Task information processing method and device, equipment, storage medium and program product

Abstract

The embodiment of the invention provides a task information processing method and device, equipment, a storage medium and a program product, and relates to the field of artificial intelligence. The method comprises the following steps: acquiring configuration information of a target task input by a user; according to the configuration information, target scheduling plan information of the target task about an execution tool is generated through the planning agent, and the target scheduling plan information comprises at least one subtask of the target task and a preset execution tool needing to be scheduled by the subtask; according to the target scheduling plan information, determining an input parameter of each preset execution tool needing to be executed through the execution agent, and according to the input parameter, controlling the preset execution tool to execute a corresponding sub-task in the target scheduling plan information; and according to the target scheduling plan information, the execution result of each preset execution tool is converted through the expression agent and then is output to the user. According to the invention, multiple different agents work cooperatively, so that the decision accuracy and applicability of the task processing process are improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and in particular to a task information processing method, electronic equipment, storage medium, and program product. Background Art

[0002] A large model agent is an entity with specific functions and decision-making capabilities developed based on a large language model. It can understand the input information, use its own knowledge and algorithms to analyze and process it, and output corresponding results or decisions.

[0003] Current financial data processing involves the application of multiple algorithmic models. Different types of business data processing may require different algorithmic model combinations and scheduling strategies. A common approach is to use large models or traditional scheduling models to determine which specific algorithmic model to use. However, due to the large number of algorithmic models used in financial data processing, a single agent faces numerous difficulties in making decisions. On the one hand, the limited capabilities of a single agent make it difficult to fully and accurately understand and process the complex logic of financial data, resulting in low decision-making accuracy. On the other hand, using a unified algorithm or scheduling model to manage the invocation of numerous algorithmic models, while coordination can be achieved to a certain extent, this approach lacks flexibility and adaptability.

[0004] With the continuous development and innovation of financial business, business complexity is increasing. The requirements for algorithm models in different business scenarios vary greatly. It is difficult to accurately adapt a unified algorithm or scheduling model, which limits the efficiency and quality of business data processing. Summary of the Invention

[0005] The present application provides a task information processing method, apparatus, equipment, storage medium and product, which realizes the collaborative work of multiple intelligent agents with different functions, fully utilizes the advantageous performance of each intelligent agent, comprehensively and accurately understands and processes the logic of complex task data, and improves the decision-making accuracy and applicability of the task processing process.

[0006] In the first aspect, the present application provides a task information processing method, comprising: obtaining configuration information of a target task input by a user; generating target scheduling plan information about an execution tool for the target task through a planning agent based on the configuration information, the target scheduling plan information including at least one subtask of the target task and a preset execution tool that needs to be scheduled for the subtask; determining input parameters of each of the preset execution tools that need to be executed through an execution agent based on the target scheduling plan information, and controlling the preset execution tool to execute the subtask corresponding to the subtask in the target scheduling plan information based on the input parameters; and converting the execution results of each of the preset execution tools through an expression agent based on the target scheduling plan information and outputting them to the user.

[0007] In one embodiment, the target scheduling plan information of the target task regarding the execution tool is generated by the planning agent based on the configuration information, including: extracting the characteristic information and resource requirement information of the target task from the configuration information by the planning agent; determining the initial tool scheduling plan information matching the target task and the first prompt word for generating the initial tool scheduling plan information based on the characteristic information and resource requirement information, wherein the initial tool scheduling plan information and the first prompt word are determined based on the task case information of a preset scenario; searching for real-time public information related to the target task through the planning agent based on the characteristic information and resource requirement information; guiding the planning agent to adjust the initial tool scheduling plan information based on the characteristic information, resource requirement information, and the real-time public information through the first prompt word to generate the target scheduling plan information.

[0008] In one embodiment, the input parameters of each of the preset execution tools that need to be executed are determined by the execution agent based on the target scheduling plan information, including: according to the target scheduling plan information, the execution agent obtains basic information corresponding to each of the subtasks, and the basic information includes one or more of the user's own information, supplier information and market dynamic information corresponding to the subtask; the execution agent converts the basic information of each of the subtasks into a preset data format to obtain the corresponding input parameters of the preset execution tool.

[0009] In one embodiment, controlling the preset execution tool to execute the subtask corresponding to the target scheduling plan information according to the input parameters includes: passing the input parameters to the corresponding preset execution tool through the execution agent, and starting the preset execution tool according to the input parameters, so that the preset execution tool executes the subtask corresponding to the target scheduling plan information according to the input parameters; and / or obtaining the execution results of each preset execution tool through the execution agent.

[0010] In one embodiment, according to the target scheduling plan information, the execution results of each of the preset execution tools are converted by the expression agent and then output to the user, including: determining a second prompt word adapted to the target task, the second prompt word being determined based on the task case information of the preset scenario; obtaining the portrait information of the user; according to the target scheduling plan information and the portrait information, guiding the expression agent through the second prompt word to analyze the execution results of each of the preset execution tools, obtain summary information corresponding to the execution results, and convert the summary information into reply information adapted to the user; and outputting the reply information to the user.

[0011] In one embodiment, it also includes: according to preset detection indicators, the execution results of each of the preset execution tools and the reply information of the expression agent are detected by the evaluation agent to determine the first detection result of each of the preset execution tools and the second detection result of the reply information; if the first detection result and / or the second detection result do not meet the standard requirements of the preset detection indicators, the reason information for not meeting the standard requirements is analyzed by the evaluation agent, and the reason information is fed back to the planning agent.

[0012] In one embodiment, it also includes: obtaining feedback information from the user; if the feedback information indicates that the user approves the reply information of the expression agent, the operation process of this tool scheduling processing is learned by the learning agent, and the parameters of the planning agent, the execution agent and the expression agent are adjusted according to the learning results.

[0013] In a second aspect, the present application provides a task information processing device, comprising:

[0014] The acquisition module is used to obtain the configuration information of the target task input by the user;

[0015] a generation module, configured to generate, through a planning agent, target scheduling plan information of the target task regarding an execution tool according to the configuration information, the target scheduling plan information including at least one subtask of the target task and a preset execution tool to be scheduled for the subtask;

[0016] An execution module, configured to determine, based on the target scheduling plan information, the input parameters of each of the preset execution tools to be executed by an execution agent, and control the preset execution tools to execute the subtasks corresponding to the target scheduling plan information according to the input parameters;

[0017] The output module is used to convert the execution results of each of the preset execution tools through the expression agent according to the target scheduling plan information and output them to the user.

[0018] In one embodiment, a generation module is used to extract the characteristic information and resource requirement information of the target task from the configuration information through a planning agent; determine the initial tool scheduling plan information matching the target task and a first prompt word for generating the initial tool scheduling plan information based on the characteristic information and resource requirement information, wherein the initial tool scheduling plan information and the first prompt word are determined based on the task case information of a preset scenario; search for real-time public information related to the target task through the planning agent based on the characteristic information and resource requirement information; guide the planning agent to adjust the initial tool scheduling plan information based on the characteristic information, resource requirement information, and the real-time public information through the first prompt word to generate the target scheduling plan information.

[0019] In one embodiment, the execution module is used to obtain basic information corresponding to each of the subtasks through an execution agent based on the target scheduling plan information, wherein the basic information includes one or more of the user's own information, supplier information and market dynamic information corresponding to the subtask; and convert the basic information of each of the subtasks into a preset data format through the execution agent to obtain the corresponding input parameters of the preset execution tool.

[0020] In one embodiment, the execution module is used to pass the input parameters to the corresponding preset execution tool through the execution agent, and start the preset execution tool according to the input parameters, so that the preset execution tool executes the subtask corresponding to the target scheduling plan information according to the input parameters; and / or obtains the execution results of each preset execution tool through the execution agent.

[0021] In one embodiment, an output module is used to determine a second prompt word that is adapted to the target task, wherein the second prompt word is determined based on the task case information of a preset scenario; obtain the portrait information of the user; based on the target scheduling plan information and the portrait information, guide the expression agent through the second prompt word to analyze the execution results of each of the preset execution tools, obtain summary information corresponding to the execution results, and convert the summary information into reply information adapted to the user; and output the reply information to the user.

[0022] In one embodiment, the device also includes: a detection module, which is used to detect the execution results of each of the preset execution tools and the reply information of the expression agent according to preset detection indicators through an evaluation agent, and determine the first detection result of each of the preset execution tools and the second detection result of the reply information; if the first detection result and / or the second detection result does not meet the standard requirements of the preset detection indicators, the reason information of non-compliance with the standard requirements is analyzed by the evaluation agent, and the reason information is fed back to the planning agent.

[0023] In one embodiment, the device also includes: a learning module for obtaining feedback information from the user; if the feedback information is that the user approves the reply information of the expression agent, the operation process of this tool scheduling processing is learned by the learning agent, and the parameters of the planning agent, the execution agent and the expression agent are adjusted according to the learning results.

[0024] In a third aspect, an embodiment of the present application provides an electronic device, including:

[0025] at least one processor; and

[0026] a memory communicatively coupled to the at least one processor;

[0027] 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 electronic device to execute the method described in any one of the above aspects.

[0028] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions. When a processor executes the computer-executable instructions, the method described in any one of the above aspects is implemented.

[0029] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which implements the method described in any of the above aspects when executed by a processor.

[0030] The task information processing method, the electronic device, the storage medium and the program product provided by the application, by obtaining the configuration information of the target task input by the user, using the planning agent to refine the target task into at least one subtask according to the configuration information, and assigning appropriate preset execution tools to each subtask, ensure the reasonable decomposition of the task and the optimized utilization of resources. Then the execution agent determines the input parameters required by each preset execution tool according to the target scheduling plan information, and controls the corresponding execution tool to execute the corresponding subtask according to the input parameters, and finally the execution result is converted and output to the user by the expression agent, so that the user can intuitively understand the results of task execution. The scheme cooperates with multiple intelligent agents with different functions, fully utilizes the advantages of each intelligent agent, comprehensively and accurately understands and processes the logic of complex task data, not only improves the decision accuracy in the task processing process, but also improves the adaptive flexibility of different scene tasks, and improves the intelligent level of the system and the user experience. BRIEF DESCRIPTION OF DRAWINGS

[0031] The accompanying drawings, which are incorporated herein and constitute part of the specification, illustrate embodiments consistent with the application and, together with the description, serve to explain the principles of the application. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0032] Figure 1 A structural schematic diagram of an electronic device provided by an embodiment of the application is shown in the following figure.

[0033] Figure 2 An application scenario schematic diagram of a task information processing system provided by an embodiment of the application is shown in the following figure.

[0034] Figure 3 A framework schematic diagram of a task information processing system provided by an embodiment of the application is shown in the following figure.

[0035] Figure 4 A flowchart of a task information processing method provided by an embodiment of the application is shown in the following figure.

[0036] Figure 5 A structural schematic diagram of a task information processing device provided by an embodiment of the application is shown in the following figure.

[0037] Through the above-mentioned drawings, the specific embodiments of the application have been shown, and will be described in more detail in the following. These drawings and textual descriptions are not intended to limit the scope of the concept of the application by any means, but to illustrate the concept of the application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0038] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0039] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data comply with the relevant laws, regulations and standards of relevant countries and regions, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0040] In addition, this application involves big data analysis of user information (including but not limited to personal biometrics, identity data, consumption data, asset data, electronic terminal operation data, etc.), and the use of artificial intelligence technology for automated decision-making, and a technical solution for making decisions that have a significant impact on personal rights and interests based on the results of automated decision-making. The application provides users with corresponding operation entrances for users to choose to agree or reject the results of automated decision-making; if the user chooses to reject, the expert decision-making process will be entered.

[0041] It should be noted that the task information processing method, device, equipment, storage medium and program product provided in this application can be used in the field of artificial intelligence, and can also be used in any field other than artificial intelligence. The application field of the task information processing method, device, equipment, storage medium and product in this application is not limited.

[0042] In order to clearly describe the technical solutions of the embodiments of the present application, the following definitions are given first:

[0043] URL: uniform resource locator, uniform resource locator system.

[0044] API: Application Programming Interface, application programming interface.

[0045] A large model agent is an entity with specific functions and decision-making capabilities developed based on a large language model. It can understand the input information, use its own knowledge and algorithms to analyze and process it, and output corresponding results or decisions.

[0046] Current financial data processing involves the application of multiple algorithmic models. Different types of business data processing may require different algorithmic model combinations and scheduling strategies. A common approach is to use large models or traditional scheduling models to determine which specific algorithmic model to use. However, due to the large number of algorithmic models used in financial data processing, a single agent faces numerous difficulties in making decisions. On the one hand, the limited capabilities of a single agent make it difficult to fully and accurately understand and process the complex logic of financial data, resulting in low decision-making accuracy. On the other hand, using a unified algorithm or scheduling model to manage the invocation of numerous algorithmic models, while coordination can be achieved to a certain extent, this approach lacks flexibility and adaptability.

[0047] With the continuous development and innovation of financial services, business complexity is increasing. Different business scenarios have vastly different requirements for algorithmic models, making it difficult to precisely adapt a unified algorithm or scheduling model. This limits the efficiency and quality of business data processing. Furthermore, each change in a business scenario requires extensive adjustments to the unified algorithm or scheduling model, resulting in a massive workload and low efficiency.

[0048] In order to solve at least one of the above problems, an embodiment of the present application provides a task information processing solution, which obtains the configuration information of the target task input by the user, uses the planning agent to refine the target task into at least one subtask according to the configuration information, and assigns a suitable preset execution tool to each subtask, thereby ensuring the reasonable decomposition of tasks and the optimal use of resources. Then, the execution agent determines the input parameters required by each preset execution tool according to the target scheduling plan information, and controls the corresponding execution tool to execute the corresponding subtask accordingly. Finally, the execution result is converted and output to the user through the expression agent, so that the user can intuitively understand the results of the task execution. This solution uses multiple agents with different functions to work together, fully utilizes the advantages of each agent, and comprehensively and accurately understands and processes the logic of complex task data. It not only improves the decision-making accuracy in the task processing process, but also improves the adaptability to tasks in different scenarios, thereby improving the intelligence level of the system and user experience.

[0049] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0050] like Figure 1 As shown, this embodiment provides an electronic device 1, including: at least one processor 11 and a memory 12, Figure 1In the example, a processor 11 and a memory 12 are connected via a bus 10. Memory 12 stores instructions executable by processor 11. The instructions are executed by processor 11 so that electronic device 1 can execute all or part of the process of the method described in the following embodiments. This allows multiple intelligent agents with different functions to work together, fully utilizing the advantages of each intelligent agent, comprehensively and accurately understanding and processing the logic of complex task data, and improving the decision-making accuracy and applicability of the task processing process.

[0051] In one embodiment, the electronic device 1 may be a mobile phone, a tablet computer, a laptop computer, a desktop computer, or a large computing system composed of multiple computers.

[0052] Figure 2 This is a schematic diagram of an application scenario 200 of a task information processing system provided in an embodiment of the present application. Figure 2 As shown, the system includes: a server 210 and a terminal 220, wherein:

[0053] The server 210 may be a data platform that provides task information processing services, such as a data platform for financial business data processing. In actual scenarios, a data platform may have multiple servers 210. Figure 2 Here, one server 210 is taken as an example.

[0054] The terminal 220 may be an electronic device for data interaction with the data platform, such as a computer, mobile phone, tablet or other device used to access the data platform. There may also be multiple terminals 220. Figure 2 Two terminals 220 are used as an example for illustration.

[0055] The terminal 220 and the server 210 can transmit information via the Internet, so that the terminal 220 can access the data on the server 210. The terminal 220 and / or the server 210 can be implemented by the electronic device 1.

[0056] The task information processing solution of the embodiment of the present application can be deployed on the server 210, can also be deployed on the terminal 220, or can be deployed partially on the server 210 and partially on the terminal 220. In actual scenarios, the choice can be based on actual needs, and this embodiment does not limit it.

[0057] When the task information processing solution is fully or partially deployed on the server 210 , a calling interface may be opened to the terminal 220 to provide algorithm support to the terminal 220 .

[0058] The method provided in the embodiments of the present application can be implemented by executing corresponding software code on electronic device 1 and by interacting with a server. The electronic device 1 can be a local terminal device. When the method is run on a server, the method can be implemented and executed based on a cloud interaction system, which includes a server and a client device.

[0059] In a possible implementation, the method provided in the embodiment of the present application provides a graphical user interface through a terminal device, wherein the terminal device can be the local terminal device mentioned above, or it can be a client device in the cloud interaction system mentioned above.

[0060] like Figure 3 As shown, a task information processing system provided by an embodiment of the present application includes: a front end and a back end, wherein the front end includes:

[0061] User interaction interface: Design a simple and intuitive user operation interface to facilitate users (such as relevant business personnel) to enter relevant information about financial business, such as business scenario type, interaction data, risk preferences, etc. The interface adopts a responsive design to adapt to the access requirements of different devices. At the same time, a real-time feedback mechanism is set up. When the user enters information, the system automatically performs preliminary verification and gives timely prompts and suggestions to help users provide business information more accurately. Provide a visual operation interface for configuring required tasks and executing intelligent agents, including input calling URLs, required parameters, the purpose of the current intelligent agent, etc., and support uploading initial verification test sets.

[0062] Information Display Module: This module displays model processing progress, intermediate results, and final decision recommendations in real time, using charts, text, and other formats for intuitive user understanding. It also provides interactive functionality, allowing users to delve deeper into detailed data and analysis processes based on their needs. It also includes a historical query log function, allowing users to easily trace and analyze past business transactions, and supports data export for further offline analysis.

[0063] The information display module also displays agent configuration information, including automatically generated plans, executions, expressions, evaluations, prompts for self-learning agents, configurations of previous and subsequent executions, and metrics such as current accuracy. A manual optimization portal is provided to facilitate manual adjustments to each agent's description. The module also displays the progress and results of automatic optimization, allowing users to understand the system's optimization status.

[0064] The system's backend utilizes a large-model multi-agent framework, an architecture that organically combines multiple large-model agents. Within this framework, different agents perform their respective functions, interacting with each other and working collaboratively to achieve the overall objective. This framework integrates the strengths of multiple agents, overcoming the limitations of a single agent in handling complex problems and improving the system's flexibility, adaptability, and accuracy in complex business scenarios. The system's backend can include planning agents, execution agents, and expression agents. Each agent specializes in a specific task area, such as planning, execution, expression, and evaluation, collaborating to achieve complex business process processing.

[0065] In the embodiment of the present application, the front end obtains the configuration information of the target task input by the user, and uses the planning agent to refine the target task into at least one subtask according to the configuration information, and assigns a suitable preset execution tool to each subtask, thereby ensuring the reasonable decomposition of tasks and the optimal use of resources. The execution agent then determines the input parameters required by each preset execution tool based on the target scheduling plan information, and controls the corresponding execution tool to execute the corresponding subtask accordingly. Finally, the expression agent converts the execution result and outputs it to the user, so that the user can intuitively understand the results of the task execution. This solution uses multiple agents with different functions to work together, fully utilizing the advantages of each agent, and comprehensively and accurately understanding and processing the logic of complex task data. It not only improves the decision-making accuracy in the task processing process, but also improves the adaptability to tasks in different scenarios, thereby enhancing the intelligence level of the system and user experience.

[0066] Optionally, the backend can also include an evaluation agent and a learning agent. The evaluation agent is used to comprehensively evaluate the scheduling results after receiving the execution results and the response information from the expression agent. When the user approves the results generated by the system, the learning agent is used to obtain detailed data such as the corresponding task configuration information, execution tool call process, parameter settings, and final results, and store them as part of the correct standard test set.

[0067] Optionally, the system can also include a management module for unified management of various plug-in tools (i.e., execution tools, such as APIs and algorithmic models) involved in financial services, including operations such as registration, updating, deletion, and version control of execution tools. A plug-in library should be established, with detailed metadata descriptions for each plug-in, documenting its functionality, scope of application, and input and output requirements. This ensures the availability and accuracy of plug-in services and provides interface specifications to facilitate interaction between execution agents and execution tools.

[0068] Optionally, the system may also include a data storage and processing module responsible for storing financial business data, intelligent agent data, user interaction data, as well as the intermediate results and learning data of the intelligent agent. A distributed storage architecture can be adopted to ensure high availability and scalability of data. This module has data cleaning and preprocessing functions, and can identify and process problems such as noise, missing values, and outliers in the data, and perform operations such as standardization and normalization on the data to improve data quality. In addition, encryption algorithms can be used to encrypt and store sensitive data to ensure data security and privacy. It can also provide efficient data query and access interfaces for different intelligent agents, support real-time data processing and analysis, and improve data access speed through technical means such as data caching and index optimization to meet the rapid demand for data by intelligent agents during business processing.

[0069] Please see Figure 4 , which is a task information processing method of an embodiment of the present application, the method can be Figure 1 The electronic device 1 shown is used to perform and can be applied to Figure 2-3 In the application scenario shown in , multiple agents with different functions can work together to fully utilize the advantages of each agent, fully and accurately understand and process the logic of complex task data, and improve the decision accuracy and applicability of the task processing process. This embodiment takes terminal 220 as the execution end as an example. The method includes the following steps:

[0070] Step 401: Obtain configuration information of the target task input by the user;

[0071] In this step, the user enters the target task's configuration information in the visual configuration area of ​​the front-end interface. This information includes, but is not limited to, the task description and business scenario. The user then fills in the execution agent's configuration information according to pre-set requirements, such as the URL to be called for the target task, required parameters, and the purpose of the execution tool. Simultaneously, the user uploads an initial validation test set for initial system learning and configuration. This initial validation test set can include task information processing cases for different scenarios, along with the corresponding validation criteria for each case, so that the various back-end agents can learn the characteristics of each scenario.

[0072] Optionally, the front-end transmits user input and test sets to the back-end. The back-end's large-scale multi-agent framework initiates initialization. Each agent automatically generates prompts and configures forward and backward execution relationships based on the large-scale model's understanding of the task and test set. For example, the planning agent generates preliminary prompts for the model call plan; the execution agent completes its own parameter settings and model docking preparation based on the configuration information; the expression agent determines prompts for information summarization and transformation; the evaluation agent establishes an evaluation indicator system and thresholds; and the learning agent develops learning strategies and optimization directions.

[0073] Optionally, the user can submit a processing request about the target task at the interactive interface of the front end, and the front end performs format checking and preliminary logical checking on the request, such as checking whether the required fields are complete and whether the data format is correct. If a problem is found, an error prompt is immediately displayed to the user to guide the user to correct it.

[0074] Step 402: According to the configuration information, the target scheduling plan information of the target task about the execution tool is generated by the planning agent, and the target scheduling plan information includes at least one subtask of the target task and a preset execution tool that needs to be scheduled by the subtask.

[0075] In this step, the execution tool refers to a computer tool for executing the target task, such as a specific API, a neural network algorithm model, or a large language model with specific functions. According to the configuration information, the planning agent refines the target task into at least one subtask and assigns appropriate preset execution tools to each subtask, ensuring reasonable decomposition of the task and optimized use of resources.

[0076] Step 403: According to the target scheduling plan information, the input parameters of each preset execution tool that needs to be executed are determined by the execution agent, and the preset execution tool is controlled to execute the corresponding subtask in the target scheduling plan information according to the input parameters.

[0077] Step 404: According to the target scheduling plan information, the execution results of each preset execution tool are converted and output to the user by the expression agent.

[0078] The above method, by obtaining the configuration information of the target task input by the user, using the planning agent to refine the target task into at least one subtask and assign appropriate preset execution tools to each subtask, ensures reasonable decomposition of the task and optimized use of resources. Then the execution agent determines the input parameters required by each preset execution tool according to the target scheduling plan information, and controls the corresponding execution tool to execute the corresponding subtask according to the input parameters, and finally the execution results are converted and output to the user by the expression agent, so that the user can intuitively understand the results of task execution. This scheme cooperates with multiple intelligent agents with different functions, fully utilizes the advantages of each intelligent agent, comprehensively and accurately understands and processes the logic of complex task data, not only improves the decision accuracy in the task processing process, but also improves the adaptive flexibility of different scene tasks, and improves the intelligent level of the system and user experience.

[0079] In one embodiment, step 402 may specifically include: extracting characteristic information and resource requirement information of the target task from the configuration information through the planning agent; determining the initial tool scheduling plan information matching the target task and the first prompt word for generating the initial tool scheduling plan information based on the characteristic information and resource requirement information, wherein the initial tool scheduling plan information and the first prompt word are determined based on the task case information of the preset scenario; searching for real-time public information related to the target task through the planning agent based on the characteristic information and resource requirement information; guiding the planning agent to adjust the initial tool scheduling plan information based on the characteristic information, resource requirement information, and real-time public information through the first prompt word to generate the target scheduling plan information.

[0080] In this embodiment, the planning agent extracts the target task's characteristic information and resource requirements from the configuration information to achieve a precise understanding of the task and a reasonable assessment of resources. Based on this information, the agent determines initial tool scheduling plan information that matches the target task and a first prompt word for generating the plan. The first prompt word guides the planning agent in generating scheduling plan information for the executing tools. The first prompt word is generated based on task case information for a pre-defined scenario. The pre-defined scenario can be a business scenario that is identical or similar to the target task. The case information for the pre-defined scenario can be relevant information from successful user-approved conversations, ensuring the initial rationality and applicability of the scheduling plan. The planning agent then searches for real-time public information related to the target task to obtain the latest external data and dynamically changing environmental factors. Combining the characteristic information, resource requirements, and real-time public information, the planning agent, guided by the first prompt word, adjusts the initial tool scheduling plan information to generate more accurate and efficient target scheduling plan information. This improves the adaptability of the final target scheduling plan information to the target task scenario and enhances decision-making accuracy.

[0081] Optionally, during the initial build phase, after the user uploads the required task and initial verification test set, the planning agent leverages the large model to understand the task and test set content, automatically generating initial tool scheduling information and the first prompt used to generate the initial tool scheduling information. For example, based on the task cases and task requirements of the pre-set scenarios in the test set, the planning agent determines the order and combination of algorithm models to prioritize in different business scenarios.

[0082] Optionally, during operation, the planning agent receives information input from the front-end user. After receiving the target task processing request submitted by the user, the planning agent uses the large model to deeply understand and analyze the target task's configuration information, extracting key task characteristics and resource requirements. The planning agent then initiates a web search to obtain relevant news information, including financial market trends, changes in relevant regulatory rules, and industry competition trends. This real-time public information is filtered, classified, and organized, extracting key information relevant to the current target task. This information is then combined with the task requirements and the automatically generated first prompt word to develop a more reasonable and realistic model deployment plan. For example, if the target task is related to corporate financing, the planning agent first determines the initial tool scheduling plan information corresponding to the target task in this business scenario and the first prompt word used to generate the initial tool scheduling plan information. It then searches for recent relevant news, obtains a real-time data stream of financial parameter updates, analyzes their impact on the company's financing costs and market capital supply, and adjusts the initial tool scheduling plan information based on the financing strategy information in the first prompt word. Specifically, based on the comprehensive analysis results, the planning agent selects appropriate execution tools from the preset tool library, determines their execution order, formulates a detailed calling plan, and passes the plan information to the execution agent and the evaluation agent.

[0083] For example, when the objective is to develop an asset data allocation strategy, the planning agent references macroeconomic data, industry development reports, regulatory guidance, and other news information. Combined with the risk assessment and return forecasting strategies in the automatically generated first prompt, it determines which risk assessment and return forecast models to invoke, and determines the order in which they should be executed to adapt to market changes and regulatory requirements. By combining real-time public information with the automatically generated first prompt to formulate target invocation plans, the planning agent ensures that the execution of execution tools is more closely aligned with the actual business environment.

[0084] In one embodiment, step 403 may specifically include: according to the target scheduling plan information, obtaining basic information corresponding to each subtask through the execution agent, the basic information including one or more of the user's own information, supplier information and market dynamic information corresponding to the subtask; converting the basic information of each subtask into a preset data format through the execution agent to obtain the input parameters of the corresponding preset execution tool.

[0085] In this implementation, the execution agent obtains basic information corresponding to each subtask based on the target scheduling plan information. This basic information includes, but is not limited to, one or more of the following: user information, supplier information, and market dynamics information, ensuring the comprehensiveness and accuracy of the background data required for task execution. The execution agent then converts this basic information into a pre-set data format, generating the corresponding input parameters for the pre-set execution tool. This not only standardizes information processing, ensures the compatibility and consistency of data from different sources, and improves the system's adaptability to different scenarios, but also enables the pre-set execution tool to directly utilize these input parameters for efficient operation.

[0086] Optionally, during initial configuration, the execution agent can be set up based on the configuration information provided by the user (including the URL to be called, required parameters, purpose, etc.). The execution agent is responsible for the execution of different execution tools. Before executing the subtasks of each tool, the execution agent determines the necessary input parameters required based on the characteristics of the execution tool and business needs. For example, it can obtain these parameters from a variety of data sources, including internal corporate databases, external data suppliers, and real-time market data. The execution agent then performs necessary format conversion and preprocessing on the parameters to ensure that they meet the input requirements of the model. For example, when executing a credit risk assessment model, the execution agent needs to obtain parameters such as the customer's credit record, financial statement data, and the industry's average credit level, and clean and standardize these data. The processed parameters are then passed to the credit risk assessment model for calculation to obtain the model's output results.

[0087] Optionally, taking the execution tool as an algorithm model as an example, after the execution agent receives the instruction from the planning agent, it determines the algorithm model to be executed according to the target call plan information. For each subtask corresponding to the algorithm model, the execution agent first determines the necessary input parameters. For example, the user's own information can be obtained by querying the internal database of the enterprise related to the target task to obtain internal enterprise data, such as financial statements, customer information, etc.; at the same time, establish a connection with external data suppliers to obtain external supplier information such as market data and industry data; it can also monitor financial market interaction data in real time to obtain the latest market dynamics information. After obtaining this basic information, the execution agent cleans, converts and verifies the basic information to ensure the accuracy and completeness of the basic information. For example, date data in different formats are uniformly converted to a standard format, and missing financial data are supplemented or estimated to ensure accurate model input;

[0088] In one embodiment, step 403 may also include: passing the input parameters to the corresponding preset execution tool through the execution agent, and starting the preset execution tool according to the input parameters, so that the preset execution tool executes the corresponding subtask in the target scheduling plan information according to the input parameters; and / or obtaining the execution results of each preset execution tool through the execution agent.

[0089] In this embodiment, the execution agent passes the generated input parameters to the corresponding preset execution tool and starts the execution tool based on these input parameters. Ensure that the execution tool can accurately receive and understand the task instructions, so as to execute the corresponding subtasks according to the specific requirements in the target scheduling plan information. Automated parameter transfer and tool startup reduce the necessity of human intervention, reduce the risk of operational errors, and improve overall execution efficiency. In addition, the execution agent can obtain the execution results of each preset execution tool, so that the system can monitor the execution status and results of the task in real time, ensuring the correct completion and timely adjustment of the task. By combining automated control and real-time feedback, not only the intelligence level of the system is improved, but also the flexibility and accuracy of task execution are enhanced.

[0090] Taking the aforementioned scenario where the execution tool is an algorithmic model as an example, the execution agent passes the processed basic information as input parameters to the corresponding algorithmic model, starts the algorithmic model to execute the corresponding subtask, and monitors the algorithmic model's operating status in real time. Optionally, after the algorithmic model completes, the execution agent can obtain the algorithmic model's output results and pass the results to the expression agent and evaluation agent, providing basic information for the implementation of the subsequent system feedback mechanism.

[0091] In one embodiment, step 404 may specifically include: determining a second prompt word that is adapted to the target task, the second prompt word being determined based on the task case information of a preset scenario; obtaining the user's portrait information; based on the target scheduling plan information and the portrait information, guiding the expression agent through the second prompt word to analyze the execution results of each preset execution tool, obtain summary information corresponding to the execution results, and convert the summary information into reply information adapted to the user; and outputting the reply information to the user.

[0092] In this embodiment, the second prompt word is used to guide the expression agent to summarize and output the execution results. First, the second prompt word that is adapted to the target task is determined. The prompt word is generated based on the task case information of the preset scenario, ensuring its pertinence and effectiveness. Then, by obtaining the user's portrait information, the user's needs and preferences can be deeply understood. On this basis, the expression agent analyzes the execution results of each preset execution tool under the guidance of the second prompt word, generates summary information of the execution results, and converts the summary information into reply information adapted to the user. By combining the user portrait information, personalized customization of the output information is achieved, ensuring that the output content is more in line with the user's reading habits. Finally, the system outputs the reply information to the user, so that the user can obtain clear, accurate and personalized feedback, enhances the system's responsiveness and user satisfaction, and realizes efficient and personalized information transmission.

[0093] Optionally, during initial construction, the Expression Agent generates secondary prompts based on the macro model's understanding of the task and test set. These prompts guide the Expression Agent in effectively summarizing the Execution Agent's execution results based on the Planning Agent's output. In actual operation, after receiving multiple execution results from the Execution Agent and the Planning Agent's target scheduling plan information, the Expression Agent summarizes and organizes this information based on the pre-generated secondary prompts, identifying key data points, trends, and conclusions. Then, based on the user's knowledge level and business needs as reflected in the user profile, the Expression Agent translates this complex information into accessible language, enhancing the user's reading experience. For example, complex mathematical model outputs can be converted into intuitive charts and textual descriptions, and professional financial terms can be translated into everyday language. The Expression Agent also integrates and analyzes this information based on business logic, providing deeper insights and recommendations. For example, when analyzing a company's financial status, it not only presents the numerical values ​​of various financial indicators but also compares them to industry averages, analyzes the company's strengths and weaknesses, and proposes corresponding improvement suggestions. Finally, the generated clear and accurate responses are fed back to the front-end for display to the user.

[0094] In one embodiment, the method further includes: according to preset detection indicators, the execution results of each preset execution tool and the response information of the expression agent are detected by the evaluation agent to determine the first detection result of each preset execution tool and the second detection result of the response information; if the first detection result and / or the second detection result do not meet the standard requirements of the preset detection indicators, the reason information for not meeting the standard requirements is analyzed by the evaluation agent, and the reason information is fed back to the planning agent.

[0095] In this embodiment, preset detection indicators are used to quantitatively test the accuracy, reliability, and completeness of the execution results output by the execution tools and the response information output by the expression agent. For example, the accuracy of the model results is evaluated by comparing them with historical data and industry standard reference values. The quality of the response is evaluated by checking whether the response contains key information and whether there are logical loopholes. The evaluation agent comprehensively tests the execution results of each preset execution tool and the response information of the expression agent based on the preset detection indicators. By determining the first detection results of each execution tool and the second detection results of the response information, potential problems and deviations are promptly identified.

[0096] Optionally, if the evaluation agent determines that the first test result and / or the second test result meet the standard requirements of the preset test indicators, such as the model accuracy exceeds the set threshold, the user task requirements are met, etc., the evaluation agent sends a task completion signal to the system, and the system records and archives the processing results. If the evaluation agent determines that the first test result and / or the second test result do not meet the standard requirements of the preset test indicators, it will analyze the cause information in detail, such as whether there is an error in model selection, improper parameter setting, data quality problems, etc., and feed back the analysis results to the planning agent. At the same time, the evaluation agent triggers the re-iteration mechanism, and the system will re-execute the task starting from the planning agent based on the feedback information until a satisfactory result is obtained. The feedback mechanism forms a closed-loop self-improvement system to ensure that the system can continuously optimize its performance and output quality.

[0097] Optionally, during system initialization, the evaluation agent analyzes the test set based on the large model to determine an evaluation index system and thresholds for testing the execution results of each execution tool. During operation, the evaluation agent receives the execution results of the execution agent and the response information of the expression agent, and then conducts a comprehensive test and evaluation of the execution results.

[0098] In one embodiment, the method further includes: obtaining feedback information from the user; if the feedback information indicates that the user approves the reply information of the expression agent, the operation process of the tool scheduling processing is learned by the learning agent, and the parameters of the planning agent, the execution agent and the expression agent are adjusted according to the learning results.

[0099] In this embodiment, the intelligence level of the system and user satisfaction can be significantly improved through the effective use of user feedback and the adaptive adjustment of the learning agent. The system first obtains user feedback information as an important basis for evaluating system performance. When the user approves the reply information of the expression agent, this positive feedback triggers the learning agent to conduct in-depth learning of the operation process of the tool scheduling process during this task processing. By analyzing successful operation cases, the learning agent identifies effective strategies and parameter settings, which not only consolidates the successful experience of the system, but also provides an optimization reference for future task execution. Based on the learning results, the system further adjusts the parameters of the planning agent, execution agent and expression agent, so that each agent can work better together and adapt to user needs and environmental changes. The feedback-based learning and adjustment mechanism ensures that the system can continuously improve its performance and responsiveness.

[0100] Optionally, during the initial system configuration, the learning agent determines its learning strategy and optimization direction based on the large model's understanding of the test set. During system operation, the learning agent can continuously collect data. When a user approves a system-generated response, the learning agent obtains detailed data, including the business information, model invocation process, parameter settings, and final output corresponding to the target task, and stores this data as a success case as part of the correct standard test set. The learning agent can periodically utilize this test set and apply machine learning algorithms to optimize the prompt words and parameters of the planning agent, execution agent, and expression agent. For example, by analyzing the commonalities of successful cases in the test set, the planning agent can adjust the keywords and strategies used when formulating target invocation plan information; based on the discrepancy between model execution results and expected results, the execution agent's parameter settings can be optimized; and based on user feedback on the response information, the expression agent's language can be improved. The learning agent continuously repeats this learning and optimization process, gradually improving the system's success rate and accuracy in processing financial transactions within a short period of time, enabling the system to better adapt to different business scenarios and user needs.

[0101] Optionally, the learning agent supports users to set an accuracy threshold. When the threshold is reached during the automatic optimization process, the learning agent stops the optimization task. If the optimization process continues for a preset period of time (for example, 1 hour) and is not completed, the system will automatically stop the optimization and retain the version with the highest accuracy, ensuring that the system obtains a better configuration within a reasonable time.

[0102] Optionally, the method may also include: unified management of various plug-in tools (i.e., execution tools, such as APIs and algorithmic models) involved in financial services, including operations such as registration, updating, deletion, and version control of the execution tools. A plug-in library is established, with detailed metadata descriptions for each plug-in, documenting its functionality, scope of application, input and output requirements, and other information. This ensures the availability and accuracy of plug-in services and provides interface specifications to facilitate interaction between execution agents and execution tools.

[0103] Optionally, the method may also include: storing financial business data, intelligent agent data, user interaction data, as well as intermediate results and learning data of the intelligent agent. A distributed storage architecture can be used to ensure high availability and scalability of data. This module has data cleaning and preprocessing functions, and can identify and process problems such as noise, missing values, and outliers in the data, and perform operations such as standardization and normalization on the data to improve data quality. In addition, encryption algorithms can be used to encrypt and store sensitive data to ensure data security and privacy. It can also provide efficient data query and access interfaces for different intelligent agents, support real-time data processing and analysis, and improve data access speed through technical means such as data caching and index optimization to meet the rapid demand for data by intelligent agents during business processing.

[0104] The solution of the embodiment of the present application realizes the specialized processing and collaborative work of each link by introducing agents with different functions such as planning agents, execution agents, expression agents, evaluation agents and learning agents. The planning agent is responsible for formulating a reasonable tool calling plan according to business needs; the execution agent performs specific tool operations according to the calling plan; the expression agent presents the processing results to the user or other systems in a clear and easy-to-understand manner; the evaluation agent detects and evaluates the entire process and results; the learning agent learns from historical scheduling data and continuously improves the scheduling strategy, so that the entire framework can adapt to the ever-changing financial business scenarios and improve the overall performance of the system. Each agent divides the work and cooperates, each doing its own job, reducing redundancy and errors in the processing process. It not only greatly improves the processing speed of financial business and improves the business response rate, but also enhances the adaptability of the system.

[0105] Please see Figure 5 , which is a task information processing device 500 of an embodiment of the present application, which can be applied to electronic equipment 1 and can be applied to Figure 2-3 In the application scenario shown in , multiple intelligent agents with different functions can work together to fully utilize the advantages of each intelligent agent, fully and accurately understand and process the logic of complex task data, and improve the decision-making accuracy and applicability of the task processing process. The device includes: an acquisition module 501, a generation module 502, an execution module 503, and an output module 504. The functional principles of each module are as follows:

[0106] The acquisition module 501 is used to acquire the configuration information of the target task input by the user;

[0107] A generation module 502 is configured to generate target scheduling plan information about execution tools for a target task through a planning agent based on the configuration information, wherein the target scheduling plan information includes at least one subtask of the target task and a preset execution tool to be scheduled for the subtask;

[0108] The execution module 503 is used to determine the input parameters of each preset execution tool to be executed according to the target scheduling plan information through the execution agent, and control the preset execution tool to execute the subtask corresponding to the target scheduling plan information according to the input parameters;

[0109] The output module 504 is used to convert the execution results of each preset execution tool through the expression agent according to the target scheduling plan information and output them to the user.

[0110] In one embodiment, the generation module 502 is used to extract the characteristic information and resource requirement information of the target task from the configuration information through the planning agent; determine the initial tool scheduling plan information matching the target task and the first prompt word for generating the initial tool scheduling plan information based on the characteristic information and resource requirement information, and the initial tool scheduling plan information and the first prompt word are determined based on the task case information of the preset scenario; search for real-time public information related to the target task through the planning agent based on the characteristic information and resource requirement information; guide the planning agent to adjust the initial tool scheduling plan information through the first prompt word based on the characteristic information, resource requirement information, and real-time public information to generate the target scheduling plan information.

[0111] In one embodiment, the execution module 503 is used to obtain basic information corresponding to each subtask through the execution agent based on the target scheduling plan information. The basic information includes one or more of the user's own information, supplier information and market dynamic information corresponding to the subtask; the basic information of each subtask is converted into a preset data format through the execution agent to obtain the input parameters of the corresponding preset execution tool.

[0112] In one embodiment, the execution module 503 is used to pass the input parameters to the corresponding preset execution tool through the execution agent, and start the preset execution tool according to the input parameters, so that the preset execution tool executes the corresponding subtask in the target scheduling plan information according to the input parameters; and / or obtains the execution results of each preset execution tool through the execution agent.

[0113] In one embodiment, the output module 504 is used to determine a second prompt word that is adapted to the target task, and the second prompt word is determined based on the task case information of the preset scenario; obtain the user's portrait information; based on the target scheduling plan information and the portrait information, guide the expression agent through the second prompt word to analyze the execution results of each preset execution tool, obtain summary information corresponding to the execution results, and convert the summary information into reply information adapted to the user; and output the reply information to the user.

[0114] In one embodiment, the device also includes: a detection module, which is used to detect the execution results of each preset execution tool and the reply information of the expression agent according to the preset detection indicators through the evaluation agent, and determine the first detection result of each preset execution tool and the second detection result of the reply information; if the first detection result and / or the second detection result does not meet the standard requirements of the preset detection indicators, the reason information of non-compliance with the standard requirements is analyzed by the evaluation agent, and the reason information is fed back to the planning agent.

[0115] In one embodiment, the device also includes: a learning module for obtaining user feedback information; if the feedback information is that the user approves the reply information of the expression agent, the operation process of this tool scheduling processing is learned by the learning agent, and the parameters of the planning agent, execution agent and expression agent are adjusted according to the learning results.

[0116] For a detailed description of the task information processing device 500 , please refer to the description of the relevant method steps in the above embodiment. The implementation principle and technical effects are similar and will not be repeated here in this embodiment.

[0117] An embodiment of the present application further provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the method of any of the aforementioned embodiments is implemented.

[0118] An embodiment of the present application also provides a computer program product, including a computer program, which implements the method of any of the aforementioned embodiments when executed by a processor.

[0119] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the module division is only a logical function division. In actual implementation, other division methods may be used. For example, multiple modules can be combined or integrated into another system, or some features can be ignored or not implemented.

[0120] The integrated modules realized in the form of software function modules can be stored in a computer readable storage medium. The software function modules are stored in a storage medium and include a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to execute part of the steps of the method of various embodiments of the present application.

[0121] It should be understood that the processor described above can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the application can be directly embodied as hardware processor execution, or executed by a combination of hardware and software modules in the processor. The memory can include a high-speed RAM (Random Access Memory) memory, and can also include a non-volatile storage NVM (Nonvolatile memory), such as at least one disk memory, and can also be a U disk, a mobile hard disk, a read-only memory, a magnetic disk or an optical disk, etc.

[0122] The storage medium described above can be realized by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk. The storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0123] An example storage medium is coupled to the processor such that the processor can read information from, and can write information to, the storage medium. Of course, the storage medium can be a part of the processor. Consistent with the teachings provided herein, the processor can execute instructions embodied by computer code stored on the storage medium, and the storage medium can store computer code which, when executed by the processor, causes the processor to perform methods as described herein. The processor and the storage medium can be supplemented by, or incorporated in, ASICs (application-specific integrated circuits).

[0124] It should be noted that, in the present document, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises a", "comprising", or "includes a", does not, without more constraints, preclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0125] The above-mentioned sequence numbers of the embodiments of the present application are only for description, and do not represent advantages or disadvantages of the embodiments.

[0126] Those skilled in the art can clearly understand the above-mentioned embodiment methods by means of software and necessary general hardware platforms, of course, they can also be implemented by hardware, but in many cases, the former is a better implementation. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a plurality of instructions for causing a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods of various embodiments of the present application.

[0127] In the technical solutions of the present application, the collection, storage, use, processing, transmission, provision and disclosure of user data and other information comply with relevant laws and regulations, and do not violate public order and good customs.

[0128] The above are only preferred embodiments of the present application, and do not limit the patent scope of the present application, and any equivalent structure or equivalent process transformation made by using the contents of the specification and drawings, or directly or indirectly applied to other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A task information processing method, characterized in that: include: Get the configuration information of the target task entered by the user; Generate, by a planning agent, target scheduling plan information of the target task regarding an execution tool according to the configuration information, wherein the target scheduling plan information includes at least one subtask of the target task and a preset execution tool to be scheduled for the subtask; According to the target scheduling plan information, the execution agent determines the input parameters of each of the preset execution tools to be executed, and controls the preset execution tools to execute the subtasks corresponding to the target scheduling plan information according to the input parameters; According to the target scheduling plan information, the execution results of each of the preset execution tools are converted by the expression agent and then output to the user.

2. The method according to claim 1, characterized in that The step of generating target scheduling plan information of the target task on the execution tool by a planning agent according to the configuration information includes: Extracting the characteristic information and resource requirement information of the target task from the configuration information by a planning agent; Determining, based on the feature information and resource requirement information, initial tool scheduling plan information matching the target task and a first prompt word for generating the initial tool scheduling plan information, wherein the initial tool scheduling plan information and the first prompt word are determined based on task case information of a preset scenario; Searching, by the planning agent, for real-time public information related to the target task based on the feature information and resource requirement information; According to the feature information, resource demand information, and the real-time public information, the planning agent is guided by the first prompt word to adjust the initial tool scheduling plan information to generate the target scheduling plan information.

3. The method according to claim 1, characterized in that The step of determining the input parameters of each of the preset execution tools to be executed by the execution agent according to the target scheduling plan information includes: According to the target scheduling plan information, the executing agent obtains basic information corresponding to each of the subtasks, wherein the basic information includes one or more of user information, supplier information, and market dynamic information corresponding to the subtask; The basic information of each subtask is converted into a preset data format by the execution agent to obtain the corresponding input parameters of the preset execution tool.

4. The method according to claim 1, wherein The controlling the preset execution tool to execute the subtask corresponding to the target scheduling plan information according to the input parameters includes: The input parameters are passed to the corresponding preset execution tool through the execution agent, and the preset execution tool is started according to the input parameters, so that the preset execution tool executes the subtask corresponding to the target scheduling plan information according to the input parameters; and / or The execution results of each of the preset execution tools are obtained through the execution agent.

5. The method according to claim 1, wherein The step of converting the execution results of each of the preset execution tools according to the target scheduling plan information through an expression agent and outputting the converted results to the user includes: Determining a second prompt word adapted to the target task, where the second prompt word is determined based on task case information of a preset scenario; Obtaining the user's portrait information; Based on the target scheduling plan information and the portrait information, guiding the expression agent to analyze the execution results of each of the preset execution tools through the second prompt word, obtain summary information corresponding to the execution results, and convert the summary information into response information adapted to the user; The reply information is output to the user.

6. The method according to any one of claims 1 to 5, characterized in that Also includes: According to the preset detection index, the evaluation agent detects the execution results of each of the preset execution tools and the response information of the expression agent, and determines a first detection result of each of the preset execution tools and a second detection result of the response information; If the first detection result and / or the second detection result does not meet the standard requirements of the preset detection indicators, the evaluation agent analyzes the reason information of not meeting the standard requirements and feeds back the reason information to the planning agent.

7. The method according to any one of claims 1 to 5, characterized in that Also includes: Obtaining feedback information from the user; If the feedback information indicates that the user approves the reply information of the expression agent, the operation process of this tool scheduling processing is learned by the learning agent, and the parameters of the planning agent, the execution agent and the expression agent are adjusted according to the learning results.

8. A task information processing device, characterized in that: include: The acquisition module is used to obtain the configuration information of the target task input by the user; a generation module, configured to generate, through a planning agent, target scheduling plan information of the target task regarding an execution tool according to the configuration information, the target scheduling plan information including at least one subtask of the target task and a preset execution tool to be scheduled for the subtask; An execution module, configured to determine, based on the target scheduling plan information, the input parameters of each of the preset execution tools to be executed by an execution agent, and control the preset execution tools to execute the subtasks corresponding to the target scheduling plan information according to the input parameters; The output module is used to convert the execution results of each of the preset execution tools through the expression agent according to the target scheduling plan information and output them to the user.

9. An electronic device, characterized in that: include: a processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 7 when executed by a processor.

11. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 7 when being executed by a processor.

Images