Search

CN-121980234-A - Popularization information generation method and device based on large model

CN121980234ACN 121980234 ACN121980234 ACN 121980234ACN-121980234-A

Abstract

The disclosure provides a popularization information generation method and device based on a large model, and relates to the artificial intelligence fields of deep learning, large model, natural language processing and the like. The method comprises the steps of obtaining target characteristics of an object to be promoted, respectively generating an attribute model and a scene model of a target promotion user of the object to be promoted, generating initial promotion information of the object to be promoted by utilizing a target large model according to the target characteristics, the attribute model and the scene model, and generating target promotion information comprising image-text content according to the initial promotion information. By applying the scheme disclosed by the disclosure, the generation efficiency, the generation quality and the like of the popularization information can be improved.

Inventors

  • CAO YAJING

Assignees

  • 百度在线网络技术（北京）有限公司

Dates

Publication Date
20260505
Application Date
20251205

Claims (15)

  1. 1. A popularization information generation method based on a large model comprises the following steps: Acquiring target characteristics of an object to be promoted; respectively generating an attribute model and a scene model of the target popularization user of the object to be popularized; generating initial popularization information of the object to be promoted by utilizing a target big model according to the target characteristics, the attribute model and the scene model; and generating target popularization information comprising image-text content according to the initial popularization information.
  2. 2. The method of claim 1, wherein the obtaining the target feature of the object to be promoted comprises: Acquiring object description information of the object to be promoted; According to the object description information, object characteristics of M different dimensions of the object to be promoted are respectively determined, M is a positive integer greater than 1, and the target characteristics comprise M object characteristics.
  3. 3. The method of claim 1, wherein generating the attribute model comprises: respectively acquiring first historical behavior information of each target popularization user in a first latest time period, and respectively determining attribute characteristics of N different dimensions of the corresponding target popularization users according to the first historical behavior information, wherein N is a positive integer greater than 1; And fusing the attribute characteristics of each target popularization user to obtain the attribute model.
  4. 4. The method of claim 1, wherein generating the scene model comprises: Respectively acquiring second historical behavior information of each target popularization user in a second latest time period, and respectively determining scene characteristics of P different dimensions of the corresponding target popularization users according to the second historical behavior information, wherein P is a positive integer larger than 1; and fusing the scene characteristics of each target popularization user to obtain the scene model.
  5. 5. The method of claim 4, wherein, The scene features include temporal scene features, location scene features, and contextual scene features.
  6. 6. The method of claim 1, wherein the generating initial promotion information for the object to be promoted using a target large model based on the target features, the attribute model, and the scene model comprises: acquiring popularization description information corresponding to the object to be popularized; Generating a prompt word according to the target feature, the attribute model, the scene model and the popularization description information; and inputting the prompt word into the target large model to obtain the output initial popularization information.
  7. 7. The method of claim 6, wherein the generating a hint word from the target feature, the attribute model, the scene model, and the promotional description information comprises: respectively determining weights corresponding to the target features, the attribute model, the scene model and the popularization description information; and generating the prompt word according to the target feature, the attribute model, the scene model, the popularization description information and the determined weights.
  8. 8. The method of claim 1, wherein the generating target promotional information comprising teletext content from the initial promotional information comprises: determining the number of the objects to be promoted, and acquiring layout optimization information matched with the number; Generating optimized popularization information according to the layout optimization information and the initial popularization information; and generating the target popularization information according to the optimized popularization information.
  9. 9. The method of claim 8, wherein, The generating the optimized popularization information according to the layout optimization information and the initial popularization information comprises: according to the layout optimization information, the number of pictures and the number of interaction components corresponding to each object to be promoted are respectively determined, and the pictures and the interaction components with the corresponding numbers are obtained; according to the layout optimization information, respectively determining the display positions of each picture, each interaction component and each text content in the initial popularization information; and generating the optimized popularization information according to the pictures, the interaction components and the initial popularization information according to the display positions.
  10. 10. The method of claim 8, wherein the generating the target promotion information from the optimized promotion information comprises: Scoring the optimized popularization information from L different dimensions respectively, wherein L is a positive integer greater than 1, and determining the comprehensive score of the optimized popularization information according to the L scores; And in response to determining that the comprehensive score meets a preset requirement, determining that the quality evaluation of the optimized popularization information is passed, and determining the optimized popularization information as the target popularization information.
  11. 11. The method of claim 10, further comprising: Generating correction suggestion information in response to the fact that the quality evaluation fails, correcting the optimized popularization information according to the correction suggestion information, and determining the corrected optimized popularization information as the target popularization information; Or carrying out quality evaluation on the corrected optimized popularization information, determining the corrected optimized popularization information as the target popularization information in response to the determination that the quality evaluation is passed, and continuing to correct the corrected optimized popularization information until the quality evaluation is passed in response to the determination that the quality evaluation is not passed.
  12. 12. The popularization information generation device based on the large model comprises a first characteristic acquisition module, a second characteristic acquisition module, a first information generation module and a second information generation module; the first feature acquisition module is used for acquiring target features of the object to be promoted; The second feature acquisition module is used for respectively generating an attribute model and a scene model of the target popularization user of the object to be promoted; The first information generating module is used for generating initial popularization information of the object to be promoted by utilizing a target big model according to the target characteristics, the attribute model and the scene model; the second information generating module is used for generating target popularization information including image-text content according to the initial popularization information.
  13. 13. An electronic device, comprising: at least one processor, and 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 1-11.
  14. 14. A non-transitory computer readable storage medium storing computer instructions for causing a computer to perform the method of any one of claims 1-11.
  15. 15. A computer program product comprising computer programs/instructions which, when executed by a processor, implement the method of any of claims 1-11.

Description

Popularization information generation method and device based on large model Technical Field The disclosure relates to the technical field of artificial intelligence, in particular to the fields of deep learning, large models, natural language processing and the like, and particularly relates to a popularization information generation method and device based on the large models. Background In the scenes of e-commerce, content marketing and the like, graphic content creation is required. Traditional graphic content creation mainly adopts a templated mode. Namely, the creator selects a corresponding template according to the commodity category, and then fills commodity attributes, pictures and the like to generate image-text content. This approach not only requires significant labor and time costs, but the quality of the generated content is often difficult to guarantee. Disclosure of Invention The disclosure provides a popularization information generation method and device based on a large model. A popularization information generation method based on a large model comprises the following steps: Acquiring target characteristics of an object to be promoted; respectively generating an attribute model and a scene model of the target popularization user of the object to be popularized; generating initial popularization information of the object to be promoted by utilizing a target big model according to the target characteristics, the attribute model and the scene model; and generating target popularization information comprising image-text content according to the initial popularization information. The popularization information generation device based on the large model comprises a first characteristic acquisition module, a second characteristic acquisition module, a first information generation module and a second information generation module; the first feature acquisition module is used for acquiring target features of the object to be promoted; The second feature acquisition module is used for respectively generating an attribute model and a scene model of the target popularization user of the object to be promoted; The first information generating module is used for generating initial popularization information of the object to be promoted by utilizing a target big model according to the target characteristics, the attribute model and the scene model; the second information generating module is used for generating target popularization information including image-text content according to the initial popularization information. An electronic device, comprising: at least one processor, and 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 as described above. A non-transitory computer readable storage medium storing computer instructions for causing a computer to perform a method as described above. A computer program product comprising computer programs/instructions which when executed by a processor implement a method as described above. 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 flowchart of a first embodiment of a large model-based promotional information generation method of the present disclosure; FIG. 2 is a flow chart of an embodiment of a method of generating merchandise characteristics for any merchandise according to the present disclosure; FIG. 3 is a flow chart of an embodiment of a method of generating attribute features for any targeted promotional user in accordance with the present disclosure; FIG. 4 is a flow chart of an embodiment of a method of generating scene features for any targeted promotional user as described in the present disclosure; FIG. 5 is a flow chart of an embodiment of a method of generating optimized popularization information according to the present disclosure; FIG. 6 is a flow chart of an embodiment of a method of generating targeted promotional information from optimized promotional information according to the present disclosure; FIG. 7 is a flowchart of a second embodiment of a large model-based promotional information generation method of the present disclosure; Fig. 8 is a schematic diagram of a composition structure of an embodiment 800 of a large model-based promotion information generation apparatus according to the present disclosure; Fig. 9 shows a schematic block diagram of an electronic device 900 that may be used to implement embodiments of the present disclosure. Detailed De