A method for generating meeting minutes based on large language models

Abstract

The present invention discloses a method for generating meeting minutes based on a large language model, including: constructing a meeting template and entering basic meeting information according to the meeting template; obtaining a meeting audio file, recording timestamps in the meeting template, and obtaining a meeting record text with timestamps; based on the meeting record text with timestamps, segmenting the meeting record text to obtain multiple text segments; inputting the text segments into a large language model to obtain meeting minutes. The present invention combines the language understanding and generation capabilities of a large language model with prompt engineering technology to establish a set of standardized methods for generating meeting minutes, which can automatically generate structured, accurate, and comprehensive meeting minutes based on meeting records or meeting recordings according to predetermined meeting sessions and types.

Description

Technical Field

[0001] The present invention belongs to the technical field of artificial intelligence, and particularly relates to a method for generating meeting minutes based on a large language model. Background Art

[0002] Large language models are generative artificial intelligence models trained based on massive text data. In recent years, with the significant improvement in computing power and the development of big data technology, large language models have made remarkable leaps in scale and performance. In particular, the proposed Transformer architecture enables the model to better capture long-distance dependencies, greatly enhancing the ability of language understanding and generation. Large language models can not only generate natural language texts but also deeply understand the meaning of texts and complete various natural language tasks such as text summarization and text extraction. Users can interact with application systems based on large language models using natural language, thus realizing various applications from understanding to generation, including question answering, classification, summarization, translation, chatting, etc. Currently, large language models and their applications in various vertical business fields have become research and application hotspots in the global artificial intelligence field, and the scale of model parameters has leaped from billions to hundreds of billions or even trillions. The increase in the number of parameters enables the model to capture detailed information in human language more precisely and understand the complexity of human language more deeply. With the continuous maturity of technology and the continuous accumulation of data, large language models have gradually become widely used artificial intelligence technology tools across industries.

[0003] Meetings, as an important form of communication, decision-making, and collaboration among people, are of great significance to various social and economic activities. Effective meetings can promote information circulation, enhance teamwork, and improve decision-making efficiency. However, traditional meetings often have problems such as low efficiency and inaccurate information recording. Relying on intelligent meeting systems can effectively improve the efficiency and quality of meetings. Existing intelligent meeting systems mainly improve work efficiency for meeting organization and summary by automating the meeting process management, recording meeting content in real time, and organizing and sharing meeting materials. With the development of artificial intelligence technology, intelligent meeting systems based on technologies such as speech recognition and natural language processing can, to a certain extent, achieve semi-automation of meeting records, as well as automatic recording and intelligent summarization of meeting content. However, most of the above technologies or methods are mainly used for recording the meeting process. Limited by algorithm capabilities, manual summarization is still adopted for post-meeting summaries, facing certain limitations.

[0004] Intelligent meeting systems typically employ advanced speech recognition technology for meeting recording, which can convert the recordings in meetings into text. Although these systems can identify different speakers and generate corresponding speech texts, they have obvious deficiencies in the recorded content. The meeting records generated by speech recognition technology are often highly loose, lacking structure, logic, and coherence, making the generated text unable to be directly used as meeting minutes. In addition, due to the complexity and diversity of language, the transcribing model still faces challenges in dealing with professional terms, accents, speech rate changes, and background noise, resulting in the accuracy and integrity of the transcribing results being affected.

[0005] Although intelligent meeting systems have made certain progress in speech transcribing, the compilation of meeting minutes still largely relies on manual work. After the meeting, it is usually necessary for specialized recorders to sort out the meeting minutes based on the recordings and transcribing texts. This process not only has low efficiency but is also affected by the personal capabilities, experience, and subjective judgments of the recorders, resulting in uneven quality of meeting minutes compiled by different recorders. In addition, the process of manually compiling meeting minutes is also time-consuming and cannot meet the requirements for quick response and decision-making in a fast-paced working environment.

[0006] Large language models have demonstrated powerful capabilities in the field of natural language processing and can summarize meeting records and form abstracts of key meeting content. However, there are also certain limitations in practical applications. The themes and agendas of meetings are diverse, and different meetings require different minutes formats and content focuses. Currently, most large language models still rely on a single prompt template or fixed format when generating meeting minutes, which limits the flexibility of the model to adapt to different meeting requirements. In addition, a single template may not accurately capture the key information and decision points in the meeting, resulting in the generated minutes lacking pertinence and practicality. More importantly, due to the limitation of large language models on the input token length, the long text of a long meeting record cannot be input into the large language model at one time to generate a complete meeting minute, which limits the efficiency and effectiveness of existing intelligent meeting systems in practical applications.

[0007] Therefore, there is an urgent need for a method for generating meeting minutes based on large language models to solve the above problems. Summary of the Invention

[0008] To solve the above technical problems, the present invention proposes a method for generating meeting minutes based on large language models, which can save labor costs, improve the efficiency of meeting recording, and can also provide more in-depth insights and suggestions for meeting participants through intelligent analysis, thereby enhancing the overall value of the meeting.

[0009] To achieve the above object, the present invention provides a method for generating meeting minutes based on a large language model, including:

[0010] Construct a meeting template and enter the basic meeting information according to the meeting template;

[0011] Obtain the meeting audio file, record the time stamps in the meeting template, and obtain the meeting record text with time stamps;

[0012] Based on the meeting record text with time stamps, segment the meeting record text to obtain multiple text segments;

[0013] Input the text segments into the large language model to obtain the meeting minutes.

[0014] Optionally, the basic meeting information includes: basic meeting information, personnel information, and meeting agenda information.

[0015] Optionally, the text segments include: first-level text blocks and second-level text fragments.

[0016] Optionally, obtaining the first-level text blocks includes:

[0017] Perform first-level segmentation on the meeting record to obtain first-level text blocks.

[0018] Optionally, obtaining the second-level text fragments includes:

[0019] Judge whether the first-level text block meets the preset length to obtain a first-level judgment result;

[0020] According to the first-level judgment result, judge whether to perform second-level segmentation on the first-level text block to obtain the second-level text fragments.

[0021] Optionally, inputting the text segments into the large language model to obtain the meeting minutes includes:

[0022] Based on the text segments, combine the meeting summary prompt template and the to-do item prompt template to obtain a meeting summary generation prompt and a to-do item generation prompt;

[0023] Based on the meeting summary generation prompt and the to-do item generation prompt, obtain the meeting summary and to-do items;

[0024] Based on the meeting summary and to-do items, combine the meeting minutes template to obtain a meeting summary text object and a to-do item text object;

[0025] Based on the meeting summary text object, the to-do item text object, and the basic meeting information, obtain the meeting minutes.

[0026] Optionally, the meeting summary prompt template is:

[0027] PA = You are a meeting record summary tool. Please summarize the content of +{spoken}+{spken_identity}+ for +{con_agenda}+{con_type}+ into a general statement with a length not exceeding +{length}+ characters. The meeting content is as follows: +{content}

[0028] Among them, PA is the meeting summary generation prompt word template object, {spoken_identity} is the speaker identity in the meeting template, {spoken} is the speaker in the meeting template, {con_agenda} is the meeting agenda in the meeting template, {con_type} is the meeting type in the meeting template, {content} is the meeting content in the meeting template, and {length} is the length limit of the meeting summary or to-do items to be refined;

[0029] The to-do item prompt word template is:

[0030] PT

[0031] PT = You are a to-do item summary tool. Please summarize the potential to-do items in the content of +{spoken}+{spoken_identity}+ for +{con_agenda}+{con_type}+. Each to-do item is in one sentence with a length not exceeding +{length}+ characters. The meeting content is as follows: +{content}

[0032] Among them, PT is the to-do item prompt word template object.

[0033] Optionally, the meeting summary is:

[0034]

[0035] Among them, is the total sum of the generated meeting summary content, is the meeting summary group obtained by processing the first-level text block or the second-level text fragment group by the large language model, PA i is the meeting summary prompt word with different variable values, is the i-th first-level text block, is the second-level text fragment, P is the total number of text fragments in each second-level text fragment group, p is the text fragment sequence in each second-level text fragment group, is the first-level text block, and judge the text type passed into the large language model through if If the current text type does not belong to the first-level text block then select the formula corresponding to else;

[0036] The to-do item is as follows:

[0037]

[0038] Among them, is to define the total of to-do items, and PT i is a to-do item prompt word with different variable values.

[0039] Optionally, the meeting summary text object is:

[0040]

[0041] Among them, H(A abs ) is the meeting summary text object, and SORT(*) is the meeting summary document information that effectively arranges each first-level text summary in the order of the meeting agenda session;

[0042] The to-do item text object is:

[0043]

[0044] Among them, H(A next ) is the to-do item text object, and ORG(*) is the to-do item that effectively segments and integrates and sorts the total of to-do items from the perspectives of language logic and semantic integrity.

[0045] Compared with the prior art, the present invention has the following advantages and technical effects:

[0046] 1. The present invention applies large language model technology to the automated generation process of meeting minutes, greatly improving the efficiency and accuracy of meeting records. By standardizing the meeting process into a meeting template, the systematicness and integrity of the meeting content are ensured. The meeting process is captured in real time through meeting recordings, and combined with the recording of timestamps, the links of the meeting audio data are structurally defined. Through the transcribing function of artificial intelligence algorithms, unstructured audio is converted into structured meeting record text, improving the readability of information. A two-level text segmentation technology is established, which not only distinguishes the content by theme according to the meeting agenda, but also takes into account the token length limit of the large language model for processing text, and ensures the logical coherence and single-theme integrity of the secondary text segments through secondary segmentation.

[0047] 2. The present invention utilizes the deep semantic understanding, reasoning, and generation capabilities of large language models. Through prompt engineering, it deeply analyzes and integrates text information to generate meeting summaries and to-do items with high generality and logical coherence. By means of text splicing technology and formatted output, it generates standardized meeting minutes with complete structures and substantial content, greatly facilitating the collation and subsequent application of meeting results. Generally speaking, while the present invention improves the automation level of meeting minutes, it also provides strong technical support for intelligent office work, with significant social and economic value. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The accompanying drawings, which form a part of this application, are used to provide a further understanding of this application. The schematic embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation to this application. In the drawings:

[0049] Figure 1 is a flowchart of a method for generating meeting minutes based on a large language model according to an embodiment of the present invention;

[0050] Figure 2 is a flowchart of text segmentation of meeting records according to an embodiment of the present invention;

[0051] Figure 3 is a flowchart of generating meeting summaries and to-do items based on a large language model according to an embodiment of the present invention;

[0052] Figure 4 is a flowchart of generating meeting minutes according to a meeting minutes template according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] It should be noted that, without conflict, the embodiments in this application and the features in the embodiments can be combined with each other. The following will refer to the accompanying drawings and combine with the embodiments to detail this application.

[0054] It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. And although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0055] The present invention abstracts the meeting process into a general and customizable template, and based on the language understanding and generation capabilities of large language models and integrating prompt engineering technology, it establishes a set of standardized methods for generating meeting minutes. It can automatically generate structured, accurate, and comprehensive meeting minutes based on meeting records or meeting recordings according to predetermined meeting sessions and types. It can not only save labor costs and improve the efficiency of meeting records, but also provide deeper insights and suggestions for meeting participants through intelligent analysis, thereby enhancing the overall value of the meeting.

[0056] The present invention proposes a method for generating meeting minutes based on a large language model, as Figure 1 shown, which specifically includes the following steps:

[0057] Construct a meeting template and enter the basic meeting information according to the meeting template;

[0058] Obtain the meeting audio file, record the time stamps in the meeting template, and obtain the meeting record text with time stamps;

[0059] Based on the meeting record text with time stamps, segment the meeting record text to obtain multiple text segments;

[0060] Input the text segments into the large language model to obtain the meeting minutes.

[0061] Furthermore, the basic meeting information includes: basic meeting information, personnel information, and meeting agenda information.

[0062] Specifically, as shown in Table 1, the present invention proposes a standard meeting template as the core element for interacting with users. This template abstracts specific basic meeting information, personnel information, and meeting agenda information into a general and universal template. The element content filled in the template will be used as variables in the prompt template to control the input of the large language model. The design of the template fully considers the diversity and complexity of meetings to adapt to the needs of different types and scales of meetings, ensuring the generality and flexibility of the template to guide subsequent text segmentation and content integration work, thereby improving the quality and consistency of the meeting minutes. The meeting template proposed by the present invention is designed as follows in the table:

[0063] Table 1

[0064]

[0065] Before the meeting starts, the user needs to enter information according to the template and can continuously modify and improve it during the meeting. The personnel information and meeting agenda information can be increased or decreased according to the actual meeting needs.

[0066] Specifically, generate meeting record text based on the meeting audio file: mainly convert unstructured information such as the audio data file recorded during the meeting into unified structured text information. The meeting audio file recording needs to adopt an efficient audio acquisition system to ensure that every detail during the meeting can be accurately captured. While recording the meeting process as a complete audio data file, record the start and end timestamps of each link in the meeting agenda in the meeting template. Use artificial intelligence algorithms to transcribe the audio data file into meeting record text with timestamps. This step needs to consider the ability to suppress environmental noise and the ability to identify and separate different sound sources to improve the clarity and accuracy of the voice data. During the transcription process, the artificial intelligence algorithm should be able to combine factors such as speech rate, accent, and professional terms to reduce transcription errors and improve the text quality.

[0067] Furthermore, the text segment includes: a first-level text block and a second-level text fragment.

[0068] Furthermore, obtaining the first-level text block includes:

[0069] Perform a first-level segmentation on the meeting record to obtain the first-level text block.

[0070] Furthermore, obtaining the second-level text fragment includes:

[0071] Judge whether the first-level text block meets the preset length to obtain the first-level judgment result;

[0072] According to the first-level judgment result, judge whether to perform a second-level segmentation on the first-level text block to obtain the second-level text fragment.

[0073] Specifically, text segmentation is the process of decomposing the long meeting transcription content into text blocks or text fragments that are easier to manage and analyze. The text segmentation tool designed in the present invention has powerful text matching and segmentation functions and can reasonably segment the text according to predefined rules.

[0074] As Figure 2 shown, the present invention designs two-level text segmentation rules. The first-level segmentation rule is to perform a first-level segmentation on the meeting record text according to the meeting agenda timestamp to form several first-level text blocks with independent subjects and contents. Each text block includes all the meeting record texts within the timestamp range. Define the meeting record text object as H(A), and the number of timestamp pairs in the meeting template as N t , the preset first-level text block where represents the information of the first-level text block composed of several word elements in the N t th block, and the expression is:

[0075]

[0076] Among them, SPLIT(H(A), N t ) means rationalizing and splitting H(A) according to N t to obtain the first-level text blocks To avoid the length of the input to the large language model exceeding the model's own token length limit, the present invention designs the optimal capacity threshold t of the tokens input to the large language model lan , as the judgment criterion for whether the first-level text blocks need to be secondarily split. Thus, the accuracy of the output is ensured:

[0077] t lan = T lan ×h

[0078] Among them, T lan is the maximum token length that the model can input, and h is the optimal capacity ratio. After a large number of verifications, the present invention designs it to be 80%, and the specific setting of h can be flexibly adjusted according to different large language models.

[0079] According to the context length that the large language model can handle, the token capacity of each first-level text block is determined, and the first-level text blocks exceeding the threshold t lan are secondarily split to generate multiple text fragments with a single theme and content and with a front-back logical order. Define the second-level text fragment group

[0080]

[0081] Among them, means splitting the first-level text block according to the optimal capacity t lan to obtain the corresponding second-level text fragment group. represents the kth text fragment in the Mth second-level text fragment group.

[0082] Combined with to expand the first-level text block expression into:

[0083]

[0084] Among them, first, the token length of the first-level text block is determined. If its length is less than the model optimal capacity t lan , then it is not split and remains as the first-level text block. Conversely, it is split through function to obtain the second-level text group

[0085] Define the total number of text blocks and text fragments that meet the optimal capacity t lan input to the large language model as N, and through the timestamp for Nt , the number M of secondary text segments, and the total number of text segments included in the secondary text segment group are calculated as follows: Perform the calculation:

[0086]

[0087] Furthermore, as Figures 3-4 shown, input the text segment into the large language model, and the meeting minutes obtained include:

[0088] Based on the text segment, combine the meeting summary prompt template and the to-do item prompt template to obtain the meeting summary generation prompt and the to-do item generation prompt;

[0089] Specifically, the present invention comprehensively considers the scale, complexity, and performance requirements of the method. The prompts for generating meeting minutes should have characteristics such as simplicity, generality, and portability. The design of the prompts in the present invention has a high degree of simplicity, integrity, flexibility, and accuracy, and can meet the requirements for summarizing and outputting meeting content while adapting to different large language models.

[0090] The goal of the prompt engineering in the present invention is to combine the first-level text blocks and the secondary text segments to complete the generation of two types of content: meeting summaries and to-do items. By merging the prompt templates with different variable values with the corresponding text blocks or text segments, they are used as the input to the large language model in sequence. Combining the header content in the meeting template, using the standardized syntactic patterns proposed by the present invention, the descriptive words and descriptive phrases in the prompt content are generated. Define PA as the meeting summary generation prompt template object. Then the syntactic description of this text input is:

[0091] PA = You are a meeting record summary tool. Please summarize the content of +{spoken}+{spoken_identity}+ for +{con_agenda}+{con_type}+ into a summarized statement with a length of no more than +{length}+ words. The meeting content is as follows: +{content}

[0092] Similarly, define PT as the to-do item prompt template object. The descriptive syntactic expression is:

[0093] PT = You are a to-do item summary tool. Please summarize the potential to-do items in the content of +{spoken}+{spoken_identity}+ for +{con_agenda}+{con_type}+

[0094] Each to-do item is in one sentence, with the content length not exceeding +{length}+ characters. The meeting content is as follows: +{contrnt}. If there are no to-do items, return "No to-do items"

[0095] In the above two syntactic patterns, PA is the prompt word template object for meeting summary generation, {spoken_identity} is the speaker identity in the meeting template, {spoken} is the speaker in the meeting template, {con_agenda} is the meeting agenda in the meeting template, {con_type} is the meeting type in the meeting template, {content} is the meeting content in the meeting template, {length} is the length limit of the meeting summary or to-do items to be refined, defined by the user, and PT is the to-do item prompt word template object.

[0096] Based on the meeting summary generation prompt word and the to-do item generation prompt word, obtain the meeting summary and to-do items;

[0097] Specifically, the present invention designs a general method and a prompt word template to form the input of a large language model for generating meeting summaries and to-do items. The present invention does not rely on the capabilities of the large language model. During engineering applications, it is necessary to collect meeting sample data, test the generation capabilities of different models, and evaluate the coherence, readability of the text generated by the models, and whether it can meet the standard format of the meeting minutes, so as to select the optimal model. For each segmented first-level text block and second-level text fragment, the present invention generates prompt words for meeting summaries and to-do items, establishes the model input, and sequentially inputs them into the large language model to generate meeting summaries and to-do items, and combines them respectively according to the following methods.

[0098] Define as the total sum of the generated meeting summary content, and the expression is:

[0099]

[0100] where PA i represents the meeting summary prompt words with different variable values. For the first-level text blocks, directly splice each text block with the corresponding PA i and use it as the input of the large language model to generate the corresponding meeting summary; for the second-level text fragment groups, respectively splice each second-level text fragment with the corresponding PA i as the input of the large language model, and sum up all the generated second-level meeting summaries to finally form the second-level text summary group under this single theme.

[0101] Similarly, define the total sum of to-do items as The expression is:

[0102]

[0103] Where PT i is a to - do reminder with different variable values. For the first - level text blocks, directly use Language(*) to splice each text block with the corresponding PT i and use it as the input of the large - language model to generate the corresponding first - level to - do items; for the second - level text fragment groups, splice each second - level text fragment with the corresponding PT i as the input of the large - language model, and sum up all the generated second - level to - do items, which will ultimately be used as the second - level text summary group under this single theme. Judge the generated content of the large - language model. If it returns "no to - do items", set the to - do items corresponding to this text block or text fragment to an empty string.

[0104] Based on the meeting summary and to - do items, combined with the meeting minutes template, obtain the meeting summary text object and to - do item text object;

[0105] Based on the meeting summary text object, to - do item text object and basic meeting information, obtain the meeting minutes;

[0106] Specifically, the present invention designs a standardized meeting minutes template, which combines the basic meeting information, the sum of meeting summaries, and the sum of to - do items into the final meeting minutes. The present invention abstracts the specific and special meeting minutes structure into a general template to adapt to different types of meeting requirements, while ensuring the professionalism, consistency, and readability of the minutes. The meeting minutes template consists of two parts: ① The meeting minutes header, including the meeting name, meeting time, meeting place, participants, meeting record, etc.; ② The meeting minutes content, which is composed of multiple structured text groups through text splicing technology to ensure the logic and coherence of the main content.

[0107] According to the meeting minutes template, using text splicing technology, obtain the final meeting summary text object H(A abs ):

[0108]

[0109] Where C abs is the number of meeting agendas in the meeting template. SORT(*) means arranging each first - level text summary in the order of the meeting agenda session to generate a meeting summary document information with rich content and unified format.

[0110] Similarly, define the to - do item text object as H(A next ):

[0111]

[0112] Indicates the total of to-do items A next From the perspectives of language logic and semantic integrity, perform effective segmentation, integration, and sorting to generate a to-do list with readability and clear structure.

[0113] Finally, combine the meeting minutes header, meeting summary object, and to-do item text object to generate a meeting minutes with complete structure, unified format, and comprehensive content.

[0114] The above is only a preferred specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed in the present application should be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for generating meeting minutes based on a large language model, characterized in that: include: Construct a meeting template and enter basic meeting information according to the meeting template; The basic information of the meeting includes: basic information of the meeting, personnel information and meeting agenda information; Obtain the conference audio file, record the timestamp in the conference template, and obtain the conference record text with the timestamp; Based on the meeting record text with the timestamp, segment the meeting record text to obtain multiple text segments; The text segment includes: a primary text block and a secondary text segment; Acquiring the first-level text block includes: Performing a first-level segmentation on the meeting minutes to obtain first-level text blocks; Determine whether the first-level text block meets a preset length, and obtain a first-level determination result; According to the first-level judgment result, determine whether to perform second-level segmentation on the first-level text block to obtain the second-level text segment; Input the text segment into a large language model to obtain the meeting minutes; The text segment is input into a large language model to obtain the meeting minutes including: Based on the text segment, in combination with the meeting summary prompt word template and the to-do list prompt word template, obtaining the meeting summary generation prompt word and the to-do list generation prompt word; The conference summary prompt word template is: PA = You are a summary tool for meeting minutes. Please sort out the content of +{spoken}+{spoken_identity}+{con_agenda}+{con_type}+ into a summary. The content should not exceed +{length}+ words. The content of the meeting is as follows: +{content} Among them, PA is the template object for generating prompt words for the conference summary, {spoken_identity} is the speaker identity in the conference template, {spoken} is the speaker in the conference template, {con_agenda} is the conference agenda in the conference template, {con_type} is the conference type in the conference template, {content} is the conference content in the conference template, and {length} is the length limit of the conference summary or to-do items that need to be refined; The to-do item prompt word template is: PT =You are a to-do summary tool. Please summarize the potential to-do items in the +{con_agenda}+{con_type}+ content based on +{spoken}+{spoken_identity}+. Each to-do item should be one sentence, and the content length should not exceed +{length}+ words. The meeting content is as follows: +{content} Among them, PT is the to-do item prompt word template object; Generate prompt words based on the meeting summary and to-do items, and obtain the meeting summary and to-do items; The summary of the meeting is: in, is the sum of the generated meeting summary contents, The conference summary group is obtained by processing the first-level text block or the second-level text fragment group by a large language model. i for conference summary prompts with different variable values, is the i-th first-level text block, is the secondary text segment, P is the total number of text segments in each secondary text segment group, p is the sequence of text segments in each secondary text segment group, For the first-level text block, through if To determine the text type passed into the large language model, if the current text type does not belong to the first-level text block Then select the formula corresponding to else, Language(*) is to convert each text block With the corresponding PA i To splice; The to-do items are: in, To define the total number of to-do items, PT i To-do prompts with different variable values, Language(*) is to add each text block With the corresponding PA i To splice; Based on the meeting summary and to-do items, combined with the meeting minutes template, obtain the meeting summary text object and the to-do items text object; The meeting summary text object is: Among them, H(A abs ) is the conference summary text object, SORT(*) is the conference summary document information that effectively arranges each first-level text summary in the order of the conference agenda, C abs The number of meeting agendas in the meeting template; The to-do text object is: Among them, H(A next ) is a to-do list text object, and ORG(*) is a to-do list that is effectively segmented, integrated and sorted according to the language logic and semantic integrity. Based on the meeting summary text object, the to-do item text object and the basic information of the meeting, the meeting minutes are obtained.

Images