Conference process resource processing result acquisition method, electronic equipment and program product

Abstract

The invention provides a conference process resource processing result acquisition method, electronic equipment, a readable storage medium and a computer program product. The conference process resource processing result acquisition method disclosed by the invention comprises the following steps: in response to a target conference triggering cache instruction, downloading conference process resources of a target conference to the local; analyzing based on the conference process resource downloaded to the local to obtain a conference process document; performing template filling on a first prompt information generation template corresponding to the target conference based on the conference process document to obtain first prompt information; inputting the first prompt information into the large language model to obtain a conference process resource processing result; storing the conference process resource processing result to a first cache; and in response to the received result query instruction, obtaining a conference process resource processing result corresponding to the result query instruction from the first cache.

Description

Technical Field

[0001] This disclosure relates to a method for obtaining the results of conference process resource processing, an electronic device, a storage medium, and a program product. Background Technology

[0002] With the development of smart office solutions and enterprise digital transformation, meeting management systems are increasingly being used in business operations. By collecting and analyzing meeting data, companies can improve operational efficiency and conduct business reviews. In relevant application scenarios, artificial intelligence (AI) technology is widely used to automate meeting content processing, such as automatically generating meeting minutes, summarizing document content, and generating reference questions for specific roles (e.g., review experts, judges). These AI-based content summarization functions aim to help users quickly extract core meeting information, simplify document processing, and provide decision support, thereby improving overall business efficiency.

[0003] In existing technologies, the processing chain for automating meeting content based on artificial intelligence technology generally includes: first, downloading meeting-related resources (such as meeting screen recordings, transcribed documents, meeting PPTs, etc.); then parsing the downloaded resources and converting them into a text format that can be processed by the model; subsequently, inputting the converted text as a Prompt into a large language model, which then performs reasoning and summarization; and finally, presenting the results output by the large language model in a streaming format, such as outputting a generated document word by word according to a preset format.

[0004] However, during the resource download phase, the large size of the resources or network transmission delays can lead to lengthy waiting times. In the resource parsing phase, converting resources into text involves complex extraction and conversion algorithms, further increasing processing time. During the model inference phase, large language models need to process vast amounts of text information for deep inference and generation, often consuming several seconds to minutes. In the streaming output phase, word-by-word output slows down the result presentation process. These issues persist throughout the entire conference content processing—before, during, and after—resulting in significant overall processing latency. Summary of the Invention

[0005] This disclosure provides a method for obtaining the results of resource processing during a meeting, an electronic device, a storage medium, and a program product.

[0006] According to one aspect of this disclosure, a method for obtaining meeting process resource processing results is provided, comprising: in response to a target meeting triggering a cache instruction, downloading the meeting process resources of the target meeting to a local machine; parsing the downloaded meeting process resources to obtain a meeting process document; filling a template for a first prompt information generation template corresponding to the target meeting based on the meeting process document to obtain a first prompt information; the first prompt information is used to guide a large language model to generate meeting process resource processing results; inputting the first prompt information into the large language model to obtain meeting process resource processing results; the meeting process resource processing results include meeting minutes, meeting document summaries, or meeting candidate questions; storing the meeting process resource processing results in a first cache; and in response to receiving a result query instruction, retrieving the meeting process resource processing results corresponding to the result query instruction from the first cache.

[0007] According to one technical solution, a templated prompt construction and result caching mechanism are introduced into the conference resource processing, so that the results of the conference process resource processing can be directly obtained from the cache. This avoids the overhead of real-time download and parsing during resource processing and reduces the computation in the model inference stage. This effectively alleviates the overall latency problem caused by large resource size, complex parsing algorithm, time-consuming model inference and streaming output, and improves the efficiency of conference resource processing.

[0008] According to at least one embodiment of the meeting process resource processing result acquisition method of the present disclosure, after obtaining the meeting process document, the method further includes: storing the meeting process document in a second cache; in response to receiving a regeneration instruction for the meeting process resource processing result, reading the meeting process document corresponding to the regeneration instruction from the second cache; and processing the meeting process document corresponding to the regeneration instruction based on the large language model to obtain the meeting process resource processing result corresponding to the regeneration instruction.

[0009] According to the technical solution of this embodiment, efficient reuse of meeting process documents can be achieved, which reduces computational latency and improves the diversity and real-time nature of result generation.

[0010] According to at least one embodiment of the meeting process resource processing result acquisition method of this disclosure, the step of processing the meeting process document corresponding to the regenerated instruction based on the large language model to obtain the meeting process resource processing result corresponding to the regenerated instruction includes: filling a template of a second prompt information generation template corresponding to the regenerated instruction based on the meeting process document corresponding to the regenerated instruction to obtain a second prompt information; the second prompt information is used to guide the large language model to regenerate the meeting process resource processing result; the second prompt information is input into the large language model to obtain the meeting process resource processing result corresponding to the regenerated instruction; the meeting process resource processing result corresponding to the regenerated instruction includes meeting minutes, meeting document summary, or meeting candidate questions.

[0011] According to the technical solution of this embodiment, large language models can more accurately understand task intent. Furthermore, different types of generation instructions can correspond to different templates, thus flexibly supporting various meeting resource processing scenarios.

[0012] According to at least one embodiment of the meeting process resource processing result acquisition method of the present disclosure, after storing the meeting process document in the second cache, the method further includes: in response to the deletion of the first meeting process resource, deleting the first meeting process document corresponding to the first meeting process resource from the second cache.

[0013] According to the technical solution of this embodiment, the misuse of invalid data left in the cache can be avoided, and the content in the cache can be kept consistent with the original resource state, thereby improving the data consistency and security of the system and reducing storage usage.

[0014] According to at least one embodiment of the meeting process resource processing result acquisition method of this disclosure, the step of filling the second prompt information generation template corresponding to the regeneration instruction with template based on the meeting process document corresponding to the regeneration instruction to obtain the second prompt information includes: obtaining the second prompt information generation template corresponding to the regeneration instruction from a preset template set; and filling the second prompt information generation template with template based on the meeting process document corresponding to the regeneration instruction to obtain the second prompt information.

[0015] According to the technical solution of this embodiment, the content of the input large language model can be guaranteed to be accurate and the format is consistent, and the accuracy and consistency of the model generation results are also improved.

[0016] According to at least one embodiment of the present disclosure, the method for obtaining meeting process resource processing results, after storing the meeting process resource processing results in a first cache, further includes: in response to the deletion of a second meeting process resource, deleting the meeting process resource processing result corresponding to the second meeting process resource from the first cache.

[0017] According to the technical solution of this embodiment, the misuse of invalid data left in the cache can be avoided, and the content in the cache can be kept consistent with the original resource state, thereby improving the data consistency and security of the system and reducing storage usage.

[0018] According to at least one embodiment of the meeting process resource processing result acquisition method of the present disclosure, before responding to the target meeting triggering cache indication, the method further includes: acquiring the meeting attributes of the target meeting to be cached; matching the meeting attributes with a preset cache strategy to obtain a matching result; the matching result includes a cache timing; and triggering the cache indication of the target meeting based on the cache timing.

[0019] According to the technical solution of this embodiment, the caching action can be automated, differentiated and precise, which can not only avoid the waste of resources for irrelevant meetings, but also ensure that key meetings are cached in a timely manner at the appropriate time, thereby improving the overall resource utilization, processing efficiency and user response experience of the system.

[0020] According to at least one embodiment of the meeting process resource processing result acquisition method of this disclosure, the step of matching the meeting attributes with a preset caching strategy to obtain a matching result includes: acquiring user behavior data and user attributes of the creator of the target meeting; and matching the meeting attributes, the user behavior data, and the user attributes with the preset caching strategy to obtain a matching result.

[0021] According to the technical solution of this embodiment, caching actions can be determined more precisely, thereby ensuring that the allocation of cache resources is more in line with actual usage needs.

[0022] According to at least one embodiment of the meeting process resource processing result acquisition method of the present disclosure, the caching strategy is based on at least one of the following: meeting type, meeting lifecycle, meeting processing purpose, document type, meeting importance, user behavior data of the meeting creator, and user attributes.

[0023] According to the technical solution of this embodiment, cache triggering can be made more refined and dynamic.

[0024] According to at least one embodiment of the meeting process resource processing result acquisition method of the present disclosure, the step of filling the template of the first prompt information generation template corresponding to the target meeting based on the meeting process document to obtain the first prompt information includes: obtaining the resource processing purpose of the target meeting; obtaining the first prompt information generation template corresponding to the resource processing purpose from a preset template set; and filling the template of the first prompt information generation template based on the meeting process document to obtain the first prompt information.

[0025] According to the technical solution of this embodiment, it can be ensured that the first prompt information not only meets the specific business needs, but also accurately reflects the meeting content, thereby improving the relevance and effectiveness of the large model input, and thus improving the quality and reliability of the model output results.

[0026] A method for obtaining meeting process resource processing results according to at least one embodiment of the present disclosure, wherein the purpose of the resource processing includes generating meeting minutes, generating document summaries, or generating reference questions.

[0027] According to the technical solution of this embodiment, it can be ensured that the generated results are highly matched with business requirements.

[0028] According to at least one embodiment of the meeting process resource processing result acquisition method of the present disclosure, the step of inputting the first prompt information into a large language model to obtain the meeting process resource processing result includes: inputting the first prompt information into the large language model; and in response to receiving the streaming output of the large language model, receiving and concatenating the characters of the streaming output to obtain the meeting process resource processing result.

[0029] According to the technical solution of this embodiment, user waiting time can be reduced, interactive experience can be improved, and the integrity and continuity of the final meeting process resource processing results can be guaranteed.

[0030] According to another aspect of this disclosure, an electronic device is provided, comprising: a memory storing execution instructions; and a processor executing the execution instructions stored in the memory, causing the processor to perform a conference process resource processing result acquisition method according to any embodiment of this disclosure.

[0031] According to another aspect of this disclosure, a readable storage medium is provided, wherein executable instructions are stored therein, which, when executed by a processor, are used to implement the meeting process resource processing result acquisition method of any embodiment of this disclosure.

[0032] According to another aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements a method for obtaining meeting process resource processing results according to any embodiment of this disclosure. Attached Figure Description

[0033] The accompanying drawings illustrate exemplary embodiments of the present disclosure and, together with the description thereof, serve to explain the principles of the present disclosure. These drawings are included to provide a further understanding of the present disclosure and are incorporated in and constitute a part of this specification.

[0034] Figure 1 This is an illustrative interactive flow of a method for obtaining meeting process resource processing results according to one embodiment of the present disclosure. Figure 1 .

[0035] Figure 2 This is an illustrative interactive flow of a method for obtaining meeting process resource processing results according to one embodiment of the present disclosure. Figure 2 .

[0036] Figure 3 yes Figure 2 The diagram shows an illustrative interactive flowchart of the model processing method in the method for obtaining resource processing results during the meeting process.

[0037] Figure 4 This is an illustrative interactive flow of a method for obtaining meeting process resource processing results according to one embodiment of the present disclosure. Figure 3 .

[0038] Figure 5 yes Figure 3 The diagram shows an illustrative interactive flowchart of the template filling method in the model processing method shown.

[0039] Figure 6 This is an illustrative interactive flow of a method for obtaining meeting process resource processing results according to one embodiment of the present disclosure. Figure 4 .

[0040] Figure 7 This is an illustrative interactive flow of a method for obtaining meeting process resource processing results according to one embodiment of the present disclosure. Figure 5 .

[0041] Figure 8 yes Figure 7 The diagram shows an illustrative interactive flowchart of the cache matching method in the method for obtaining the resource processing results of the meeting process.

[0042] Figure 9 yes Figure 1 The diagram shows an illustrative interactive flowchart of the template filling method in the meeting process resource processing result acquisition method.

[0043] Figure 10 yes Figure 1 The diagram shows an illustrative interactive flowchart of the model processing method in the method for obtaining resource processing results during the meeting process.

[0044] Figure 11 This is an illustrative interactive flowchart of a method for obtaining meeting process resource processing results according to one embodiment of the present disclosure.

[0045] Figure 12 This is a schematic structural block diagram of a meeting process resource processing result acquisition device according to one embodiment of the present disclosure.

[0046] Figure 13 This is a schematic structural block diagram of an electronic device according to one embodiment of the present disclosure. Detailed Implementation

[0047] The present disclosure will now be described in further detail with reference to the accompanying drawings and examples. It should be understood that the specific examples described herein are for illustrative purposes only and are not intended to limit the scope of the disclosure. Furthermore, it should be noted that, for ease of description, only the parts relevant to the present disclosure are shown in the accompanying drawings.

[0048] It should be noted that, where there is no conflict, the embodiments and features described in this disclosure can be combined with each other. The technical solutions of this disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0049] Taking a company's business management debriefing meeting as an example, this meeting lasted for more than two hours and involved resources such as screen recording files of hundreds of megabytes, transcribed documents of tens of thousands of words, and meeting PPTs containing dozens of pages of content. Automating the processing of this meeting content first requires a considerable amount of time to download and parse the screen recording files, transcribed documents, and meeting PPTs. Then, the parsed text is used to generate a Prompt, which is then input into a large language model for inference and summarization. Due to the large size of the text, the model inference process often takes several minutes; and in the streaming output stage, the word-by-word generation of results further lengthens the waiting time. The entire processing process can take more than ten minutes in total, making it impossible to generate output quickly enough and failing to meet the requirements of real-time performance and efficiency.

[0050] To this end, the present disclosure proposes the following technical solution, which can reduce the latency of the overall conference resource processing process.

[0051] Figure 1 A schematic diagram of the overall flow of a method for obtaining meeting process resource processing results according to one embodiment of this disclosure is shown. Figure 1 The method M100 for obtaining the results of meeting process resource processing includes steps S110 to S160. This method can be executed by an electronic device such as a server.

[0052] In step S110, in response to the target meeting triggering a cache instruction, the meeting process resources of the target meeting are downloaded to the local machine.

[0053] In some embodiments of this disclosure, the cache indication in step S110 can be generated by the meeting management system or by monitoring the meeting management system locally. Step S110 can download meeting process resources in the following ways: by pulling from the meeting management system, obtaining through the meeting management system interface, or by using a resource list (such as a Uniform Resource Locator (URL), file ID, etc.). The meeting process resources downloaded in step S110 can include meeting recordings, transcripts, meeting PPTs, and other meeting-related content.

[0054] In step S120, the meeting process resources downloaded to the local machine are parsed to obtain the meeting process document.

[0055] In some embodiments of this disclosure, the meeting process document obtained through step S120 can be a unified format document, such as Markdown, which is easy for large language models to process. Step S120 can employ different parsing processes for different types of resources. For example, for meeting screen recordings, step S120 can use audio-video demultiplexing technology to separate the audio track from the video track. On the one hand, the audio stream is extracted for speech recognition and analysis; on the other hand, video frames are extracted for image recognition and visual analysis. Specifically, the extracted audio stream can be subjected to speech recognition to obtain a transcribed document. In particular, during speech recognition, speaker techniques (such as feature extraction + clustering methods, deep neural network-based methods, etc.) can be combined to distinguish the speeches of different participants. During speech recognition, classification can also be performed based on acoustic features and deep speech representation, combined with convolutional neural networks, recurrent neural networks, or Transformer models, to identify the speaker's emotional state, enabling the transcribed text to carry emotional tags and improve the comprehensibility of the meeting content. For the extracted video frames, image analysis can be performed on the video frames to identify presentation materials, whiteboard writing, screen-shared content, etc., to obtain a structured video conference process document. Specifically, to reduce storage and computational overhead, image parsing can be performed only on keyframes within the video frame. The image parsing process can be based on Optical Character Recognition (OCR) technology to recognize text. During image parsing, face detection and identity matching can also be performed on participant images to verify participant identity and attendance statistics, thereby enhancing the completeness and reliability of the meeting analysis.

[0056] For transcribed documents, text extraction and structuring can be performed directly to obtain structured text. During text extraction and structuring, for charts, chart description text extraction and numerical analysis can be performed; for images, text extraction and structuring can be achieved through artificial intelligence models.

[0057] For conference PPTs, since PPTs generally contain a lot of images, you can directly extract the content of each page of the conference PPT as an image and then input the image into an artificial intelligence model to achieve text extraction and structured processing.

[0058] In step S130, the template for the first prompt information corresponding to the target meeting is generated based on the meeting process document and then filled to obtain the first prompt information.

[0059] In some embodiments of this disclosure, the first prompt information generation template in step S130 can be related to the processing purpose of the meeting. Step S130 allows the meeting process document to be filled into the corresponding position of the first prompt information generation template to obtain the first prompt information. This first prompt information is used to guide the large language model in generating the meeting process resource processing results; that is, the first prompt information is the large language model's prompt, used for subsequent large language model processing.

[0060] The first prompt information obtained through step S130 may include meeting process documents and specific processing instructions, such as "summarize meeting minutes based on the uploaded file content, summarize document summaries, or generate reviewer reference questions, etc." In particular, to improve the processing capabilities of the large language model, the first prompt information may also include key information contained in the document, such as document title and nature, presenter information, presentation schedule, and presentation structure.

[0061] In step S140, the first prompt information is input into the large language model to obtain the meeting process resource processing results.

[0062] In some embodiments of this disclosure, after the first prompt information is input into the large language model in step S140, the large language model performs semantic parsing and content integration on the first prompt information, identifying key participants, discussion focus, and analysis objectives of the meeting, and obtaining parsing results; subsequently, logical reasoning and information summarization are performed based on the parsing results; finally, natural language generation is performed and output is performed. The large language model can generate text step by step in a streaming manner, or it can generate structured text according to a preset template or format. The meeting process resource processing results obtained through step S140 include meeting minutes, meeting document summaries, or meeting candidate questions.

[0063] In step S150, the results of the meeting process resource processing are stored in the first cache.

[0064] In some embodiments of this disclosure, the first cache in step S150 can be a memory cache, a distributed cache, or a local process cache, etc. This first cache can generate a cache key using a unique meeting identifier and set a cache expiration policy to support fast access and high-concurrency reading of meeting process resource processing results.

[0065] In step S160, in response to receiving a result query instruction, the meeting process resource processing result corresponding to the result query instruction is retrieved from the first cache.

[0066] In some embodiments of this disclosure, the result query instruction in step S160 can be triggered by user requests, system background tasks, API calls, etc. This result query instruction may include details such as the meeting's unique identifier, query content, and result format. A cache key can be generated based on the meeting's unique identifier to retrieve the corresponding meeting process resource processing result from the first cache. When a cache hit occurs, the meeting process resource processing result is returned directly; when the cache miss occurs, it can trigger a regeneration from the original resource or meeting process document and write it to the first cache, thereby supporting fast access and high-concurrency queries. The meeting process resource processing result obtained through step S160 can be in a Markdown or other format for subsequent user retrieval.

[0067] The method for obtaining meeting process resource processing results disclosed herein introduces a templated prompt construction and a result caching mechanism in meeting resource processing, enabling the meeting process resource processing results to be obtained directly from the cache. This avoids the overhead of real-time downloading and parsing during resource processing and reduces the computation in the model inference stage, thereby effectively alleviating the overall latency problem caused by large resource size, complex parsing algorithms, time-consuming model inference, and streaming output, and improving the efficiency of meeting resource processing.

[0068] Furthermore, the method for obtaining the processing results of meeting process resources provided in this disclosure may further include, after step S120, the following steps: Figure 2 Steps S170 to S190 are shown.

[0069] In step S170, the meeting process document is stored in the second cache.

[0070] In some embodiments of this disclosure, the second cache in step S170 can be a memory cache, a distributed cache, or a local process cache, etc. This second cache can generate a cache key using a unique meeting identifier and set a cache expiration policy to support fast access and high-concurrency reading of meeting process documents. The second cache in step S170 can be the same type of cache as the first cache in step S160, or it can be a different type.

[0071] In step S180, in response to receiving a regeneration instruction for the processing results of the meeting process resources, the meeting process document corresponding to the regeneration instruction is read from the second cache.

[0072] In some embodiments of this disclosure, the regeneration instruction in step S180 may originate from user operations, system scheduling, business linkage, or cache replenishment. This regeneration instruction may include a unique meeting identifier, the type of document to be generated, etc.

[0073] Step S180 can generate a cache key based on the unique meeting identifier and then retrieve the corresponding meeting process document from the second cache. When a cache hit occurs, the meeting process document is returned directly; when a cache miss occurs, the original resource can be re-parsed and written to the second cache, thereby supporting fast access and high-concurrency queries.

[0074] In step S190, the meeting process document corresponding to the regenerated instruction is processed based on the large language model to obtain the meeting process resource processing result corresponding to the regenerated instruction.

[0075] In some embodiments of this disclosure, step S190 can use the regenerated meeting process document corresponding to the instruction as model input, and generate meeting process resource processing results through processes such as prompt construction, retrieval enhancement, model reasoning, post-processing and verification.

[0076] Steps S160 to S190 achieve efficient reuse of meeting process documents by caching meeting process documents and calling large language models on demand, which reduces computational latency and improves the diversity and real-time performance of result generation.

[0077] Regarding step S190, in some embodiments of this disclosure, it may include, for example... Figure 3 Steps S191 and S192 are shown.

[0078] In step S191, the template for generating the second prompt information corresponding to the regenerated instruction is filled based on the meeting process document corresponding to the regenerated instruction, and the second prompt information is obtained.

[0079] In some embodiments of this disclosure, the second prompt information generation template in step S191 can be related to the purpose of the generation instruction. The purpose of this generation instruction can be meeting minutes generation, document summary generation, or reference question generation, etc. Through step S191, documents can be filled into the corresponding positions of the second prompt information generation template to obtain the second prompt information. This second prompt information is used to guide the large language model to regenerate the meeting process resource processing results; that is, the second prompt information is the large language model's prompt, used for subsequent large language model processing.

[0080] In step S192, the second prompt information is input into the large language model to obtain the meeting process resource processing result corresponding to the regenerated instruction.

[0081] In some embodiments of this disclosure, after the second prompt information is input into the large language model in step S192, the large language model performs semantic parsing and content integration on the second prompt information, identifying key participants, discussion focus, and analysis objectives of the meeting, and obtaining parsing results; subsequently, logical reasoning and information summarization are performed based on the parsing results; finally, natural language generation is performed and output. The large language model can generate text step by step in a streaming manner, or it can generate structured text according to a preset template or format. The regenerated meeting process resource processing results corresponding to the instructions include meeting minutes, meeting document summaries, or meeting candidate questions.

[0082] Steps S191 to S192 generate templates using the second prompt information to standardize the input, enabling the large language model to more accurately understand the task intent. Furthermore, different types of generation instructions can correspond to different templates, thus flexibly supporting various meeting resource processing scenarios.

[0083] Furthermore, in some embodiments of this disclosure, after step S170, the following steps may also be included: Figure 4 The step S200 is shown.

[0084] In step S200, in response to the deletion of the first meeting process resource, the first meeting process document corresponding to the first meeting process resource is deleted from the second cache.

[0085] In some embodiments of this disclosure, the instruction to delete the first meeting process resource can be triggered by user operation, system preset policies, administrator intervention, or exception handling logic. When the first meeting process resource is deleted, the corresponding meeting process document can be directly deleted from the second cache; alternatively, the validity of the corresponding event can be verified, and the corresponding meeting process document can be deleted from the second cache if it is valid.

[0086] Step S200 synchronously deletes the meeting process documents in the second cache when the first meeting process resource is deleted. This can prevent invalid data left in the cache from being misused, ensure that the content in the cache is consistent with the original resource state, thereby improving the data consistency and security of the system and reducing storage usage.

[0087] Regarding step S191, in some embodiments of this disclosure, it may include, for example... Figure 5 Steps S1911 to S1912 are shown.

[0088] In step S1911, the second prompt information generation template corresponding to the regeneration instruction is obtained from the preset template set.

[0089] In some embodiments of this disclosure, the preset template set can be configured with different types of templates for different types of processing tasks. When a regeneration instruction is received, a template can be generated by selecting appropriate second prompt information from the preset template set according to the regeneration instruction in step S1911.

[0090] In step S1912, the second prompt information generation template is filled in based on the meeting process document corresponding to the regenerated instruction to obtain the second prompt information.

[0091] Steps S1911 to S1912 ensure that the content of the input large language model is accurate and the format is consistent, and also improve the accuracy and consistency of the model's generated results.

[0092] Furthermore, the method for obtaining the results of meeting process resource processing provided in this disclosure may further include, after step S160, the following steps: Figure 6 Step S210 is shown.

[0093] In step S210, in response to the deletion of the second conference process resource, the conference process resource processing result corresponding to the second conference process resource is deleted from the first cache.

[0094] In some embodiments of this disclosure, the instruction to delete the second meeting process resource can be triggered by user operation, system preset policy, administrator intervention, or exception handling logic. When the second meeting process resource is deleted, the corresponding meeting process resource processing result can be directly deleted from the first cache; alternatively, the validity of the corresponding event can be verified, and the corresponding meeting process resource processing result can be deleted from the first cache if it is valid.

[0095] Step S210 can prevent invalid data left in the cache from being misused, ensure that the content in the cache is consistent with the original resource state, thereby improving the data consistency and security of the system and reducing storage usage.

[0096] Furthermore, the method for obtaining meeting process resource processing results provided in this disclosure may further include, before step S110, the following steps: Figure 7 Steps S220 to S240 are shown.

[0097] In step S220, the meeting attributes of the target meeting to be triggered for caching are obtained.

[0098] In some embodiments of this disclosure, the meeting attributes obtained through step S220 may include basic attributes (such as meeting type, meeting lifecycle (such as before, during, and after the meeting)), document attributes (such as document type), importance, and other content.

[0099] In step S230, the meeting attributes are matched with the preset caching strategy to obtain the matching result; the matching result includes the caching timing.

[0100] In some embodiments of this disclosure, the preset caching strategy consists of triggering conditions and caching timing; different triggering conditions can be set with different caching conditions. For example: triggering conditions A, B, and C, with immediate caching timing; triggering conditions D, B, and C, with caching timing after a preset X hours; triggering conditions D, E, and C, with caching timing within a preset uniform time period, etc. Step S230 can match the conditions in the preset caching strategy item by item based on meeting attributes. The caching strategy can be set based on at least one of the following: meeting type, meeting lifecycle, meeting processing purpose, document type, meeting importance, user behavior data of the meeting creator, and user attributes.

[0101] In step S240, a cache indication for the target meeting is triggered based on the cache timing.

[0102] In some embodiments of this disclosure, the local system can detect in real time whether the caching time has arrived and trigger a caching indication when it does.

[0103] Steps S220 to S240 enable automated, differentiated, and precise control of caching actions, which avoids wasting resources on irrelevant meetings and ensures that key meetings are cached in a timely manner at the appropriate time, thereby improving the overall resource utilization, processing efficiency, and user response experience of the system.

[0104] Regarding step S230, in some embodiments of this disclosure, it may include, for example... Figure 8 Steps S231 to S232 are shown.

[0105] In step S231, user behavior data and user attributes of the creator of the target meeting are obtained.

[0106] In some embodiments of this disclosure, the user behavior data obtained through step S231 may include the frequency with which a user initiates meetings, whether they frequently call functions such as meeting summaries, whether they repeatedly view meeting content, and typical usage time periods. User attributes obtained through step S231 may include the user's department, job level, permission scope, and preferred usage scenarios.

[0107] It is understood that before using the technical solutions disclosed in the embodiments of this disclosure, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this disclosure in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained. For example, in response to receiving a user's active request, a prompt message can be sent to the user to clearly inform the user that the requested operation will require the acquisition and use of the user's personal information. This allows the user to choose whether to provide personal information to the software or hardware such as electronic devices, applications, servers, or storage media that perform the operations of the technical solutions of this disclosure, based on the prompt message. As an optional but non-limiting implementation, the way to send a prompt message to the user in response to receiving a user's active request can be, for example, a pop-up window, in which the prompt message can be presented in text form. Furthermore, the pop-up window can also include a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0108] It is understood that the above notification and user authorization acquisition process are merely illustrative and do not constitute a limitation on the implementation of this disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this disclosure. The data involved in the technical solution of this disclosure (including but not limited to the data itself, the acquisition or use of data) shall comply with the requirements of relevant laws, regulations and related provisions.

[0109] In step S232, the meeting attributes, user behavior data, and user attributes are matched with the preset caching strategy to obtain the matching results.

[0110] Steps S231 to S232 can make more precise decisions on caching actions, thereby ensuring that the allocation of cache resources is more in line with actual usage needs.

[0111] Regarding step S130, in some embodiments of this disclosure, it may include, for example... Figure 9 Steps S131 to S133 are shown.

[0112] In step S131, the resource processing objective of the target meeting is obtained.

[0113] In some embodiments of this disclosure, the resource processing objective in step S131 can be obtained through meeting metadata, documents, instructions, or inference from an artificial intelligence model. The resource processing objective obtained through step S131 may include meeting minutes generation, document summary generation, or reference question generation.

[0114] In step S132, a template is generated by obtaining the first prompt information corresponding to the resource processing purpose from the preset template set.

[0115] In step S133, the template for generating the first prompt information is filled based on the meeting process document to obtain the first prompt information.

[0116] Steps S131 to S133 ensure that the first prompt information not only meets specific business needs but also accurately reflects the meeting content, thereby improving the relevance and effectiveness of the large model input and thus improving the quality and reliability of the model output results.

[0117] Regarding step S140, in some embodiments of this disclosure, it may include, for example... Figure 10 Steps S141 to S142 are shown.

[0118] In step S141, the first prompt information is input into the large language model.

[0119] In step S142, in response to receiving the streaming output of the large language model, the characters in the streaming output are received and concatenated to obtain the resource processing result of the conference process.

[0120] In some embodiments of this disclosure, the large language model does not return the complete meeting process resource processing result all at once, but rather outputs it in a streaming manner, i.e., it is generated step by step by character or small fragments. During the receiving process, the system continuously concatenates the content of the streaming output through step S142 until the large language model output ends, thus obtaining the complete meeting process resource processing result.

[0121] Steps S141 to S142 can reduce user waiting time, improve the interactive experience, and at the same time ensure the integrity and continuity of the final meeting process resource processing results.

[0122] The method for obtaining meeting process resource processing results disclosed herein allows for caching during document parsing. This enables meeting resource processing to skip the resource download and document parsing processes, directly retrieving the meeting process document from the cache for processing using a large language model. By caching the meeting process resource processing results, results can be retrieved almost instantly when the user queries the results. This method reduces the time spent on pre-processing and post-processing in traditional meeting process resource processing result acquisition methods, improving overall efficiency by 80% and significantly enhancing the user experience.

[0123] Figure 11 An exemplary flowchart of the meeting process resource processing result acquisition method based on this disclosure is shown.

[0124] Figure 11 The flowchart shown includes steps S310 to S420 for obtaining the results of meeting process resource processing.

[0125] Let’s say there’s a quarterly project review meeting for the R&D department this afternoon. This review meeting is also known as the target meeting. The target meeting is initiated by the R&D director and involves multiple project reports and a recording of the meeting.

[0126] In step S310, the meeting attributes of the target meeting to be triggered for caching are obtained.

[0127] In some embodiments of this disclosure, the target meeting to be triggered for caching in step S310 can be a meeting in the meeting management system that did not generate meeting process resource processing results after the meeting ended. The meeting attributes obtained through step S310 may include: meeting type: quarterly project review meeting, importance level: high, meeting resources: meeting screen recording, meeting PPT.

[0128] In step S320, the meeting attributes are matched with the preset caching strategy to obtain the matching result.

[0129] In some embodiments of this disclosure, the caching timing of the matching result obtained through step S320 can be immediate caching.

[0130] In step S330, a cache indication for the target meeting is triggered.

[0131] In step S340, in response to the cache instruction triggered by the target meeting, the meeting process resources of the target meeting are downloaded to the local machine.

[0132] In step S350, the meeting process resources downloaded to the local machine are parsed to obtain the meeting process document.

[0133] In step S360, the meeting process document is stored in the second cache.

[0134] In step S370, the template for the first prompt information corresponding to the target meeting is generated based on the meeting process document and then filled to obtain the first prompt information.

[0135] In step S380, the first prompt information is input into the large language model to obtain the meeting process resource processing results.

[0136] In step S390, the results of the meeting process resource processing are stored in the first cache.

[0137] In step S400, in response to receiving a result query instruction, the meeting process resource processing result corresponding to the result query instruction is retrieved from the first cache.

[0138] In step S410, in response to receiving a regeneration instruction for the meeting process resource processing result, the meeting process document corresponding to the regeneration instruction is read from the second cache.

[0139] In step S420, the document corresponding to the regenerated instruction is processed based on the large language model to obtain the meeting process resource processing result corresponding to the generated instruction.

[0140] In some embodiments of this disclosure, when the user inputs a regenerated instruction, the meeting process document in the second cache is read and processed by the large model in step S410, eliminating the time-consuming document parsing operation. When the user inputs a result to obtain an instruction, the result is directly read from the first cache in step S400, and the result is available instantly.

[0141] This disclosure also provides a device for obtaining the results of meeting process resource processing (corresponding to a method for obtaining the results of meeting process resource processing). Figure 12 A schematic diagram of the hardware implementation using the processing system is shown.

[0142] like Figure 13 As shown, the hardware architecture of electronic devices / devices can be implemented using a bus architecture. The bus architecture can include any number of interconnect buses and bridges, depending on the specific application and overall design constraints of the hardware. Bus 1100 connects various circuits including one or more processors 1200, memory 1300, and / or hardware modules. Bus 1100 can also connect various other circuits 1400 such as peripherals, voltage regulators, power management circuits, external antennas, etc. Bus 1100 can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Component (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, only one connection line is used in this figure, but this does not indicate that there is only one bus or one type of bus.

[0143] For ease of explanation, certain steps of the above method are described in relation to modules. It should be understood that the corresponding module for performing one or more steps of the above method may be one or more hardware modules specifically configured to perform the corresponding step, or implemented by a processor configured to perform the corresponding step, or stored in a computer-readable medium for implementation by a processor, or implemented by some combination thereof.

[0144] like Figure 12 As shown, the meeting process resource processing result acquisition device includes a resource download module 1010, a resource parsing module 1020, a template filling module 1030, a model processing module 1040, a result caching module 1050, and a result acquisition module 1060.

[0145] The resource download module 1010 is used to download the meeting process resources of the target meeting to the local machine in response to the target meeting triggering a cache instruction.

[0146] The resource parsing module 1020 is used to parse the meeting process resources downloaded to the local machine to obtain the meeting process document.

[0147] The template filling module 1030 is used to fill the template generated by the first prompt information corresponding to the target meeting based on the meeting process document, and obtain the first prompt information; the first prompt information is used to guide the large language model to generate the meeting process resource processing results.

[0148] The model processing module 1040 is used to input the first prompt information into the large language model to obtain the meeting process resource processing results; the meeting process resource processing results include meeting minutes, meeting document summaries or meeting candidate questions.

[0149] The result caching module 1050 is used to store the results of the meeting process resource processing into the first cache.

[0150] The result acquisition module 1060 is used to retrieve the meeting process resource processing result corresponding to the result query instruction from the first cache in response to receiving the result query instruction.

[0151] The specific implementation of each module in the above-mentioned device can be referred to the implementation process of the corresponding steps in the above-mentioned method implementation method of this disclosure, and will not be repeated here.

[0152] This disclosure also provides a readable storage medium storing a computer program that, when executed by a processor, is used to implement the methods described above. A "readable storage medium" can be any means capable of containing, storing, communicating, propagating, or transmitting a program for use by or in conjunction with an instruction execution system, apparatus, or device. More specific examples of a readable storage medium include: an electrical connection with one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and portable read-only memory (CDROM), etc.

[0153] This disclosure also provides a computer program product, the methods of which can be implemented wholly or partially through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented wholly or partially as a computer program product. The computer program product includes one or more computer programs or instructions. When the computer program or instructions are loaded and executed, all or part of the processes or functions of this disclosure are performed.

[0154] Computer programs or instructions can be stored in a readable storage medium or transferred from one readable storage medium to another. For example, the computer program or instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The readable storage medium can be any available medium capable of access, or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium, such as a floppy disk, hard disk, or magnetic tape; an optical medium, such as a digital video optical disc; or a semiconductor medium, such as a solid-state drive. The computer-readable storage medium can be a volatile or non-volatile storage medium, or it can include both volatile and non-volatile types of storage media.

[0155] Those skilled in the art will understand that embodiments of this disclosure can be provided as methods, systems, or computer program products. Therefore, this disclosure can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this disclosure can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0156] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0157] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0158] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0159] In the description of this specification, the references to terms such as "one embodiment / mode," "some embodiments / modes," "example," "specific example," or "some examples," etc., refer to specific features, structures, or characteristics described in connection with that embodiment / mode or example, which are included in at least one embodiment / mode or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment / mode or example. Moreover, the specific features, structures, or characteristics described may be combined in any suitable manner in one or more embodiments / modes or examples. Furthermore, without contradiction, those skilled in the art can combine and integrate the different embodiments / modes or examples described in this specification, as well as the features of different embodiments / modes or examples.

[0160] Those skilled in the art should understand that the above embodiments are merely for illustrating the present disclosure and are not intended to limit the scope of the disclosure. Those skilled in the art can make other changes or modifications based on the above disclosure, and these changes or modifications still fall within the scope of the present disclosure.

Claims

1. A method for obtaining the results of resource processing during a meeting, characterized in that, include: In response to a cache indication triggered by the target meeting, the meeting process resources of the target meeting are downloaded to the local machine; The meeting process resources downloaded to the local machine are parsed to obtain the meeting process document; Based on the meeting process document, a template for generating first prompt information corresponding to the target meeting is populated to obtain the first prompt information; The first prompt message is used to guide the large language model to generate the resource processing results of the meeting process; The first prompt information is input into the large language model to obtain the resource processing results of the meeting process; The results of the meeting process resource processing include meeting minutes, meeting document summaries, or meeting candidate questions; Store the resource processing results of the meeting process in the first cache; as well as In response to receiving a result query instruction, the meeting process resource processing result corresponding to the result query instruction is retrieved from the first cache.

2. The method for obtaining meeting process resource processing results as described in claim 1, characterized in that, After obtaining the meeting proceedings document, the method further includes: Store the meeting process document in the second cache; In response to receiving a regeneration instruction for meeting process resource processing results, the meeting process document corresponding to the regeneration instruction is read from the second cache; and The meeting process document corresponding to the regenerated instruction is processed based on the large language model to obtain the meeting process resource processing result corresponding to the regenerated instruction.

3. The method for obtaining meeting process resource processing results as described in claim 2, characterized in that, The process of processing the meeting process document corresponding to the regenerated instruction based on the large language model to obtain the meeting process resource processing result corresponding to the regenerated instruction includes: Based on the meeting process document corresponding to the regeneration instruction, the template for generating the second prompt information corresponding to the regeneration instruction is filled to obtain the second prompt information; the second prompt information is used to guide the large language model to regenerate the meeting process resource processing results; and The second prompt information is input into the large language model to obtain the meeting process resource processing result corresponding to the regenerated instruction; the meeting process resource processing result corresponding to the regenerated instruction includes meeting minutes, meeting document summary or meeting candidate questions.

4. The method for obtaining meeting process resource processing results as described in claim 3, characterized in that, The second prompt information is generated by filling the template of the second prompt information corresponding to the regenerated instruction based on the meeting process document corresponding to the regenerated instruction, and the template is filled to obtain the second prompt information, including: Obtain the second prompt information generation template corresponding to the regeneration instruction from the preset template set; and Based on the meeting process document corresponding to the regenerated instruction, the template for generating the second prompt information is filled to obtain the second prompt information.

5. The method for obtaining meeting process resource processing results as described in any one of claims 1 to 3, characterized in that, After storing the meeting process resource processing results in the first cache, the method further includes: In response to the deletion of the second conference process resource, the conference process resource processing result corresponding to the second conference process resource is deleted from the first cache.

6. The method for obtaining meeting process resource processing results as described in any one of claims 1 to 3, characterized in that, Prior to triggering a cache indication in response to the target meeting, the method further includes: Retrieve the meeting attributes of the target meeting to be triggered in the cache; The meeting attributes are matched with a preset caching strategy to obtain a matching result; the matching result includes the caching timing; and The cache instruction for the target meeting is triggered based on the cache timing.

7. The method for obtaining meeting process resource processing results as described in claim 6, characterized in that, The step of matching the meeting attributes with a preset caching strategy to obtain a matching result includes: Obtain user behavior data and user attributes of the creator of the target meeting; and The meeting attributes, user behavior data, and user attributes are matched with a preset caching strategy to obtain the matching results.

8. The method for obtaining meeting process resource processing results as described in any one of claims 1 to 3, characterized in that, The first prompt information is generated by filling the template corresponding to the target meeting based on the meeting process document, and includes: To obtain the resource processing objectives of the target meeting; Generate a template by retrieving the first prompt information corresponding to the resource processing purpose from a preset template set; and The first prompt message is generated by filling the template based on the meeting process document.

9. An electronic device, characterized in that, include: The memory stores execution instructions; as well as A processor that executes the execution instructions stored in the memory, causing the processor to perform the meeting process resource processing result acquisition method according to any one of claims 1 to 8.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for obtaining the meeting process resource processing results as described in any one of claims 1 to 8.

Images