Multi-agent cooperative data processing method and device, storage medium and electronic equipment

Abstract

The application provides a multi-agent cooperative data processing method and device, a storage medium and an electronic device. The method is applied to the field of computer technology and includes the following steps: a dialogue agent acquires target demand text input by a user, inputs the target demand text into a large language model, and acquires target structured data output by the large language model; the dialogue agent transmits the target structured data to a processing agent; the processing agent writes all structured data in the form of parallel stream data into a cache in real time, wherein the all structured data includes the target structured data; a control component acquires the target structured data from the cache in a polling calling mode, generates snapshot expression data corresponding to the target structured data, and displays the snapshot expression data corresponding to the target demand text in a front-end page.

Description

Technical Field

[0001] This application relates to the field of computer technology, and more specifically, to a multi-agent collaborative data processing method, apparatus, storage medium, and electronic device in the field of computer technology. Background Technology

[0002] With the development of internet and artificial intelligence technologies, models have been applied to all aspects of people's daily lives. For example, Large Language Models (LLMs) can engage in question-and-answer sessions to solve everyday problems. Using LLMs to improve efficiency and solve complex problems has become a trend. However, LLMs themselves have problems such as illusions, lack of real-time performance and knowledge update delays, and limitations on context length. Also, because users may have domain-specific expertise, they may not be able to accurately describe their needs, or the LLM's answers may be difficult for users to understand quickly. However, if users re-enter their needs, they need to wait for the LLM to regenerate the answer, resulting in a long wait and wasting time and resources, which is inconvenient for users. Therefore, there is a need to provide a more convenient method for processing large model data. Summary of the Invention

[0003] This application provides a multi-agent collaborative data processing method, apparatus, storage medium, and electronic device. This method enables users to continue inputting required text during the data display process through the collaborative work of dialogue agents and processing agents, reducing the waiting time when users input multiple required texts into a large language model and avoiding waste of time resources.

[0004] In a first aspect, embodiments of this application provide a multi-agent collaborative data processing method, applied to a multi-agent collaborative system, the multi-agent collaborative system including a dialogue agent, a processing agent, and a control component, the method comprising:

[0005] The dialogue agent obtains the target requirement text input by the user, inputs the target requirement text into the large language model, and obtains the target structured data output by the large language model.

[0006] The dialogue agent transmits the target structured data to the processing agent;

[0007] The processing agent writes all structured data into the buffer in real time as parallel streaming data, and all structured data includes the target structured data.

[0008] The control component uses a polling method to retrieve the target structured data from the cache and generate snapshot representation data corresponding to the target structured data.

[0009] The control component displays the snapshot expression data on the front-end page in response to the target requirement text.

[0010] The above technical solution enables users to continue inputting their desired text while the data is being displayed, through the collaborative work of the dialogue agent and the processing agent. This reduces the waiting time when users input multiple desired texts into a large language model and avoids wasting time resources.

[0011] In conjunction with the first aspect, in some possible implementations, the dialogue agent transmits the target structured data to the data processing agent, including:

[0012] If the dialogue agent detects that the target structured data does not meet the preset transmission conditions, it will improve the prompt information for the user's display needs on the front-end page and continue to execute the step of obtaining the target demand text entered by the user.

[0013] If the dialogue agent detects that the target structured data meets the preset transmission conditions, it transmits the target structured data to the processing agent.

[0014] The above technical solution can prevent structured data that does not meet the preset transmission conditions from being transmitted and displayed incorrectly, reduce the occurrence of errors in large model responses, and improve the efficiency of users in using large language models.

[0015] In conjunction with the first aspect, in some possible implementations, before the dialogue agent transmits the target structured data to the data processing agent, it further includes:

[0016] If the dialogue agent detects a return question in the structured data, it generates a request improvement prompt based on the return question.

[0017] The above technical solution can display the returned question to the user, allowing the user to improve the request text in a timely manner, and further improve the response efficiency of the large language model.

[0018] In conjunction with the first aspect, in some possible implementations, the processing agent includes an intermediate storage component and a data processing component;

[0019] The processing agent writes all structured data into a buffer in real time as parallel streaming data, including:

[0020] The intermediate storage component stores the target structured data;

[0021] The intermediate storage component transmits all structured data to the data processing component;

[0022] The data processing component writes all the structured data into the cache in real time as parallel streaming data.

[0023] The above technical solutions enable the intermediate storage component to act as a data stream buffer, allowing the data processing component to ensure the stability and uninterrupted transmission of the data stream.

[0024] In conjunction with the first aspect, in some possible implementations, before the dialogue agent obtains the target request text input by the user, it further includes:

[0025] The dialogue agent obtains the current request text input by the user and obtains the historical request text previously input by the user.

[0026] The dialogue agent integrates the current request text and the historical request text to obtain the target request text.

[0027] The above technical solution can integrate the user's previously input request text, further improve the user's input question description, improve the accuracy of the large language model's response, and ensure the efficiency of the large language model's response.

[0028] In conjunction with the first aspect, in some possible implementations, the dialogue agent acquires the user's current request text, including:

[0029] The dialogue agent obtains the improvement request text input by the user based on the requirement improvement prompt information, and confirms the improvement request text as the current requirement text.

[0030] In conjunction with the first aspect, in some possible implementations, the method further includes:

[0031] If the control component receives a confirmation instruction from the user regarding the snapshot expression data, it will persistently display the snapshot expression data for the target requirement text.

[0032] If the control component receives a denial instruction from the user regarding the snapshot expression data, it then obtains the modification request text from the denial instruction.

[0033] The control component transmits the modification request text to the dialogue agent;

[0034] The dialogue agent acquires the user's current request text, including:

[0035] The dialogue agent confirms the modified request text as the current request text.

[0036] The above technical solution enables users to promptly input modification requests when they receive responses that do not meet their expectations, instructing the large language model to respond again, thus improving the communication efficiency between users and the large language model.

[0037] Secondly, embodiments of this application provide a multi-agent collaborative data processing device, the device including a dialogue agent, a processing agent, and a control component;

[0038] The dialogue agent is used to obtain the target demand text input by the user, input the target demand text into the large language model, and obtain the target structured data output by the large language model.

[0039] The target structured data is transmitted to the processing agent;

[0040] The processing agent is used to write all structured data into a buffer in real time in the form of parallel streaming data, wherein all structured data includes the target structured data.

[0041] The control component is used to retrieve the target structured data from the cache using a polling method, and generate snapshot representation data corresponding to the target structured data;

[0042] On the front-end page, the snapshot expression data is displayed in response to the target requirement text.

[0043] Thirdly, embodiments of this application provide a computer storage medium storing a plurality of instructions adapted for loading by a processor and executing the above-described method steps.

[0044] Fourthly, embodiments of this application provide a computer program product that stores multiple instructions adapted for loading by a processor and executing the above-described method steps.

[0045] Fifthly, embodiments of this application provide an electronic device that may include: a processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and to execute the above-described method steps.

[0046] In one or more embodiments of this application, the dialogue agent acquires the target requirement text input by the user, inputs the target requirement text into a large language model, and acquires the target structured data output by the large language model. The dialogue agent transmits the target structured data to the processing agent, which writes all structured data into a buffer in real time as parallel streaming data. All structured data includes the target structured data. The control component uses a polling method to retrieve the target structured data from the buffer and generates snapshot expression data corresponding to the target structured data. The control component displays the snapshot expression data for the target requirement text on the front-end page. Through the collaborative work of the dialogue agent and the processing agent, the user can continue to input requirement text during data display, reducing the waiting time when the user inputs multiple requirement texts into the large language model and avoiding waste of time resources. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a system architecture diagram of a multi-agent cooperative system provided in an embodiment of this application;

[0049] Figure 2 This is a flowchart illustrating a multi-agent collaborative data processing method provided in an embodiment of this application;

[0050] Figure 3 This is a flowchart illustrating a multi-agent collaborative data processing method provided in an embodiment of this application;

[0051] Figure 4 This is a schematic diagram illustrating an example of structured data writing provided in an embodiment of this application;

[0052] Figure 5 This is a schematic diagram of the structure of a multi-agent collaborative data processing device provided in an embodiment of this application;

[0053] Figure 6 This is a schematic diagram of the structure of a processing intelligent agent provided in an embodiment of this application;

[0054] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0055] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0056] With the development of internet and artificial intelligence technologies, people have become accustomed to using various models to solve problems encountered in life and work. For example, large language models possess semantic understanding capabilities, enabling them to understand user input and output answers accordingly. However, due to varying levels of expertise among users in different professional fields, users' descriptions of problems within those fields may be inaccurate, and the large language model's answers within those fields may not be well understood by users. Consequently, users and the large language model inevitably need to refine their questions and modify their answers multiple times. However, in existing technologies, large language models can only process one question from the user at a time. While answering the text, users cannot input other questions or adjust previously submitted questions; they can only wait for the large language model to output its answer before continuing to ask questions, resulting in unnecessary waste of time and resources. This application provides a multi-agent collaborative data processing device that can be applied to multi-agent collaborative systems. Through multi-agent collaborative work, users can continue to input their desired text while the large language model is displaying its answer, thus avoiding wasted time and resources.

[0057] The multi-agent cooperative data processing method provided in this application can be implemented using a computer program and can run on a multi-agent cooperative data processing device based on the von Neumann architecture. This computer program can be integrated into an application or run as a standalone utility application. Please also refer to... Figure 1This application provides a system architecture diagram for a multi-agent collaborative system. The multi-agent collaborative system may include a dialogue agent, a processing agent, and a control component, and may also include a large language model and a buffer. Both the dialogue agent and the processing agent are agents. An agent is a computational entity or system capable of perceiving its environment, making autonomous decisions, and executing specific actions to achieve a specific goal. It is not merely a program passively responding to instructions, but an autonomous unit with goal orientation, proactive adaptability, and tool-using capabilities, capable of completing complex tasks such as multi-step reasoning and execution. Dialogue agents can be applied to the question-and-answer process between users and large language models. For example, a dialogue agent can obtain the user's input request text on the front-end page, input the request text into the large language model, and obtain the structured data returned by the large language model. A processing agent writes the structured data obtained by the dialogue agent into a cache, and a control component retrieves the structured data from the cache, generates snapshot representation data based on the structured data, and displays it on the front-end page. This allows users to see the large language model's response to the request text on the front-end page. Because the dialogue agent and the processing agent work together, the process of inputting the request text and receiving the large language model's response is decoupled from the process of displaying the large language model's response. This allows users to continue inputting request text on the front-end page while the response is being displayed, reducing the waiting time when users input multiple request texts into the large language model and avoiding wasted time resources.

[0058] Among them, the multi-agent collaborative data processing device can construct a domain knowledge base, enabling the large language model to generate structured data based on the requirement text. The domain knowledge base can include form domain fields, domain prompt words, etc., and can be stored in the form of documents, tables, etc. The multi-agent collaborative data processing device can obtain past transactions in different professional fields. Domain fields are the characteristics and attributes of these transactions, and domain prompt words are the prompt words (Prompt) in the professional field. Prompt is the instruction or guidance information that the user inputs to the large language model, which is used to trigger the model to generate specific content, perform tasks, or adjust the output style. It can guide the large language model to understand the intent and control the generated results through language description. The domain knowledge base allows users to obtain relatively complete requirement results through simple input.

[0059] The multi-agent cooperative data processing method provided in this application will be described in detail below with reference to specific embodiments.

[0060] Please see Figure 2 This is a flowchart illustrating a multi-agent collaborative data processing method provided in an embodiment of this application. Figure 1As shown, the method described in this application embodiment may include the following steps S102-S110.

[0061] S102, the dialogue agent obtains the target requirement text input by the user, inputs the target requirement text into the large language model, and obtains the target structured data output by the large language model.

[0062] Specifically, the multi-agent collaborative data processing device can display an input window on the front-end page for users to input their desired text. This desired text is the text the user wants to input into the large language model for understanding and response. The dialogue agent can obtain the target desired text input by the user from the front-end page. This target desired text is a text containing the user's input. The dialogue agent can then input this target desired text into the large language model. The large language model can understand and learn from the target desired text and respond, returning target structured data to the dialogue agent. This target structured data is structured data generated by the large language model based on the target desired text. The structured data is in a preset format, which can be specified by the user or preset by the multi-agent collaborative data processing device, such as JSON format. This preset formatted structured data can be quickly transmitted between components such as the dialogue agent, processing agent, and buffer, and can be parsed by the front-end page to display the large language model's response in text or table format that the user can understand.

[0063] S104, the dialogue agent transmits the target structured data to the processing agent.

[0064] Specifically, after the dialogue agent obtains the target structured data returned by the large language model, it can transmit the target structured data to the processing agent. The user may input multiple request texts, and the large language model will return multiple sets of structured data. Therefore, once the dialogue agent obtains the structured data returned by the large language model, it can transmit it to the processing agent.

[0065] S106, the processing agent writes all structured data into the buffer in real time as parallel streaming data, and all structured data includes the target structured data.

[0066] Specifically, the processing agent can write all the structured data obtained from the dialogue agent into the buffer in real time as parallel stream data. That is, multiple structured data streams can be written into the buffer in real time in a concurrent thread manner. All the structured data is the same as all the structured data received by the processing agent, including the target structured data.

[0067] S108, the control component uses a polling method to obtain the target structured data from the cache and generate snapshot representation data corresponding to the target structured data.

[0068] Specifically, the control component can use a polling method to retrieve structured data from the cache, including the target structured data. Then, it generates corresponding snapshot representation data based on the target structured data. The snapshot representation data is a lightweight view data package that can be directly rendered without recalculation. After the front-end page obtains the snapshot representation data, it can generate page content without recalculation or request.

[0069] S110, the control component in the front-end page, displays snapshot data to express the target requirement text.

[0070] Specifically, the control component allows the front-end page to display snapshot representation data based on the target requirement text. When displaying the snapshot representation data, the front-end page can directly restore it to the page content, such as restoring it to text content that the user can understand. In particular, when displaying the snapshot representation data generated based on the target structured data, it can reference the target requirement text previously entered by the user, thus establishing a correspondence between the snapshot representation data and the target requirement text on the front-end page. This prevents users from being unable to distinguish the corresponding text content when entering multiple requirement texts.

[0071] In this embodiment, the dialogue agent acquires the target requirement text input by the user, inputs the target requirement text into a large language model, and obtains the target structured data output by the large language model. The dialogue agent transmits the target structured data to the processing agent, which writes all structured data into a buffer in real time as parallel streaming data. All structured data includes the target structured data. The control component uses a polling method to retrieve the target structured data from the buffer and generates snapshot expression data corresponding to the target structured data. The control component displays the snapshot expression data for the target requirement text on the front-end page. Through the collaborative work of the dialogue agent and the processing agent, the user can continue to input requirement text during the data display process, reducing the waiting time when the user inputs multiple requirement texts into the large language model and avoiding waste of time resources.

[0072] Please see Figure 3 This is a flowchart illustrating a multi-agent collaborative data processing method provided in an embodiment of this application. Figure 3 As shown, the method described in this application embodiment may include the following steps S202-S224.

[0073] S202, the dialogue agent obtains the current request text input by the user, and obtains the historical request text input by the user previously.

[0074] Specifically, the intelligent agent collaborative data processing device can display an input window on the front-end page for users to input their request text. This request text is the text the user wants to input into the large language model for understanding and response. The dialogue agent can obtain the user's current request text from the front-end page. This current request text is the most recently entered request text by the user on the front-end page. It's understandable that the user might input multiple request texts for the same question, allowing for modification and refinement of the question description. To further refine the user's description, the dialogue agent can also retrieve the user's historical request texts previously entered on the front-end page.

[0075] Optionally, the front-end page displayed by the intelligent agent collaborative data processing device can be a dialogue window page. When a user consults the large language model about a question in the same domain, they will enter the request text multiple times in one dialogue window page. When they need to consult about questions in other domains, they can open a new dialogue window page to enter the request text, thus ensuring the consistency of the text content of the large language model's response. Therefore, the dialogue agent can obtain the user's previously entered historical request text in the dialogue window page corresponding to the current request text.

[0076] S204, the dialogue agent integrates the current request text and the historical request text to obtain the target request text.

[0077] Specifically, the dialogue agent can integrate the current request text and the historical request text to obtain the target request text. The integration process can be a splicing process, which combines the current request text and the historical request text into a single text.

[0078] S206, the dialogue agent obtains the target requirement text input by the user, inputs the target requirement text into the large language model, and obtains the target structured data output by the large language model.

[0079] Specifically, the dialogue agent can input the target request text into the large language model, which can understand and learn the target request text and respond, returning the target structured data to the dialogue agent. The target structured data is the structured data generated by the large language model for the target request text. The structured data is data in a preset format, which can be specified by the user or preset by the multi-agent collaborative data processing device, such as JSON format. The preset format structured data can be quickly transmitted between components such as the dialogue agent, the processing agent, and the cache, and can be parsed by the front-end page to display the large language model's response in text or table form that the user can understand.

[0080] S208, the dialogue agent determines whether the target structured data meets the preset transmission conditions.

[0081] Specifically, after receiving the target structured data, the dialogue agent does not directly transmit it to the processing agent. Instead, it first determines whether the target structured data meets preset transmission conditions. These preset transmission conditions can be the initial settings of the multi-agent collaborative data processing device or can be set by the user. They can be used to determine whether the target structured data can be transmitted. For example, if the data format or content of the target structured data is incorrect, it can be confirmed that it does not meet the preset transmission conditions. If the target structured data does not meet the preset transmission conditions, step S210 is executed; if the target structured data meets the preset transmission conditions, step S212 is executed.

[0082] Optionally, when the target requirement text entered by the user cannot accurately describe the question the user wants to ask, and the large language model cannot respond based on the target requirement text, the large language model can send a return question to the dialogue agent. The return question can be information used by the large language model to prompt the user to improve or clarify the question. If the dialogue agent detects that there is a return question in the structured data, it can generate requirement improvement prompts based on the return question. That is, the requirement improvement prompts can include the return question.

[0083] Optionally, when the dialogue agent detects that the target structured data does not conform to the preset format, it can be confirmed that the target structured data does not meet the preset transmission conditions. It is understandable that structured data that does not conform to the preset format is difficult to quickly parse into snapshot representation data, difficult to display quickly on the front-end page, or may result in display errors. Based on the preset format, a request improvement prompt can be generated to instruct the large language model to generate structured data according to the preset format when the user re-enters the request text. Alternatively, the dialogue agent can generate prompt words based on the preset format, and input the prompt words and the target request text back into the large language model, allowing the large language model to regenerate the target structured data.

[0084] S210, the dialogue agent improves the prompts on the front-end page according to the user's display needs.

[0085] Specifically, the dialogue agent can display a request improvement prompt on the front-end page in response to the user's input of the target request text. The request improvement prompt is used to remind the user to further improve and clarify the question they want to ask. After seeing the request improvement prompt, the user can re-enter the request text on the front-end page.

[0086] The dialogue agent can obtain the improved requirement text input by the user based on the requirement improvement prompt information, then confirm the improved requirement text as the current requirement text, and continue to execute step S202.

[0087] S212, the dialogue agent transmits the target structured data to the processing agent.

[0088] Specifically, if the target structured data meets the preset transmission conditions, the dialogue agent can transmit the target structured data to the processing agent. Users may input multiple request texts, and the large language model will return multiple sets of structured data. Therefore, as long as the dialogue agent obtains structured data that meets the preset transmission conditions, it can transmit it to the processing agent.

[0089] S214, the processing agent writes all structured data into the buffer in real time as parallel streaming data, and all structured data includes the target structured data.

[0090] Specifically, the processing agent can write all the structured data obtained from the dialogue agent into the buffer in real time as parallel stream data. That is, multiple structured data streams can be written into the buffer in real time in a concurrent thread manner. All the structured data is the same as all the structured data received by the processing agent, including the target structured data.

[0091] Optional, please see also Figure 4 This embodiment of the application provides an example of writing structured data. The processing agent may include an intermediate storage component and a data processing component. The intermediate storage component can receive and store the structured data transmitted by the dialogue agent. For example, it can cache the received structured data in the form of a message queue and transmit all the stored structured data to the data processing component in a first-in-first-out manner. The data processing component can write all the structured data into the cache in real time in the form of parallel streaming data. Optionally, the data processing component can also perform CRUD operations on the structured data. The intermediate storage component can act as a data stream buffer, so that the data processing component can ensure the stability and uninterrupted transmission of the data stream.

[0092] S216, the control component uses a polling method to obtain the target structured data from the cache and generate snapshot representation data corresponding to the target structured data.

[0093] Specifically, the control component can use a polling method to retrieve structured data from the cache, including the target structured data. Then, it generates corresponding snapshot representation data based on the target structured data. The snapshot representation data is a lightweight view data package that can be directly rendered without recalculation. After the front-end page obtains the snapshot representation data, it can generate page content without recalculation or request.

[0094] S218, the control component in the front-end page displays snapshot data to meet the target requirements of the text.

[0095] Specifically, the control component allows the front-end page to display snapshot representation data based on the target requirement text. When displaying the snapshot representation data, the front-end page can directly restore it to the page content, such as restoring it to text content that the user can understand. In particular, when displaying the snapshot representation data generated based on the target structured data, it can reference the target requirement text previously entered by the user, thus establishing a correspondence between the snapshot representation data and the target requirement text on the front-end page. This prevents users from being unable to distinguish the corresponding text content when entering multiple requirement texts.

[0096] S220, the control component determines whether it has received a confirmation instruction from the user regarding the snapshot data.

[0097] Specifically, after the snapshot expression data is displayed on the front-end page, feedback from the user regarding the text content can be awaited. If the user feels the text content matches their initially entered target text, they can send a confirmation command to the multi-agent collaborative data processing device via the front-end page. If the user feels the text content does not match their initially entered target text and does not resolve their issue, they can send a denial command to the multi-agent collaborative data processing device via the front-end page. If the control component confirms that it has received a denial command instead of a confirmation command from the user regarding the snapshot expression data, it continues with step S222. If the control component confirms that it has received a confirmation command from the user regarding the snapshot expression data, it proceeds to step S224.

[0098] Optionally, while displaying snapshot data on the front-end page, the control component can also display a confirmation button and a denial button. When the user clicks the confirmation button, a confirmation command is sent; when the user clicks the denial button, a denial command is sent.

[0099] S222, the control component obtains the modification request text in the denial instruction, and the control component transmits the modification request text to the dialogue agent.

[0100] Specifically, when a user sends a denial command, they can also input a modification request text within the denial command. The modification request text is used to instruct the large language model to improve and modify the response. The control component can transmit the modification request text to the dialogue agent, which can confirm the modification request text as the current request text and continue to execute step S202.

[0101] S224, The control component persists the snapshot data to the target requirement text for display.

[0102] Specifically, if a user confirms the snapshot data, the control component can persistently display the snapshot data for the target text requirement. For example, the snapshot data can be written to local storage, cloud storage, or a database, and a display format that can be directly rendered without secondary calculation can be provided, so that the front-end page can directly restore and display the complete text content at any time for the user to view.

[0103] In this embodiment, the dialogue agent acquires the current requirement text input by the user and the historical requirement text previously input by the user. The current requirement text and the historical requirement text are integrated and processed to obtain the target requirement text, thereby further completing the user's problem description and improving the accuracy of the large language model's response. The dialogue agent acquires the target requirement text input by the user, inputs the target requirement text into the large language model, and acquires the target structured data output by the large language model. If the target structured data does not meet the preset transmission conditions, a requirement improvement prompt is displayed to the user on the front-end page. The agent acquires the improved requirement text input by the user based on the requirement improvement prompt and then confirms the improved requirement text as the current requirement text, so that the user can improve the requirement text in a timely manner, further improving the response efficiency of the large language model. If the target structured data meets the preset transmission conditions, the target structured data is transmitted to the processing agent. The processing agent writes all structured data into a buffer in real time in the form of parallel streaming data. All structured data includes the target structured data. The processing agent may include an intermediate storage component and a data processing component. The intermediate storage component can act as a data stream buffer, so that the data processing component can ensure the stability and uninterrupted transmission of the data stream. The control component uses a polling approach to retrieve the target structured data from the cache and generate corresponding snapshot representation data. On the front-end page, the control component displays this snapshot representation data for the target request text. Through the collaborative work of the dialogue agent and the processing agent, users can continue to input their request text while the data is being displayed, reducing the waiting time when users input multiple request texts into the large language model and avoiding wasted time resources. Furthermore, if the control component confirms receipt of a user's confirmation instruction for the snapshot representation data, it can persistently display the snapshot representation data for the target request text, facilitating direct restoration of this content on the front-end page and allowing users to view it at any time. If the control component receives a denial instruction, it retrieves the modified request text from the denial instruction and transmits the modified request text to the dialogue agent, allowing users to promptly input modified request text when receiving an unexpected response, instructing the large language model to respond again.

[0104] The following will be combined with the appendix Figure 5 - Appendix Figure 6This application provides a detailed description of the multi-agent collaborative data processing device provided in its embodiments. It should be noted that the appendix... Figure 5 - Appendix Figure 6 The multi-agent collaborative data processing device in the present application is used to execute the data. Figures 1-4 The methods shown in the embodiments are illustrated for ease of explanation, showing only the parts relevant to the embodiments of this application. For specific technical details not disclosed, please refer to this application. Figures 1-4 The example shown.

[0105] Please see Figure 5 This illustration shows a schematic diagram of a multi-agent collaborative data processing device provided in an exemplary embodiment of this application. This multi-agent collaborative data processing device can be implemented as all or part of a device through software, hardware, or a combination of both. The device 1 includes a dialogue agent 11, a processing agent 12, and a control component 13.

[0106] The dialogue agent 11 is used to obtain the target demand text input by the user, input the target demand text into the large language model, and obtain the target structured data output by the large language model.

[0107] The target structured data is transmitted to the processing agent;

[0108] The processing agent 12 is used to write all structured data into the buffer in real time in the form of parallel streaming data, and the all structured data includes the target structured data.

[0109] The control component 13 is used to obtain the target structured data from the cache using a polling method, and generate snapshot representation data corresponding to the target structured data;

[0110] On the front-end page, the snapshot expression data is displayed in response to the target requirement text.

[0111] Optionally, the dialogue agent 11 is specifically used to, if it is detected that the target structured data does not meet the preset transmission conditions, display the user's display needs improvement prompt information on the front-end page, and continue to execute the step of obtaining the target needs text input by the user.

[0112] If the target structured data is detected to meet the preset transmission conditions, the target structured data is transmitted to the processing agent.

[0113] Optionally, the dialogue agent 11 is further configured to generate a request improvement prompt based on the return question if a return question is detected in the structured data.

[0114] Optional, please see also Figure 6This provides a schematic diagram of the structure of a processing intelligent agent according to an embodiment of this application. Figure 6 As shown, the processing agent 12 may include: an intermediate storage component 121 and a data processing component 122;

[0115] The intermediate storage component 121 is used to store the target structured data and transfer all structured data to the data processing component.

[0116] The data processing component 122 is used to write all the structured data into the buffer in real time in the form of parallel streaming data.

[0117] Optionally, the dialogue can also be used to obtain the current request text input by the user and to obtain the historical request text previously input by the user.

[0118] The current requirement text and the historical requirement text are integrated to obtain the target requirement text.

[0119] Optionally, the dialogue agent 11 is specifically used to obtain the improvement request text input by the user based on the requirement improvement prompt information, and confirm the improvement request text as the current requirement text.

[0120] Optionally, the control component 13 is further configured to, if a confirmation instruction from the user regarding the snapshot expression data is received, persistently display the snapshot expression data for the target requirement text;

[0121] If a denial instruction from the user regarding the snapshot data is obtained, then the modification request text in the denial instruction is obtained;

[0122] The modification request text is transmitted to the dialogue agent;

[0123] Optionally, the dialogue agent 11 is specifically used to confirm the modified request text as the current request text.

[0124] In this embodiment, the dialogue agent acquires the current requirement text input by the user and the historical requirement text previously input by the user. The current requirement text and the historical requirement text are integrated and processed to obtain the target requirement text, thereby further completing the user's problem description and improving the accuracy of the large language model's response. The dialogue agent acquires the target requirement text input by the user and inputs it into the large language model, then acquires the target structured data output by the large language model. If the target structured data does not meet the preset transmission conditions, a requirement improvement prompt is displayed to the user on the front-end page. The agent acquires the improved requirement text input by the user based on the requirement improvement prompt and then confirms the improved requirement text as the current requirement text, allowing the user to improve the requirement text in a timely manner, further improving the response efficiency of the large language model. If the target structured data meets the preset transmission conditions, the target structured data is transmitted to the processing agent. The processing agent writes all structured data into a buffer in real time in the form of parallel streaming data. All structured data includes the target structured data. The processing agent may include an intermediate storage component and a data processing component. The intermediate storage component acts as a data stream buffer, enabling the data processing component to ensure the stability and uninterrupted transmission of the data stream. The control component uses a polling approach to retrieve the target structured data from the cache and generate corresponding snapshot representation data. On the front-end page, the control component displays this snapshot representation data for the target request text. Through the collaborative work of the dialogue agent and the processing agent, users can continue to input their request text while the data is being displayed, reducing the waiting time when users input multiple request texts into the large language model and avoiding wasted time resources. Furthermore, if the control component confirms receipt of a user's confirmation instruction for the snapshot representation data, it can persistently display the snapshot representation data for the target request text, facilitating direct restoration of this content on the front-end page and allowing users to view it at any time. If the control component receives a denial instruction, it retrieves the modified request text from the denial instruction and transmits the modified request text to the dialogue agent, allowing users to promptly input modified request text when receiving an unexpected response, instructing the large language model to respond again.

[0125] It should be noted that the multi-agent collaborative data processing device provided in the above embodiments is only illustrated by the division of the above functional modules when executing the multi-agent collaborative data processing method. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the multi-agent collaborative data processing device and the multi-agent collaborative data processing method embodiments provided in the above embodiments belong to the same concept, and the implementation process is detailed in the method embodiments, which will not be repeated here.

[0126] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0127] This application also provides a computer storage medium that can store multiple instructions, which are adapted to be loaded and executed by a processor as described above. Figures 1-4 The multi-agent cooperative data processing method described in the illustrated embodiment can be found in the following documentation for its specific execution process. Figures 1-4 The specific details of the illustrated embodiments will not be elaborated here.

[0128] This application also provides a computer program product storing at least one instruction, which is loaded and executed by the processor as described above. Figures 1-4 The multi-agent cooperative data processing method described in the illustrated embodiment can be found in the following documentation for its specific execution process. Figures 1-4 The specific details of the illustrated embodiments will not be elaborated here.

[0129] Please refer to Figure 7 This diagram illustrates a structural block diagram of an electronic device provided in an exemplary embodiment of this application. The electronic device in this application may include one or more components such as a processor 110, a memory 120, an input device 130, an output device 140, and a bus 150. The processor 110, memory 120, input device 130, and output device 140 may be connected via the bus 150.

[0130] Processor 110 may include one or more processing cores. Processor 110 connects to various parts of the electronic device using various interfaces and lines, and executes various functions of terminal 100 and processes data by running or executing instructions, programs, code sets, or instruction sets stored in memory 120, and by calling data stored in memory 120. Optionally, processor 110 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). Processor 110 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user page, and applications; the GPU is responsible for rendering and drawing the displayed content; and the modem handles wireless communication. It is understood that the modem may also not be integrated into processor 110 and may be implemented separately using a communication chip.

[0131] The memory 120 may include random access memory (RAM) or read-only memory (ROM). Optionally, the memory 120 may include non-transitory computer-readable storage medium. The memory 120 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 120 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for implementing at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the various method embodiments described above, etc. The operating system may be the Android system, including systems deeply developed based on the Android system, the iOS system developed by Apple Inc., including systems deeply developed based on the iOS system, or other systems.

[0132] The memory 120 can be divided into operating system space and user space. The operating system runs in the operating system space, while native and third-party applications run in user space. To ensure that different third-party applications can achieve good running performance, the operating system allocates corresponding system resources for each application. However, different application scenarios within the same third-party application have different requirements for system resources. For example, in local resource loading scenarios, third-party applications have high requirements for disk read speed; in animation rendering scenarios, third-party applications have high requirements for GPU performance. Since the operating system and third-party applications are independent of each other, the operating system often cannot promptly perceive the current application scenario of a third-party application, resulting in the operating system's inability to adapt system resources accordingly.

[0133] In order for the operating system to distinguish the specific application scenarios of third-party applications, it is necessary to establish data communication between the third-party applications and the operating system. This would allow the operating system to obtain the current scenario information of the third-party applications at any time, and then perform targeted system resource adaptation based on the current scenario.

[0134] The input device 130 is used to receive input instructions or data, and includes, but is not limited to, a keyboard, mouse, camera, microphone, or touch device. The output device 140 is used to output instructions or data, and includes, but is not limited to, a display device and a speaker. In one example, the input device 130 and the output device 140 can be combined, and the input device 130 and the output device 140 can be a touch display screen.

[0135] The touch display screen can be designed as a full-screen, curved screen, or irregularly shaped screen. It can also be designed as a combination of a full-screen and a curved screen, or a combination of an irregularly shaped screen and a curved screen; however, this application does not limit the specific design in this regard.

[0136] In addition, those skilled in the art will understand that the structure of the electronic device shown in the above figures does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown, or combine certain components, or have different component arrangements. For example, the electronic device may also include radio frequency circuits, input units, sensors, audio circuits, Wireless Fidelity (WiFi) modules, power supplies, Bluetooth modules, etc., which will not be described in detail here.

[0137] exist Figure 7 In the illustrated electronic device, the processor 110 can be used to call the multi-agent collaborative data processing application stored in the memory 120, and specifically perform the following operations:

[0138] The dialogue agent obtains the target requirement text input by the user, inputs the target requirement text into the large language model, and obtains the target structured data output by the large language model.

[0139] The dialogue agent transmits the target structured data to the processing agent;

[0140] The processing agent writes all structured data into the buffer in real time as parallel streaming data, and all structured data includes the target structured data.

[0141] The control component uses a polling method to retrieve the target structured data from the cache and generate snapshot representation data corresponding to the target structured data.

[0142] The control component displays the snapshot expression data on the front-end page in response to the target requirement text.

[0143] In one embodiment, when the processor 110 executes the dialogue agent to transmit the target structured data to the data processing agent, it specifically performs the following operations:

[0144] If the dialogue agent detects that the target structured data does not meet the preset transmission conditions, it will improve the prompt information for the user's display needs on the front-end page and continue to execute the step of obtaining the target demand text entered by the user.

[0145] If the dialogue agent detects that the target structured data meets the preset transmission conditions, it transmits the target structured data to the processing agent.

[0146] In one embodiment, before the dialogue agent transmits the target structured data to the data processing agent, the processor 110 also performs the following operations:

[0147] If the dialogue agent detects a return question in the structured data, it generates a request improvement prompt based on the return question.

[0148] In one embodiment, the processing agent includes an intermediate storage component and a data processing component;

[0149] When the processor 110 executes the processing agent to write all structured data into the buffer in real time as parallel streaming data, it specifically performs the following operations:

[0150] The intermediate storage component stores the target structured data;

[0151] The intermediate storage component transmits all structured data to the data processing component;

[0152] The data processing component writes all the structured data into the cache in real time as parallel streaming data.

[0153] In one embodiment, before the processor 110 executes the dialogue agent to obtain the target request text input by the user, it also performs the following operations:

[0154] The dialogue agent obtains the current request text input by the user and obtains the historical request text previously input by the user.

[0155] The dialogue agent integrates the current request text and the historical request text to obtain the target request text.

[0156] In one embodiment, when the processor 110 executes the current request text input by the user by the dialogue agent, it specifically performs the following operations:

[0157] The dialogue agent obtains the improvement request text input by the user based on the requirement improvement prompt information, and confirms the improvement request text as the current requirement text.

[0158] In one embodiment, when executing the method, the processor 110 also performs the following operations:

[0159] If the control component receives a confirmation instruction from the user regarding the snapshot expression data, it will persistently display the snapshot expression data for the target requirement text.

[0160] If the control component receives a denial instruction from the user regarding the snapshot expression data, it then obtains the modification request text from the denial instruction.

[0161] The control component transmits the modification request text to the dialogue agent;

[0162] When the dialogue agent obtains the user's current request text, the following operations are performed:

[0163] The dialogue agent confirms the modified request text as the current request text.

[0164] In this embodiment, the dialogue agent acquires the current requirement text input by the user and the historical requirement text previously input by the user. The current requirement text and the historical requirement text are integrated and processed to obtain the target requirement text, thereby further completing the user's problem description and improving the accuracy of the large language model's response. The dialogue agent acquires the target requirement text input by the user and inputs it into the large language model, then acquires the target structured data output by the large language model. If the target structured data does not meet the preset transmission conditions, a requirement improvement prompt is displayed to the user on the front-end page. The agent acquires the improved requirement text input by the user based on the requirement improvement prompt and then confirms the improved requirement text as the current requirement text, allowing the user to improve the requirement text in a timely manner, further improving the response efficiency of the large language model. If the target structured data meets the preset transmission conditions, the target structured data is transmitted to the processing agent. The processing agent writes all structured data into a buffer in real time in the form of parallel streaming data. All structured data includes the target structured data. The processing agent may include an intermediate storage component and a data processing component. The intermediate storage component acts as a data stream buffer, enabling the data processing component to ensure the stability and uninterrupted transmission of the data stream. The control component uses a polling approach to retrieve the target structured data from the cache and generate corresponding snapshot representation data. On the front-end page, the control component displays this snapshot representation data for the target request text. Through the collaborative work of the dialogue agent and the processing agent, users can continue to input their request text while the data is being displayed, reducing the waiting time when users input multiple request texts into the large language model and avoiding wasted time resources. Furthermore, if the control component confirms receipt of a user's confirmation instruction for the snapshot representation data, it can persistently display the snapshot representation data for the target request text, facilitating direct restoration of this content on the front-end page and allowing users to view it at any time. If the control component receives a denial instruction, it retrieves the modified request text from the denial instruction and transmits the modified request text to the dialogue agent, allowing users to promptly input modified request text when receiving an unexpected response, instructing the large language model to respond again.

[0165] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory, or random access memory, etc.

[0166] The above-disclosed embodiments are merely preferred embodiments of this application and should not be construed as limiting the scope of this application. Therefore, any equivalent variations made in accordance with the claims of this application shall still fall within the scope of this application.

[0167] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.), and signals involved in the embodiments of this specification are all authorized by the user or fully authorized by all parties, and the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. For example, the target requirement text, target structured data, and snapshot expression data involved in this specification were all obtained with full authorization.

Claims

1. A multi-agent collaborative data processing method applied to a multi-agent collaborative system, characterized in that, The multi-agent collaborative system includes a dialogue agent, a processing agent, and a control component; the method includes: The dialogue agent obtains the target requirement text input by the user, inputs the target requirement text into the large language model, and obtains the target structured data output by the large language model based on the domain knowledge base for the target requirement text; The dialogue agent transmits the target structured data to the processing agent; The processing agent writes all structured data into the buffer in real time in the form of parallel streaming data. All structured data includes the target structured data. All structured data refers to all structured data received by the processing agent, including the target structured data. The control component uses a polling method to retrieve the target structured data from the cache and generate snapshot representation data corresponding to the target structured data. The control component displays the snapshot expression data on the front-end page in response to the target requirement text.

2. The method of claim 1, wherein, The dialogue agent transmits the target structured data to the data processing agent, including: If the dialogue agent detects that the target structured data does not meet the preset transmission conditions, it will improve the prompt information for the user's display needs on the front-end page and continue to execute the step of obtaining the target demand text entered by the user. If the dialogue agent detects that the target structured data meets the preset transmission conditions, it transmits the target structured data to the processing agent.

3. The method of claim 2, wherein, Before the dialogue agent transmits the target structured data to the data processing agent, it also includes: If the dialogue agent detects a return question in the structured data, it generates a request improvement prompt based on the return question.

4. The method of claim 1, wherein, The processing agent includes an intermediate storage component and a data processing component; The processing agent writes all structured data into a buffer in real time as parallel streaming data, including: The intermediate storage component stores the target structured data; The intermediate storage component transmits all structured data to the data processing component; The data processing component writes all the structured data into the cache in real time as parallel streaming data.

5. The method of claim 2, wherein, Before the dialogue agent obtains the target requirement text input by the user, it also includes: The dialogue agent obtains the current request text input by the user and obtains the historical request text previously input by the user. The dialogue agent integrates the current request text and the historical request text to obtain the target request text.

6. The method according to claim 5, characterized in that, The dialogue agent acquires the user's current request text, including: The dialogue agent obtains the improvement request text input by the user based on the requirement improvement prompt information, and confirms the improvement request text as the current requirement text.

7. The method according to claim 5, characterized in that, Also includes: If the control component receives a confirmation instruction from the user regarding the snapshot expression data, it will persistently display the snapshot expression data for the target requirement text. If the control component receives a denial instruction from the user regarding the snapshot expression data, it then obtains the modification request text from the denial instruction. The control component transmits the modification request text to the dialogue agent; The dialogue agent acquires the user's current request text, including: The dialogue agent confirms the modified request text as the current request text.

8. A multi-agent collaborative data processing device, characterized in that, The device includes a dialogue agent, a processing agent, and a control component; The dialogue agent is used to obtain the target demand text input by the user, input the target demand text into the large language model, and obtain the target structured data output by the large language model based on the domain knowledge base for the target demand text; The target structured data is transmitted to the processing agent; The processing agent is used to write all structured data into a buffer in real time in the form of parallel streaming data. All structured data includes the target structured data. All structured data refers to all structured data received by the processing agent, including the target structured data. The control component is used to retrieve the target structured data from the cache using a polling method, and generate snapshot representation data corresponding to the target structured data; On the front-end page, the snapshot expression data is displayed in response to the target requirement text.

9. A computer storage medium, characterized in that, The computer storage medium stores a plurality of instructions, which are adapted to be loaded by a processor and executed as method steps as claimed in any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product stores multiple instructions, which are adapted to be loaded by a processor and executed as the method steps of any one of claims 1 to 7.

11. An electronic device, characterized in that, include: A processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and executed the method steps as claimed in any one of claims 1 to 7.

Images