Method for automatically generating PowerPoint based on large language model

Abstract

The invention provides a method for automatically generating a PowerPoint based on a large language model. The method comprises the following steps: extracting a PowerPoint demand text from an original file provided by a user; inputting the demand text of the PowerPoint into the large model, obtaining key information and a theme of the demand text through the model capability of the large model, and generating a PowerPoint outline; carrying out detailed extension on each part of content in the presentation file outline by utilizing the large model to obtain a presentation file text; automatically filling the presentation file text into a preset presentation file template by utilizing the large model, and generating a target presentation file corresponding to the presentation file demand text; according to the method for automatically generating the PowerPoint, the working efficiency of a user can be remarkably improved, and the tedious degree of manual operation is reduced; according to the method, the content of the original file provided by the user can be intelligently analyzed, and automatic acquisition, arrangement and generation of the content of the presentation file are realized, so that the operation complexity of the user is reduced, and the working efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of AI (Artificial Intelligence), and more specifically, to a method for automatically generating presentations based on a large language model (LLM). Background Art

[0002] With the rapid development of information technology, the amount of data is increasing, and the demand for creating and optimizing presentations is also rising. Although there are many presentation creation tools on the market, users still need to perform a lot of manual operations to organize content and design documents. For example, although Microsoft's PowerPoint (PPT) and Google's Slides provide certain template support, they lack intelligent support in terms of logical integration of content and chart generation.

[0003] In recent years, with the development of artificial intelligence, large language models have been widely used in the automatic generation of PPTs. Relying on their powerful natural language processing capabilities, large language models automatically generate PPTs according to user needs, effectively improving the efficiency of PPT generation.

[0004] For example, an existing patent document discloses a method, device, electronic device and storage medium for generating a presentation, and the scheme is: obtaining input information associated with a presentation to be generated; displaying multiple presentation templates, and in response to monitoring a template selection operation, determining a target template from multiple presentation templates; displaying a presentation outline based on the input information; and displaying a presentation based on the presentation outline and the target template; however, most of the existing technologies, including this scheme, generate PPTs by having users manually edit the required information and match it to a predetermined template. When it is necessary to process complex data and extract a large amount of information, users still need to spend a lot of time manually organizing and designing prompt information, and the efficiency is still low. Summary of the invention

[0005] In order to overcome the defect that the above-mentioned prior art still requires a lot of manual operations by the user to generate a presentation, the present invention provides a method for automatically generating presentations based on a large language model. By introducing the large model of the LangChain framework, the automatic acquisition, organization and generation of presentation content are realized, thereby reducing the user's operation complexity and improving work efficiency.

[0006] In order to solve the above technical problems, the technical solution of the present invention is as follows: A method for automatically generating presentations based on a large language model comprises the following steps: S1: Extract presentation requirement text from the original file provided by the user; S2: inputting the presentation requirement text into the big model, obtaining key information and themes of the requirement text through the model capability of the big model, and generating a presentation outline; S3: using the large model to expand each part of the presentation outline in detail to obtain the presentation text; S4: using the large model to automatically fill the presentation text into a preset presentation template, and generate a target presentation corresponding to the presentation requirement text.

[0007] Preferably, in step S1, the original file provided by the user includes an unstructured demand file.

[0008] Preferably, the unstructured demand file is specifically a PDF file or a picture file.

[0009] Preferably, an OCR optical character recognition algorithm is used to extract the presentation requirement text from the unstructured requirement file.

[0010] Preferably, in step S1, the original file provided by the user also includes a structured demand file.

[0011] Preferably, the structured demand file is specifically an Excel file.

[0012] Preferably, when the original file provided by the user also includes a structured requirement file, extracting the presentation requirement text from the unstructured requirement file; At the same time, DuckDB is used to perform advanced data processing on the structured demand documents and extract structured data from the structured demand documents.

[0013] Preferably, the step S2 further includes: The structured data extracted from the structured requirement file is input into the big model, the big model is used to automatically generate the corresponding matplotlib drawing code, and the chart corresponding to the structured data is drawn according to the matplotlib drawing code.

[0014] Preferably, the step S4 also includes: using the large model to automatically fill the presentation text and charts into a preset presentation template to generate a target presentation corresponding to the presentation requirement text.

[0015] Preferably, the large model is specifically GPT-3 or BERT.

[0016] Compared with the prior art, the technical solution of the present invention has the following beneficial effects: The present invention provides a method for automatically generating presentations based on a large language model. First, a presentation requirement text is extracted from an original file provided by a user. Then, the presentation requirement text is input into the large model, and key information and themes of the requirement text are obtained through the model capability of the large model, and a presentation outline is generated. Then, the large model is used to perform detailed expansion on each part of the presentation outline to obtain the presentation text. Finally, the large model is used to automatically fill the presentation text into a preset presentation template to generate a target presentation corresponding to the presentation requirement text. The automatic presentation generation method proposed in the present invention can significantly improve the work efficiency of users and reduce the tediousness of manual operations; the present invention can intelligently analyze the content of the original file provided by the user and quickly generate a structured presentation. Especially when processing a large amount of data and complex information, the present invention also demonstrates a powerful data processing and integration capability without the need for manual sorting and operation by the user. In addition, the present invention can also reduce the dependence on professional knowledge and reduce training costs, so that non-professional users can also quickly obtain professional-level presentations. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of a method for automatically generating presentations based on a large language model provided in Example 1.

[0018] Figure 2 This is a schematic diagram of the generation outline provided in Example 1.

[0019] Figure 3 This is the outline rendering generated in Example 1.

[0020] Figure 4 This is a rendering of a target presentation generated in Example 1.

[0021] Figure 5 This is a rendering of another target presentation generated in Example 1.

[0022] Figure 6 This is a specific flow chart of a method for automatically generating presentations based on a large language model provided in Example 2. DETAILED DESCRIPTION

[0023] The drawings are for illustrative purposes only and should not be construed as limiting the present patent; In order to better illustrate the present embodiment, some parts in the drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product; It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0024] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0025] Example 1 like Figure 1 As shown, this embodiment provides a method for automatically generating a presentation based on a large language model, comprising the following steps: S1: Extract presentation requirement text from the original file provided by the user; S2: inputting the presentation requirement text into the big model, obtaining key information and themes of the requirement text through the model capability of the big model, and generating a presentation outline; S3: using the large model to expand each part of the presentation outline in detail to obtain the presentation text; S4: using the large model to automatically fill the presentation text into a preset presentation template, and generate a target presentation corresponding to the presentation requirement text.

[0026] In the specific implementation process, taking the generation of a PPT with the theme of "Concept Report on Productivity" as an example, firstly, the presentation requirement text is extracted from the original file provided by the user to determine the theme of the PPT; Then, the presentation requirement text is input into the big model, and the key information and themes of the requirement text are obtained through the model capability of the big model, and the presentation outline is generated, such as Figure 2 The following is a schematic diagram of generating an outline. Figure 3 This is the rendering of the generated outline; Then, the large model is used to expand each part of the presentation outline in detail to obtain the presentation text; Finally, the large model is used to automatically fill the presentation text into the preset presentation template to generate the target presentation corresponding to the presentation requirement text, such as Figure 4 and 5 Shown are the resulting renderings of the two target presentations; This method can significantly improve the user's work efficiency and reduce the tediousness of manual operations; at the same time, this method can intelligently analyze the content of the original file provided by the user and quickly generate a structured presentation, especially when processing large amounts of data and complex information, it also demonstrates powerful data processing and integration capabilities without the need for manual organization and operation by the user; in addition, this method can also reduce dependence on professional knowledge and reduce training costs, so that non-professional users can also quickly obtain professional-level presentations.

[0027] Example 2 This embodiment provides a method for automatically generating a presentation based on a large language model, comprising the following steps: S1: Extract presentation requirement text from the original file provided by the user; The original file provided by the user includes an unstructured demand file and a structured demand file; in this embodiment, the unstructured demand file is specifically a PDF file or a picture file, and the structured demand file is specifically an Excel file; Use OCR optical character recognition algorithm to extract presentation requirement text from unstructured requirement documents; use DuckDB to perform advanced data processing on structured requirement documents and extract structured data from structured requirement documents; S2: inputting the presentation requirement text into the big model, obtaining key information and themes of the requirement text through the model capability of the big model, and generating a presentation outline; Input the structured data extracted from the structured requirement file into the big model, use the big model to automatically generate the corresponding matplotlib drawing code, and draw the chart corresponding to the structured data according to the matplotlib drawing code; S3: using the large model to expand each part of the presentation outline in detail to obtain the presentation text; S4: using the large model to automatically fill the presentation text and charts into a preset presentation template, and generate a target presentation corresponding to the presentation requirement text; In this embodiment, the large model is specifically GPT-3 or BERT.

[0028] In the specific implementation process, Figure 6 As shown, the specific process of this embodiment is as follows: 1) Data input and conversion: In the first step of the entire presentation automatic generation process, various unstructured data files (such as PDF, pictures, etc.) provided by the user need to be converted into a processable text format; to achieve this process, this embodiment uses advanced text recognition technologies such as optical character recognition (OCR); OCR technology converts image data into machine-encoded text by scanning documents, recognizing characters in the documents. This process includes not only simple character recognition, but also complex processes such as layout recognition and font recognition, ensuring that the converted text data maintains the format and structural information of the original file; 2) Content summarization and outline generation: After converting the document content into text, the next step is to summarize the content and generate an outline. This method uses a large language model based on deep learning (such as GPT-3 or BERT) to conduct in-depth semantic understanding and summary of the text content, and automatically extract the central idea and key points of the text by filtering key information and topics through algorithms. On this basis, this method will automatically construct a logically clear presentation outline to ensure that the final presentation has a reasonable structure and well-organized content; the generation of the outline relies on the natural language processing capabilities of the language model, which can hierarchically sort information according to the inherent logic and importance of the text; In addition, based on the generated outline, this method further uses a large language model to expand each part of the outline in detail; in this step, in addition to adding necessary explanations and examples, the system will also appropriately reorganize and optimize the information to make the content richer and more complete; Content expansion is more than simple text generation, it also involves ensuring the logic and coherence of the text, so that the generated presentation text is both informative and easy to understand; 3) Presentation construction: After the content expansion is completed, the next step is to construct the presentation. This embodiment automatically fills the generated content into the preset presentation template by integrating Python's PPTX library. The PPTX library provides rich functions to support text formatting, image insertion, and dynamic effect addition, so that the final presentation is both beautiful and professional. 4) Advanced data processing: For documents containing complex Excel spreadsheet data, this method uses DuckDB for advanced data processing; DuckDB is an embedded SQL database engine that supports complex data query and analysis; by importing data tables into DuckDB, the system can build a data analysis model, automatically execute SQL queries, and accurately and quickly extract the required data; 5) Chart generation and insertion: The last step is chart generation and insertion. The corresponding matplotlib drawing code is automatically generated using the large language model to draw various charts (such as bar charts, line charts, pie charts, etc.) based on the extracted data. The charts are then inserted into the presentation in picture format, making the data presentation more intuitive and more attractive to the audience. This embodiment takes the generation of a PPT with the theme of "market analysis" as an example. The user first uploads an Excel file containing annual sales data and a PDF document of several pages of market analysis; first, the PDF file is converted into text using OCR technology, and then the large language model is used to extract key information from the text and generate a presentation outline; the Excel file is processed through DuckDB, an SQL query is constructed to obtain a specific data set, and then the automatically generated matplotlib code is used to draw a sales trend chart; finally, all the information is integrated and filled into the predetermined PPT template to generate a complete presentation; this process requires almost no human intervention, greatly improving work efficiency and document quality.

[0029] The same or similar reference numerals correspond to the same or similar components; The terms used in the drawings to describe positional relationships are only used for illustrative purposes and should not be construed as limiting this patent; Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A method for automatically generating presentations based on a large language model, characterized in that: The following steps are involved: S1: Extract presentation requirement text from the original file provided by the user; S2: inputting the presentation requirement text into the big model, obtaining key information and themes of the requirement text through the model capability of the big model, and generating a presentation outline; S3: using the large model to expand each part of the presentation outline in detail to obtain the presentation text; S4: using the large model to automatically fill the presentation text into a preset presentation template, and generate a target presentation corresponding to the presentation requirement text.

2. The method for automatically generating presentations based on a large language model according to claim 1, characterized in that: In step S1, the original file provided by the user includes an unstructured demand file.

3. The method for automatically generating presentations based on a large language model according to claim 2, characterized in that: The unstructured demand file is specifically a PDF file or a picture file.

4. The method for automatically generating presentations based on a large language model according to claim 2, characterized in that: Use OCR optical character recognition algorithm to extract presentation requirement text from unstructured requirement documents.

5. The method for automatically generating presentations based on a large language model according to claim 2, characterized in that: In step S1, the original file provided by the user also includes a structured demand file.

6. The method for automatically generating presentations based on a large language model according to claim 5, characterized in that: The structured demand file is specifically an Excel file.

7. The method for automatically generating presentations based on a large language model according to claim 5, characterized in that: When the original file provided by the user also includes a structured requirement file, extracting the presentation requirement text from the unstructured requirement file; At the same time, DuckDB is used to perform advanced data processing on the structured demand documents and extract structured data from the structured demand documents.

8. The method for automatically generating presentations based on a large language model according to claim 7, characterized in that: The step S2 further comprises: The structured data extracted from the structured requirement file is input into the big model, the big model is used to automatically generate the corresponding matplotlib drawing code, and the chart corresponding to the structured data is drawn according to the matplotlib drawing code.

9. The method for automatically generating presentations based on a large language model according to claim 8, characterized in that: The step S4 also includes: using the large model to automatically fill the presentation text and charts into a preset presentation template to generate a target presentation corresponding to the presentation requirement text.

10. A method for automatically generating presentations based on a large language model according to any one of claims 1 to 9, characterized in that: The large model is specifically GPT-3 or BERT.

Images