality of product identifiers based on the plurality of code groups to obtain an identifier group at each of the N packet sizes comprises: Determining M target code groups with the same first j code results from the code groups, wherein j is more than or equal to 1 and less than or equal to N, j is an integer, M is more than or equal to 1, and M is an integer; determining M first target product identifiers corresponding to the M target code groups one by one from the plurality of product identifiers; And constructing an identification group under the j-th grouping granularity in the N grouping granularities based on the M first target product identifications.
  5. 5. The method of claim 1, wherein grouping the plurality of product identifiers to obtain an identifier group at each of N grouping granularities comprises: Clustering the product identifiers to obtain a tree structure, wherein the tree structure has N clustering layers from thick to thin; And constructing an identification group under the kth grouping granularity in the N grouping granularities based on a clustering result under the kth clustering hierarchy in the N clustering hierarchies, wherein k is more than or equal to 1 and less than or equal to N, and k is an integer.
  6. 6. The method according to any one of claims 1-5, wherein the performing, by using the first large model, the identifier integration for the first identifier group to obtain a first identifier integration result includes: Acquiring an identifier to be integrated corresponding to the first identifier group; constructing an identification integration prompt; And integrating the identification to be integrated according to the identification integration prompt by using the first large model to obtain the first identification integration result.
  7. 7. The method of claim 6, wherein the obtaining the identifier to be integrated corresponding to the first identifier group comprises: taking the product identifiers in the first identifier group as the identifiers to be integrated under the condition that the ith packet granularity is the finest packet granularity in the N packet granularities; Or under the condition that the ith grouping granularity is not the finest grouping granularity in the N grouping granularities, obtaining the identification to be integrated based on a second identification integration result, wherein the second identification integration result is obtained by utilizing the first large model to integrate the identifications of a second identification group, and the second identification group is the identification group in the (i+1) th grouping granularity in the N grouping granularities, and the product identifications in the second identification group are contained in the first identification group.
  8. 8. The method of claim 7, wherein the obtaining the identifier to be integrated based on the second identifier integration result includes: Under the condition that the second identifier integration result is commonly referred to as an integral identifier, commonly referred to as the identifier to be integrated in the second identifier integration result; or under the condition that the second identifier integration result is not commonly called by the whole identifier, obtaining the identifier to be integrated based on the product level classification information in the second identifier integration result.
  9. 9. The method of claim 6, wherein the constructing an identification integration hint comprises: acquiring an identification integration example; acquiring an identification integration requirement; and constructing the identification integration prompt based on the identification integration example and the identification integration requirement.
  10. 10. The method of claim 9, wherein the identity integration requirement comprises: Under the condition that the marks to be integrated can be generalized into a whole mark collective term, generating a first mark integration result comprising the whole mark collective term; And/or generating a first identification integration result comprising the product level classification information under the condition that the product level classification information can be obtained based on the identification to be integrated.
  11. 11. The method of claim 1, wherein the obtaining a plurality of product identifications comprises: acquiring a plurality of user search words from a target platform; and obtaining the product identifiers based on the user search words.
  12. 12. The method of claim 11, wherein the deriving the plurality of product identifications based on the plurality of user search terms comprises: constructing an entity extraction instruction; aiming at each user search word in the plurality of user search words, extracting prompts according to the entity by utilizing a second large model, and obtaining candidate identifications based on the user search words; And obtaining the product identifiers based on the candidate identifiers corresponding to the user search words one by one.
  13. 13. The method of claim 12, wherein the building entity extracts hints comprising: Obtaining an entity extraction example; acquiring industry knowledge related to the plurality of user search terms; and constructing the entity extraction prompt based on the entity extraction examples and the industry knowledge.
  14. 14. A large model-based product architecture information generation apparatus comprising: The identification acquisition unit is used for acquiring a plurality of product identifications; the identification grouping unit is used for grouping the plurality of product identifications to obtain identification groups under each grouping granularity in N grouping granularities, wherein N is more than or equal to 2 and is an integer; The identification integration unit is used for carrying out identification integration aiming at a first identification group by utilizing a large model to obtain a first identification integration result, wherein the first identification group is an identification group with the i-th grouping granularity in the N grouping granularities, i is more than or equal to 1 and less than or equal to N, and i is an integer; And the information acquisition unit is used for acquiring product architecture information related to the plurality of product identifiers based on the first identifier integration result.
  15. 15. An electronic device, comprising: At least one processor; a memory communicatively coupled to the at least one processor; Wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-13.
  16. 16. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-13.
  17. 17. A computer program product comprising a computer program, wherein the computer program is capable of implementing the method of any one of claims 1 to 13 when executed by a processor.

Description

Product architecture information generation method and device based on large model and electronic equipment Technical Field The disclosure relates to the technical field of computers, in particular to the technical fields of artificial intelligence, large models, information retrieval, data mining and the like, and particularly relates to a method and a device for generating product architecture information based on large models and electronic equipment. Background The product architecture information is a digital center of an enterprise product system, is used for systematically organizing product nodes and mounting digital assets such as advertisement materials, landing pages and the like on the product architecture information, and has a core value of providing an understandable and retrievable framework for dynamic advertisement strategies or electronic commerce platforms, so that the accurate recall of the matched digital assets based on the intention of a user is possible, and the product marketing is driven to be changed from static configuration to intelligent matching. Disclosure of Invention The disclosure provides a method and a device for generating product architecture information based on a large model and electronic equipment. According to a first aspect of the present disclosure, there is provided a large model-based product architecture information generation method, including: acquiring a plurality of product identifiers; Grouping the product identifiers to obtain an identifier group under each grouping granularity in N grouping granularities, wherein N is more than or equal to 2 and is an integer; Performing identification integration aiming at a first identification group by using a first large model to obtain a first identification integration result, wherein the first identification group is an identification group with the i-th grouping granularity in N grouping granularities, i is more than or equal to 1 and less than or equal to N, and i is an integer; And obtaining product architecture information related to the plurality of product identifiers based on the first identifier integration result. According to a second aspect of the present disclosure, there is provided a large model-based product architecture information generation apparatus, including: The identification acquisition unit is used for acquiring a plurality of product identifications; the identification grouping unit is used for grouping a plurality of product identifications to obtain identification groups under each grouping granularity in N grouping granularities, wherein N is more than or equal to 2 and is an integer; The identification integration unit is used for integrating the identifications of the first identification group by utilizing the large model to obtain a first identification integration result, wherein the first identification group is an identification group with the i-th grouping granularity in N grouping granularities, i is more than or equal to 1 and less than or equal to N, and i is an integer; and the information acquisition unit is used for acquiring product architecture information related to the plurality of product identifiers based on the first identifier integration result. According to a third aspect of the present disclosure, there is provided an electronic device comprising: At least one processor; a memory communicatively coupled to the at least one processor; Wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method provided in the first aspect of the present disclosure. According to a fourth aspect of the present disclosure there is provided a non-transitory computer readable storage medium storing computer instructions for causing a computer to perform the method provided by the first aspect of the present disclosure. According to a fifth aspect of the present disclosure, there is provided a computer program product comprising a computer program, wherein the computer program, when executed by a processor, is capable of implementing the method provided by the first aspect of the present disclosure. By adopting the method and the device, the accuracy and the acquisition efficiency of the product architecture information can be improved. It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification. Drawings The drawings are for a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein: Fig. 1 is a flow chart of a method for generating product architecture information based on a large model according to an embodiment of the disclosure; Fig. 2 is a schematic diagram of an acquisition flow of a code