Remote cooperation data analysis method and device based on large-model multi-agent

Abstract

The invention discloses a remote cooperation data analysis method and device based on a large model and multiple agents. The device comprises a large model, a coordination agent, a data discovery agent, an analysis agent, a report generation agent and a report review agent. Through the data discovery agent deployed at each remote node, the data resources in the local database are automatically scanned and reported, full automation is realized, and the discovery efficiency and real-time performance of the remote data are greatly improved. And the data discovery agent can screen and integrate useful data in advance, so that the burden of the central processing node is reduced. According to the method, the processing flow of the distributed data is optimized, the data integration speed and quality are improved through remote automatic data discovery and preliminary evaluation, and a complex and dynamic investment and research data environment is effectively supported.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and in particular to a remote collaborative data analysis method and device based on large-model multi-agents. Background Art

[0002] In modern corporate investment research and analysis, the integrity, timeliness and diversity of data are crucial. However, in reality, data is often scattered in different geographical locations, and data resources in different locations are unaware of each other, forming "data islands". This data isolation makes it impossible to fully utilize data resources in multiple locations, limiting the depth and breadth of investment research and analysis. At the same time, data resources often change, further increasing the difficulty of data integration.

[0003] Existing methods of distributed data processing and analysis usually require that data be centralized in a single location for processing. This leads to problems such as high data transmission costs, data privacy and security risks, and a lack of intelligent collaboration mechanisms. Summary of the invention

[0004] The present disclosure provides a remote collaborative data analysis method, device, equipment and storage medium based on a large model and multiple intelligent agents to at least solve the above technical problems existing in the prior art.

[0005] According to a first aspect of the present application, a remote collaborative data analysis method based on a large model multi-agent is provided, the method comprising:

[0006] Constructing a large model, a coordination agent, a data discovery agent, an analysis agent, a report generation agent, and a report review agent; wherein the large model, the coordination agent, the report generation agent, and the report review agent are arranged on a central server, and the data discovery agent and the analysis agent are arranged on a remote server, wherein the remote server includes a plurality of remote servers, and the plurality of remote servers are connected to the central server;

[0007] The large model receives user requirements, understands the user requirements, and generates a target report task;

[0008] The coordination agent determines the analysis target and required data according to the target reporting task, and sends a data acquisition request to the data discovery agent;

[0009] After receiving the data request, the data discovery agent obtains data meta information from the local database and sends it to the coordination agent, and the coordination agent identifies the target data discovery agent with the required data in the data meta information, and sends a request to obtain the required data to the target data discovery agent; wherein the target data discovery agent includes multiple agents, which are arranged in multiple remote servers;

[0010] Based on the demand data request, the target data discovery agent obtains the demand data from a local database;

[0011] The analysis agent analyzes the demand data and generates analysis results;

[0012] The coordination agent receives the analysis results from multiple remote servers and performs data integration to obtain integrated data;

[0013] The report generation agent generates an initial report based on the integrated data;

[0014] The report review agent reviews the initial report, revises the initial report according to the review result, and obtains the target report.

[0015] In one possible implementation, the large model is communicatively connected to the coordinating agent, and the coordinating agent is communicatively connected to the data discovery agent, the analysis agent, and the report generation agent respectively; the data discovery agent is communicatively connected to the analysis agent; and the report generation agent is communicatively connected to the report review agent.

[0016] In one possible implementation, after the coordination agent receives the target reporting task, the method further includes:

[0017] Determining whether the coordination agent communicates normally with the data discovery agents in each remote server;

[0018] After confirming normal communication, a request for obtaining data is sent to the data discovery agent.

[0019] In one possible implementation, the coordination agent identifies the target data discovery agent having the required data in the data meta information, including:

[0020] Identifying demand data from the data meta information; the demand data includes multiple types;

[0021] Determine the data discovery agent corresponding to the identified demand data, and determine the corresponding data discovery agent as the target data discovery agent.

[0022] In one possible implementation manner, the step of determining the data discovery agent corresponding to the identified demand data and determining the corresponding data discovery agent as the target data discovery agent includes:

[0023] Processing the acquired demand data, matching the processed demand data with the analysis target, and obtaining matching data;

[0024] The coordinating agent corresponding to the matching data is determined as the target coordinating agent.

[0025] In one possible implementation manner, the analysis agent analyzes the demand data to generate analysis results, including:

[0026] receiving an analysis goal of the coordination agent;

[0027] Processing outliers and missing values ​​in the demand data;

[0028] Based on the preset prompt text requirements, multiple types of demand data are analyzed in multiple dimensions to obtain analysis results.

[0029] In one possible implementation manner, the coordination agent receives the analysis results from multiple remote servers and performs data integration to obtain integrated data, including:

[0030] Check whether the data in the received analysis results are complete, and if complete, verify whether the data in the analysis results are within a reasonable range. If they are within a reasonable range, determine whether the analysis results meet the requirements of the analysis objectives, and save the analysis results that meet the requirements as integrated data.

[0031] In one embodiment, the report generation agent generates an initial report based on the integrated data, including:

[0032] The integrated data is received, a text portion of an initial report is generated based on the integrated data and a preset report template, and a picture position symbol is inserted into the initial report to obtain the initial report.

[0033] In one possible implementation manner, the reviewing of the initial report includes:

[0034] Check the correctness of the data and conclusions of the initial report;

[0035] Check the logic of the initial report;

[0036] The initial report is checked for syntax.

[0037] According to a second aspect of the present application, a remote collaborative data analysis device based on a large model and multiple agents is provided, comprising:

[0038] A construction module is used to construct a large model, a coordination agent, a data discovery agent, an analysis agent, a report generation agent, and a report review agent; wherein the large model, the coordination agent, the report generation agent, and the report review agent are arranged on a central server, and the data discovery agent and the analysis agent are arranged on a remote server, wherein the remote server includes a plurality of remote servers, and the plurality of remote servers are connected to the central server;

[0039] An understanding module, used for the large model to receive user requirements, understand the user requirements, and generate a target report task;

[0040] A determination module, used for the coordination agent to determine the analysis target and required data according to the target reporting task, and send a data acquisition request to the data discovery agent;

[0041] an identification module, for the data discovery agent to obtain data meta information from a local database and send it to the coordination agent after receiving the data request, the coordination agent to identify the target data discovery agent with the required data in the data meta information, and send a request to obtain the required data to the target data discovery agent; wherein the target data discovery agent includes a plurality of agents, which are arranged in a plurality of remote servers;

[0042] An acquisition module, configured to enable the target data discovery agent to acquire the required data from a local database based on the required data request;

[0043] An analysis module, used for the analysis agent to analyze the demand data and generate analysis results;

[0044] An integration module, used for the coordination agent to receive the analysis results of multiple remote servers to perform data integration to obtain integrated data;

[0045] A generating module, used for the report generating agent to generate an initial report based on the integrated data;

[0046] The review module is used for the report review agent to review the initial report, revise the initial report according to the review result, and obtain the target report.

[0047] According to a third aspect of the present application, an electronic device is provided, including:

[0048] at least one processor; and

[0049] a memory communicatively connected to the at least one processor; wherein,

[0050] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described in the present application.

[0051] According to a fourth aspect of the present application, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute the method described in the present application.

[0052] According to a fifth aspect of the present application, a computer program product is provided, comprising a computer program or instructions, which implement the method described in the present application when executed by a processor.

[0053] By utilizing the technical solution of the present application, the original audio and the new audio can be played separately.

[0054] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present application will become readily understood. In the accompanying drawings, several embodiments of the present application are shown in an exemplary and non-limiting manner, wherein:

[0056] In the drawings, the same or corresponding reference numerals represent the same or corresponding parts.

[0057] Figure 1 A schematic diagram of the implementation process of the remote collaborative data analysis method based on a large model and multiple agents in an embodiment of the present application is shown;

[0058] Figure 2 A schematic diagram showing the structure of a central server and a remote server in an embodiment of the present application is shown;

[0059] Figure 3 A schematic diagram of the structure of a remote collaborative data analysis device based on a large model multi-agent in an embodiment of the present application is shown;

[0060] Figure 4 A schematic diagram of the structure of an electronic device in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0061] In order to make the purpose, features, and advantages of the present application more obvious and easy to understand, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0062] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0063] In the following description, reference is made to “some embodiments”, which describe a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0064] In the following description, the terms "first\second" involved are merely used to distinguish similar objects and do not represent a specific ordering of the objects. It can be understood that "first\second" can be interchanged with a specific order or sequence where permitted, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0065] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0066] It should be understood that in the various embodiments of the present application, the size of the serial number of each implementation process does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0067] In the related technologies, in existing large-scale multi-robot collaborative tasks, the communication cost between agents is high, the policy control efficiency is low, and it is difficult to adapt to the policy changes of other agents. The existing multi-agent large model interaction method is to obtain user input data and perform collaborative processing through multiple large language model agents to simplify the interaction process. In the process of agent collaboration, task planning and task execution plans are used to allocate and process tasks, but the existing industrial Internet platform interaction is complex, user operation is difficult, and generalization ability is weak.

[0068] In summary, although the distributed system in the prior art obtains data from each sub-node to the central node, each sub-node obtains data according to a set degree, that is, all relevant data are collected to the central node, which are all fixed procedures or fixed processes.

[0069] The following describes a remote collaborative data analysis method and device based on a large model and multiple agents provided by the present application in conjunction with the accompanying drawings.

[0070] Figure 1The schematic diagram shows the implementation process of the remote collaborative data analysis method based on a large model multi-agent in the embodiment of the present application. Figure 1 .like Figure 1 As shown, the method includes:

[0071] S (Step) 101: construct a large model, a coordination agent, a data discovery agent, an analysis agent, a report generation agent, and a report review agent; wherein the large model, the coordination agent, the report generation agent, and the report review agent are arranged on a central server, and the data discovery agent and the analysis agent are arranged on a remote server, wherein the remote server includes a plurality of remote servers, and the plurality of remote servers are connected to the central server;

[0072] It should be noted that the attached Figure 2 As shown, in the present application, a plurality of remote nodes and central nodes are provided, wherein the remote nodes are provided with remote servers, the central nodes are provided with central servers, and the remote servers are communicatively connected with the central servers; wherein both the central servers and the remote servers are provided with local databases, and the data contents stored in the local databases of each remote server are different or partially different; in the present application, the large model is communicatively connected with the coordinating agent, and the coordinating agent is communicatively connected with the data discovery agent, the analysis agent, and the report generation agent respectively, and the data discovery agent is communicatively connected with the analysis agent; and the report generation agent is communicatively connected with the report review agent.

[0073] For example, in this application, there are four remote servers, and the local databases in the four remote servers can be the first database, the second database, the third database and the fourth database. The local databases in this application can use MySQL, PostgreSQL, etc. This application can also store data in the local databases of each remote server in advance. The data can be financial, market or other related data, and the storage method can be a CSV or Excel file. The central server in this application is provided with a large language model, referred to as the large model. The large model provided in this application can understand the request input by the user and output the target text corresponding to the input request.

[0074] Specifically, in this application, a remote collaborative data analysis method is implemented by constructing a large model, a coordination agent, a data discovery agent, an analysis agent, a report generation agent, and a report review agent.

[0075] S102: The large model receives user requirements, understands the user requirements, and generates a target report task;

[0076] In this application, the big model is first used to receive user needs. After the big model understands the user needs, the target report task can be obtained. It can be understood that the user needs in this application can be that the user publishes the investment research data analysis report task of enterprise A through the system interface or API, specifying the analysis goals and requirements. It should be noted that the user needs to provide information such as the company name (such as "Company A"), analysis scope, and report delivery time. Data format: Task information is stored in JSON format.

[0077] S103: The coordination agent determines the analysis target and required data according to the target reporting task, and sends a data acquisition request to the data discovery agent;

[0078] It is understandable that in this application, the coordination agent is started on the central server: it is ensured that the coordination agent listens on the specified port and terminal, the coordination agent can receive the target report task generated by the large model, and can also send a data acquisition request DataDiscoveryRequest message to each data discovery agent. It should be noted that at this time, the coordination agent needs to obtain all the data in the local database of all remote servers.

[0079] S104: After receiving the data request, the data discovery agent obtains data meta information from the local database and sends it to the coordination agent, the coordination agent identifies the target data discovery agent with the required data in the data meta information, and sends a request to obtain the required data to the target data discovery agent; wherein the target data discovery agent includes multiple agents, which are arranged in multiple remote servers;

[0080] Specifically, after receiving the data request, the data discovery agents in all remote servers can obtain the data elements in the corresponding local database and send the data elements to the coordination agent. The coordination agent identifies whether there is required data in the data element information. It can be understood that the required data can be data available in the local databases of multiple remote servers. Therefore, the required data may need to be obtained from several local databases, or it may only need to be obtained from one local server. The data discovery agent corresponding to the local database where the required data exists is used as the target data discovery agent, and the required data request is sent to these target data discovery agents, thereby specifying the target data discovery agent to obtain the required data. In the present application, the data discovery agent can also send a data discovery report to the coordination agent.

[0081] S105: Based on the demand data request, the target data discovery agent obtains the demand data from the local database;

[0082] After receiving the required data request, the target data discovery agent explores the connected local database, traverses the table, obtains columns and other detailed information, classifies the obtained table, and can also evaluate each table based on integrity, number of null values, update time, etc., and finally obtains the required data. In this application, the target data discovery agent can also send a data discovery report to the coordination agent.

[0083] S106: The analysis agent analyzes the demand data and generates analysis results;

[0084] The analysis agent is connected to the data discovery agent. After receiving the demand data, it analyzes the demand data and generates corresponding analysis reports, including financial analysis, industry and market analysis, etc., and sends the analysis results to the coordination agent.

[0085] S107: The coordination agent receives the analysis results from multiple remote servers and performs data integration to obtain integrated data;

[0086] The coordination agent can collect the analysis results of each remote node, integrate the data, and pass the integrated data to the report generation agent.

[0087] S108: The report generating agent generates an initial report based on the integrated data;

[0088] After receiving the data integrated by the coordination agent, the report generation agent generates preliminary investment research report content based on the data and passes the preliminary report to the report content review agent. Among them, the report generation tool in this application can use LaTeX, Markdown, Jupyter Notebook and other tools to generate report content.

[0089] S109: The report review agent reviews the initial report, and revises the initial report according to the review result to obtain a target report.

[0090] Specifically, after the report review agent in this application receives the preliminary report of the report generating agent, it reviews the accuracy, logic and consistency of the report content, and then revise and optimize the report, and finally generates the final report in PDF format for final delivery.

[0091] The present application provides a remote collaborative data analysis method based on a large model and multiple agents. After the user inputs the requirements, the large model understands the user requirements and generates a target report task. Then the coordinating agent determines the analysis target and the required data according to the target report task; the coordinating agent sends a data acquisition request to the data discovery agent, and then the data discovery agent obtains the data metadata from the local database according to the data request. The coordinating agent identifies whether the data metadata contains the required data, thereby determining which data discovery agent can provide the required data, and sends the required data request to the data discovery agent, so that the target data discovery agent obtains the required data from the local database; the analysis agent analyzes the required data and generates analysis results. The coordinating agent receives the analysis results of each remote server and performs data integration to obtain integrated data. The report generation agent generates an initial report based on the integrated data. The report review agent reviews and revises the initial report to obtain the target report.

[0092] This application achieves full automation by automatically scanning and reporting data resources in the local database through data discovery agents deployed on each remote node, greatly improving the efficiency and real-time nature of remote data discovery. The data discovery agent can screen and integrate useful data in advance, reducing the burden on central processing nodes. In this application, data is transmitted between agents through a predefined communication protocol, ensuring the security and integrity of data between remote nodes, while improving the response speed and accuracy of data processing. This application optimizes the processing flow of distributed data, improves the speed and quality of data integration through remote automated data discovery and preliminary evaluation, and effectively supports complex and dynamic investment research data environments.

[0093] In some embodiments, after the coordination agent receives the target reporting task, the method further includes:

[0094] Determining whether the coordination agent communicates normally with the data discovery agents in each remote server;

[0095] After confirming normal communication, a request for obtaining data is sent to the data discovery agent.

[0096] It should be noted that after the coordination agent receives the target report task, it first starts the data discovery agent on each remote server, runs each data discovery agent, and ensures that each agent is connected to the coordination agent at startup. Then, the data acquisition request is sent to the data discovery agent.

[0097] In some embodiments, the coordination agent identifies the target data discovery agent having the required data in the data meta information, including:

[0098] Identifying demand data from the data meta information; the demand data includes multiple types;

[0099] Determine the data discovery agent corresponding to the identified demand data, and determine the corresponding data discovery agent as the target data discovery agent.

[0100] In some embodiments, the step of determining a data discovery agent corresponding to the identified demand data and determining the corresponding data discovery agent as a target data discovery agent includes:

[0101] Processing the acquired demand data, matching the processed demand data with the analysis target, and obtaining matching data;

[0102] The coordinating agent corresponding to the matching data is determined as the target coordinating agent.

[0103] In this application, the data discovery agent connects to the local database and extracts metadata after receiving a request to obtain data, including traversing the table, obtaining columns and other detailed information; and classifying the table, and evaluating each table based on completeness, number of null values, update time, etc. The metadata is then structured into a predefined JSON format, and a data discovery report is sent to the coordination agent. The coordination agent thereby identifies the demand data from the data metadata; determines the data discovery agent corresponding to the identified demand data, and determines the corresponding data discovery agent as the target data discovery agent. Specifically, identifying the demand data includes processing the acquired demand data, matching the processed demand data with the analysis target, obtaining matching data, and finally determining the coordination agent corresponding to the matching data as the target coordination agent.

[0104] In some embodiments, the analysis agent analyzes the demand data to generate analysis results, including:

[0105] receiving an analysis goal of the coordination agent;

[0106] Processing outliers and missing values ​​in the demand data;

[0107] Based on the preset prompt text requirements, multiple types of demand data are analyzed in multiple dimensions to obtain analysis results.

[0108] Specifically, the analysis agent in this application obtains and analyzes local data according to the task instructions of the coordination agent's analysis target. Specifically, it can use Pandas, NumPy and other libraries for data processing, specifically processing outliers and missing values ​​in the demand data, and then based on the preset prompt text requirements, conduct multi-dimensional analysis of various types of demand data, generate corresponding analysis reports, including financial analysis, industry and market analysis, etc., and send the analysis results to the coordination agent. Among them, the analysis results include data tables, charts (drawn using Matplotlib), text analysis, etc. It should be noted that the analysis agent of this application can implement analysis models and algorithms in their respective fields.

[0109] It can be understood that the analysis agent in the present application can be a financial analysis agent, an industry and market analysis agent, a competitiveness analysis agent, a management and governance structure analysis agent, and a risk analysis agent.

[0110] In some embodiments, the coordination agent receives the analysis results from multiple remote servers and integrates the data to obtain integrated data, including:

[0111] Check whether the data in the received analysis results are complete, and if complete, verify whether the data in the analysis results are within a reasonable range. If they are within a reasonable range, determine whether the analysis results meet the requirements of the analysis objectives, and save the analysis results that meet the requirements as integrated data.

[0112] Specifically, after receiving the analysis results, the coordination agent in this application evaluates the analysis results. The evaluation criteria include: data integrity, specifically checking whether all required indicators are included. Data accuracy, specifically verifying whether the data is within a reasonable range. Result applicability, specifically judging whether the analysis results meet the task requirements. Finally, the best result is selected. For multiple results: compare data quality scores and analysis depth, and select the best result. Record the decision-making process and save the basis for selection for easy traceability.

[0113] In some embodiments, the report generating agent generates an initial report based on the integrated data, including:

[0114] The integrated data is received, a text portion of an initial report is generated based on the integrated data and a preset report template, and a picture position symbol is inserted into the initial report to obtain the initial report.

[0115] In this application, the coordination agent sends the integrated analysis results to the report generation agent, and the report generation agent writes the report: specifically including receiving the analysis results: the analysis results of each analysis dimension in JSON format, including the analysis content of all dimensions. Generate report content: call the local Ollama model to generate the text part of the report. And insert image location symbols in the text. Among them, a predefined report template is used in this application, which includes a cover, table of contents, abstract, main text, conclusion and other parts. Insert content into the report template, and embed the generated text content and charts into the template. Finally, a preliminary report file is generated, specifically using Markdown or LaTeX format to organize the content, which is convenient for subsequent conversion to PDF.

[0116] In some embodiments, the reviewing of the initial report includes:

[0117] Check the correctness of the data and conclusions of the initial report;

[0118] Check the logic of the initial report;

[0119] The initial report is checked for syntax.

[0120] Specifically, the content review agent of this social situation report reviews the report and uses the big model to check and polish the content. LLM should not be required to output structured modification point information, including: accuracy check, check whether the data and conclusions are correct. Logical check, ensure that the report structure is clear and the argument is strong. Language quality check, correct grammatical errors, and optimize expression. Modify the report and let Ollama fix it one by one according to the output modification point information.

[0121] Final report generation and delivery, generate PDF report, convert report to PDF format, convert Markdown or LaTeX to PDF. Report delivery is when the report content review agent sends the final PDF report to the coordination agent, and finally the coordination agent delivers the report to the user.

[0122] The remote collaborative data analysis method based on large model multi-agent provided in this application has the following beneficial effects:

[0123] This application uses data discovery agents deployed in each remote node to automatically scan and report data resources in the local database. This process is fully automated, greatly improving the efficiency and real-time performance of remote data discovery. These agents can evaluate data quality, screen and integrate useful data in advance, and reduce the burden on central processing nodes.

[0124] Data is transmitted between agents through predefined communication protocols, ensuring the security and integrity of data between remote nodes while improving the response speed and accuracy of data processing.

[0125] It optimizes the processing flow of distributed data, improves the speed and quality of data integration through remote automated data discovery and preliminary evaluation, and effectively supports complex and dynamic investment research data environments.

[0126] This application realizes dynamic collaboration of agents in multiple locations. The coordination agent can monitor the status of data discovery agents and analysis agents in various locations in real time, and dynamically adjust task allocation and resource utilization to adapt to the challenges and needs of remote work.

[0127] The coordinated intelligent agent also has the ability to handle complex data analysis tasks across time zones, providing strong support for global corporate investment research.

[0128] Through flexible remote multi-agent collaboration, it is possible to quickly respond to data changes and analysis needs, improve the efficiency and effectiveness of task management, and ensure data security and processing efficiency in remote collaboration.

[0129] This application also introduces a data discovery agent, which has the functions of automatically scanning local databases, extracting data metadata, and evaluating data quality. This allows the agent to dynamically discover and collect data according to task requirements, solving the problem of discovering data scattered in multiple locations. By collaborating with the coordination agent, the data discovery agent can achieve real-time integration of data resources.

[0130] In addition, this application significantly reduces the need for manual intervention and improves the speed and quality of data integration through intelligent data discovery and collection. Compared with existing technologies, it can provide basic data support for investment research and analysis more quickly and accurately, especially for scenarios with variable data types and frequent updates.

[0131] In this application, a coordination agent is designed to dynamically allocate analysis tasks by real-time evaluation of metadata and analysis requirements reported by the data discovery agent. The coordination agent optimizes the task allocation strategy based on factors such as data quality and task priority, and can monitor and adjust in real time during task execution to ensure efficient use of resources.

[0132] Dynamic task allocation improves the system's adaptability to sudden tasks and data changes, ensuring that the analysis process is always carried out efficiently and accurately. Compared with existing technologies, it can realize dynamic adjustment of tasks during the analysis process, effectively avoiding resource conflicts and repeated analysis problems, and improving overall efficiency.

[0133] This application optimizes the communication protocol between intelligent agents. Through the combination of central coordination and regional distribution, unnecessary global communication is reduced and communication delay is reduced. At the same time, the use of regional coordination mechanism and local decision-making reduces the pressure on the central node and improves the overall communication efficiency.

[0134] In addition, the hierarchical management of data and tasks in this application allows the agent to complete most of the data processing and pre-analysis locally, and only upload key results to the coordinating agent, which greatly reduces the amount of data transmission.

[0135] By optimizing communication paths and reducing redundant data transmission, this application significantly improves the efficiency of collaboration between agents, ensuring smoother data processing and task allocation. Compared with existing technologies, the system has a faster response speed and is suitable for large-scale, multi-agent complex investment research scenarios.

[0136] This application uses data discovery agents to conduct preliminary quality assessments on data, and further confirms the integrity, accuracy, and applicability of the data through multiple data validation rules at the coordination agent. At the same time, the report generation and review agent conducts multi-level reviews of the generated reports, including checks on semantic consistency, logic, and language quality, to ensure the professionalism and accuracy of the reports.

[0137] This multi-review mechanism combines the automation capabilities of large language models and the rigor of manual rules, providing more comprehensive protection for data and analysis results.

[0138] Multi-level data quality control and review effectively improve the credibility and scientificity of the report. Compared with existing technologies, it can better adapt to high-demand investment research and analysis scenarios, ensuring that the final output report is of higher quality and professionalism.

[0139] like Figure 3 As shown, the embodiment of the present application provides a remote collaborative data analysis device based on a large model multi-agent, the device comprising:

[0140] A construction module 301 is used to construct a large model, a coordination agent, a data discovery agent, an analysis agent, a report generation agent, and a report review agent; wherein the large model, the coordination agent, the report generation agent, and the report review agent are arranged on a central server, and the data discovery agent and the analysis agent are arranged on a remote server, wherein the remote server includes a plurality of remote servers, and the plurality of remote servers are connected to the central server;

[0141] An understanding module 302 is used for the large model to receive user requirements, understand the user requirements, and generate a target report task;

[0142] Determination module 303, used for the coordination agent to determine the analysis target and required data according to the target reporting task, and send a data acquisition request to the data discovery agent;

[0143] The identification module 304 is used for the data discovery agent to obtain data meta information from the local database and send it to the coordination agent after receiving the data request. The coordination agent identifies the target data discovery agent with the required data in the data meta information and sends the request to obtain the required data to the target data discovery agent; wherein the target data discovery agent includes multiple agents, which are arranged in multiple remote servers;

[0144] An acquisition module 305 is configured to acquire the required data from a local database based on the required data request by the target data discovery agent;

[0145] Analysis module 306, used for the analysis agent to analyze the demand data and generate analysis results;

[0146] Integration module 307, used for the coordination agent to receive the analysis results of multiple remote servers to perform data integration to obtain integrated data;

[0147] A generating module 308, for the report generating agent to generate an initial report based on the integrated data;

[0148] The review module 309 is used for the report review agent to review the initial report, revise the initial report according to the review result, and obtain the target report.

[0149] The embodiment of the present application provides a remote collaborative data analysis device based on a large model and multiple agents, wherein a large model, a coordination agent, a data discovery agent, an analysis agent, a report generation agent and a report review agent are constructed through a construction module 301; wherein the large model, the coordination agent, the report generation agent and the report review agent are arranged on a central server, and the data discovery agent and the analysis agent are arranged on a remote server, and the remote server includes a plurality of remote servers connected to the central server; the large model in the understanding module 302 receives user requirements, understands the user requirements, and generates a target report task; the coordination agent in the determination module 303 determines the analysis target and the required data according to the target report task, and sends a data acquisition request to the data discovery agent; the data discovery agent in the identification module 304 obtains the data from the local database after receiving the data request. Get the data meta information and send it to the coordination agent, the coordination agent identifies the target data discovery agent with the demand data in the data meta information, and sends the request for obtaining the demand data to the target data discovery agent; wherein the target data discovery agent includes multiple agents, which are arranged in multiple remote servers; the acquisition module 305, based on the demand data request, the target data discovery agent obtains the demand data from the local database; the analysis module 306, the analysis agent analyzes the demand data and generates analysis results; the integration module 307, the coordination agent receives the analysis results of multiple remote servers, performs data integration, and obtains integrated data; the generation module 308, the report generation agent generates an initial report according to the integrated data; the review module 309, the report review agent reviews the initial report, and revises the initial report according to the review results to obtain the target report.

[0150] According to an embodiment of the present application, the present application also provides an electronic device and a readable storage medium.

[0151] The electronic device includes at least one processor; and a memory connected to the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the remote collaborative data analysis method based on large model multi-agents described in the present application. The computer instructions are used to enable the computer to execute the remote collaborative data analysis method based on large model multi-agents described in the present application.

[0152] The present application also provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the large-model multi-agent based remote collaborative data analysis method of the present application.

[0153] Figure 4 A schematic block diagram of an example electronic device 400 that can be used to implement an embodiment of the present application is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present application described and / or required herein.

[0154] like Figure 4 As shown, the device 400 includes a computing unit 401, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 402 or a computer program loaded from a storage unit 408 into a random access memory (RAM) 403. In the RAM 403, various programs and data required for the operation of the device 400 can also be stored. The computing unit 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0155] A number of components in the device 400 are connected to the I / O interface 405, including: an input unit 406, such as a keyboard, a mouse, etc.; an output unit 407, such as various types of displays, speakers, etc.; a storage unit 408, such as a disk, an optical disk, etc.; and a communication unit 409, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 409 allows the device 400 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0156] The computing unit 401 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 401 performs the various methods and processes described above, such as a remote collaborative data analysis method based on a large model multi-agent. For example, in some embodiments, a remote collaborative data analysis method based on a large model multi-agent may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 408. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 400 via the ROM 402 and / or the communication unit 409. When the computer program is loaded into the RAM 403 and executed by the computing unit 401, one or more steps of the remote collaborative data analysis method based on the large model multi-agent described above may be performed. Alternatively, in other embodiments, the computing unit 401 may be configured in any other appropriate manner (eg, by means of firmware) to execute a large-model multi-agent-based remote collaborative data analysis method.

[0157] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0158] The program code for implementing the method of the present application can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, implements the functions / operations specified in the flow chart and / or block diagram. The program code can be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0159] In the context of the present application, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0160] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0161] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0162] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0163] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A remote collaborative data analysis method based on large model multi-agent, characterized in that: The method comprises: Constructing a large model, a coordination agent, a data discovery agent, an analysis agent, a report generation agent, and a report review agent; wherein the large model, the coordination agent, the report generation agent, and the report review agent are arranged on a central server, and the data discovery agent and the analysis agent are arranged on a remote server, wherein the remote server includes a plurality of remote servers, and the plurality of remote servers are connected to the central server; The large model receives user requirements, understands the user requirements, and generates a target report task; The coordination agent determines the analysis target and required data according to the target reporting task, and sends a data acquisition request to the data discovery agent; After receiving the data request, the data discovery agent obtains data meta information from the local database and sends it to the coordination agent, and the coordination agent identifies the target data discovery agent with the required data in the data meta information, and sends a request to obtain the required data to the target data discovery agent; wherein the target data discovery agent includes multiple agents, which are arranged in multiple remote servers; Based on the demand data request, the target data discovery agent obtains the demand data from a local database; The analysis agent analyzes the demand data and generates analysis results; The coordination agent receives the analysis results from multiple remote servers and performs data integration to obtain integrated data; The report generation agent generates an initial report based on the integrated data; The report review agent reviews the initial report, revises the initial report according to the review result, and obtains the target report.

2. The method according to claim 1, characterized in that The large model is communicatively connected to the coordination agent, and the coordination agent is communicatively connected to the data discovery agent, analysis agent, and report generation agent respectively. The data discovery agent is communicatively connected to the analysis agent; the report generation agent is communicatively connected to the report review agent.

3. The method according to claim 1, characterized in that After the coordination agent receives the target report task, the method further includes: Determining whether the coordination agent communicates normally with the data discovery agents in each remote server; After confirming normal communication, a request for obtaining data is sent to the data discovery agent.

4. The method according to claim 1, characterized in that: The coordination agent identifies a target data discovery agent having required data in the data meta information, including: Identifying demand data from the data meta information; the demand data includes multiple types; Determine the data discovery agent corresponding to the identified demand data, and determine the corresponding data discovery agent as the target data discovery agent.

5. The method according to claim 4, characterized in that The step of determining a data discovery agent corresponding to the identified demand data and determining the corresponding data discovery agent as a target data discovery agent comprises: Processing the acquired demand data, matching the processed demand data with the analysis target, and obtaining matching data; The coordinating agent corresponding to the matching data is determined as the target coordinating agent.

6. The method according to claim 4, characterized in that The analysis agent analyzes the demand data and generates analysis results, including: receiving an analysis goal of the coordination agent; Processing outliers and missing values ​​in the demand data; Based on the preset prompt text requirements, multiple types of demand data are analyzed in multiple dimensions to obtain analysis results.

7. The method according to claim 6, characterized in that The coordination agent receives the analysis results from multiple remote servers and integrates the data to obtain integrated data, including: Check whether the data in the received analysis results are complete, and if complete, verify whether the data in the analysis results are within a reasonable range. If they are within a reasonable range, determine whether the analysis results meet the requirements of the analysis objectives, and save the analysis results that meet the requirements as integrated data.

8. The method according to claim 1, characterized in that The report generating agent generates an initial report based on the integrated data, including: The integrated data is received, a text portion of an initial report is generated based on the integrated data and a preset report template, and a picture position symbol is inserted into the initial report to obtain the initial report.

9. The method according to claim 1, characterized in that: The review of the initial report includes: Check the correctness of the data and conclusions of the initial report; Check the logic of the initial report; The initial report is checked for syntax.

10. A remote collaborative data analysis device based on a large model and multiple agents, characterized in that: The device comprises: A construction module is used to construct a large model, a coordination agent, a data discovery agent, an analysis agent, a report generation agent, and a report review agent; wherein the large model, the coordination agent, the report generation agent, and the report review agent are arranged on a central server, and the data discovery agent and the analysis agent are arranged on a remote server, wherein the remote server includes a plurality of remote servers, and the plurality of remote servers are connected to the central server; An understanding module, used for the large model to receive user requirements, understand the user requirements, and generate a target report task; A determination module, used for the coordination agent to determine the analysis target and required data according to the target report task, and send a data acquisition request to the data discovery agent; an identification module, for the data discovery agent to obtain data meta information from a local database and send it to the coordination agent after receiving the data request, the coordination agent to identify the target data discovery agent with the required data in the data meta information, and send a request to obtain the required data to the target data discovery agent; wherein the target data discovery agent includes a plurality of agents, which are arranged in a plurality of remote servers; An acquisition module, configured to enable the target data discovery agent to acquire the required data from a local database based on the required data request; An analysis module, used for the analysis agent to analyze the demand data and generate analysis results; An integration module, used for the coordination agent to receive the analysis results of multiple remote servers to integrate data and obtain integrated data; A generating module, used for the report generating agent to generate an initial report based on the integrated data; The review module is used for the report review agent to review the initial report, revise the initial report according to the review result, and obtain the target report.

Images