Intelligent official document generation method based on large model and knowledge base

Abstract

The invention discloses an intelligent official document generation method based on a large model and a knowledge base, and relates to the technical field of automatic office. Comprising the steps of 1, automatically collecting data files, 2, carrying out large model training and optimization, and 3, constructing a knowledge base; step 4, intelligently generating an official document outline, step 5, intelligently generating an official document, and step 6, visually displaying the generated official document. A large model and knowledge base technology is utilized, various knowledge sources such as internal data, user accumulation data and external data are integrated, rich material and information support is provided, context continuity of articles is improved, specification of official documents and depth and breadth of content are enhanced, office automation level and efficiency are greatly improved, and the method is suitable for popularization and application. Good popularization and application values are realized.

Description

Technical Field

[0001] The invention discloses a method for generating intelligent official documents based on a large model and a knowledge base, and relates to the technical field of automated office technology. Background Art

[0002] Document processing is an important daily task in government agencies, enterprises, institutions and other organizations. The existing document processing methods rely on manual work, which is inefficient and error-prone. In recent years, with the development of natural language processing technology, especially the application of large-scale pre-trained models (large models), new opportunities have been brought to document processing, and some automatic text generation technologies have emerged. However, these technologies often have the following limitations and shortcomings:

[0003] 1. Real-time retrieval and updating are not timely: The existing text generation system lacks an effective knowledge updating mechanism, which causes the generated content to be outdated or inaccurate. Especially for government agencies and enterprises and institutions, it is unable to reflect the latest policy changes and information in a timely manner.

[0004] 2. Poor knowledge base integration and material matching: The existing system has limited capabilities in integrating internal data, user accumulation and external data. It is difficult to accurately match relevant materials according to the subject and requirements of official documents, and it is also difficult to make full use of multi-source data.

[0005] 3. Poor contextual coherence: Automatically generated text often lacks contextual coherence, resulting in unclear logic in the generated document content.

[0006] 4. Insufficient user experience and interaction: Existing technologies or systems are insufficient in terms of user interaction and experience, making it difficult for users to conveniently modify and optimize the document outline. Summary of the invention

[0007] In response to the problems of the prior art, the present invention provides a smart document generation method based on a big model and a knowledge base. The method utilizes the big model and knowledge base technology to integrate multiple knowledge sources such as internal data, user accumulated data and external data, provides rich material and information support, improves the contextual coherence of the article, enhances the standardization of the official document, and the depth and breadth of the content, greatly improves the level and efficiency of office automation, and has good promotion and application value.

[0008] The specific scheme proposed by the present invention is:

[0009] A method for generating smart official documents based on a large model and a knowledge base, comprising:

[0010] Step 1: Automatically collect data files:

[0011] Preset keywords to filter data files.

[0012] Set up periodic tasks to collect data files,

[0013] Set keywords and filter conditions based on custom options and adjust collection strategies according to specific needs;

[0014] Step 2: Train and optimize the big model: Train the big model with user-provided data, external data, and public document data on the Internet. Regularly fine-tune and optimize the big model, continuously improve the model performance based on user feedback and new data, and ensure that the generated documents meet the regulatory requirements.

[0015] Step 3: Build a knowledge base: Store the metadata of official documents into the knowledge base. The metadata includes the document name, confidentiality level, owner, version number, update time, private or shared attributes of the document, data source, subject classification, and keywords; set the authorization level, and allow users to perform viewing, uploading, and delisting operations according to the authorization level; set the search conditions of keywords and time range for accurate search, generate material tables, and meet the needs of different users; record the source of each official document, including uploader information and data source, for tracing and auditing;

[0016] Step 4: Intelligently generate document outlines: Refer to the documents uploaded by users, identify and extract key information, integrate and rewrite the outlines, form new outlines, and analyze the document content, including identifying core ideas, key data and arguments, and analyzing the logical relationship between core ideas, key data and arguments.

[0017] Step 5: Intelligently generate official documents: Intelligently match the keywords, titles and dates entered by the user, and combine the official document outline to find the document excerpt that best matches the user's needs. Generate a draft document through context analysis, polish the language and adjust the format of the draft document to ensure that it meets the writing standards and style requirements of the official document.

[0018] Step 6: Visually display the generated documents.

[0019] Furthermore, in step 1 of the method for generating smart official documents based on a large model and a knowledge base, a periodic task is set to regularly check and update the files in the knowledge base to maintain the timeliness of the information.

[0020] Furthermore, in step 1 of the smart document generation method based on a large model and a knowledge base, an instant trigger mechanism is set. After a new policy is released, the document collection process is immediately triggered to ensure the timely inclusion of the latest information.

[0021] Furthermore, in step 6 of the method for generating smart official documents based on a large model and a knowledge base, the generated official documents are visually displayed on a web page in a streaming output form, so that users can intuitively view and edit the content of the official documents.

[0022] The present invention also provides a smart document generation device based on a large model and a knowledge base, comprising a collection module, a large model management module, a knowledge base management module, a document generation management module and a display module.

[0023] The collection module automatically collects data files:

[0024] Preset keywords to filter data files.

[0025] Set up periodic tasks to collect data files,

[0026] Set keywords and filter conditions based on custom options and adjust collection strategies according to specific needs;

[0027] The large model management module trains and optimizes the large model: the large model is trained by combining the data provided by users, external data, and the official document data publicly available on the Internet. The large model is regularly fine-tuned and optimized. The model performance is continuously improved based on user feedback and new data to ensure that the generated official documents meet the regulatory requirements.

[0028] The knowledge base management module builds a knowledge base: the metadata of official documents is stored in the knowledge base, including the document name, confidentiality level, owner, version number, update time, private or shared attributes of the document, data source, subject classification and keywords; the authorization level is set, and users are allowed to perform viewing, uploading and delisting operations according to the authorization level; the search conditions of keywords and time range are set for accurate search and generation of material tables to meet the needs of different users; the source of each official document is recorded, including uploader information and data source, for tracing and auditing;

[0029] The document generation management module intelligently generates document outlines: referring to documents uploaded by users, identifying and extracting key information, integrating and rewriting outlines, forming new outlines, and analyzing document content, including identifying core ideas, key data and arguments, and analyzing the logical relationship between core ideas, key data and arguments.

[0030] The document generation management module generates documents intelligently: it intelligently matches keywords, titles and dates entered by users, finds the document excerpt that best matches the user's needs in combination with the document outline, generates a document draft through context analysis, and performs language polishing and formatting adjustments on the document draft to ensure that it complies with the writing standards and style requirements of the document.

[0031] The display module will visually display the generated official documents.

[0032] Furthermore, the collection module of the intelligent document generation device based on a large model and a knowledge base sets periodic tasks to regularly check and update the files in the knowledge base to maintain the timeliness of the information.

[0033] Furthermore, the collection module of the intelligent document generation device based on a large model and a knowledge base is provided with an instant trigger mechanism, which immediately triggers the document collection process after a new policy is issued, thereby ensuring the timely inclusion of the latest information.

[0034] Furthermore, the display module of the intelligent document generation device based on the large model and the knowledge base visually displays the generated document in the form of streaming output on a Web page, allowing users to intuitively view and edit the document content.

[0035] The benefits of the present invention are:

[0036] By efficiently and accurately automatically generating the outline and content of official documents, the workload of manual writing is reduced and the efficiency of official document processing is improved. Especially for report-type official documents, the automated file collection function and the intelligent knowledge base management mechanism not only ensure the accuracy and latest of the referenced knowledge, but also further strengthen the protection of information security through classified and graded authority management. In addition, for official documents such as notices, requests, letters, minutes, etc., even if they do not rely on the support of the knowledge base, they can independently produce high-standard and high-quality official documents that meet regulatory requirements with their built-in intelligent generation capabilities. It can play a key role in a variety of official document writing scenarios, ensuring the flexibility and efficiency of official document processing, and has good application prospects and promotion value. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a schematic flow chart of the method of the present invention. DETAILED DESCRIPTION

[0038] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0039] Example 1

[0040] The present invention provides a method for generating smart documents based on a large model and a knowledge base, comprising:

[0041] Step 1: Automatically collect data files:

[0042] Preset keywords to filter data files, and automatically filter out relevant files based on preset keywords to ensure that the collected information is closely related to the subject of the official document report.

[0043] Set periodic tasks to collect data files, regularly check and update files in the knowledge base, and keep the information up to date.

[0044] Set keywords and filter conditions based on custom options, and adjust collection strategies according to specific needs to improve the pertinence and efficiency of data collection.

[0045] An instant trigger mechanism is also set up to trigger the document collection process immediately after the release of new policies to ensure the timely inclusion of the latest information.

[0046] Step 2: Train and optimize the big model: Train the big model with user-provided data, external data, and public document data on the Internet. Regularly fine-tune and optimize the big model, continuously improve model performance based on user feedback and new data, and ensure that the generated documents meet regulatory requirements.

[0047] Step 3: Build a knowledge base: Store the metadata of official documents into the knowledge base. The metadata includes the document name, confidentiality level, owner, version number, update time, private or shared attributes of the document, data source, subject classification, and keywords; set the authorization level, and allow users to perform viewing, uploading, and delisting operations according to the authorization level; set the search conditions of keywords and time range for accurate search, generate material tables, and meet the needs of different users; record the source of each official document, including uploader information and data source, for traceability and auditing. This ensures the security, integrity, and traceability of documents in the knowledge base, and provides users with a reliable and efficient knowledge base management method.

[0048] Step 4: Intelligently generate official document outlines: Refer to the documents uploaded by users, identify and extract key information, integrate and rewrite the outlines, form new outlines, and analyze the document content, including identifying core ideas, key data and arguments, and analyzing the logical relationship between core ideas, key data and arguments. It is particularly suitable for scenarios that require rapid response and processing of urgent official document tasks, greatly improving the efficiency of official document writing. For the newly generated outline, users can make further adjustments and rewrite it according to their specific requirements to ensure that the final official document content not only meets the standard requirements, but also reflects personalized characteristics.

[0049] Step 5: Intelligently generate official documents: Intelligently match the keywords, titles and dates entered by the user, and combine the official document outline to find the document excerpt that best matches the user's needs. Generate a draft document through context analysis, polish the language and adjust the format of the draft document to ensure that it meets the writing standards and style requirements of the official document. When using the knowledge base, it can intelligently recommend knowledge and cases related to the topic of the official document, and support knowledge tracing, tracing the source and basis of the content generated by the model, so as to help users judge whether the generated content is authentic and reliable, and ensure the authority and accuracy of the official document.

[0050] Step 6: Visually display the generated official document. Visually display the generated official document in the form of streaming output on the web page, so that users can intuitively view and edit the content of the official document.

[0051] Example 2

[0052] The present invention also provides a smart document generation device based on a large model and a knowledge base, comprising a collection module, a large model management module, a knowledge base management module, a document generation management module and a display module.

[0053] The collection module automatically collects data files:

[0054] Preset keywords to filter data files.

[0055] Set up periodic tasks to collect data files,

[0056] Set keywords and filter conditions based on custom options and adjust collection strategies according to specific needs;

[0057] The large model management module trains and optimizes the large model: the large model is trained by combining the data provided by users, external data, and the official document data publicly available on the Internet. The large model is regularly fine-tuned and optimized. The model performance is continuously improved based on user feedback and new data to ensure that the generated official documents meet the regulatory requirements.

[0058] The knowledge base management module builds a knowledge base: the metadata of official documents is stored in the knowledge base, including the document name, confidentiality level, owner, version number, update time, private or shared attributes of the document, data source, subject classification and keywords; the authorization level is set, and users are allowed to perform viewing, uploading and delisting operations according to the authorization level; the search conditions of keywords and time range are set for accurate search and generation of material tables to meet the needs of different users; the source of each official document is recorded, including uploader information and data source, for tracing and auditing;

[0059] The document generation management module intelligently generates document outlines: referring to documents uploaded by users, identifying and extracting key information, integrating and rewriting outlines, forming new outlines, and analyzing document content, including identifying core ideas, key data and arguments, and analyzing the logical relationship between core ideas, key data and arguments.

[0060] The document generation management module generates documents intelligently: it intelligently matches keywords, titles and dates entered by users, finds the document excerpt that best matches the user's needs in combination with the document outline, generates a document draft through context analysis, and performs language polishing and formatting adjustments on the document draft to ensure that it complies with the writing standards and style requirements of the document.

[0061] The display module will visually display the generated official documents.

[0062] As the information interaction and execution process between the modules in the above-mentioned device are based on the same concept as the embodiment of the method of the present invention, the specific contents can be found in the description of the embodiment of the method of the present invention and will not be repeated here.

[0063] Similarly, the device of the present invention reduces the workload of manual writing and improves the efficiency of document processing by efficiently and accurately automatically generating the outline and content of official documents. Especially for report-type official documents, the automated file collection function and the intelligent knowledge base management mechanism not only ensure the accuracy and currency of the referenced knowledge, but also further strengthen the protection of information security through classified and graded authority management. In addition, for official documents such as notices, requests, letters, minutes, etc., even if they do not rely on the support of the knowledge base, they can independently produce high-standard and high-quality official documents that meet regulatory requirements with their built-in intelligent generation capabilities. It can play a key role in a variety of official document writing scenarios, ensuring the flexibility and efficiency of official document processing, and has good application prospects and promotion value.

[0064] It should be noted that not all steps and modules in the above-mentioned processes and device structures are necessary, and some steps or modules can be ignored according to actual needs. The execution order of each step is not fixed and can be adjusted as needed. The system structure described in the above-mentioned embodiments can be a physical structure or a logical structure, that is, some modules may be implemented by the same physical entity, or some modules may be implemented by multiple physical entities, or some components in multiple independent devices may be implemented together.

[0065] The above-described embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Equivalent substitutions or changes made by those skilled in the art based on the present invention are within the protection scope of the present invention. The protection scope of the present invention shall be subject to the claims.

Claims

1. A method for generating smart documents based on a large model and a knowledge base, characterized by: include: Step 1: Automatically collect data files: Preset keywords to filter data files. Set up periodic tasks to collect data files, Set keywords and filter conditions based on custom options and adjust collection strategies according to specific needs; Step 2: Train and optimize the big model: Train the big model with user-provided data, external data, and public document data on the Internet. Regularly fine-tune and optimize the big model, continuously improve the model performance based on user feedback and new data, and ensure that the generated documents meet the regulatory requirements. Step 3: Build a knowledge base: Store the metadata of official documents into the knowledge base. The metadata includes the document name, confidentiality level, owner, version number, update time, private or shared attributes of the document, data source, subject classification, and keywords; set the authorization level, and allow users to perform viewing, uploading, and delisting operations according to the authorization level; set the search conditions of keywords and time range for accurate search, generate material tables, and meet the needs of different users; record the source of each official document, including uploader information and data source, for tracing and auditing; Step 4: Intelligently generate document outlines: Refer to the documents uploaded by users, identify and extract key information, integrate and rewrite the outlines, form new outlines, and analyze the document content, including identifying core ideas, key data and arguments, and analyzing the logical relationship between core ideas, key data and arguments. Step 5: Intelligently generate official documents: Intelligently match the keywords, titles and dates entered by the user, and combine the official document outline to find the document excerpt that best matches the user's needs. Generate a draft official document through context analysis, polish the language and adjust the format of the draft official document to ensure that it meets the writing standards and style requirements of the official document. Step 6: Visually display the generated documents.

2. According to claim 1, a method for generating smart documents based on a large model and a knowledge base is characterized by: In step 1, set periodic tasks to regularly check and update the files in the knowledge base to keep the information up to date.

3. The method for generating smart documents based on a large model and a knowledge base according to claim 1, characterized in that In step 1, an immediate trigger mechanism is set up to trigger the document collection process immediately after the new policy is released to ensure the timely inclusion of the latest information.

4. The method for generating smart documents based on a large model and a knowledge base according to claim 1, characterized in that In step 6, the generated official document is visually displayed on a web page in a streaming output form, so that the user can intuitively view and edit the content of the official document.

5. A smart document generation device based on a large model and a knowledge base, characterized by It includes collection module, large model management module, knowledge base management module, document generation management module and display module. The collection module automatically collects data files: Preset keywords to filter data files. Set up periodic tasks to collect data files, Set keywords and filter conditions based on custom options and adjust collection strategies according to specific needs; The large model management module trains and optimizes the large model: the large model is trained by combining the data provided by users, external data, and the official document data publicly available on the Internet. The large model is regularly fine-tuned and optimized. The model performance is continuously improved based on user feedback and new data to ensure that the generated official documents meet the regulatory requirements. The knowledge base management module builds a knowledge base: the metadata of official documents is stored in the knowledge base, including the document name, confidentiality level, owner, version number, update time, private or shared attributes of the document, data source, subject classification and keywords; the authorization level is set, and users are allowed to perform viewing, uploading and delisting operations according to the authorization level; the search conditions of keywords and time range are set for accurate search and generation of material tables to meet the needs of different users; the source of each official document is recorded, including uploader information and data source, for tracing and auditing; The document generation management module intelligently generates document outlines: referring to documents uploaded by users, identifying and extracting key information, integrating and rewriting outlines, forming new outlines, and analyzing document content, including identifying core ideas, key data and arguments, and analyzing the logical relationship between core ideas, key data and arguments. The document generation management module generates documents intelligently: it intelligently matches keywords, titles and dates entered by users, finds the document excerpt that best matches the user's needs in combination with the document outline, generates a document draft through context analysis, and performs language polishing and formatting adjustments on the document draft to ensure that it complies with the writing standards and style requirements of the document. The display module will visually display the generated official documents.

6. The device for generating smart documents based on a large model and a knowledge base according to claim 1, characterized in that The collection module sets periodic tasks to regularly check and update files in the knowledge base to keep the information current.

7. The device for generating smart documents based on a large model and a knowledge base according to claim 5, characterized in that The collection module is equipped with an instant trigger mechanism. When a new policy is released, the document collection process is triggered immediately to ensure the timely inclusion of the latest information.

8. The device for generating smart documents based on a large model and a knowledge base according to claim 5, characterized in that The display module displays the generated official documents visually on the Web page in the form of streaming output, allowing users to intuitively view and edit the content of the official documents.

Images