Key point information generation method and device, equipment and storage medium

Abstract

The invention provides a key point information generation method and device, equipment and a storage medium, and relates to the technical field of artificial intelligence, in particular to the technical fields of deep learning, natural language processing, intelligent documents and the like. The method comprises the steps of generating a segmentation outline according to document contents of paragraphs in a to-be-processed document; according to preset theme information and the segmentation outline, paragraphs in the to-be-processed document are aggregated, and multiple aggregated theme contents are obtained; the theme content is segmented according to preset sub-themes, multiple segmented sub-theme content is obtained, and the sub-themes are sub-classes of themes obtained by dividing the theme information; and generating key point information for the to-be-processed document according to the information of the paragraphs in the multiple sub-theme contents. According to the method, the generation time of the document key point information is shortened, and the generation efficiency of the key point information is improved.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, specifically to the fields of deep learning, natural language processing and intelligent documents, and in particular to a method, apparatus, device and storage medium for generating key information. Background Technology

[0002] With the popularization of digital learning and document management, multi-format documents such as Word, PPT, and PDF have become the core carriers of knowledge transfer. The need to quickly extract core logic from massive documents and generate structured notes to meet the needs of enterprise document management, educational lesson preparation and other scenarios is becoming increasingly urgent. Summary of the Invention

[0003] This disclosure provides a method, apparatus, device, and storage medium for generating key information.

[0004] According to a first aspect of this disclosure, a method for generating key information is provided, comprising: generating a segmentation outline based on the document content of paragraphs in a document to be processed; aggregating the paragraphs in the document to be processed based on preset topic information and the segmentation outline to obtain multiple aggregated topic contents; segmenting the topic contents based on preset subtopics to obtain multiple segmented subtopic contents, wherein a subtopic is a sub-category topic obtained by dividing the topic information; and generating key information for the document to be processed based on the information of the paragraphs in the multiple subtopic contents.

[0005] According to a second aspect of this disclosure, a key information generation apparatus is provided, comprising: an outline generation module configured to generate a segmented outline based on the document content of paragraphs in a document to be processed; an aggregation module configured to aggregate the paragraphs in the document to be processed based on preset topic information and the segmented outline to obtain multiple aggregated topic contents; a segmentation module configured to segment the topic contents based on preset subtopics to obtain multiple segmented subtopic contents, wherein the subtopics are subcategories of topics obtained by dividing the topic information; and an information generation module configured to generate key information for the document to be processed based on the information of the paragraphs in the multiple subtopic contents.

[0006] According to a third aspect of this disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform a method as described in any implementation of the first aspect.

[0007] According to a fourth aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions for causing a computer to perform a method as described in any implementation of the first aspect.

[0008] According to a fifth aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the method as described in any implementation of the first aspect.

[0009] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0010] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein: Figure 1 This is an exemplary system architecture diagram to which this disclosure can be applied; Figure 2 This is a flowchart of an embodiment of a method for generating key information according to this disclosure; Figure 3 This is a flowchart of another embodiment of the method for generating key information according to this disclosure; Figure 4 This is a flowchart of one embodiment of updating document content; Figure 5 This is a flowchart of yet another embodiment of the method for generating key information according to this disclosure; Figure 6 This is a schematic diagram of an application process for generating key information based on this disclosure; Figure 7 This is a schematic diagram of the structure of an embodiment of the apparatus for generating key information according to this disclosure; Figure 8 This is a block diagram of an electronic device used to implement the method for generating key information according to embodiments of the present disclosure. Detailed Implementation

[0011] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0012] It should be noted that, unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other. This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0013] Figure 1 An exemplary system frame 100 is shown, which can be used to illustrate an embodiment of the method or apparatus for generating key information of this disclosure.

[0014] like Figure 1 As shown, system architecture 100 may include terminal devices 101, 102, and 103, a network 104, and a server 105. Network 104 serves as the medium for providing communication links between terminal devices 101, 102, and 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0015] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send information, etc. Various client applications can be installed on terminal devices 101, 102, and 103.

[0016] Terminal devices 101, 102, and 103 can be either hardware or software. When terminal devices 101, 102, and 103 are hardware, they can be various electronic devices, including but not limited to smartphones, tablets, laptops, and desktop computers. When terminal devices 101, 102, and 103 are software, they can be installed in the aforementioned electronic devices. They can be implemented as multiple software programs or software modules, or as a single software program or software module. No specific limitations are made here.

[0017] Server 105 can provide various services. For example, server 105 can analyze and process documents to be processed obtained from terminal devices 101, 102, and 103, and generate processing results (such as key information of the documents to be processed).

[0018] It should be noted that server 105 can be either hardware or software. When server 105 is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or as a single server. When server 105 is software, it can be implemented as multiple software programs or software modules (for example, used to provide distributed services), or as a single software program or software module. No specific limitations are made here.

[0019] It should be noted that the method for generating key information provided in this embodiment is generally executed by server 105, and correspondingly, the device for generating key information is generally located in server 105.

[0020] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0021] Continue to refer to Figure 2 The diagram illustrates a flow 200 of an embodiment of a method for generating key information according to the present disclosure. This method for generating key information includes the following steps: Step 201: Generate a segment outline based on the document content of the paragraphs in the document to be processed.

[0022] In this embodiment, the execution entity of the method for generating key information (e.g.) Figure 1 The server 105 shown first acquires the document to be processed, which is the document to be studied and for which key information needs to be generated. The document can be in formats such as Word, PPT, or PDF. After acquiring the document, the aforementioned execution entity parses it to obtain the content of each paragraph. Specifically, the execution entity first determines the document format and then uses a parsing tool corresponding to that format to parse the document, thereby obtaining the content of each paragraph. For example, if the document format is PDF, the execution entity will use the PDFBox tool to parse the document, thus obtaining its content.

[0023] Next, the aforementioned execution entity will generate segment outlines based on the document content of each paragraph of the document to be processed; that is, it extracts segment outlines from the document content using paragraphs as nodes. For example, the execution entity can extract multiple keywords from each paragraph and generate segment outlines based on these keywords; alternatively, it can use a large language model to extract segment outlines from the document content.

[0024] In addition, the aforementioned executing entity will use the original heading levels in the document to be processed as a benchmark to determine the hierarchical relationship of each section outline, thereby ensuring that the hierarchy of the section outline is clear.

[0025] Step 202: Based on the preset topic information and segmentation outline, aggregate the paragraphs in the document to be processed to obtain multiple aggregated topic contents.

[0026] In this embodiment, the execution entity presets multiple topic information. This topic information is a broad category, which may also include multiple sub-topics. For example, the topic information could be: the background of AI (Artificial Intelligence), the applications of AI, the challenges of AI, etc. The execution entity determines the topic information corresponding to each paragraph based on the content of the segmented outline, and then aggregates paragraphs with the same topic information together, thus obtaining multiple aggregated topic contents. For example, a clustering algorithm can be used to aggregate paragraphs with the same topic information; alternatively, a pre-trained clustering model can also be used to aggregate paragraphs with the same topic information.

[0027] It should be noted that there can be one or more documents to be processed. When there is only one document to be processed, the execution entity will aggregate multiple paragraphs in that document. When there are multiple documents to be processed, the execution entity will determine the topic information of each paragraph in each document to be processed, and then aggregate all paragraphs in the multiple documents to be processed, thereby achieving the aggregation of multiple documents.

[0028] Step 203: Divide the topic content according to the preset subtopics to obtain multiple subtopic contents.

[0029] In this embodiment, the execution entity segments the topic content according to preset sub-topics, obtaining multiple sub-topic contents. A sub-topic is a sub-category topic obtained by dividing the topic information. Since the aforementioned steps cluster the topic content based on the topic information, for each topic content, the execution entity first determines the sub-topic corresponding to the topic information. A sub-topic is a sub-category topic or subdivided topic obtained by dividing the topic information. It can be pre-generated. After setting the topic information, it can be further subdivided according to the domain, type, and other information of the topic information, thereby obtaining multiple sub-categories of topics under the topic information, i.e., sub-topics. For example, if the topic information is about AI applications, then the topic information can be divided according to different application domains to obtain multiple sub-topics, i.e., sub-topics can include: AI applications in manufacturing, AI applications in finance, AI applications in healthcare, etc.

[0030] After determining the sub-topics, the aforementioned execution entity will segment the topic content according to the sub-topics, that is, divide paragraphs with the same sub-topics into a part, thereby further segmenting the content under a major topic into multiple sub-topic contents.

[0031] Step 204: Generate key information for the document to be processed based on the information from paragraphs in multiple subtopics.

[0032] In this embodiment, the aforementioned execution entity generates key information for the document to be processed based on the information of paragraphs in multiple sub-topics. This key information can be a learning path, which refers to the path planning of the learning process. The learning path is a closed-loop learning path generated based on the outline to locate the association of knowledge points. Specifically, it can include path information such as learning order, key modules, prior knowledge requirements, and time planning. It can be in the form of text path descriptions, flowcharts, or mind maps. The key information can also be a summary of all important knowledge points in the document to be processed, which can be in the form of text path descriptions, flowcharts, or mind maps.

[0033] Specifically, for each subtopic, the aforementioned execution entity generates key information for that subtopic based on the information of the paragraphs contained within it. Finally, it summarizes the key information from multiple subtopics to obtain the key information for the document to be processed.

[0034] For each subtopic, the execution entity first determines the original content of the paragraphs included in the subtopic and the starting position of these paragraphs in the document to be processed. Then, it uses the large model to generate key information for the subtopic. After determining the key information for each subtopic, it uses the large model to generate key information for the document to be processed.

[0035] The method for generating key information provided in this embodiment first generates a segmentation outline based on the document content of paragraphs in the document to be processed; then, based on preset topic information and the segmentation outline, the paragraphs in the document to be processed are aggregated to obtain multiple aggregated topic contents; next, the topic contents are segmented according to preset subtopics to obtain multiple segmented subtopic contents, where a subtopic is a subclass of topic information; finally, key information for the document to be processed is generated based on the information of paragraphs in the multiple subtopic contents. This method for generating key information first aggregates paragraphs with the same topic information using the segmentation outline of the document to be processed, then segments the aggregated content according to the subtopics under the topic information, thereby improving the topic focus of the document information and the accuracy of aggregation, solving problems such as scattered information and topic deviation in document processing; finally, it generates key information for the document to be processed based on the segmented content, thus shortening the generation time of document key information and improving the generation efficiency of key information. This method for generating key information can be widely applied to scenarios such as educational lesson preparation, enterprise document management, and personal learning.

[0036] Furthermore, the acquisition, storage, use, processing, transportation, provision, and disclosure of user personal information (such as documents to be processed in this disclosure) involved in the technical solutions disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0037] Continue to refer to Figure 3 , Figure 3 A flow 300 of another embodiment of the method for generating key information according to this disclosure is shown. The method for generating key information includes the following steps: Step 301: Determine the document format of the document to be processed, and use a parsing tool corresponding to the document format to parse the document to obtain the document content.

[0038] In this embodiment, the execution entity of the method for generating key information (e.g.) Figure 1 The server 105 shown first determines the document format of the document to be processed, and then uses a parsing tool corresponding to the document format to parse the document to obtain the document content of each paragraph in the document to be processed. That is, different document formats correspond to different parsing tools. For example, if the document format of the document to be processed is PDF, the above-mentioned execution entity will use the PDFBox tool to parse the document to be processed, thereby obtaining the document content of each paragraph in the document to be processed.

[0039] Step 302: Generate a segmented outline based on the document content.

[0040] In this embodiment, the execution entity generates segmented outlines based on the document content of the document to be processed, that is, extracting segmented outlines from the document content using paragraphs as nodes. For example, the execution entity can extract multiple keywords from each paragraph and generate segmented outlines based on these keywords; alternatively, a large language model can be used to extract segmented outlines from the document content. Furthermore, the execution entity also uses the original heading levels in the document to determine the hierarchical relationship of each segmented outline, thereby ensuring a clear hierarchy in the segmented outlines.

[0041] By first parsing the document to be processed to generate the content of each paragraph, and then generating a segment outline based on the document content, the efficiency and accuracy of segment outline generation are improved.

[0042] Step 303: Determine the topic information of the paragraphs in the document to be processed based on the segmentation outline.

[0043] In this embodiment, the execution entity will preset multiple topic information. The topic information here is a broad category, which may also include multiple sub-topics. For example, the topic information may be: the background of AI, the application of AI, the challenges of AI, etc. The execution entity will determine the topic information corresponding to each paragraph according to the content of the segment outline.

[0044] Step 304: Use a clustering algorithm to aggregate paragraphs with the same topic information to obtain multiple aggregated topic contents.

[0045] In this embodiment, the execution entity uses a clustering algorithm to group paragraphs with the same topic information together, thereby obtaining multiple aggregated topic contents. The clustering algorithm used here can be an existing clustering algorithm, such as the K-means clustering algorithm, which will not be elaborated upon here.

[0046] It should be noted that there can be one or more documents to be processed. When there is only one document to be processed, the execution entity will aggregate multiple paragraphs in that document. When there are multiple documents to be processed, the execution entity will determine the topic information of each paragraph in each document to be processed, and then aggregate all paragraphs in the multiple documents to be processed, thereby achieving the aggregation of multiple documents.

[0047] This allows the paragraphs of the document to be processed to be aggregated based on the topic information, thereby improving the relevance between the aggregated paragraphs and the accuracy of the aggregation results.

[0048] Step 305: Input each topic content into the pre-trained large model, and use the large model to verify the relevance of paragraphs in the topic content.

[0049] In this embodiment, the aforementioned execution entity will also input each aggregated topic content into a pre-trained large model, so that the large model can verify the relevance of each paragraph in each topic content, that is, the large model will verify the accuracy of the aggregation result. Specifically, the large model will identify whether the topic of each paragraph in the topic content deviates from the topic content, determine the correlation, coherence and completeness between each paragraph, and thus obtain the verification result.

[0050] Step 306: Based on the verification results, identify the problematic paragraphs that do not meet the preset conditions and remove them from the main content.

[0051] In this embodiment, if the execution entity determines that the subject content contains paragraphs that do not meet the preset conditions, it will record them as problematic paragraphs and remove them from the subject content. The preset conditions here may be that the topic of the paragraph deviates from the topic of the subject content, or that the correlation, coherence and completeness between paragraphs do not meet the preset conditions.

[0052] This allows for secondary verification of the aggregation results through a large model, and the removal of questionable content, thereby improving the relevance of the aggregated paragraphs.

[0053] Step 307: Divide the topic content according to the preset subtopics to obtain multiple subtopic contents.

[0054] Step 308: Generate key information for the document to be processed based on the information from paragraphs in multiple subtopics.

[0055] Steps 307-308 are basically the same as steps 203-204 in the aforementioned embodiments. For specific implementation methods, please refer to the aforementioned description of steps 203-204, which will not be repeated here.

[0056] from Figure 3 It can be seen from this that, with Figure 2 Compared to the corresponding embodiments, the method for generating key information in this embodiment emphasizes the steps of generating segmented outlines and aggregating paragraphs of the document to be processed. First, the document to be processed is parsed to generate the content of each paragraph. Then, a segmented outline is generated based on the document content, thereby improving the efficiency and accuracy of the segmented outline generation. Next, the paragraphs of the document to be processed are aggregated according to the topic information, and the aggregation results are verified a second time using a large model, removing questionable content. This improves the relevance of the aggregated paragraphs and enhances the accuracy of the aggregation results.

[0057] Continue to refer to Figure 4 , Figure 4 A process 400 illustrating one embodiment of generating document content includes: Step 401: Extract the text content from each page of the document to be processed.

[0058] For each page of the document to be processed, the aforementioned execution entity will extract the text content of that page. Specifically, the execution entity will call a format parsing tool to extract the text content of each page sequentially by page number and count the number of words extracted.

[0059] Step 402: Determine the target page document based on the number of words in the text content and the preset word count threshold.

[0060] The aforementioned execution entity compares the number of words in the current page document with a preset word count threshold, and determines the target page document based on the comparison result. The target page document is a chart / graph type document; that is, the comparison result determines whether the current page document is a chart / graph type document. If the number of words in the current page document is less than the preset word count threshold, then the page document can be determined to be a chart / graph type document.

[0061] Step 403: Update the document content of the target page document using the multimodal large model.

[0062] If the executing entity determines that the current page document is the target page document (i.e., a chart-type document), it will use a multimodal large model to generate the document content for the target page document and update the previous document content of the target page document with the generated document content. Conversely, if the number of characters in the text content of the current page document exceeds a preset character threshold, it is determined that the current page document is not a chart-type document, and the document content can be directly generated based on the parsing results of that page document.

[0063] In some optional implementations of this embodiment, step 403 further includes: Step 4031: Obtain the first document content of the previous page of the document adjacent to the target page and the second document content of the next page of the document to be processed.

[0064] For the target page document, the aforementioned execution entity will obtain the document content of the previous page and the next page of the document to be processed that are adjacent to the target page document, thereby obtaining the first document content and the second document content.

[0065] In addition, the aforementioned executing entities can also obtain the segment outlines of the current page document, the previous page document, and the next page document.

[0066] Step 4032: Convert the target page document into an image, and input the image, the content of the first document, and the content of the second document into the multimodal large model, and output the target document content of the target page document.

[0067] If the current page document is determined to be a chart-type document, the aforementioned execution entity will convert the target page document into an image, and then input the converted image, the document content of the previous page document, the segment outline, and the document content and segment outline of the next page document into the multimodal large model, so that the multimodal large model can generate and output the document content of the current page document based on the input information.

[0068] Step 4033: Update the document content of the target page document using the target document content.

[0069] The aforementioned execution entity updates the original document content of the target page document using the target document content generated by the multimodal large model. First, it uses a parsing tool to parse the document to be processed, generating the document content for each page. Further, the execution entity determines whether the current page document is a chart type document based on the number of characters in each page. If so, it uses the multimodal large model to generate the target document content for the current page document, and finally uses this target document content to update the original document content of the target page document. This improves the accuracy of chart type document content generation by using the multimodal large model and the context information of the current page document.

[0070] This allows us to determine whether a document is a chart or graph type by analyzing the number of words in its text. If it is determined to be a chart or graph type document, we can generate the document content using a multimodal large model and contextual information, thereby improving the completeness and accuracy of the document content.

[0071] Continue to refer to Figure 5 , Figure 5 A flowchart 500 is shown as another embodiment of a method for generating key information according to the present disclosure. This method for generating key information includes the following steps: Step 501: Generate a segment outline based on the document content of the paragraphs in the document to be processed.

[0072] Step 502: Based on the preset topic information and segmentation outline, aggregate the paragraphs in the document to be processed to obtain multiple aggregated topic contents.

[0073] Steps 501-502 are basically the same as steps 201-202 in the aforementioned embodiments. For specific implementation methods, please refer to the aforementioned description of steps 201-202, which will not be repeated here.

[0074] Step 503: Determine the subtopics corresponding to the paragraphs in the main content based on the segmentation outline.

[0075] In this embodiment, the execution entity of the method for generating key information (e.g.) Figure 1 The server 105 shown will determine the subtopics corresponding to the paragraphs in the main content based on the segmentation outline. The aforementioned steps cluster the main content based on the main information; therefore, for each main content, the executing entity will first determine the subtopics corresponding to the main content's main information. These subtopics are the subclasses or sub-categories of the main information. For example, if the main information is "Applications of AI," then the corresponding subtopics could include: "Applications of AI in Manufacturing," "Applications of AI in Finance," and "Applications of AI in Healthcare," etc.

[0076] Step 504: Segment the paragraphs with the same sub-topic in the main content to obtain multiple sub-topic contents.

[0077] In this embodiment, after determining the sub-topics, the aforementioned execution entity segments the topic content according to the sub-topics, that is, it divides paragraphs with the same sub-topic into a single part, thereby finely segmenting the content under a large topic category to obtain multiple sub-topic contents. This improves the topic focus of document information and the accuracy of aggregation, solving problems such as scattered information and topic deviation in document processing.

[0078] Step 505: Generate key information for the subtopic content based on the document content corresponding to each paragraph in the subtopic content and the position information of each paragraph in the document to be processed.

[0079] In this embodiment, for each subtopic content, the execution entity first determines the original content of the paragraphs included in the subtopic content, the document content corresponding to the paragraphs, and the starting position of these paragraphs in the document to be processed. Then, it uses the large model to generate key information for the subtopic content, such as a learning route.

[0080] Step 506: Summarize the key information of multiple sub-topics to obtain the key information for the document to be processed.

[0081] In this embodiment, the aforementioned execution entity summarizes the key information of multiple sub-topics to obtain key information for the document to be processed. For example, after determining the learning path for each sub-topic, the execution entity uses a large model to generate a learning path for the document to be processed. Here, the learning path refers to the path planning of the learning process. The learning path is a closed-loop learning path generated based on the outline's knowledge point connections. Specifically, it can include path information such as learning order, key modules, prior knowledge requirements, and time planning. It can be in the form of textual path descriptions, flowcharts, or mind maps. Thus, by first generating learning paths for sub-topics, and then generating the learning path for the document to be processed based on the learning paths of multiple sub-topics, the generation time of the document's learning path is shortened, and the generation efficiency of the learning path is improved.

[0082] from Figure 5 It can be seen from this that, with Figure 3 Compared with the corresponding embodiments, the method for generating key information in this embodiment emphasizes the steps of segmenting the topic content and generating key information of the document to be processed. This method generates key information for the document to be processed based on the segmented content, thereby shortening the generation time of key information and improving the generation efficiency of key information.

[0083] In some optional implementations of this embodiment, the method for generating the above-mentioned key information further includes: generating key notes corresponding to the document to be processed based on the key information, wherein the key notes include: a pilot page, a details page, and a summary page.

[0084] In this implementation, the aforementioned execution entity can generate key notes for the document to be processed based on the key information of the document to be processed. These key notes can include a pilot page, a detail page, and a summary page.

[0085] Specifically, the introductory page includes four sub-modules: Prior Knowledge, Learning Objectives and Gains, Mind Map, and Learning Navigation. The Prior Knowledge sub-module uses relatable, everyday scenarios to attract user attention. The Learning Objectives and Gains sub-module clearly outlines what knowledge users will gain after learning the material. The Mind Map sub-module provides the overall framework. The Learning Navigation sub-module details the learning methods for this document page.

[0086] The details page comprises five sub-modules: Key Points, Concept Cards, Case Study Cards, Question Cards, and Comparison Tables. The Key Points sub-module presents the core knowledge, outlining it in an easy-to-understand way. The Concept Cards sub-module supplements the Key Points, providing additional information. The Case Study Cards elaborate on relevant case studies from the document that relate to the core Key Points. The Question Cards address potential questions users may have while learning the document. The Comparison Tables present similar or contrasting concepts in tabular form.

[0087] The summary page comprises three sub-modules: Conclusion First, Summary of Learning Outline, and Content Enhancement. The Conclusion First sub-module presents the user with the conclusions after completing the document, enhancing the learning experience. The Summary of Learning Outline helps users re-organize their learning process. The Content Enhancement sub-module elevates the content learned by the user.

[0088] Finally, the aforementioned execution entity can generate HTML (Hyper Text Markup Language) and TTS (Text To Speech) based on the generated key notes. The HTML is used to generate a PPT for the document to be processed, and the TTS is used to generate a speech script for the document to be processed.

[0089] By following the steps above, complete key points notes are generated for users, and the entire learning process is presented to users through the introductory page, details page, and summary page, thus improving the user experience.

[0090] Continue to refer to Figure 6 , Figure 6 An application flow 600 illustrating a method for generating key information according to this disclosure is shown, comprising: Step 601: Obtain the document to be processed.

[0091] Step 602: Determine the document content and store it.

[0092] Step 603, updating the document content, further includes: Step 6031: Extract the text content of a single page of the document; Step 6032: Compare the number of characters in the text content with a preset character count threshold; If the value is below the threshold, proceed to steps 6033-6035; if the value is above the threshold, proceed to step 6036. Step 6033: If the value is below the threshold, convert the document page into an image; Step 6034: Input the images of this page and the document content of the adjacent pages into the multimodal large model; Step 6035: Output the speech transcript for this page. Step 6036: Add the generated speech draft from the single-page document to the document content to obtain the updated document.

[0093] Step 604: Generate a segmented outline using the updated document content.

[0094] Step 605: Aggregate documents on the same topic together using a combination of large and small models.

[0095] Small model: Using segmented outlines and paths, clustering algorithms are used to aggregate multiple documents on the same topic together; Large Model: For documents aggregated by the small model, the correctness of the clustering is verified a second time through the large model, and questionable documents are removed.

[0096] Step 606, HTML segmentation: This involves segmenting the aggregated content of multiple documents or multiple paragraphs from a single document into subtopics, grouping content with the same subtopic into a single paragraph. Segmentation logic: The segmentation outline is directly used to generate subtopics through a large model.

[0097] Step 607: Learning route generation.

[0098] For each HTML document generated from the segmented outline, the starting position of the original document content is obtained, the original content of each HTML document is acquired, and the original content of each HTML document is used to generate an easy-to-understand learning path with certain closed-loop requirements through the large model.

[0099] Step 608: Developer page generation.

[0100] The introductory page includes: prior knowledge, learning objectives and gains, mind map, and learning navigation.

[0101] Prior knowledge: To better attract user attention, this module incorporates everyday life scenarios in an easy-to-understand way; Learning objectives and gains: In a simple and direct way, it shows users what knowledge they can acquire after learning this material; Mind map: Provides users with the overall framework of this HTML; Learning navigation: Shows how to learn from this HTML.

[0102] Step 609: Detail page generated.

[0103] This module includes five sub-modules: key information, concept cards, case cards, question cards, and comparison tables.

[0104] Key Information: This module outlines the core knowledge points of a page, presenting them to the user in an easy-to-understand way; Concept Cards: These supplement the key information points by providing additional knowledge; Case Study Cards: These elaborate on case studies that relate to the core content of the key information points; Question Cards: These address potential questions users may have while learning the document; Comparison Tables: These present similar or contrasting concepts in a tabular format.

[0105] Step 610: Generate the summary page.

[0106] This module includes three sub-modules: Conclusion First, Summary of Context, and Content Enhancement.

[0107] Conclusion First: Present the conclusions to the user after learning this HTML to enhance the user learning experience; Structure Recap: Help the user reorganize the learning structure; Content Enhancement: Enhance the content learned.

[0108] Step 611: Generate key points notes.

[0109] Generate key points notes based on the above content, and generate HTML and TTS from the key points notes. You can generate PPT from HTML and a speech draft from TTS.

[0110] This generates learning paths, key points notes, and other information for users, thereby reducing their learning costs and helping them to understand relevant knowledge points more comprehensively and accurately.

[0111] Further reference Figure 7 As an implementation of the methods shown in the above figures, this disclosure provides an embodiment of a device for generating key information, which is similar to... Figure 2Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.

[0112] like Figure 7 As shown, the key information generation device 700 of this embodiment includes: an outline generation module 701, an aggregation module 702, a segmentation module 703, and an information generation module 704. The outline generation module 701 is configured to generate a segmented outline based on the document content of paragraphs in the document to be processed; the aggregation module 702 is configured to aggregate paragraphs in the document to be processed based on preset topic information and the segmented outline, obtaining multiple aggregated topic contents; the segmentation module 703 is configured to segment the topic contents based on preset subtopics, obtaining multiple segmented subtopic contents, wherein a subtopic is a sub-category topic obtained by dividing the topic information; and the information generation module 704 is configured to generate key information for the document to be processed based on the information of paragraphs in the multiple subtopic contents.

[0113] In this embodiment, the specific processing of the outline generation module 701, aggregation module 702, segmentation module 703, and information generation module 704 in the key information generation device 700, and the resulting technical effects, can be found in the following references: Figure 2 The relevant descriptions of steps 201-204 in the corresponding embodiments will not be repeated here.

[0114] In some optional implementations of this embodiment, the outline generation module 701 is further configured to: determine the document format of the document to be processed, and use a parsing tool corresponding to the document format to parse the document to be processed to obtain the document content; and generate a segmented outline based on the document content.

[0115] In some optional implementations of this embodiment, the above-mentioned key information generation device 700 further includes: an extraction module configured to extract text content from each page of the document to be processed; a judgment module configured to determine the target page document based on the number of characters in the text content and a preset character threshold, wherein the target page document is a document of the chart type; and a document content generation module configured to update the document content of the target page document using a multimodal large model.

[0116] In some optional implementations of this embodiment, the document content generation module is further configured to: obtain the first document content of the previous page document adjacent to the target page document and the second document content of the next page document in the document to be processed; convert the target page document into an image, and input the image, the first document content and the second document content into the multimodal large model, and output the target document content of the target page document; and update the document content of the target page document using the target document content.

[0117] In some optional implementations of this embodiment, the aggregation module 702 is further configured to: determine the topic information of paragraphs in the document to be processed according to the segmentation outline; and aggregate paragraphs with the same topic information using a clustering algorithm to obtain multiple aggregated topic contents.

[0118] In some optional implementations of this embodiment, the above-mentioned key information generation device 700 further includes: a verification module, configured to input each topic content into a pre-trained large model, and verify the relevance of paragraphs in the topic content through the large model; and a removal module, configured to determine problematic paragraphs that do not meet preset conditions based on the verification results, and remove problematic paragraphs from the topic content.

[0119] In some optional implementations of this embodiment, the segmentation module 703 is further configured to: determine the sub-topics corresponding to the paragraphs in the main content according to the segmentation outline; segment the paragraphs with the same sub-topics in the main content to obtain multiple sub-topic contents after segmentation.

[0120] In some optional implementations of this embodiment, the information generation module 704 is further configured to: generate key information for the subtopic content based on the document content corresponding to each paragraph in the subtopic content and the position information of each paragraph in the document to be processed; and summarize the key information of multiple subtopic contents to obtain key information for the document to be processed.

[0121] In some optional implementations of this embodiment, the above-mentioned key information generation device 700 further includes: a note generation module, configured to generate key notes corresponding to the document to be processed based on the key information, wherein the key notes include: a pilot page, a details page, and a summary page.

[0122] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0123] Figure 8 A schematic block diagram of an example electronic device 800 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0124] like Figure 8As shown, device 800 includes a computing unit 801, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 802 or a computer program loaded from storage unit 808 into random access memory (RAM) 803. RAM 803 may also store various programs and data required for the operation of device 800. The computing unit 801, ROM 802, and RAM 803 are interconnected via bus 804. Input / output (I / O) interface 805 is also connected to bus 804.

[0125] Multiple components in device 800 are connected to I / O interface 805, including: input unit 806, such as keyboard, mouse, etc.; output unit 807, such as various types of monitors, speakers, etc.; storage unit 808, such as disk, optical disk, etc.; and communication unit 809, such as network card, modem, wireless transceiver, etc. Communication unit 809 allows device 800 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0126] The computing unit 801 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 801 performs the various methods and processes described above, such as the method for generating key information. For example, in some embodiments, the method for generating key information may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 808. In some embodiments, part or all of the computer program may be loaded and / or installed on device 800 via ROM 802 and / or communication unit 809. When the computer program is loaded into RAM 803 and executed by the computing unit 801, one or more steps of the method for generating key information described above may be performed. Alternatively, in other embodiments, the computing unit 801 may be configured to perform the method for generating key information by any other suitable means (e.g., by means of firmware).

[0127] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0128] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0129] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0130] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0131] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0132] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0133] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0134] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for generating key information, comprising: Generate a segment outline based on the document content of the paragraphs in the document to be processed; Based on the preset topic information and the segmentation outline, the paragraphs in the document to be processed are aggregated to obtain multiple aggregated topic contents; The topic content is segmented according to a preset subtopic to obtain multiple segmented subtopic contents, wherein the subtopic is a sub-category topic obtained by dividing the topic information; Based on the information from the paragraphs in the multiple subtopics, generate key information for the document to be processed.

2. The method according to claim 1, wherein, The process of generating a segmented outline based on the document content of paragraphs in the document to be processed includes: The document format of the document to be processed is determined, and the document to be processed is parsed using a parsing tool corresponding to the document format to obtain the document content; The segment outline is generated based on the document content.

3. The method according to claim 2, further comprising: Extract the text content of each page of the document to be processed; The target page document is determined based on the number of words in the text content and a preset word count threshold. The target page document is a document of the chart type. Update the document content of the target page document using a multimodal large model.

4. The method according to claim 3, wherein, The step of updating the document content of the target page document using a multimodal large model includes: Obtain the first document content of the previous page document and the second document content of the next page document in the document to be processed that are adjacent to the target page document; The target page document is converted into an image, and the image, the content of the first document, and the content of the second document are input into a multimodal large model to output the target document content of the target page document; Update the document content of the target page document using the content of the target document.

5. The method according to claim 1, wherein, The step involves aggregating paragraphs in the document to be processed based on preset topic information and the segmentation outline to obtain multiple aggregated topic contents, including: Determine the topic information of the paragraphs in the document to be processed based on the segmentation outline; Clustering algorithms are used to aggregate paragraphs with the same topic information, resulting in multiple aggregated topic contents.

6. The method according to claim 1, further comprising: Each topic content is input into a pre-trained large model, and the relevance of paragraphs in the topic content is verified by the large model. Based on the verification results, problematic paragraphs that do not meet the preset conditions are identified and removed from the topic content.

7. The method according to claim 1, wherein, The step of segmenting the topic content according to preset sub-topics to obtain multiple segmented sub-topic contents includes: Based on the segmentation outline, determine the sub-topics corresponding to the paragraphs in the main content; The paragraphs with the same sub-topic in the main content are segmented to obtain multiple sub-topic contents.

8. The method according to claim 7, wherein, The step of generating key information for the document to be processed based on the information of paragraphs in the multiple sub-topics includes: Based on the document content corresponding to each paragraph in the subtopic content and the position information of each paragraph in the document to be processed, generate key information for the subtopic content; By summarizing the key information of the multiple sub-topics, key information for the document to be processed is obtained.

9. The method according to any one of claims 1-8, further comprising: Generate key notes corresponding to the document to be processed based on the key information, wherein the key notes include: a pilot page, a details page, and a summary page.

10. A device for generating key information, comprising: The outline generation module is configured to generate segmented outlines based on the document content of paragraphs in the document to be processed; The aggregation module is configured to aggregate paragraphs in the document to be processed based on preset topic information and the segmentation outline to obtain multiple aggregated topic contents; The segmentation module is configured to segment the topic content according to preset subtopics to obtain multiple segmented subtopic contents, wherein the subtopics are sub-categories of topics obtained by dividing the topic information; The information generation module is configured to generate key information for the document to be processed based on information from paragraphs in the multiple subtopic contents.

11. An electronic device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-9.

12. A non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-9.

13. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1-9.

Images