Information processing method and system, intelligent terminal and computer readable storage medium

Abstract

The invention discloses an information processing method and system, an intelligent terminal and a computer readable storage medium, and relates to the technical field of data processing.The method comprises the steps that an information processing workflow and to-be-processed information are obtained, and the information processing workflow comprises a plurality of work nodes; determining a target category corresponding to the to-be-processed information from a plurality of preset information categories; determining a target working node corresponding to the to-be-processed information from a plurality of working nodes of the information processing workflow according to the target category; and processing the to-be-processed information according to the target working node and working nodes connected behind the target working node in the information processing workflow, and outputting an information processing result. Therefore, the simplicity and the processing efficiency of information processing can be improved.

Description

Technical Field

[0001] This application relates to the technical field of data processing, and particularly to an information processing method, system, intelligent terminal, and computer-readable storage medium. Background Art

[0002] With the progress of science and technology, more and more data information needs to be processed. For example, in a work scenario, various types of information such as meeting reservations, meeting notifications, meeting minutes, and item management need to be processed.

[0003] In the prior art, different types of information processing rely on various different and independent information processing systems. Therefore, in a scenario where there may be multiple different types of information, it is necessary to set up and call multiple independent information processing systems, and the processing process is cumbersome and the efficiency is low, which is not conducive to improving the simplicity and processing efficiency of information processing.

[0004] Therefore, the related technology still needs to be improved and developed. Summary of the Invention

[0005] The main purpose of this application is to provide an information processing method, system, intelligent terminal, and computer-readable storage medium, aiming to solve the technical problem that in the related technology, different types of information processing rely on various different and independent information processing systems, the processing process is cumbersome and the efficiency is low, which is not conducive to improving the simplicity and efficiency of information processing.

[0006] To achieve the above purpose, in the first aspect of this application, an information processing method is provided, where the information processing method includes:

[0007] Obtain an information processing workflow and information to be processed, where the information processing workflow includes multiple work nodes;

[0008] Determine the target category corresponding to the information to be processed from a preset variety of information categories;

[0009] According to the target category, determine the target work node corresponding to the information to be processed from the multiple work nodes of the information processing workflow;

[0010] Process the information to be processed according to the target work node and the work nodes connected after the target work node in the information processing workflow, and output an information processing result.

[0011] Optionally, the work node of the information processing workflow includes a start node, and the start node is used to trigger the information processing workflow to start working;

[0012] The obtaining of the information processing workflow and the information to be processed includes:

[0013] Obtain an information processing workflow;

[0014] Through the start node of the above information processing workflow, obtain the information to be processed input by the target object.

[0015] Optionally, the working nodes of the above information processing workflow include an information classification node;

[0016] Determining the target category corresponding to the information to be processed from a preset variety of information categories includes:

[0017] Input the above information to be processed into the above information classification node;

[0018] Through the above information classification node, perform condition discrimination on the above information to be processed, and determine the target category corresponding to the above information to be processed from the preset variety of information categories according to the condition discrimination result.

[0019] Optionally, the above information to be processed includes meeting room reservation information, the above target category is meeting room reservation, and the above target working node is a meeting room reservation node;

[0020] Processing the above information to be processed according to the above target working node and the working nodes connected after the above target working node in the above information processing workflow, and outputting an information processing result, including:

[0021] Through the above meeting room reservation node, transmit the above meeting room reservation information to a large language model processing node to trigger the large language model processing node to determine a meeting room arrangement result according to the above meeting room reservation information, and transmit the above meeting room arrangement result to a reply node as the above information processing result;

[0022] Output the above meeting room arrangement result through the above reply node.

[0023] Optionally, the above information to be processed includes meeting minutes processing information, the above target category is meeting minutes processing, and the above target working node is a meeting minutes processing node;

[0024] Processing the above information to be processed according to the above target working node and the working nodes connected after the above target working node in the above information processing workflow, and outputting an information processing result, including:

[0025] Through the above meeting minutes processing node, transmit the above meeting minutes processing information to a large language model processing node to trigger the large language model processing node to generate a target meeting minutes according to the above meeting minutes processing information, and transmit the above target meeting minutes to a reply node as the above information processing result;

[0026] Output the above-mentioned target meeting minutes through the above-mentioned reply node.

[0027] Optionally, there are multiple target working nodes corresponding to the above-mentioned information to be processed;

[0028] Process the above-mentioned information to be processed according to the above-mentioned target working nodes and the working nodes connected after the above-mentioned target working nodes in the above-mentioned information processing workflow, and output the information processing result, including:

[0029] For each target working node, process the above-mentioned information to be processed according to the above-mentioned target working node and the working nodes connected after the above-mentioned target working node in the above-mentioned information processing workflow, obtain the candidate processing result corresponding to the above-mentioned target working node, and transmit the above-mentioned candidate processing result to the variable aggregation node;

[0030] Perform variable aggregation processing on the candidate processing results provided by each of the above-mentioned target working nodes through the above-mentioned variable aggregation node to obtain the information processing result, and transmit the above-mentioned information processing result to the reply node;

[0031] Output the above-mentioned information processing result through the above-mentioned reply node.

[0032] Optionally, the above-mentioned information processing workflow includes a knowledge retrieval node;

[0033] The above-mentioned method further includes: retrieving supplementary information corresponding to the above-mentioned information to be processed through the above-mentioned knowledge retrieval node;

[0034] Process the above-mentioned information to be processed according to the above-mentioned target working nodes and the working nodes connected after the above-mentioned target working node in the above-mentioned information processing workflow, and output the information processing result, including:

[0035] Obtain the above-mentioned supplementary information from the above-mentioned knowledge retrieval node through the above-mentioned target working node, process the above-mentioned information to be processed and the above-mentioned supplementary information according to the above-mentioned target working node and the working nodes connected after the above-mentioned target working node in the above-mentioned information processing workflow, and output the information processing result.

[0036] The second aspect of the present application provides an information processing system, wherein the above-mentioned information processing system includes:

[0037] A data acquisition module, configured to acquire an information processing workflow and information to be processed, wherein the above-mentioned information processing workflow includes multiple working nodes;

[0038] An information classification module, configured to determine the target category corresponding to the above-mentioned information to be processed from a preset variety of information categories;

[0039] A target working node determination module, configured to determine a target working node corresponding to the to-be-processed information from multiple working nodes of the information processing workflow according to the above-mentioned target category;

[0040] An information processing module, configured to process the to-be-processed information according to the above-mentioned target working node and the working nodes connected after the above-mentioned target working node in the information processing workflow, and output an information processing result.

[0041] A third aspect of this application provides an intelligent terminal, where the intelligent terminal includes a memory, a processor, and an information processing program stored on the memory and executable on the processor. When the information processing program is executed by the processor, the steps of any one of the above-mentioned information processing methods are implemented.

[0042] A fourth aspect of this application provides a computer-readable storage medium, on which an information processing program is stored. When the information processing program is executed by a processor, the steps of any one of the above-mentioned information processing methods are implemented.

[0043] As can be seen from the above, in the solution of this application, an information processing workflow and to-be-processed information are obtained, where the information processing workflow includes multiple working nodes; a target category corresponding to the to-be-processed information is determined from a preset variety of information categories; according to the above-mentioned target category, a target working node corresponding to the to-be-processed information is determined from multiple working nodes of the information processing workflow; the to-be-processed information is processed according to the above-mentioned target working node and the working nodes connected after the above-mentioned target working node in the information processing workflow, and an information processing result is output.

[0044] Compared with the prior art, in the solution corresponding to the information processing method provided by this application, when performing information processing, only one information processing workflow needs to be called, instead of relying on multiple different and independent information processing systems. Specifically, after classifying the to-be-processed information to determine the target category, a target working node for processing the target category is determined from the information processing workflow, so that the to-be-processed information is processed and the corresponding information processing result is output based on the target working node and the working nodes connected after the target working node. In this way, even in a scenario where there are multiple different types of information, only one information processing workflow needs to be set and called, which is beneficial to improving the simplicity and processing efficiency of information processing. Description of the Drawings

[0045] To more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other accompanying drawings can be obtained based on these drawings without creative efforts.

[0046] Figure 1 It is a schematic flowchart of an information processing method provided by an embodiment of the present application;

[0047] Figure 2 It is a schematic diagram of a work node of an information processing workflow provided by an embodiment of the present application;

[0048] Figure 3 It is a schematic diagram of the usage scenario of a problem classification node provided by an embodiment of the present application;

[0049] Figure 4 It is a schematic diagram of the usage scenario of an http request node provided by an embodiment of the present application;

[0050] Figure 5 It is a schematic diagram of the usage scenario of a conditional branch node provided by an embodiment of the present application;

[0051] Figure 6 It is a schematic diagram of the usage scenario of a knowledge retrieval node provided by an embodiment of the present application;

[0052] Figure 7 It is a schematic flowchart of a specific information processing process provided by an embodiment of the present application;

[0053] Figure 8 It is a schematic diagram of the component modules of an information processing system provided by an embodiment of the present application;

[0054] Figure 9 It is a schematic diagram of the internal structure principle of an intelligent terminal provided by an embodiment of the present application. Detailed implementation manners

[0055] In the following description, specific details such as specific system structures and technologies are proposed for the purpose of illustration rather than limitation, so as to thoroughly understand the embodiments of the present application. However, those skilled in the art should clearly understand that the present application can also be implemented in other embodiments without these specific details. In other cases, the detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid unnecessary details from interfering with the description of the present application.

[0056] It should be understood that when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integers, steps, operations, elements and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their groups.

[0057] It should also be understood that the terminology used in the specification of this application is for the purpose of describing particular embodiments only and is not intended to limit this application. As used in the specification of this application and the appended claims, unless the context clearly dictates otherwise, the singular forms "a", "an" and "the" are intended to include the plural forms.

[0058] It should be further understood that the term "and / or" used in the specification of this application and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0059] As used in this specification and the appended claims, the term "if" may be construed, depending on the context, as "when" or "once" or "in response to determining" or "in response to classifying to". Similarly, the phrase "if determined" or "if classified to [the described condition or event]" may be construed, depending on the context, as meaning "once determined" or "in response to determining" or "once classified to [the described condition or event]" or "in response to classifying to [the described condition or event]".

[0060] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this application without creative efforts shall fall within the scope of protection of this application.

[0061] Many specific details are set forth in the following description in order to provide a thorough understanding of this application, but this application may also be implemented in other ways different from those described herein. Those skilled in the art may make similar extensions without departing from the connotation of this application, so this application is not limited by the specific embodiments disclosed below.

[0062] Currently, there is an increasing amount of data information that needs to be processed. For example, when performing operations such as organizational development, meeting room reservation, meeting notice, meeting minutes collation, and meeting minutes sending in daily work, corresponding information processing is required to achieve related functions. In the prior art, different types of information processing rely on various different and independent information processing systems. Therefore, in scenarios where there may be multiple different types of information, it is necessary to set up and call multiple independent information processing systems, and the processing process is cumbersome and inefficient, which is not conducive to improving the simplicity and processing efficiency of information processing.

[0063] In an application scenario, a meeting system includes various subsystems and related functions such as meeting reservation, meeting notice, meeting minutes, and email assistant, but these subsystems and functions are all independent of each other. For example, for meeting room reservation, users need to manually screen the time and available meeting rooms that meet the meeting conditions. Specifically, it is necessary to manually select a time period and screen for a suitable meeting room during that time period, such as the need to support video conferencing and the need to have a projector, etc. Even if some systems support voice reservation based on artificial intelligence (AI, Artificial Intelligence), it is difficult for users to describe various requirements in a single voice.

[0064] Another example is that when meeting minutes need to be collated after a meeting, the related technology can only record the meeting voice and then convert the voice into text. However, during the conversion process, due to reasons such as lack of context understanding and low speech recognition accuracy, the text conversion effect may be poor. Further, it is necessary to extract information from it and organize it into meeting minutes. This information extraction workload is large and key information is easily lost. Moreover, the meeting minutes formats for different meeting types are different, and the current meeting minutes format is fixed and not flexible enough. Therefore, it is also necessary to manually perform structured collation according to the minutes template and finally select a suitable email format to send it out. The entire process is very cumbersome, inefficient, and requires a large amount of time and labor costs. In addition, different links need to be completed by different people, which is prone to errors and omissions.

[0065] To solve at least one of the above multiple technical problems, in the solution of this application, an information processing workflow and information to be processed are obtained, where the above information processing workflow includes multiple work nodes; a target category corresponding to the information to be processed is determined from a preset variety of information categories; according to the above target category, a target work node corresponding to the information to be processed is determined from the multiple work nodes of the above information processing workflow; according to the above target work node and the work nodes connected after the above target work node in the above information processing workflow, the information to be processed is processed, and an information processing result is output.

[0066] Compared with the prior art, in the solution corresponding to the information processing method provided by this application, when performing information processing, only one information processing workflow needs to be invoked, rather than relying on multiple different and independent information processing systems. Specifically, after classifying the information to be processed to determine the target category, the target work node for processing this target category is determined from the information processing workflow, so as to process the information to be processed based on the target work node and the work nodes connected after the target work node, and output the corresponding information processing result. In this way, even in a scenario where there are multiple different types of information, only one information processing workflow needs to be set and invoked, which is beneficial to improving the simplicity and processing efficiency of information processing.

[0067] As Figure 1 shown, an embodiment of this application provides an information processing method. Specifically, the above method includes the following steps:

[0068] Step S100, obtain an information processing workflow and the information to be processed, where the above information processing workflow includes multiple work nodes.

[0069] Among them, the above information processing workflow is a pre-set workflow, and the above information to be processed is the information that needs to be processed. Specifically, the above information processing workflow includes multiple work nodes connected in a logical order, and each work node and the connection relationship between each work node can be set and adjusted according to actual needs, and no specific limitation is made here.

[0070] The above information to be processed can be input by the user, obtained from a preset other information platform, or obtained through a preset information interface, and no specific limitation is made here.

[0071] Specifically, the work nodes of the above information processing workflow include a start node, and the start node is used to trigger the start of the above information processing workflow;

[0072] The above obtaining the information processing workflow and the information to be processed includes:

[0073] Obtain an information processing workflow;

[0074] Through the start node of the above information processing workflow, obtain the information to be processed input by the target object.

[0075] It should be noted that the information processing method provided in the embodiments of the present application can be integrated into an information processing system or an information processing application, and information processing is performed based on the information processing system or the information processing application. At this time, the start node can be used as the entry of the information processing application, and all user requirements are uniformly input at the start node. For example, when a user inputs a problem requirement as the information to be processed, the system can be allowed to make a conference room reservation, and it is required that the conference room must be able to accommodate 20 people, support video conferencing, support projector equipment, etc., and the list of participants needs to be input. These information are transmitted by the start node. After the meeting, the user can input text, upload files such as audio, video, pictures, and text to generate a meeting minutes, and the format of the meeting minutes can be required. Specifically, the text and files input by the user will be transmitted to the next node, such as transmitted to an information classification node (or a problem classification node) for information classification and subsequent processing.

[0076] Step S200, determine the target category corresponding to the to-be-processed information from a variety of preset information categories.

[0077] Among them, a variety of preset information categories can be preset and adjusted according to actual needs. For example, the above-mentioned variety of preset information categories can include conference room reservation, meeting minutes processing, platform access protocol-related issues, item release-related issues, etc., and can also include other categories, which are not specifically limited here.

[0078] In the embodiments of the present application, the entire information processing process is completed based on an information processing workflow. Specifically, the working nodes of the above-mentioned information processing workflow include an information classification node;

[0079] The above-mentioned determining the target category corresponding to the to-be-processed information from a variety of preset information categories includes:

[0080] Input the to-be-processed information into the information classification node;

[0081] Through the information classification node, perform condition discrimination on the to-be-processed information, and determine the target category corresponding to the to-be-processed information from the above-mentioned variety of preset information categories according to the condition discrimination result.

[0082] Specifically, discrimination conditions matching a variety of preset information categories are preset in the information classification node, so as to perform condition discrimination to determine the target category. For example, if the to-be-processed information contains keywords related to "conference room reservation", then the conference room reservation is taken as the target category.

[0083] It should be noted that if there is no information category in the preset multiple information categories that matches the information to be processed, feedback prompt information can be output to the target object (i.e., the user) to prompt that the currently input information to be processed is not within the processable range, facilitating the target object to adjust the input of the information to be processed in a timely manner and improving the processing efficiency and the probability of successful processing.

[0084] Step S300: According to the above target category, determine the target working node corresponding to the information to be processed from multiple working nodes of the above information processing workflow.

[0085] In the embodiments of the present application, the start node is respectively connected to multiple working nodes. According to the target category, at least one working node that matches the target category is determined from the multiple working nodes connected to the start node as the target working node.

[0086] Step S400: Process the information to be processed according to the above target working node and the working nodes connected after the above target working node in the above information processing workflow, and output the information processing result.

[0087] Specifically, for different working nodes, different working nodes may be connected according to the information categories that they may process and their corresponding processing methods, so as to implement the information processing process. The connection method between working nodes can be set and adjusted according to actual needs, and no specific limitation is made here.

[0088] In some scenarios of the embodiments of the present application, the information to be processed includes meeting room reservation information, the target category is meeting room reservation, and the target working node is the meeting room reservation node;

[0089] The processing of the information to be processed according to the above target working node and the working nodes connected after the above target working node in the above information processing workflow, and outputting the information processing result includes:

[0090] Transmit the above meeting room reservation information to the large language model processing node through the above meeting room reservation node to trigger the large language model processing node to determine the meeting room arrangement result according to the above meeting room reservation information, and transmit the above meeting room arrangement result to the reply node as the above information processing result;

[0091] Output the above meeting room arrangement result through the above reply node.

[0092] In some scenarios of the embodiments of the present application, the information to be processed includes meeting minutes processing information, the target category is meeting minutes processing, and the target working node is the meeting minutes processing node;

[0093] The above-mentioned information to be processed is processed according to the above-mentioned target working node and the working node connected after the above-mentioned target working node in the above-mentioned information processing workflow, and the information processing result is output, including:

[0094] The meeting minutes processing information is transmitted to the large language model processing node through the meeting minutes processing node, so as to trigger the large language model processing node to generate target meeting minutes according to the meeting minutes processing information, and transmit the target meeting minutes as the information processing result to the reply node;

[0095] The above target meeting minutes are output through the above reply node.

[0096] Among them, the above-mentioned large language model processing node performs information processing based on a pre-trained large language model, thereby utilizing the processing power of the large language model (LLM, Large Language Model) to solve complex information processing problems, for example, solving the complex conference room reservation, the difficulty in generating meeting minutes, and the cumbersome problems of sending meeting notices and meeting minutes, thereby improving the user experience.

[0097] It should be noted that, in an application scenario, multiple different large language model processing nodes can be set to perform targeted information processing based on each large language model node to improve the accuracy of information processing.

[0098] In the embodiment of the present application, a large language model processing node is set in the information processing workflow, and different functions of the large language model processing node are triggered based on different working nodes connected before the large language model processing node. In this way, the complexity of working node setting can be reduced, and the complexity of the entire system in the information processing process can be reduced.

[0099] Furthermore, in the embodiment of the present application, the information processing result is fed back to the user based on a unified reply node. The reply format corresponding to the reply node can be customized in advance, and plain text, charts, such as the meeting minutes format required by the user, can be output. Furthermore, different reply formats can be set for the information to be processed of different information categories, so that the user experience can be improved.

[0100] In some scenarios of the embodiments of the present application, the above-mentioned information to be processed corresponds to multiple target working nodes;

[0101] The above-mentioned information to be processed is processed according to the above-mentioned target working node and the working node connected after the above-mentioned target working node in the above-mentioned information processing workflow, and the information processing result is output, including:

[0102] For each target working node, based on the above target working node and the working nodes connected after the above target working node in the above information processing workflow, process the above information to be processed, obtain the candidate processing result corresponding to the above target working node, and transmit the above candidate processing result to the variable aggregation node;

[0103] Through the above variable aggregation node, perform variable aggregation processing on the candidate processing results provided by each of the above target working nodes to obtain the information processing result, and transmit the above information processing result to the reply node;

[0104] Output the above information processing result through the above reply node.

[0105] Specifically, the information to be processed can be information of a composite type, that is, corresponding to multiple information categories, or corresponding to a preset information category for indicating a composite type. At this time, the information to be processed can correspond to multiple target working nodes.

[0106] Furthermore, based on multiple target nodes and the downstream working nodes connected to each target node, perform information processing to implement the processing of information of a composite type. And further, perform data aggregation based on the variable aggregation node to implement result output based on a final reply node.

[0107] Among them, the above variable aggregation node can aggregate variables of multiple branches into one variable to achieve unified configuration of downstream nodes. Specifically, the variable aggregation node can integrate the output results of different branches to ensure that no matter which branch is executed, its result can be referenced and accessed through a unified variable. Through variable aggregation, multiple outputs such as problem classification or conditional branches can be aggregated into a single path for use and operation by nodes downstream in the process, simplifying the management of the data flow.

[0108] Further, the above information processing workflow includes a knowledge retrieval node;

[0109] The above method further includes: retrieving, through the above knowledge retrieval node, supplementary information corresponding to the above information to be processed;

[0110] The above processing the above information to be processed according to the above target working node and the working nodes connected after the above target working node in the above information processing workflow, and outputting the information processing result includes:

[0111] Obtain the above supplementary information from the above knowledge retrieval node through the above target working node, process the above information to be processed and the above supplementary information according to the above target working node and the working nodes connected after the above target working node in the above information processing workflow, and output the information processing result.

[0112] Specifically, the above knowledge retrieval node can perform information retrieval based on a pre-set database to obtain supplementary information corresponding to the information to be processed, so as to assist the information processing process based on the supplementary information.

[0113] It should be noted that the content provided in the above pre-set database can be pre-set and adjusted according to actual needs. For example, enterprise employee position information, industry-related reference information, etc. can be provided, and specific limitations are not made here.

[0114] In an application scenario, providing supplementary information to the large language model processing node can enable the large language model to obtain sufficient and complete knowledge to answer questions, so as to obtain more reliable results and better information processing effects.

[0115] As can be seen from the above, in the solution corresponding to the information processing method provided in the embodiments of the present application, when performing information processing, only one information processing workflow needs to be called, rather than relying on multiple different and independent information processing systems. Specifically, after classifying the information to be processed to determine the target category, the target work node for processing the target category is determined from the information processing workflow, so as to process the information to be processed and output the corresponding information processing result based on the target work node and the work nodes connected after the target work node. In this way, even in a scenario where there are multiple different types of information, only one information processing workflow needs to be set and called, which is beneficial to improving the simplicity and processing efficiency of information processing.

[0116] In the embodiments of the present application, the above information processing method is also specifically described based on a specific application scenario. The information processing method provided in the embodiments of the present application can be implemented based on an information processing system. In an application scenario, based on a large language model management platform, through the design and development of a workflow, an AI intelligent conference application system can be implemented. The system defines each function such as meeting reservation, meeting notice, meeting minutes generation, and meeting minutes sending at different nodes of the workflow to implement a conversational intelligent conference application system. Each of the above nodes has an embedded large language model and a Retrieval-augmented Generation (RAG) system to implement information processing, so as to implement functions such as rapid meeting room reservation, accurate meeting minutes generation, formatted meeting minutes generation, and sending.

[0117] In the embodiments of the present application, the above-mentioned AI intelligent conference application system is provided with a dialogue window where requirements can be input and files can be uploaded. The system embeds various system functions such as meeting reservation, meeting notice, and meeting minutes in different nodes. After a question is input in the input box, it will be classified. For example, if only a meeting room needs to be reserved, the question will be assigned to the meeting reservation node for processing. If a meeting minutes needs to be generated, it will be processed by the meeting minutes generation node. If all functions of meeting reservation, meeting notice, meeting minutes generation, and meeting minutes sending are required, they are implemented by 4 nodes respectively. For example, after the user reserves a meeting room, they can request a meeting notice. After the meeting ends, a voice file or text file can be uploaded in the dialogue window, and then the meeting minutes generation node will generate the meeting minutes, and the meeting minutes sending node will send the content. In this way, the workflow is designed based on the large language model management platform, and it is docked with the meeting reservation system, email system, etc. based on the large language model management platform to achieve conversational intelligent information processing.

[0118] In this way, the large language model management platform is docked with the meeting reservation system, email system, etc. to achieve conversational intelligent meeting management, and to build a fast, convenient, efficient and accurate meeting management system, which can solve the problems of cumbersome meeting room reservation and difficult meeting minutes collation.

[0119] It should be noted that the naming of the above-mentioned AI intelligent conference application system, as well as the division, connection and naming of each work node in the workflow of the AI intelligent conference application system, can be set and adjusted according to actual needs, and are not specifically limited here.

[0120] In a specific application scenario, based on the large language model management platform, the construction of a conversational intelligent conference application system is realized. Complex tasks such as meeting reservation, meeting notice, meeting minutes collation, and meeting minutes sending are decomposed into different step nodes. Based on the capabilities of the large language model system and the RAG system, multiple functions are integrated and implemented in one application system, and are realized based on a unified simple conversational window. In this way, based on the capabilities of the large language model and the RAG system, the problems of complex meeting room reservation, difficult meeting minutes generation, and cumbersome meeting notice and meeting minutes sending are solved, and the user experience is improved.

[0121] Figure 2 It is a schematic diagram of the work nodes of an information processing workflow provided by the embodiments of the present application, as Figure 2 shown, the above-mentioned information processing workflow includes a start node, a question classification node, a condition classification node, a knowledge retrieval node, a variable aggregation node, an http request node, a template conversion node, a large language model node, and a reply node. It should be noted that Figure 2The working nodes shown and the connection methods between the working nodes are only for illustration and not for specific limitation. During actual use, the working nodes can be designed and connected according to each subtask to be implemented, and different tasks can be achieved by combining different nodes. It should be further noted that the functions corresponding to the information classification nodes in the information processing method provided in the embodiments of the present application can be assigned to the problem classification nodes and the condition classification nodes for implementation, which is not specifically limited herein.

[0122] In some application scenarios of the embodiments of the present application, the development node serves as the entry of the entire system application and provides necessary initial information for the subsequent nodes and the normal operation of the application system. The condition classification node can split the entire system application into multiple branches based on preset conditions through if / else judgments. For example, issues related to the conference system application are taken as one branch, and other issues are taken as another branch. The problem classification node can integrate a large language model, make reasoning and judgment based on the user's input, and output the same type of problems to the next node. The knowledge retrieval node, as a supplement to the knowledge data of the entire system application, can contain various supplementary information such as professional content and enterprise internal data. If the user's question involves relevant professional issues within the enterprise or issues related to enterprise private data, relevant content is obtained from the knowledge retrieval node and then handed over to the large language model node for use and processing. The variable aggregation node is used to aggregate the variables of multiple branches into one variable, and the output results of different branches are unified as one variable for reference and access. In the case of multiple branches, for example, if the knowledge retrieval node has multiple knowledge bases and each knowledge base has a retrieval result, the results can be aggregated through variables and then uniformly given to the next large language model node for processing, without each retrieval result corresponding to a large language model node. The http request node is used to dock with a third-party platform to obtain and synchronize relevant data, thereby expanding the usage scenarios of the system. The template conversion node can convert some fragmented data into the required data format for output. For example, it can be output in a preset format such as JSON format or markdown format. The reply node is used to display the output results to the user, and all the question answers that the user wants are displayed through the reply node. It should be noted that the above-mentioned various nodes can be flexibly combined according to different application scenarios, which is not specifically limited herein.

[0123] Figure 3 is a schematic diagram of the usage scenario of a problem classification node provided by the embodiments of the present application, as Figure 3As shown, a problem classifier is set in the problem classification node, which is used to classify the user's requirements (i.e., the problems input by the user), and assign different problems to the corresponding node streams for processing. For example, if the user's problem is "Help me reserve a meeting room", then this problem will be assigned to the meeting room reservation branch node; if it is to generate a meeting minutes, it will be sent to the meeting minutes processing branch node. In the embodiment of the present application, the problem classification node is classified based on the preset large language model, but this is not specifically limited.

[0124] Figure 4 This is a schematic diagram of the usage scenario of an http request node provided by an embodiment of the present application. Based on the http node, http requests can be used to send feedback to the company's internal system. Specifically, server requests are sent through the http protocol to communicate with other systems to obtain relevant data. For example, to implement the meeting room reservation function and obtain the reservation information of the current meeting room, it is necessary to obtain the meeting room information through the meeting room management background of this node, obtain the data of the currently available meeting rooms for reservation, and hand over the meeting room reservation requirements of the user to the large language model for calculation and judgment, give the optimal result, and output one or more solutions for the user to choose. After the user determines the reserved meeting room information, this node also needs to submit the relevant data to the meeting room management background through a post request. When the user needs to send a meeting notice, it is necessary to notify the mail system, the phone or SMS notification api of the cloud platform through the http method to send relevant messages.

[0125] Figure 5 This is a schematic diagram of the usage scenario of a conditional branch node provided by an embodiment of the present application. The conditional branch node splits the entire intelligent meeting system application process into multiple branches according to conditional discrimination branches such as If / else / elif. For example, problems containing keywords such as meeting minutes, meeting reservation, and meeting notice in the user's questions can be processed as one branch respectively, and other problems as another branch.

[0126] Figure 6It is a schematic diagram of the usage scenario of a knowledge retrieval node provided by an embodiment of the present application. The knowledge retrieval node performs knowledge retrieval based on a preset knowledge base. The knowledge base is a way for the large language model to obtain the latest external knowledge and enterprise internal professional knowledge. For example, when directly providing the meeting content text or voice to the large language model for content generation in the meeting minutes generation function, there will be problems such as inaccurate recognition of professional terms, lack of understanding of context information and key information, and possible omission of important information. However, if the user's question is first searched in the knowledge base through vector search, relevant content is retrieved through semantic similarity matching, and then the meeting minutes content and the retrieved relevant knowledge are provided to the large language model, enabling the large language model to obtain sufficient and complete knowledge to answer the question, then more reliable results can be obtained. For example, a knowledge base can be constructed based on leadership position information, employee position information, etc. within the enterprise as a supplement to the relevant content information of the entire system.

[0127] The template conversion node is used to perform data format conversion to ensure that the finally output content meets the preset format requirements. In one application scenario, the template conversion node allows for lightweight and flexible data conversion in the intelligent application system workflow through the Python template language of Jinja2, and can perform text processing, JSON conversion, etc. For example, information such as a person's job title, level, department, etc. can be concatenated into complete information, and the meeting theme, time, content, participants, etc. can be concatenated into a meeting minutes content.

[0128] The large language model node is the core node in the workflow. All questions need to be finally processed by the large language model. The core of the entire system is to leverage the generation, classification, processing, etc. capabilities of the large language model. For example, when generating meeting minutes, it is necessary to use the knowledge base through the RAG system, reorganize the retrieved relevant knowledge and the user's question and reply to the question to generate the meeting minutes content in the format required by the user. When booking a meeting room, after obtaining the information of the meeting room, it needs to be passed to the large language model together with the user's requirements for calculation and processing, and the optimal solution is given to reply to the user, including the list of participants and the list of meeting minutes recipients required by the user.

[0129] The variable aggregation node is used to aggregate variables from multiple branches into one variable to achieve unified configuration of downstream nodes. The usage scenario of the variable aggregation node can refer to Figure 4。The variable aggregation node is a key node in the workflow. It is responsible for integrating the output results of different branches to ensure that regardless of which branch is executed, its results can be referenced and accessed through a unified variable. In the case of multiple branches, variables with the same function under different branches can be mapped to an output variable to avoid duplicate definition by downstream nodes. Through variable aggregation, multiple outputs such as problem classification or conditional branches can be aggregated into a single path for use and operation by downstream nodes in the process, simplifying the management of data streams.

[0130] The reply node is used for the grayscale content presented to the user, and the reply format can be customized. For example, it can output plain text, charts, or the meeting minutes format required by the user.

[0131] In the embodiments of the present application, based on the inference analysis and summary ability of the large language model, functions such as meeting room reservation, meeting notification, meeting minutes generation, and meeting minutes sending are divided into different branches and nodes through the architecture design of the workflow to implement a conversational intelligent meeting system and realize the processing of meeting-related information. Each function of the present application is executed in parallel in the background, providing a unified entry for users. Users can interact based on conversational information input without implementing various functions through complex UI designs. Based on the solution of the embodiments of the present application, data can be processed intelligently and efficiently, improving the work efficiency of users.

[0132] Figure 7 is a schematic diagram of a specific information processing process provided by the embodiments of the present application. It should be noted that Figure 7 The following uses scenarios such as user reservation of meeting rooms, meeting notification, meeting minutes generation, and meeting minutes sending for specific description. It should be further noted that based on the conditional branches corresponding to the background research, the entire processing system can be pre-trained in the background to improve information processing capabilities and efficiency. When implementing the background storage solution, one process looks up information in the background, and another large language model understands the situation and helps users answer questions according to the situation. In the actual application process, a processing step is as follows.

[0133] Step 1: The user enters the requirement "Reserve a 20-person meeting room at 10:00 am on Monday, which needs to support video conferencing and projection" in the dialogue window. After the start node receives the user's requirement, it makes a conditional judgment to determine which branch node this problem belongs to.

[0134] Step 2: Conditional judgment. Multiple conditions can be set for problem classification in conditional judgment. For example, questions containing keywords related to "meeting" in the question are transferred to the meeting problem classifier for classification processing.

[0135] Step 3: By analyzing that this problem belongs to the meeting room reservation problem, it is transferred to the meeting reservation branch node for processing.

[0136] Step 4: The meeting room reservation branch needs to obtain the reservation information of the current meeting room. At this time, the management background will communicate with the meeting system via HTTP, obtain all the available meeting room data currently, and return it to the management platform.

[0137] Step 5: After the management platform obtains the meeting room data, it passes it together with the keywords such as "10 am on Monday", "20-person meeting room", "video conference", and "projector" in the user's question to the large language model for meeting reservation. This model has the ability to handle meeting room reservation problems. After reasoning and calculation, the large language model gives multiple meeting room solutions for the user to choose. After the user selects a solution, the management platform returns this information to the meeting management background via HTTP.

[0138] The above steps 1 to 5 complete an independent function of meeting reservation. The meeting reservation function can also be combined with other steps to complete all the process functions of the entire meeting. For example, the meeting notification can be implemented based on the following steps 6 to 7.

[0139] Step 6: The user enters the list of participants who need to attend the meeting in the dialogue window. The start node determines this problem as querying meeting personnel information and gives the information of the participants to the enterprise internal personnel information knowledge base.

[0140] Step 7: The enterprise internal personnel information knowledge base performs a data vector query to obtain the information of relevant personnel, and passes the relevant information to the large language model. After obtaining the relevant information, the large language model processes the next meeting notification according to the defined format, and informs the email system or the SMS and phone notification cloud platform in the form of HTTP to conduct the meeting notification.

[0141] Step 8: The user enables the meeting recording function during the meeting to perform voice recording of the meeting content throughout the meeting. After the meeting ends, the user ends the meeting recording and saves the current voice file with a name. The system will automatically upload the meeting voice file to the meeting minutes document management background. After obtaining the user's requirements at the following node, the system automatically obtains the current meeting minutes audio file from the meeting minutes document management background for processing.

[0142] Step 9: Through keyword extraction and analysis, it is confirmed that the current requirement is to generate meeting minutes and is related to the meeting, and then it is transferred to the meeting problem classifier.

[0143] Step 10: After the problem classifier identifies the problem, it is determined as the branch for generating meeting minutes and is transferred to the branch for generating meeting minutes for processing.

[0144] Step 11: Knowledge base creation and invocation. The generation of meeting minutes is related to many professional and privatized contents within an enterprise. The large language model lacks knowledge in this field. By creating a knowledge base in the relevant field, the problem of data missing in the large language model can be supplemented.

[0145] Step 12: When the large language model generates meeting minutes, it will determine whether it lacks the current professional knowledge content. After confirming the lack, it will automatically retrieve from the knowledge base. After completing the vector retrieval of the knowledge base, it will restart the generation of the entire meeting minutes and output according to the format requirements.

[0146] The above steps 8 to 12 are the meeting minute generation steps. When sending the meeting minutes, they can be sent to relevant personnel in the standard format of an email.

[0147] Step 13: The user enters the requirement to send the meeting minutes in the input window.

[0148] Step 14: Condition judgment. By extracting that the user requirement is to send the meeting minutes, it is transferred to the meeting minute sending branch for processing.

[0149] Step 15: After the large language model for meeting minutes receives the request to send the meeting minutes, it will transfer the content format of the meeting minutes to the email system in the required email format in the way of http, and the email system will perform relevant sending and notification.

[0150] It should be further noted that the above steps can be independently completed by a certain functional module, or can be executed together by the entire meeting system, and multiple functions can also be realized simultaneously. In this way, functions such as meeting room reservation, meeting notification, generation and sending of meeting minutes can be supported. Each function can be completed independently, or meeting support can be provided during the entire meeting process to realize the functions required in each stage of the meeting, which can improve the processing efficiency and processing effect of relevant information and is beneficial to improving the user experience.

[0151] As Figure 8 shown in

[0152] A data acquisition module 810, configured to acquire an information processing workflow and information to be processed, where the information processing workflow includes multiple work nodes;

[0153] An information classification module 820, configured to determine a target category corresponding to the information to be processed from a preset variety of information categories;

[0154] A target working node determination module 830, configured to determine a target working node corresponding to the to-be-processed information from multiple working nodes of the information processing workflow according to the above-mentioned target category;

[0155] An information processing module 840, configured to process the to-be-processed information according to the above-mentioned target working node and the working nodes connected after the target working node in the information processing workflow, and output an information processing result.

[0156] In this way, when performing information processing, only one information processing workflow needs to be called, rather than relying on multiple different and independent information processing systems. Specifically, after classifying the to-be-processed information to determine the target category, the target working node for processing this target category is determined from the information processing workflow, so as to process the to-be-processed information based on the target working node and the working nodes connected after the target working node and output the corresponding information processing result. In this way, even in a scenario where there are multiple different types of information, only one information processing workflow needs to be set up and called, which is beneficial to improving the simplicity and processing efficiency of information processing.

[0157] It should be noted that the specific structure and implementation manner of the above-mentioned information processing system and its various modules or units can refer to the corresponding descriptions in the above method embodiments, and will not be elaborated here.

[0158] It should be noted that the division method of each module of the above-mentioned information processing system is not unique and will not be specifically limited here.

[0159] Based on the above embodiments, the present application further provides an intelligent terminal, and its principle block diagram can be as Figure 9 shown. The above intelligent terminal includes a processor, a memory, a network interface, and a display screen connected through a system bus. Among them, the processor of the intelligent terminal is used to provide computing and control capabilities. The memory of the intelligent terminal includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and an information processing program. The internal memory provides an environment for the operation of the operating system and the information processing program in the non-volatile storage medium. The network interface of the intelligent terminal is used to communicate with an external terminal through a network connection. When the information processing program is executed by the processor, it implements the steps of any one of the above information processing methods. The display screen of the intelligent terminal can be a liquid crystal display screen or an electronic ink display screen.

[0160] Those skilled in the art can understand, Figure 9The principle block diagram shown only shows the block diagram of some structures related to the solution of this application, and does not constitute a limitation on the intelligent terminal to which the solution of this application is applied. The specific intelligent terminal may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0161] In one embodiment, an intelligent terminal is provided. The intelligent terminal includes a memory, a processor, and an information processing program stored on the memory and executable on the processor. When the information processing program is executed by the processor, the steps of any one of the information processing methods provided by the embodiments of this application are implemented.

[0162] The embodiments of this application also provide a computer-readable storage medium. An information processing program is stored on the computer-readable storage medium. When the information processing program is executed by a processor, the steps of any one of the information processing methods provided by the embodiments of this application are implemented.

[0163] It should be understood that the sequence numbers of the steps in the above embodiments do not mean the order of execution. The execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0164] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the above-mentioned division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the above device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated into a processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit. In addition, the specific names of each functional unit and module are only for the convenience of mutual distinction and do not limit the protection scope of this application. The specific working process of the units and modules in the above device can refer to the corresponding process in the foregoing method embodiments and will not be repeated here.

[0165] In the above embodiments, the descriptions of each embodiment have their own emphases. For the parts not detailed or recorded in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0166] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0167] In the embodiments provided in this application, it should be understood that the disclosed systems / terminal devices and methods can be implemented in other ways. For example, the system / terminal device embodiments described above are merely illustrative. For example, the above-mentioned division of modules or units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0168] If the above-mentioned integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, to implement all or part of the processes in the above-mentioned method embodiments of this application, it can also be completed by a computer program instructing relevant hardware. The above-mentioned computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the above-mentioned computer program includes computer program code, and the above-mentioned computer program code can be in the form of source code, object code, executable file or some intermediate form, etc. The above-mentioned computer-readable medium can include: any entity or device that can carry the above-mentioned computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, and software distribution medium, etc. It should be noted that the content included in the above-mentioned computer-readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction.

[0169] The above-mentioned embodiments are only used to illustrate the technical solutions of this application, rather than to limit them; although this application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not deviate from the spirit and scope of the technical solutions of the various embodiments of this application, and should all be included in the protection scope of this application.

Claims

1. An information processing method, characterized in that: The method comprises: Acquire an information processing workflow and information to be processed, wherein the information processing workflow includes multiple work nodes; Determining a target category corresponding to the information to be processed from a plurality of preset information categories; According to the target category, determining a target working node corresponding to the information to be processed from a plurality of working nodes of the information processing workflow; The information to be processed is processed according to the target working node and the working node connected after the target working node in the information processing workflow, and an information processing result is output.

2. The information processing method according to claim 1, characterized in that: The working nodes of the information processing workflow include a start node, and the start node is used to trigger the information processing workflow to start working; The information acquisition processing workflow and the information to be processed include: Get information processing workflow; The information to be processed input by the target object is obtained through the start node of the information processing workflow.

3. The information processing method according to claim 1, characterized in that: The working nodes of the information processing workflow include information classification nodes; The step of determining a target category corresponding to the information to be processed from a plurality of preset information categories includes: Inputting the information to be processed into the information classification node; The information classification node performs conditional judgment on the information to be processed, and determines the target category corresponding to the information to be processed from the preset multiple information categories according to the conditional judgment result.

4. The information processing method according to claim 1, characterized in that: The information to be processed includes conference room reservation information, the target category is conference room reservation, and the target working node is a conference room reservation node; The processing of the information to be processed according to the target working node and the working node connected after the target working node in the information processing workflow, and outputting the information processing result, includes: The conference room reservation information is transmitted to the large language model processing node through the conference room reservation node, so as to trigger the large language model processing node to determine the conference room arrangement result according to the conference room reservation information, and transmit the conference room arrangement result as the information processing result to the reply node; The conference room arrangement result is outputted through the reply node.

5. The information processing method according to claim 1, characterized in that: The information to be processed includes meeting minutes processing information, the target category is meeting minutes processing, and the target working node is a meeting minutes processing node; The processing of the information to be processed according to the target working node and the working node connected after the target working node in the information processing workflow, and outputting the information processing result, includes: The meeting minutes processing information is transmitted to the large language model processing node through the meeting minutes processing node to trigger the large language model processing node to generate a target meeting minutes according to the meeting minutes processing information, and transmit the target meeting minutes as the information processing result to the reply node; The target meeting minutes are outputted through the reply node.

6. The information processing method according to claim 1, characterized in that: The information to be processed corresponds to multiple target working nodes; The processing of the information to be processed according to the target working node and the working node connected after the target working node in the information processing workflow, and outputting the information processing result, includes: For each target working node, the information to be processed is processed according to the target working node and the working node connected to the target working node in the information processing workflow, to obtain a candidate processing result corresponding to the target working node, and the candidate processing result is transmitted to the variable aggregation node; Performing variable aggregation processing on the candidate processing results provided by each of the target working nodes through the variable aggregation node to obtain an information processing result, and transmitting the information processing result to the reply node; The information processing result is outputted through the reply node.

7. The information processing method according to any one of claims 1 to 6, characterized in that: The information processing workflow includes a knowledge retrieval node; The method further comprises: retrieving through the knowledge retrieval node to obtain supplementary information corresponding to the information to be processed; The processing of the information to be processed according to the target working node and the working node connected after the target working node in the information processing workflow, and outputting the information processing result, includes: The supplementary information is obtained from the knowledge retrieval node through the target working node, and the information to be processed and the supplementary information are processed according to the target working node and the working node connected to the target working node in the information processing workflow, and the information processing result is output.

8. An information processing system, characterized in that: The system comprises: A data acquisition module, used to acquire an information processing workflow and information to be processed, wherein the information processing workflow includes a plurality of working nodes; An information classification module, used to determine a target category corresponding to the information to be processed from a plurality of preset information categories; A target working node determination module, used to determine the target working node corresponding to the information to be processed from multiple working nodes of the information processing workflow according to the target category; The information processing module is used to process the information to be processed according to the target working node and the working node connected after the target working node in the information processing workflow, and output the information processing result.

9. An intelligent terminal, characterized in that: The intelligent terminal includes a memory, a processor, and an information processing program stored in the memory and executable on the processor. When the information processing program is executed by the processor, the steps of the information processing method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores an information processing program, and when the information processing program is executed by the processor, the steps of the information processing method according to any one of claims 1 to 7 are implemented.

Images