Task execution method and device, equipment and storage medium

Abstract

The invention discloses a task execution method and device, equipment and a storage medium, and relates to the technical field of computers. The method comprises the steps that input query information is displayed, the query information is used for generating a query task, and the query task is a task for outputting a query result corresponding to the query information; displaying at least two decomposition tasks generated based on the query information in a timeline form; wherein the at least two decomposition tasks are used for completing the query task; displaying state information corresponding to the at least two decomposition tasks, wherein the state information corresponding to the decomposition tasks is used for indicating execution states of the decomposition tasks; and after the query task is executed, displaying a query result corresponding to the query information. According to the method, the vertical linear list layout is adopted, the at least two decomposition tasks are displayed in the timeline form, the vertical screen space of the terminal equipment is fully utilized, and the difficulty of interactive operation is reduced.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a task execution method, apparatus, device, and storage medium. Background Technology

[0002] Users can initiate queries and obtain answers in AI (Artificial Intelligence) applications.

[0003] In related technologies, the client typically uses a mind map to display the execution process of the query task. The query task is broken down, and the resulting sub-tasks are further broken down to form a network structure with the query task node as the starting point, radiating outwards to the sub-task nodes. The dependencies between the nodes are represented by lines, and the client supports interactive operations such as zooming and panning on the mind map.

[0004] However, the above method requires a large display space to expand and display task nodes, which may cause the device screen to fail to fully display the execution flow of the query task. Summary of the Invention

[0005] This application provides a task execution method, apparatus, device, and storage medium. The technical solutions provided by this application are as follows: According to one aspect of the embodiments of this application, a task execution method is provided, the method comprising: The input query information is displayed, which is used to generate a query task. The query task is a task that outputs the query results corresponding to the query information. The query information is presented in a timeline format, showing at least two decomposed tasks generated based on the query information; wherein the at least two decomposed tasks are used to complete the query task. Display the status information corresponding to the at least two decomposition tasks respectively, wherein the status information corresponding to the decomposition tasks is used to indicate the execution status of the decomposition tasks; After the query task is completed, the query results corresponding to the query information are displayed.

[0006] According to one aspect of the embodiments of this application, a task execution method is provided, the method comprising: Obtain the input query information, which is used to generate a query task, and the query task is a task that outputs the query results corresponding to the query information; The AI ​​model generates a list of decomposed tasks corresponding to the query information based on the query information. The list of decomposed tasks includes at least two decomposed tasks, which are used to complete the query task. The AI ​​model executes each of the decomposition tasks in the decomposition task list to obtain the execution results corresponding to the at least two decomposition tasks respectively. The AI ​​model generates query results corresponding to the query information based on the execution results of the at least two decomposed tasks.

[0007] According to one aspect of the embodiments of this application, a task execution apparatus is provided, the apparatus comprising: An input display module is used to display input query information, which is used to generate a query task, and the query task is a task that outputs the query results corresponding to the query information. The task display module is used to display at least two decomposed tasks generated based on the query information in the form of a timeline; wherein the at least two decomposed tasks are used to complete the query task; A status display module is used to display the status information corresponding to the at least two decomposition tasks respectively, and the status information corresponding to the decomposition tasks is used to indicate the execution status of the decomposition tasks; The results display module is used to display the query results corresponding to the query information after the query task is completed.

[0008] According to one aspect of the embodiments of this application, a task execution apparatus is provided, the apparatus comprising: The input acquisition module is used to acquire input query information, which is used to generate a query task, and the query task is a task that outputs the query results corresponding to the query information. The task generation module is used to generate a list of decomposed tasks corresponding to the query information based on the query information using the AI ​​model. The list of decomposed tasks includes at least two decomposed tasks, which are used to complete the query task. The task execution module is used to execute each of the decomposed tasks in the decomposed task list through the AI ​​model, and obtain the execution results corresponding to the at least two decomposed tasks respectively. The result generation module is used to generate query results corresponding to the query information based on the execution results corresponding to the at least two decomposed tasks using the AI ​​model.

[0009] According to one aspect of the embodiments of this application, a computer device is provided, the computer device including a processor and a memory, the memory storing a computer program, the computer program being loaded and executed by the processor to implement the above-described task execution method.

[0010] According to one aspect of the embodiments of this application, a computer-readable storage medium is provided, wherein a computer program is stored in the computer-readable storage medium, and the computer program is loaded and executed by a processor to implement the above-described task execution method.

[0011] According to one aspect of the embodiments of this application, a computer program product is provided, the computer program product including a computer program, the computer program being loaded and executed by a processor to implement the above-described task execution method.

[0012] The technical solution provided in this application can bring the following beneficial effects: By displaying at least two decomposed tasks generated based on the query information in a timeline format after the query is entered into the user interface, the system clearly demonstrates how it breaks down complex query tasks into multiple decomposed tasks. This makes the task decomposition process transparent, facilitating user understanding of the system's working logic and enhancing user trust in the AI ​​system. Furthermore, the vertical linear list layout, displaying at least two decomposed tasks in a timeline format, fully utilizes the vertical screen space of the terminal device. Users can easily browse the complete task execution process by simply scrolling up and down. Compared to displaying mind maps on the terminal device, the technical solution provided in this application does not require a large display space to show each task node, reducing the display space needed for the task execution process. This facilitates adaptation to different terminal device screens, eliminating the need for users to perform complex operations such as zooming and panning to browse the complete task execution process, thus reducing the difficulty of interactive operation. Additionally, by displaying the status information corresponding to at least two decomposed tasks during the execution of the query task, the execution status of the decomposed tasks is visualized, allowing users to intuitively perceive the execution status of each decomposed task based on the status information, reducing the difficulty of identifying the execution status. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of a computer system provided in one embodiment of this application; Figure 2 This is a flowchart of a task execution method provided in one embodiment of this application; Figure 3 This is a schematic diagram of the interface of a task execution process provided in one embodiment of this application; Figure 4 This is a schematic diagram illustrating the misaligned display of decomposed tasks at different levels, provided in one embodiment of this application. Figure 5 This is a flowchart of a task execution method provided in one embodiment of this application; Figure 6 This is a flowchart illustrating the AI ​​model's processing of query information according to one embodiment of this application; Figure 7 This is a flowchart illustrating task breakdown and execution visualization provided in one embodiment of this application; Figure 8 This is a timing diagram of the real-time status push between the mobile terminal and the server provided in one embodiment of this application; Figure 9 This is a block diagram of a task execution device provided in one embodiment of this application; Figure 10 This is a block diagram of a task execution device provided in one embodiment of this application; Figure 11 This is a structural block diagram of a computer device provided in one embodiment of this application. Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0015] Please refer to Figure 1 This illustration shows a schematic diagram of a computer system provided in one embodiment of this application. The computer system may include: a terminal device 10 and a server 20.

[0016] There may be one or more terminal devices 10. Terminal devices 10 may be electronic devices such as mobile phones, tablets, laptops, desktop computers, game consoles, e-book readers, multimedia playback devices, wearable devices, smart voice interaction devices, smart home appliances, vehicle terminals, aircraft, etc.

[0017] The terminal device 10 may have a client application for the target application installed, which is used to answer user-input queries. This application does not limit the type of the target application; for example, the target application is an AI application. Optionally, the target application may be an application that requires downloading and installation, or it may be an application that can be used instantly; this application does not limit the type of application.

[0018] In some embodiments, the terminal device is a mobile terminal.

[0019] Server 20 provides background services for clients of the target application installed and running on terminal device 10. For example, server 20 can be a background server for the aforementioned target application. Server 20 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms, but is not limited to these. Optionally, server 20 can simultaneously provide background services for target applications on multiple terminal devices 10. Terminal devices 10 and server 20 can communicate with each other via a network.

[0020] In this embodiment, the user inputs query information into the target application's interface. This query information is used to generate a query task, which outputs the query results corresponding to the query information. The server obtains the input query information and, based on the query information, generates a list of decomposed tasks corresponding to the query information using an AI model. This list includes at least two decomposed tasks, each used to complete the query task. The AI ​​model executes each decomposed task in the list, obtaining execution results for each of the at least two decomposed tasks. Based on these execution results, the AI ​​model generates the query results corresponding to the query information. The terminal device, based on the list of decomposed tasks corresponding to the query information, displays the at least two decomposed tasks generated based on the query information in a timeline format on the user interface, along with their respective status information. This status information indicates the execution status of the decomposed tasks. After the query task is completed, the query results corresponding to the query information are displayed.

[0021] Please refer to Figure 2 The diagram illustrates a flowchart of a task execution method according to an embodiment of this application. The executing entity for each step of the method can be a terminal device. The method may include at least one of the following steps 210-240: Step 210: Display the input query information. The query information is used to generate a query task, which is a task that outputs the query results corresponding to the query information.

[0022] For example, an input box is displayed in the user interface, and in response to an input operation on the input box, the entered query information is displayed in the user interface. The input operation on the input box can be an operation in which the user manually enters query information in the input box, an operation in which the user enters query information in the input box via voice command, or a combination of operations in which the user manually enters query information in the input box and operations in which the user enters query information in the input box via voice command.

[0023] The query information may include at least one of the following formats: text, images, audio, video, tables, and charts. If the input operation for the input box is a user manually entering the query information into the input box, the query information includes, but is not limited to, text, images, audio, video, tables, and charts; if the input operation for the input box is a user entering the query information into the input box via voice command, the query information only includes audio.

[0024] Each input query is used to generate a query task. For example, an AI model generates the query task based on the query information. If the query information is a question, asking for an answer to a question, such as "Will it rain tomorrow?", then the query task retrieves the corresponding question result; that is, the query result is the question result corresponding to the question information. If the query information is a request, requesting relevant information, such as "Please introduce the specialties of location A", then the query task retrieves the corresponding request content; that is, the query result is the request content corresponding to the request information. If the query information is a statement, such as "seagull", then the query task retrieves the corresponding descriptive content; that is, the query result is the descriptive content corresponding to the statement information.

[0025] Step 220: Display at least two decomposed tasks generated based on the query information in a timeline format; wherein at least two decomposed tasks are used to complete the query task.

[0026] During the execution of a query task, the AI ​​model breaks it down into at least two independently executable decomposed tasks. These at least two decomposed tasks are structured task plans generated by the AI ​​model based on its analysis of the user's query intent. Each decomposed task is used to implement a step within the query task. By executing at least two decomposed tasks, the query task is executed, and after the completion of at least two decomposed tasks, the query task is completed, generating the corresponding query results.

[0027] For example, at least two decomposed tasks generated based on query information are displayed in the form of a linear timeline. The linear timeline is a vertical list layout adapted to mobile terminal screens, used to arrange the at least two decomposed tasks generated based on query information in a single column from top to bottom according to the execution order of the at least two decomposed tasks, and to connect the decomposed tasks together with connecting lines.

[0028] The timeline format shows at least two decomposed tasks that constitute the task execution flow of the query task. Displaying at least two decomposed tasks in the timeline format intuitively shows the execution order of at least two decomposed tasks generated based on the query information, making each execution step of the query task more independent and transparent.

[0029] Figure 3 A schematic diagram of the task execution process is shown. Figure 3 Figure (1) shows the initial decomposition results of the query information. The initial decomposition interface 310 of the query information displays four decomposition tasks obtained based on the initial decomposition of the query information. These four decomposition tasks are arranged from top to bottom in the form of a linear timeline. Figure 3 Figures (2) to (5) show the results of multiple decompositions of the query information. The multiple decomposition interface 320 of the query information displays multiple decomposition tasks obtained from multiple decompositions of the query information. These multiple decomposition tasks are arranged from top to bottom in the form of a linear timeline.

[0030] Step 230: Display the status information corresponding to at least two decomposition tasks. The status information corresponding to the decomposition tasks is used to indicate the execution status of the decomposition tasks.

[0031] For example, during the execution of a query task, status information corresponding to at least two decomposition tasks is displayed. The process of executing a query task is the overall process of executing at least two decomposition tasks, and during the overall execution of at least two decomposition tasks, status information corresponding to at least two decomposition tasks is displayed. The at least two decomposition tasks may include the decomposition tasks that define the at least two decomposition tasks.

[0032] During the execution of different decomposition tasks, the status information corresponding to at least two decomposition tasks is updated and displayed. For example, during the execution of the first decomposition task, the status information for the first decomposition task is displayed to indicate that its execution status is "in progress," while the status information for the other decomposition tasks is displayed to indicate that their execution status is "not yet started." During the execution of the second decomposition task, the status information for the first decomposition task is displayed to indicate that its execution status is "completed," the status information for the second decomposition task is displayed to indicate that its execution status is "in progress," and the status information for the other decomposition tasks is displayed to indicate that their execution status is "not yet started."

[0033] Step 240: After the query task is completed, display the query results corresponding to the query information.

[0034] The query results corresponding to the query information are the result obtained by integrating the execution results of at least two decomposed tasks, and represent the AI ​​model's response to the query information. For example, the query results for the query information include at least one of the following forms of information: text, images, audio, video, tables, and charts. When the query results for the query information include multiple forms of information, they can also be considered as a comprehensive result report.

[0035] like Figure 3 Figure (6) shows the query results of the query information. The query results interface 330 of the query information displays the query results 315 corresponding to the query information.

[0036] The technical solution provided in this application displays at least two decomposed tasks generated based on the query information in a timeline format after the query information is entered in the user interface. This allows users to clearly see how the system breaks down complex query tasks into multiple decomposed tasks, making the task decomposition process transparent, facilitating user understanding of the system's working logic, and enhancing user trust in the AI ​​system. Furthermore, the use of a vertical linear list layout to display at least two decomposed tasks in a timeline format fully utilizes the vertical screen space of the terminal device, allowing users to easily browse the complete task execution process by simply scrolling up and down. Compared to displaying mind maps on the terminal device, the technical solution provided in this application does not require a large display space to show each task node, reducing the display space needed for the task execution process and making it easier to adapt to different terminal device screens. Therefore, users can browse the complete task execution process without performing complex operations such as zooming or panning, reducing the difficulty of interactive operations. In addition, by displaying the status information corresponding to at least two decomposed tasks during the execution of the query task, the execution status of the decomposed tasks is visualized, allowing users to intuitively perceive the execution status of each decomposed task based on the status information, reducing the difficulty of identifying the execution status.

[0037] Step 220 will be described below.

[0038] In some embodiments, at least two decomposition task information are displayed sequentially based on the execution order of at least two decomposition tasks, with each decomposition task information used to indicate a decomposition task.

[0039] The execution order of at least two decomposed tasks is obtained by the AI ​​model through analysis of the task content of at least two decomposed tasks. This order indicates the sequential execution order of the at least two decomposed tasks. Based on this execution order, the at least two decomposed tasks are displayed in a timeline format. This is achieved by displaying the corresponding decomposed task information, ensuring that the information for each of the at least two decomposed tasks is arranged according to their execution order. For example, if decomposed tasks a, b, and c are generated based on query information, then the timeline will display the decomposed task information for task a, b, and c in that order, respectively.

[0040] Task decomposition information is used to indicate the decomposition task and its execution status, which can be one of the following: completed, in progress, or not yet started. Task decomposition information can consist solely of text information, indicating the task and its execution status; alternatively, it can include both text and graphic information, using a combination of both to indicate the task and its execution status. For example, it can use text information to indicate the task and graphic information to indicate its execution status, or it can use both text and graphic information to indicate the task's execution status.

[0041] By displaying information about at least two decomposed tasks sequentially in the user interface based on their execution order, the timeline display logic is demonstrated, making the execution steps in the task execution process more intuitive and transparent.

[0042] In some embodiments, task node icons corresponding to at least two decomposed task information are displayed, as well as connecting lines connecting the task node icons; wherein, the task node icons are used to mark the decomposed tasks, and the connecting lines are used to indicate the execution order of at least two decomposed tasks.

[0043] Each decomposed task includes a task node icon within its decomposed task information. The task node icon is a graphic representation of the decomposed task and is used to mark it. Each connecting line connects two task node icons. In the timeline, the decomposed task corresponding to the task node icon above the connecting line is executed before the decomposed task corresponding to the task node icon below the connecting line.

[0044] like Figure 3 As shown in Figure (1), the decomposition task information includes the name of the decomposition task and the corresponding task node icon. The task node icon is located before the name of the decomposition task and is used to mark the decomposition task. The preliminary decomposition interface 310 for querying information displays the task node icons corresponding to the four decomposition tasks and three connecting lines connecting the four task node icons. These three connecting lines are used to indicate the execution order of the four decomposition tasks. For example, the connecting line 303 between the task node icon 301 of "Planning" and the task node icon 302 of "Route Planning" is used to indicate that the execution order of the decomposition task of "Planning" is before the execution order of the decomposition task of "Route Planning".

[0045] By displaying task node icons to mark decomposed tasks, the connecting lines that link the icons in the timeline visually demonstrate the execution order of two decomposed tasks. These connecting lines then form the entire query process for the query task, improving the transparency of the decomposed tasks.

[0046] In some embodiments, task node icons have different display styles, which are used to indicate different execution states of the decomposed task.

[0047] The display style of a task node icon includes its shape and color. The shape of a task node icon includes, but is not limited to, circles, squares, and triangles. Task node icons can have different display styles, including different shapes, different colors, or both.

[0048] Different display styles are used to indicate different execution states of the decomposed task; that is, the decomposed task will display different styles depending on its execution state. Optionally, the execution state of the decomposed task can be indicated solely by the display shape of the task node icon. For example, a circular task node icon indicates that the execution state of the decomposed task is complete; a square task node icon indicates that the execution state is in progress; and a triangle task node icon indicates that the execution state has not yet started. Optionally, the execution state of the decomposed task can also be indicated solely by the display color of the task node icon. For example, a task node icon in the first color indicates that the execution state of the decomposed task is complete; a task node icon in the second color indicates that the execution state is in progress; and a task node icon in the third color indicates that the execution state has not yet started. Optionally, the execution state of the decomposed task can also be indicated by both the display shape and the display color of the task node icon. For example, a task node icon is displayed as a circle in the first color to indicate that the task is completed; a task node icon is displayed as a square in the second color to indicate that the task is being executed; and a task node icon is displayed as a triangle in the third color to indicate that the task has not yet started.

[0049] Typically, task node icons have the same display shape, and their display color indicates different execution states of the decomposed task; that is, the execution state of the decomposed task is indicated solely by the display color of the task node icon. For example... Figure 3As shown in Figure (1), the task node icons corresponding to the four decomposed tasks are all circular. However, the display color of the task node icon 301 for "making a plan" is different from that of the task node icon 302 for "making a route". This indicates that the execution status of the decomposed task for "making a plan" is different from that of the decomposed task for "making a route".

[0050] In some embodiments, at least two decomposition tasks have a hierarchical structure. Task node icons belonging to the same level have the same display shape, while task node icons belonging to different levels have different display shapes. The display color of the task node icons is used to indicate different execution states of the decomposition tasks. That is, the execution state of a decomposition task is indicated only by its display color, while its level is indicated by its display shape. For example, task node icons for first-level decomposition tasks can be displayed as circular icons, those for second-level tasks as square icons, and those for third-level tasks as triangular icons.

[0051] By designing different display styles for task node icons, users can intuitively see the execution status of each decomposed task based on the display style of each task node icon in the timeline, making the execution status of decomposed tasks visible and improving the display effect of querying the task execution process.

[0052] In some embodiments, the connecting lines have different display styles, which are used to indicate that the decomposed tasks corresponding to the task node icons pointed to by the connecting lines are in different execution states.

[0053] For example, the connecting lines can be displayed as solid or dashed lines. When the connecting line is solid, the execution status of the decomposed task corresponding to the task node icon pointed to by the connecting line can be either completed or in progress. When the connecting line is dashed, the execution status of the decomposed task corresponding to the task node icon pointed to by the connecting line is not yet started. A solid connecting line preceding a dashed connecting line indicates that the execution status of the decomposed task corresponding to the task node icon is in progress, while a solid connecting line preceding a solid connecting line indicates that the execution status of the decomposed task corresponding to the task node icon is completed. Figure 3As shown in Figure (1), the solid line 303 is located before the dashed line 304. The execution status of the decomposition task corresponding to the task node icon 302 of "Itinerary Route Formulation" pointed to by the solid line 303 is "In Progress". The execution status of the decomposition task corresponding to the task node icon of "Generate Travel Plan Results" pointed to by the dashed line 304 is "Not Started". The execution status of the decomposition task corresponding to the task node icon 301 of "Formulate Plan" pointed to by the implementation line before the solid line 303 is "Completed".

[0054] In some embodiments, at least two decomposition task information are displayed sequentially based on the execution order and hierarchical structure of at least two decomposition tasks; wherein, the hierarchical structure is used to define the parent-child relationship between decomposition tasks, the decomposition task information corresponding to decomposition tasks belonging to the same level is displayed in alignment, the decomposition task information corresponding to decomposition tasks belonging to different levels is displayed out of alignment, and below the decomposition task information corresponding to the parent task, the decomposition task information corresponding to at least one of the child tasks of the parent task is displayed respectively.

[0055] At least two decomposition tasks include at least one level, and each level includes at least one decomposition task. The first-level decomposition task is obtained by breaking down the query information; the second-level decomposition task is obtained by breaking down the first-level decomposition task; the third-level decomposition task is obtained by breaking down the second-level decomposition task, and so on, resulting in at least two decomposition tasks containing multiple levels. The first level is higher than the second level, the second level is higher than the third level, and the third level is higher than the levels obtained from subsequent decompositions.

[0056] The parent-child relationship between decomposition tasks exists between two levels. If at least two decomposition tasks include only one level, then there is no parent-child relationship between them. If at least two decomposition tasks include multiple levels, then there is a parent-child relationship between them. The decomposition task in the higher level of the two levels is taken as the parent task, and the decomposition tasks in the lower level are taken as child tasks. The decomposition task in the higher level that is the parent task is further broken down to obtain the child tasks of the parent task, including at least one decomposition task. At least one of the child tasks of the parent task is located in the lower level. For example, the first level includes decomposition task a and decomposition task b. Decomposition task a is broken down to obtain decomposition task c and decomposition task d. Decomposition tasks c and d are decomposition tasks of the second level, where decomposition task a is the parent task of decomposition tasks c and d, and decomposition tasks c and d are child tasks of decomposition task a.

[0057] For example, after a user enters query information, the query information is initially broken down into first-level decomposed tasks, which are displayed on the user interface in a timeline format. During the execution of a first-level decomposed task, that task is further broken down into subtasks, and second-level decomposed tasks are added to the user interface in a timeline format. Similarly, during the execution of a second-level decomposed task, it is broken down into subtasks, and third-level decomposed tasks are added to the user interface in a timeline format. This process continues until the lowest-level decomposed task cannot be further broken down. Based on the execution order and hierarchical structure of at least two decomposed tasks, the user interface displays information for at least two decomposed tasks sequentially. Newly added decomposed tasks are displayed sequentially in the middle of the timeline, allowing users to directly see the process of "decomposing as the task is executed," enhancing the real-time feel and transparency of the interaction.

[0058] like Figure 3 As shown in Figure (2), the first-level decomposition task 308 is decomposed to obtain the second-level decomposition task 309 and decomposition task 311. The second-level decomposition task 309 is decomposed to obtain the third-level decomposition task 312. The third-level decomposition task 312 is a subtask of the second-level decomposition task 309. The second-level decomposition task 311 is decomposed to obtain the third-level decomposition task 313. The third-level decomposition task 313 is a subtask of the second-level decomposition task 311. Figure 3 Figure (3) shows a schematic diagram of the interface when the decomposition tasks of each level of the structure are first executed. Figure 3 Figure (4) shows a schematic diagram of the interface when the decomposition tasks of each hierarchical structure are about to be completed. Figure 3 Figure (5) shows a schematic diagram of the interface when the decomposition tasks of each level are completed.

[0059] Based on the hierarchical structure of at least two decomposed tasks, information on at least two decomposed tasks is displayed. Decomposed task information belonging to the same level is displayed aligned, while decomposed task information belonging to different levels is displayed in a staggered manner. For example, decomposed task information belonging to a sub-task is indented by a preset distance compared to decomposed task information belonging to a parent task, and is displayed below it. Thus, the first-level decomposed task is displayed without indentation, the second-level decomposed task is indented by a preset distance, the third-level decomposed task is indented by double the preset distance, and so on. For example, if the first level includes decomposition tasks a and b, and decomposition task a is further broken down into decomposition tasks c and d, and decomposition task c is further broken down into decomposition tasks e and f, then decomposition tasks a and b are displayed aligned, decomposition tasks c and d are displayed aligned, decomposition tasks e and f are displayed aligned, decomposition tasks c and d are indented one unit to the right, and decomposition tasks e and f are indented two units to the right.

[0060] like Figure 4 As shown, the task decomposition display interface 400 displays at least two decomposition tasks for the query task. The query task is initially decomposed to obtain a first-level decomposition task 401. The first-level decomposition task 401 is further decomposed to obtain a second-level decomposition task 402. The second-level decomposition task is then further decomposed to obtain a third-level decomposition task 403. The first-level decomposition task 401 is the parent task of the second-level decomposition task 402, and the second-level decomposition task 402 is a child task of the first-level decomposition task 401. The second-level decomposition task 402 is the parent task of the third-level decomposition task 403, and the third-level decomposition task 403 is a child task of the second-level decomposition task 402. The decomposed tasks at the first level are displayed in alignment, the decomposed tasks at the second level are displayed in alignment, and the decomposed tasks at the third level are displayed in alignment. Furthermore, the decomposed tasks at the second level are displayed with indentation based on the decomposed tasks at the first level, and the decomposed tasks at the third level are displayed with indentation based on the decomposed tasks at the second level.

[0061] By using a hierarchical structure based on at least two decomposed tasks, information on at least two decomposed tasks is displayed in a staggered manner. This makes it easy to see which decomposed tasks belong to different levels and the parent-child relationship between them. The hierarchical structure of each decomposed task is clearer, making it easier for users to understand the execution logic and progress of the task execution process from the timeline.

[0062] In some embodiments, the display format of decomposition task information corresponding to decomposition tasks belonging to different levels is different. For example, the display font size of the decomposition task information corresponding to the parent task is larger than the display font size of the decomposition task information corresponding to the child task of the parent task.

[0063] In some embodiments, the sub-tasks of the parent task have corresponding expand / collapse controls. When the sub-task information corresponding to at least one sub-task of the parent task is hidden, in response to a first operation on the expand / collapse controls, the sub-task information corresponding to at least one sub-task of the parent task is displayed below the sub-task information corresponding to the parent task. When the sub-task information corresponding to at least one sub-task of the parent task is displayed, in response to a second operation on the expand / collapse controls, the sub-task information corresponding to at least one sub-task of the parent task is de-displayed below the sub-task information corresponding to the parent task.

[0064] The expand / collapse control allows you to expand and collapse the information of at least one sub-task corresponding to a child task of a parent task. When the sub-task information of at least one sub-task corresponding to a child task of a parent task is hidden, the expand / collapse control only allows you to expand this information. In this case, the first operation is to initiate the expansion of the sub-task information, which can be a user clicking the expand / collapse control. Conversely, when the sub-task information of at least one sub-task corresponding to a child task of a parent task is displayed, the expand / collapse control only allows you to de-display this information. In this case, the second operation is to initiate the de-display of this information, which can also be a user clicking the expand / collapse control.

[0065] By providing expand and collapse controls, the subtask breakdown information can be expanded when the user needs it and hidden when the user doesn't need it. This adapts to the information storage needs of mobile devices with limited screen space and avoids directly displaying all subtask breakdown information, which would occupy too much screen space.

[0066] Step 230 will be described below.

[0067] In some embodiments, the status information corresponding to the decomposed task includes at least one of the following: a task node icon and a status text label; the task node icon is used to mark the decomposed task, and the display style of the task node icon is used to indicate the execution status of the decomposed task; the status text label is used to indicate the execution status of the decomposed task in text form.

[0068] The display style of task node icons includes their shape and color. Optionally, the execution status of a task can be indicated solely by its shape, color, or both. Please refer to the above description for details.

[0069] The text displayed in the status text label indicates the execution status of the decomposed task, such as... Figure 3 As shown in Figure (1), the text displayed in the status text label 305, which reads "Completed", indicates that the execution status of the decomposition task is completed; the text displayed in the status text label 306, which reads "In Progress", indicates that the execution status of the decomposition task is being executed; and the text displayed in the status text label 307, which reads "Not Started", indicates that the execution status of the decomposition task is not started.

[0070] By displaying task node icons and status text labels in the status information corresponding to the decomposed tasks, users can not only know the execution status of the decomposed tasks based on the display style of the task node icons, but also intuitively see the execution status of the decomposed tasks based on the text displayed in the status text labels. Thus, even if users forget the execution status corresponding to different display styles of task node icons, they can still identify the execution status of each decomposed task through the status text labels, reducing the difficulty of task status identification and improving the convenience of user use.

[0071] In some embodiments, the color of the task node icon matches the execution status indicated by the status text label. When the task node icon corresponding to the decomposed task is a first color, the status text label indicates that the execution status of the decomposed task is completed; when the task node icon corresponding to the decomposed task is a second color, the status text label indicates that the execution status of the decomposed task is being executed; and when the task node icon corresponding to the decomposed task is a third color, the status text label indicates that the execution status of the decomposed task has not started.

[0072] The display color of the task node icon matches the execution status of the corresponding sub-task, and the status text label also matches the execution status of the corresponding sub-task. Therefore, the display color of the task node icon matches the status text label. For example, when the task node icon corresponding to the sub-task is in the first color, the status text label for the sub-task is "Completed"; when the task node icon is in the second color, the status text label indicates that the execution status of the sub-task is "In Progress"; and when the task node icon is in the third color, the status text label is "Not Started".

[0073] This application does not limit the first, second, and third colors. For example, if the first color is green, the second color is blue, and the third color is gray, then when the task node icon corresponding to the decomposed task is green, the status text label corresponding to the decomposed task is used to indicate that the execution status of the decomposed task is completed; when the task node icon corresponding to the decomposed task is blue, the status text label corresponding to the decomposed task is used to indicate that the execution status of the decomposed task is being executed; and when the task node icon corresponding to the decomposed task is gray, the status text label corresponding to the decomposed task is used to indicate that the execution status of the decomposed task has not started.

[0074] During the execution of different decomposition tasks, at least two decomposition tasks are displayed with different status information, that is, the task node icons corresponding to at least two decomposition tasks are displayed in different colors and with different status text labels. For example, during the execution of the first decomposition task, the task node icon corresponding to the first decomposition task is displayed in blue, and the status text label is "In Progress," while the task node icons corresponding to other decomposition tasks are displayed in gray, and the status text label is "Not Started." During the execution of the second decomposition task, the task node icon corresponding to the first decomposition task is displayed in green, and the status text label is "Completed," while the task node icon corresponding to the second decomposition task is displayed in blue, and the status text label is "In Progress," while the task node icons corresponding to other decomposition tasks are displayed in gray, and the status text label is "Not Started."

[0075] By matching the display color of the task node icon with the status text label, the execution status of the task can be identified based on the display color of the task node icon or the status text label. Furthermore, the display color of the task node icon and the status text label can be updated synchronously during the query of task execution to avoid user misidentification.

[0076] In some embodiments, the status information corresponding to the decomposition task further includes: progress information, which is used to indicate the execution progress of the decomposition task.

[0077] When the status information corresponding to a decomposed task is used to indicate that the task is in execution, the status information may further include progress information. Progress information includes at least one of the following: a progress icon and a progress text label. The progress icon can be a circular progress bar or a long bar, indicating the progress of the decomposed task through the progress percentage displayed on the progress bar. The progress text label can indicate the progress of the decomposed task through a percentage text. When the status information corresponding to a decomposed task is used to indicate that the task is completed or has not yet started execution, the status information corresponding to the decomposed task does not include progress information.

[0078] Optionally, when the status information corresponding to the decomposed task indicates that the execution status of the decomposed task is "in progress," the progress information corresponding to the decomposed task is displayed. The progress information can range from 0 to 100%. When the progress information is 100%, the status information indicates that the execution status of the decomposed task is updated to "completed." Alternatively, when the status information indicates that the decomposed task is "in progress" and the decomposed task has achieved progress, the progress information is displayed. Therefore, when the execution progress of the decomposed task is 0, the progress information is not displayed; once the decomposed task begins to achieve progress, the execution progress is not 0, and the progress information can be displayed.

[0079] like Figure 3 As shown in Figure (3), progress information 314 corresponding to the decomposed task is displayed in the multiple decomposition interface 320 of the query information. Progress information 314 is used to indicate that the execution progress of the decomposed task is 86%.

[0080] By displaying progress information in the status information corresponding to the decomposed tasks, users can accurately perceive the execution progress of each decomposed task, avoiding anxiety during task execution and affecting the user experience.

[0081] For example, the matching results between the state information corresponding to the decomposition tasks can be seen in Table 1 below.

[0082] Table 1

[0083] When the task node icon corresponding to a decomposed task is a solid green circle and the status label is "Completed," it indicates that the decomposed task has been successfully executed, and the user can perceive that this step has been completed. When the task node icon corresponding to a decomposed task is a solid blue circle and the status label is "In Progress," it indicates that the decomposed task is being executed, and the user can perceive that the system is processing this step. When the task node icon corresponding to a decomposed task is a solid blue circle and the status label is "In Progress XX%," it indicates that the decomposed task is being executed and its progress is displayed, and the user can perceive that this step is nearing completion. When the task node icon corresponding to a decomposed task is a hollow gray circle and the status label is "Not Started," it indicates that the decomposed task is awaiting execution, and the user can perceive that this step has not yet started. The connecting lines on the left side of the timeline are used to connect adjacent task node icons, and the user can perceive that the task node icons connected by the left connecting lines constitute a complete task execution flow.

[0084] Please refer to Figure 5 The diagram illustrates a flowchart of a task execution method provided in one embodiment of this application. The execution entity for each step of the method can be a server. The method may include at least one of the following steps 510-540: Step 510: Obtain the input query information. The query information is used to generate a query task, which is a task that outputs the query results corresponding to the query information.

[0085] Step 520: Based on the query information, the AI ​​model generates a list of decomposed tasks corresponding to the query information. The list of decomposed tasks includes at least two decomposed tasks, which are used to complete the query task.

[0086] For example, AI models include, but are not limited to, large language models (LLM), visual models, multimodal models, and other large models.

[0087] The decomposed task list includes execution and display information for at least two decomposed tasks. The AI ​​model will complete the query task based on the execution information of at least two decomposed tasks, and the terminal device will display at least two decomposed tasks in a timeline format based on the display information of at least two decomposed tasks.

[0088] In some embodiments, the decomposition task list includes at least one of the following: identification information corresponding to at least two decomposition tasks; names corresponding to at least two decomposition tasks; description text corresponding to at least two decomposition tasks; agents invoked by at least two decomposition tasks during execution; input parameters when at least two decomposition tasks invoke agents; and execution order corresponding to at least two decomposition tasks.

[0089] The name of the decomposition task is used to identify and display the task in the user interface. The description text of the decomposition task provides a detailed description of its content, which the user can view in the user interface. When a decomposition task is executed, the corresponding agent is invoked based on the input parameters to perform the task and obtain the execution result. The execution order of the decomposition tasks indicates which decomposition task a task depends on.

[0090] For example, the code structure for breaking down a task list can be as follows: type PlanItem struct { Idint`json:"id,omitempty"` / / Identifier information Namestring `json:"name"` / / Name Description string `json:"description"` / / Description text Agentstring `json:"agent,omitempty"` / / The intelligent agent to be invoked Inputsmap[string]any `json:"inputs,omitempty"` / / Input parameters AfterList[] string `json:"after_list,omitempty"` / / Execution order } type Plans []*PlanItem The data structure indicated by the code above can describe the hierarchical structure of the decomposed tasks and the calling dependencies.

[0091] The above-described task list decomposition allows the server to execute query tasks based on the execution information, and the client to display a timeline based on the display information, forming the basis for task execution and display.

[0092] Step 530: Execute each decomposition task in the decomposition task list using the AI ​​model to obtain the execution results corresponding to at least two decomposition tasks.

[0093] When decomposing tasks in the task list, a task node is automatically generated when a task is executed. During the execution of each task, tools may be called, such as online retrieval and image generation. After the tools retrieve data, the AI ​​model will integrate the retrieved data and generate the execution result corresponding to the task.

[0094] Step 540: Generate query results corresponding to the query information using an AI model based on the execution results of at least two decomposed tasks.

[0095] The AI ​​model aggregates the execution results of at least two decomposed tasks to generate query results corresponding to the query information.

[0096] By performing the above steps to break down the query task into at least two decomposed tasks, the execution results corresponding to the at least two decomposed tasks are obtained. Then, the at least two execution results are integrated to obtain the query results corresponding to the query information. This allows the final query results to fully obtain the retrieval data corresponding to each decomposed task, thereby improving the completeness and accuracy of the query results.

[0097] In some embodiments, step 520 includes at least one of sub-steps 521 to 523.

[0098] Sub-step 521: Generate new query information using an AI model based on the query information and historical query records. The historical query records are the query records that precede the current query information during the current query process.

[0099] Historical query records are the query records generated during the user's current conversation with the AI, including historical query information and the corresponding historical query results, but excluding query records from other conversations initiated by the user. Query information is the query information entered into the AI ​​during further conversations based on historical query records.

[0100] The new query information is generated by the AI ​​model combining historical query records to understand the query intent. It is used to correct typos and semantic order in the query information and to describe the user's query intent more comprehensively and completely.

[0101] Sub-step 522: Based on the new query information, the AI ​​model obtains the analysis results of the new query information. The analysis results are used to indicate whether the new query information depends on the retrieved data to generate the corresponding query results.

[0102] For example, the AI ​​model can categorize new query information to determine its query domain, such as scientific knowledge, geographical knowledge, or biological knowledge. Then, based on the query domain, the model analyzes the new query information to determine whether data retrieval is required to generate corresponding query results. If the new query information does not rely on retrieved data to generate its query results, it can be classified as a simple query; if it does rely on retrieved data to generate its query results, it can be classified as a complex query.

[0103] Sub-step 523: If the analysis results are used to indicate that new query information depends on the retrieved data to generate the query results corresponding to the query information, the AI ​​model generates a list of decomposed tasks corresponding to the query information based on the new query information.

[0104] This means that when the new query information is a complex query, it is not possible to directly generate the query results corresponding to the query information. The query task corresponding to the query information needs to be broken down into at least two decomposition tasks. Therefore, the AI ​​model generates a list of decomposition tasks corresponding to the query information based on the new query information.

[0105] By reconstructing query information by combining historical query records, the new query information can more comprehensively and completely describe the user's query intent. This improves the accuracy of analysis results when analyzing new query information, avoids analysis errors, prevents the execution of complex query processes on simple queries from affecting query efficiency, and avoids the execution of simple query processes on complex queries from leading to inaccurate query results.

[0106] In some embodiments, where the analysis results are used to indicate that new query information does not rely on retrieved data to generate query results corresponding to the query information, the AI ​​model generates query results corresponding to the query information based on the new query information.

[0107] This means that when the new query information is a simple query, the query results corresponding to the query information can be generated directly through the AI ​​model without having to break down the query task corresponding to the query information into at least two decomposed tasks. This simplifies the query steps to be performed for simple queries and improves the query efficiency of simple queries.

[0108] In some embodiments, the task decomposition list includes at least one hierarchical list of decomposed tasks, and the hierarchical structure in the task decomposition list is used to define the parent-child relationship between the decomposed tasks. Step 520 then further includes at least one sub-step from sub-steps 524 to 525.

[0109] Sub-step 524: Based on the query information, the AI ​​model generates a list of first-level decomposed tasks corresponding to the query information.

[0110] The first-level decomposition task list is used to provide the decomposition tasks that are initially displayed on the user interface.

[0111] Sub-step 525: During the execution of the decomposition task that serves as the parent task, the AI ​​model generates a list of decomposition tasks corresponding to the child tasks of the parent task based on the decomposition task that serves as the parent task.

[0112] During the execution of the decomposition task that serves as the parent task, an AI model is used to break down the parent task into its subtasks, generating a list of decomposition tasks corresponding to the subtasks of the parent task. After each generation of the list of decomposition tasks corresponding to the subtasks of the parent task, the user interface will add and display the subtasks of the parent task on the timeline based on the list of decomposition tasks corresponding to the subtasks of the parent task. These subtasks of the parent task include at least one decomposition task indicated by the list of decomposition tasks corresponding to the subtasks of the parent task.

[0113] By generating subtasks of the parent decomposition task during each execution of the parent task, a deeper understanding and refinement of each decomposition task can be achieved, enriching the task execution process. This allows for the retrieval of more data, improving the accuracy and completeness of query results.

[0114] In some embodiments, step 540 includes at least one of sub-steps 541 to 544.

[0115] Sub-step 541: Using an AI model, generate summary results corresponding to at least two decomposition tasks based on the execution results of each of the at least two decomposition tasks.

[0116] The AI ​​model integrates the execution results of at least two decomposed tasks to generate a summary result corresponding to the query information. The summary result is used to indicate the comprehensive execution result of at least two decomposed tasks.

[0117] Sub-step 542: Based on the summary results corresponding to at least two decomposition tasks, the AI ​​model generates reflection results for at least two decomposition tasks. The reflection results are used to indicate which decomposition tasks should be re-executed among the at least two decomposition tasks.

[0118] During the execution of at least two decomposition tasks, there may be instances where a decomposition task fails, data retrieval is incomplete, or the task is not properly broken down. In such cases, these decomposition tasks need to be re-executed. For cases of failed decomposition tasks or incomplete data retrieval, an AI model is used to generate new execution results for the re-execution task. For cases of inadequate task breakdown, an AI model is used to further break down the task, generating new sub-tasks that serve as the parent task.

[0119] Using an AI model, based on the aggregated results of at least two decomposition tasks, analyze the decomposition tasks with unsatisfactory execution results and generate reflection results for at least two decomposition tasks. These reflection results indicate which decomposition tasks need to be re-executed to generate the desired results and which need to be re-decomposed.

[0120] Sub-step 543: Based on the reflection results of at least two decomposed tasks, the AI ​​model generates new execution results corresponding to at least two decomposed tasks respectively.

[0121] Based on the reflection results, the AI ​​model generates new execution results for each of the at least one decomposed task that needs to be re-executed. Then, based on the reflection results, the AI ​​model generates new sub-tasks corresponding to each of the at least one decomposed task that serves as the parent task. Finally, the AI ​​model generates execution results for each of these new sub-tasks corresponding to the at least one decomposed task that serves as the parent task.

[0122] Based on the new execution results corresponding to at least one decomposed task generated from the re-execution result, the execution results of the new subtasks corresponding to at least one decomposed task decomposed from the re-execution task, and the execution results corresponding to the decomposed tasks that do not need to be re-executed, new execution results corresponding to at least two decomposed tasks are obtained.

[0123] Sub-step 544: Using an AI model, generate query results corresponding to the query information based on the new execution results corresponding to at least two decomposed tasks.

[0124] By reflecting on the execution results of at least two decomposition tasks through the above steps, and analyzing which decomposition tasks need to be re-executed, the result generation steps of some decomposition tasks can be re-executed, or the task decomposition steps can be re-executed. This improves the accuracy of the decomposition results of at least two decomposition tasks, and at the same time improves the accuracy of the execution results of each decomposition task, ultimately improving the accuracy of the query results.

[0125] Figure 6 This diagram illustrates the AI ​​model's processing flow for query information, showcasing an information query and processing process from receiving query information to finally outputting the query results. The system first understands the query information based on the query history, clarifying the user's needs. Based on this understanding, the query information is categorized to allow for different processing methods. Further understanding of the categorized query information leads to two paths: for simple or common questions, the AI ​​model can directly provide answers; for complex questions, a detailed mind map plan is needed, breaking down tasks into smaller steps. Developing the mind map plan requires decomposing the query task, breaking down complex problems into multiple sub-tasks (Decomposition Task 1, Decomposition Task 2, ..., Decomposition Task N). Some decomposed tasks can be further subdivided (e.g., Decomposition Task 2 can be subdivided into Decomposition Task 2.1, Decomposition Task 2.2, Decomposition Task 2.3, etc.). For each decomposed task, the corresponding tool is invoked for execution, and the execution results are obtained. The left side of the diagram lists the tools that may be invoked, including: document search: searching for relevant information in documents; web search: obtaining the latest information from the internet; browser webpage reading: using a browser to read webpage content and extract useful information; report generation: generating structured reports; chart generation: creating charts to aid in understanding and displaying information; image generation: generating relevant images; and webpage generation: generating webpage content. The results of each task are summarized to form a complete analysis result, which is then reflected upon and evaluated to check if it meets the query requirements. Finally, the processed result is output as the query result and fed back to the user. This flowchart details a systematic process for information query and processing, ensuring efficient and accurate answers to various user query needs through reasonable task decomposition and tool invocation.

[0126] In some embodiments, a directed acyclic graph (DAG) corresponding to the query information is constructed based on the decomposition task list corresponding to the query information. The DAG is used to indicate the dependency relationship between at least two decomposition tasks. Each task node in the DAG indicates a decomposition task, and directed edges in the DAG indicate the dependency relationship between two decomposition tasks connecting the directed edges. Each task node in the DAG is traversed. After the decomposition tasks corresponding to at least one of the preceding dependent nodes of the task node have been completed, the decomposition task corresponding to the task node is added to the ready queue. The decomposition task corresponding to the preceding dependent node is the preceding task of the decomposition task corresponding to the task node. An agent is then invoked to synchronously execute at least one decomposition task in the ready queue.

[0127] Based on the dependencies and execution order of the decomposition tasks in the decomposition task list corresponding to the query information, a Directed Acyclic Graph (DAG) is constructed. A DAG is a loop-free graph composed of task nodes and directed edges, used for hierarchical decomposition and dependency modeling of complex tasks. Each task node indicates a decomposition task, each directed edge indicates a dependency between two decomposition tasks connected by the edge, and each directed edge defines the execution order of the two decomposition tasks connected by the edge. The decomposition task corresponding to the task node pointed to by a directed edge depends on the completion of the decomposition task corresponding to the starting node of the directed edge. That is, the decomposition task corresponding to the task node pointed to by the directed edge can only be executed after the decomposition task corresponding to the starting node of the directed edge is completed. Therefore, the decomposition task corresponding to the starting node of the directed edge can be considered a prerequisite task for the decomposition task corresponding to the task node pointed to by the directed edge. For example, if a directed edge points from task node 1 to task node 2, task node 1 corresponds to decomposition task 1, and task node 2 corresponds to decomposition task 2, then decomposition task 1 is a prerequisite task for decomposition task 2.

[0128] Traverse each task node in the directed acyclic graph (DAG) and obtain the execution status of the decomposition tasks corresponding to at least one of the preceding dependent nodes for each task node; that is, obtain the execution status of at least one preceding task for each decomposition task. If at least one preceding task for a decomposition task has been completed, add this decomposition task to the ready queue. At this point, the ready queue may contain only one decomposition task or multiple decomposition tasks. Invoke the agent to synchronously execute at least one decomposition task in the ready queue.

[0129] By constructing a directed acyclic graph, the dependencies and execution order between various decomposed tasks are presented more clearly, making it easier to obtain decomposed tasks with dependencies from the directed acyclic graph. This allows for the synchronous execution of at least one decomposed task whose prerequisite dependencies are satisfied, ensuring the orderliness and efficiency of task execution.

[0130] Figure 7The flowchart illustrates the task decomposition and execution visualization process, primarily involving the use of LLM to decompose tasks and the execution of tasks by an agent. The user submits a complex task, which is then parsed using LLM to decompose the task intent into multiple decomposed tasks, generating a list of decomposed tasks. The system further parses the decomposed task list, determining the dependencies between the decomposed tasks and generating a Directed Acyclic Graph (DAG). Nodes in the DAG represent decomposed tasks, and directed edges represent the dependencies between them. All task nodes are initialized to a "waiting" state, indicating that these decomposed tasks have not yet started execution. The main execution loop then begins, continuously checking for incomplete decomposed tasks until all decomposed tasks are completed. During the execution loop, the system traverses the DAG, checking if all prerequisite tasks for each task node have been completed. If all prerequisite tasks are completed, the corresponding decomposed task is added to the ready queue; otherwise, it continues to wait for the prerequisite tasks to complete. The system checks if there are any tasks waiting to be executed in the ready queue. If the ready queue is not empty, it checks if all task nodes are completed. If the ready queue is empty but not all task nodes are completed, the system continues to wait or performs other processing. It then checks if all task nodes have completed execution. If all task nodes are completed, the process ends, and the structured final result is displayed. If there are still unfinished task nodes, it checks if there are any idle agents. It checks if there are any available agents to execute the decomposed tasks. If there are idle agents, it retrieves at least one task from the ready queue, assigns it to an agent for synchronous execution, and updates the task node's status to "in execution." If there are no idle agents, the system continues to wait until an agent becomes available. The agent executes the assigned task, returns the execution result and status upon completion, and updates the task node's status to "execution completed" or "execution failed" based on the agent's return result. The result is stored in the shared context as a result of a prerequisite for other decomposed tasks. A linear timeline is initialized, generating nodes according to the time sequence of task dependencies to display the time sequence of task execution, and synchronizing the real-time status of the task nodes to the linear timeline. This flowchart displays the state of each task node on a linear timeline, and presents a structured final result after all tasks are completed. It details the entire process of a complex task from submission to final result display, ensuring efficient and orderly task completion through steps such as task decomposition, dependency management, agent execution, and state synchronization.

[0131] In some embodiments, task status information corresponding to at least two decomposition tasks is generated. The task status information of the decomposition tasks includes at least one of the following: the creation time of the decomposition task, the execution status of the decomposition task, and the identification information corresponding to the subtasks of the decomposition task. During the execution of at least two decomposition tasks, the task status information corresponding to at least two decomposition tasks is updated.

[0132] The identification information corresponding to the subtasks of a decomposed task refers to the identification information corresponding to the subtasks obtained by breaking down the decomposed task as a parent task. The identification information corresponding to the subtasks refers to the identification information corresponding to at least one decomposed task contained in the subtask. Therefore, the identification information corresponding to the subtasks of a decomposed task refers to the identification information corresponding to at least one decomposed task obtained by breaking down the decomposed task as a parent task.

[0133] First, generate initial task status information corresponding to at least two decomposed tasks. Then, during the execution of one decomposed task, update the task status information corresponding to the at least two decomposed tasks. Then, during the execution of the next decomposed task, continue to update the task status information corresponding to the at least two decomposed tasks until all decomposed tasks are completed. The final task status information corresponding to the at least two decomposed tasks is used to indicate that the at least two decomposed tasks have been completed.

[0134] The task status information of the decomposed tasks can also include the execution duration and progress of the decomposed tasks. In the code, the status display of each node is updated in real time according to the running status of each decomposed task, and the status information of the front end is maintained.

[0135] For example, the code for the task breakdown process can be as follows: type Task struct { IDstring `json:"id"` / / Identifier information RootTaskID (string `json:"root_task_id"`) / / Identifier of the root task stack ParentId string `json:"parent_id"` / / Identifier information of the parent task Titlestring `json:"title"` / / Title Description string `json:"description,omitempty"` / / Description text CreatedAtint64`json:"created_at"` / / Creation time WaitTasks[]string `json:"wait_tasks,omitempty"` / / Wait tasks Status *TaskStatusInfo `json:"status,omitempty"` / / Task status ChildrenIds []string`json:"children_ids"` / / Identification information for subtasks } For example, the code for the task execution status can be as follows: type TaskStatus string const ( TaskStatusPending TaskStatus = "pending" / / Waiting for execution TaskStatusReady TaskStatus = "ready" / / Prerequisite tasks completed, ready. TaskStatusRunning TaskStatus = "running" / / Executing TaskStatusSucTaskStatus = "suc" / / Execution successful TaskStatusFailedTaskStatus = "failed" / / Execution failed TaskStatusStopped TaskStatus = "stopped" / / User stopped ) type TaskStatusInfo struct { StatusTaskStatus `json:"status"` Detailstring `json:"detail,omitempty"` / / Detailed description during execution BeginAtint64`json:"begin_at"` FinishedAt int64`json:"finished_at"` `Progressstring `json:"progress"` / / Progress percentage, such as "95%` } By continuously updating and generating task status information corresponding to at least one other decomposed task during the execution of at least two decomposed tasks, the mobile terminal can update and display the status information of each decomposed task in real time, and inform the user synchronously, thus ensuring the timeliness and accuracy of the front-end update display.

[0136] Figure 8 This diagram illustrates the real-time status push sequence between the mobile and server sides, primarily focusing on the processing of user query tasks, including task generation, execution, and progress updates. The mobile device initiates a network connection request, which the server receives and establishes, preparing for subsequent communication. Then, the mobile device sends a user query request to the server. The server receives the query request and uses LLM (Limited Language Management) to perform intent parsing of the query information, breaking down the query task into multiple decomposed tasks and generating a decomposed task list. The server pushes this initial task list to the mobile device. Upon receiving the initial task list, the mobile device renders an initial linear timeline based on the received list, displaying the initial status and execution order of each decomposed task. During task execution, the server dynamically generates new decomposed tasks and pushes the newly generated task nodes to the mobile device, which appends the new task nodes to the end of the linear timeline. During task execution, the server periodically calculates and pushes the task progress percentage to the mobile device. The mobile device updates the display of the corresponding task node based on the received progress percentage, labeling it "In Progress XX%". When a task node completes or fails, the server pushes a status change message to the mobile device. The mobile device then updates the color of the corresponding task node based on the received status change message to visually display the completion or failure status of each decomposed task. This sequence diagram details the interaction flow between the mobile device and the server when processing user-queried tasks. Through mechanisms such as task decomposition, dynamic execution, and progress updates, it ensures that users can understand the execution status of tasks in real time.

[0137] In some embodiments, the system establishes a persistent connection via WebSocket to push the task status to the mobile device in real time, as shown in the code below: type TaskSocketData struct { SessionIdstring`json:"session_id"` TaskList[]Task`json:"task_list,omitempty"` / / Newly added task node TaskPatches []JSONPatch `json:"task_patches,omitempty"` / / Incremental state update } By employing an incremental update mechanism based on JSON Patch, only data fragments related to status changes and progress updates are pushed, significantly reducing mobile network traffic consumption and achieving efficient network transmission, making it particularly suitable for mobile network environments. For example, the status update mechanism for task nodes can be seen in Table 2 below.

[0138] Table 2

[0139] When adding a new task node, the server pushes the completed task node to the mobile device. The amount of data required for network transmission is moderate. When the progress percentage of the decomposed task needs to be updated, the server pushes progress information to the mobile device. The amount of data required for network transmission is small. When the status of the decomposed task needs to be changed, the server pushes status information to the mobile device. The amount of data required for network transmission is small.

[0140] The following are embodiments of the apparatus described in this application, which can be used to execute the embodiments of the method described in this application. For details not disclosed in the apparatus embodiments of this application, please refer to the embodiments of the method described in this application.

[0141] Please refer to Figure 9 This diagram illustrates a block diagram of a task execution apparatus according to an embodiment of this application. The apparatus has the function of implementing the above-described task execution method; this function can be implemented in hardware or by hardware executing corresponding software. The apparatus can be the terminal device described above, or it can be installed within a terminal device. For example... Figure 9 As shown, the device 900 may include: an input display module 910, a task display module 920, a status display module 930, and a result display module 940.

[0142] The input display module 910 is used to display the input query information, which is used to generate a query task, and the query task is a task that outputs the query results corresponding to the query information.

[0143] The task display module 920 is used to display at least two decomposed tasks generated based on the query information in the form of a timeline; wherein the at least two decomposed tasks are used to complete the query task.

[0144] The status display module 930 is used to display the status information corresponding to the at least two decomposition tasks, and the status information corresponding to the decomposition tasks is used to indicate the execution status of the decomposition tasks.

[0145] The result display module 940 is used to display the query results corresponding to the query information after the query task is completed.

[0146] In some embodiments, the task display module 920 is configured to: Based on the execution order of the at least two decomposed tasks, the information of at least two decomposed tasks is displayed in sequence, and each decomposed task information is used to indicate one of the decomposed tasks.

[0147] In some embodiments, the task display module 920 is configured to: Display the task node icons corresponding to the at least two decomposed task information respectively, and the connecting lines connecting each of the task node icons; The task node icon is used to mark the decomposed task, and the connecting line is used to indicate the execution order of the at least two decomposed tasks.

[0148] In some embodiments, the task node icons have different display styles, and the different display styles are used to indicate different execution states of the decomposed tasks.

[0149] In some embodiments, the task display module 920 is configured to: Based on the execution order and hierarchical structure of the at least two decomposed tasks, the information of the at least two decomposed tasks is displayed in sequence; The hierarchical structure is used to define the parent-child relationship between the decomposition tasks. Decomposition task information corresponding to decomposition tasks at the same level is displayed in alignment, while decomposition task information corresponding to decomposition tasks at different levels is displayed out of alignment. Below the decomposition task information corresponding to the parent task, decomposition task information corresponding to at least one child task of the parent task is displayed.

[0150] In some embodiments, the decomposed tasks of the parent task are displayed with corresponding expand / collapse controls; the task display module 920 is used for: When the decomposition task information corresponding to at least one decomposition task of the parent task is hidden, in response to the first operation on the expand / collapse control, the decomposition task information corresponding to at least one decomposition task of the parent task is displayed below the decomposition task information corresponding to the parent task; or, When the decomposition task information corresponding to at least one decomposition task that is a child task of the parent task is displayed, in response to the second operation on the expand / collapse control, the decomposition task information corresponding to at least one decomposition task that is a child task of the parent task is undisplayed below the decomposition task information corresponding to the decomposition task of the parent task.

[0151] In some embodiments, the status information corresponding to the decomposed task includes at least one of the following: task node icon, status text label; The task node icon is used to mark the decomposed task, and the display style of the task node icon is used to indicate the execution status of the decomposed task; The status text label is used to indicate the execution status of the decomposed task in text form.

[0152] In some embodiments, the color of the task node icon matches the execution status indicated by the status text label; When the task node icon corresponding to the decomposed task is the first color, the status text label corresponding to the decomposed task is used to indicate that the execution status of the decomposed task is completed. When the task node icon corresponding to the decomposed task is the second color, the status text label corresponding to the decomposed task is used to indicate that the execution status of the decomposed task is being executed. When the task node icon corresponding to the decomposed task is in the third color, the status text label corresponding to the decomposed task is used to indicate that the execution status of the decomposed task is not started.

[0153] In some embodiments, the status information corresponding to the decomposed task further includes: progress information, which is used to indicate the execution progress of the decomposed task.

[0154] Please refer to Figure 10 This diagram illustrates a block diagram of a task execution apparatus according to an embodiment of this application. The apparatus has the function of implementing the above-described task execution method; this function can be implemented in hardware or by hardware executing corresponding software. The apparatus can be the server described above, or it can be located in a terminal device. Figure 10 As shown, the device 1000 may include: an input acquisition module 1010, a task generation module 1020, a task execution module 1030, and a result generation module 1040.

[0155] The input acquisition module 1010 is used to acquire input query information, which is used to trigger the large language model AI model to execute a query task, and the query task is to output the query result corresponding to the query information.

[0156] The task generation module 1020 is used to generate a list of decomposed tasks corresponding to the query information based on the query information using the AI ​​model. The list of decomposed tasks includes at least two decomposed tasks, which are used to complete the query task.

[0157] The task execution module 1030 is used to execute each of the decomposition tasks in the decomposition task list through the AI ​​model to obtain the execution results corresponding to the at least two decomposition tasks respectively.

[0158] The result generation module 1040 is used to generate query results corresponding to the query information based on the execution results corresponding to the at least two decomposed tasks through the AI ​​model.

[0159] In some embodiments, the task generation module 1020 is configured to: The AI ​​model generates new query information based on the query information and historical query records, where the historical query records are those that precede the current query information during the current query process. The AI ​​model obtains an analysis result of the new query information based on the new query information. The analysis result is used to indicate whether the new query information depends on the retrieved data to generate the query result corresponding to the query information. When the analysis results indicate that the new query information depends on the retrieved data to generate the query results corresponding to the query information, the AI ​​model generates a list of decomposed tasks corresponding to the query information based on the new query information.

[0160] In some embodiments, the task generation module 1020 is configured to: When the analysis results indicate that the new query information does not rely on the retrieved data to generate the query results corresponding to the query information, the AI ​​model generates the query results corresponding to the query information based on the new query information.

[0161] In some embodiments, the decomposition task list includes at least one of the following: The identification information corresponding to each of the at least two decomposition tasks; The names corresponding to the at least two decomposition tasks; The description text corresponding to each of the at least two decomposed tasks; The intelligent agents invoked respectively when the at least two decomposition tasks are executed; The input parameters when the at least two decomposition tasks call the agent respectively; The execution order corresponding to the at least two decomposed tasks.

[0162] In some embodiments, the task decomposition list includes at least one decomposed task list corresponding to each level, and the hierarchical structure of the task decomposition list is used to define the parent-child relationship between the decomposed tasks; the task generation module 1020 is used to: The AI ​​model generates a first-level decomposed task list corresponding to the query information based on the query information. During the execution of the decomposition task that serves as the parent task, the AI ​​model generates a list of decomposition tasks corresponding to the child tasks of the parent task based on the decomposition task that serves as the parent task.

[0163] In some embodiments, the result generation module 1040 is configured to: The AI ​​model generates a summary result corresponding to the at least two decomposed tasks based on the execution results of the at least two decomposed tasks respectively. The AI ​​model generates reflection results for the at least two decomposition tasks based on the summary results corresponding to the at least two decomposition tasks. The reflection results are used to indicate which decomposition tasks among the at least two decomposition tasks need to be re-executed. The AI ​​model generates new execution results corresponding to the at least two decomposed tasks based on the reflection results of the at least two decomposed tasks. The AI ​​model generates query results corresponding to the query information based on the new execution results corresponding to the at least two decomposed tasks.

[0164] In some embodiments, the task execution module 1030 is configured to: Based on the decomposed task list corresponding to the query information, a directed acyclic graph corresponding to the query information is constructed. The directed acyclic graph is used to indicate the dependency relationship between the at least two decomposed tasks. Each task node in the directed acyclic graph is used to indicate a decomposed task. The directed edges in the directed acyclic graph are used to indicate the dependency relationship between two decomposed tasks connected by the directed edges. Traverse each task node in the directed acyclic graph. After the decomposition tasks corresponding to at least one of the preceding dependent nodes of the task node have been completed, add the decomposition task corresponding to the task node to the ready queue. The decomposition task corresponding to the preceding dependent node is the preceding task of the decomposition task corresponding to the task node. Invoke the agent to synchronously execute at least one decomposition task in the ready queue.

[0165] In some embodiments, the task execution module 1030 is configured to: Generate task status information corresponding to the at least two decomposed tasks respectively. The task status information of the decomposed tasks includes at least one of the following: the creation time of the decomposed task, the execution status of the decomposed task, and the identification information corresponding to the subtasks of the decomposed task. During the execution of the at least two decomposed tasks, the task status information corresponding to the at least two decomposed tasks is updated respectively.

[0166] It should be noted that the apparatus provided in the above embodiments is only illustrated by the division of the above functional modules when implementing its functions. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the content structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0167] Please refer to Figure 11 This diagram illustrates a structural block diagram of a computer device 1100 provided in one embodiment of this application. The computer device 1100 can be any electronic device capable of data computation, processing, and storage. The computer device 1100 can be used to implement the task execution method provided in the above embodiments.

[0168] Typically, computer device 1100 includes a processor 1101 and a memory 1102.

[0169] Processor 1101 may include one or more processing cores, such as a quad-core processor, an octa-core processor, etc. Processor 1101 may be implemented using at least one hardware form selected from DSP (Digital Signal Processing), FPGA (Field Programmable Gate Array), and PLA (Programmable Logic Array). Processor 1101 may also include a main processor and a coprocessor. The main processor, also known as a CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, processor 1101 may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, processor 1101 may also include an AI (Artificial Intelligence) processor, which is used to handle computational operations related to machine learning.

[0170] The memory 1102 may include one or more computer-readable storage media, which may be non-transitory. The memory 1102 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In some embodiments, the non-transitory computer-readable storage media in the memory 1102 are used to store a computer program configured to be executed by one or more processors to implement the above-described task execution method.

[0171] Those skilled in the art will understand that Figure 11 The structure shown does not constitute a limitation on the computer device 1100 and may include more or fewer components than shown, or combine certain components, or use different component arrangements.

[0172] In an illustrative embodiment, a computer-readable storage medium is also provided, wherein a computer program is stored in the storage medium, and the computer program, when executed by a processor of a computer device, implements the above-described task execution method. Optionally, the above-described computer-readable storage medium may be ROM (Read-Only Memory), RAM (Random Access Memory), CD-ROM (Compact Disc Read-Only Memory), magnetic tape, floppy disk, and optical data storage device, etc.

[0173] In an exemplary embodiment, a computer program product is also provided, comprising a computer program stored in a computer-readable storage medium. A processor of a computer device reads the computer program from the computer-readable storage medium and executes the computer program, causing the computer device to perform the task execution method described above.

[0174] It should be noted that this application may display prompt interfaces, pop-ups, or output voice prompts before and during the collection of user data. These prompt interfaces, pop-ups, or voice prompts are used to inform users that their data is being collected. This ensures that the application only begins the steps for collecting user data after receiving confirmation from the user regarding the prompt interface or pop-up; otherwise (i.e., without user confirmation), the steps for collecting user data end, meaning no user data is collected. In other words, all user data collected by this application is processed strictly in accordance with the requirements of relevant national laws and regulations. The informed consent or separate consent of the data subject is obtained only with the user's consent and authorization. Subsequent data use and processing are conducted within the scope of laws, regulations, and the data subject's authorization, and the collection, use, and processing of relevant user data must comply with the relevant laws, regulations, and standards of the relevant countries and regions.

[0175] It should be understood that "multiple" as used herein refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. Furthermore, the step numbers described herein are merely illustrative of one possible execution order. In some other embodiments, the steps may not be executed in numerical order, such as two steps with different numbers being executed simultaneously, or two steps with different numbers being executed in the reverse order of the illustration. This application does not limit this.

[0176] The above description is merely an exemplary embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A task execution method, characterized in that, The method includes: The input query information is displayed, which is used to generate a query task. The query task is a task that outputs the query results corresponding to the query information. The query information is presented in a timeline format, showing at least two decomposed tasks generated based on the query information; wherein the at least two decomposed tasks are used to complete the query task. Display the status information corresponding to the at least two decomposition tasks respectively, wherein the status information corresponding to the decomposition tasks is used to indicate the execution status of the decomposition tasks; After the query task is completed, the query results corresponding to the query information are displayed.

2. The method according to claim 1, characterized in that, The at least two decomposed tasks generated based on the query information are displayed in a timeline format, including: Based on the execution order of the at least two decomposed tasks, the information of at least two decomposed tasks is displayed in sequence, and each decomposed task information is used to indicate one of the decomposed tasks.

3. The method according to claim 2, characterized in that, The method further includes: Display the task node icons corresponding to the at least two decomposed task information respectively, and the connecting lines connecting each of the task node icons; The task node icon is used to mark the decomposed task, and the connecting line is used to indicate the execution order of the at least two decomposed tasks.

4. The method according to claim 3, characterized in that, The task node icons have different display styles, and the different display styles are used to indicate different execution states of the decomposed tasks.

5. The method according to any one of claims 2 to 4, characterized in that, The step of sequentially displaying information about at least two decomposed tasks based on their execution order includes: Based on the execution order and hierarchical structure of the at least two decomposed tasks, the information of the at least two decomposed tasks is displayed in sequence; The hierarchical structure is used to define the parent-child relationship between the decomposition tasks. Decomposition task information corresponding to decomposition tasks at the same level is displayed in alignment, while decomposition task information corresponding to decomposition tasks at different levels is displayed out of alignment. Below the decomposition task information corresponding to the parent task, decomposition task information corresponding to at least one child task of the parent task is displayed.

6. The method according to claim 5, characterized in that, As a decomposed task of the parent task, it displays corresponding expand / collapse controls; the method further includes: When the decomposition task information corresponding to at least one decomposition task of the parent task is hidden, in response to the first operation on the expand / collapse control, the decomposition task information corresponding to at least one decomposition task of the parent task is displayed below the decomposition task information corresponding to the parent task; or, When the decomposition task information corresponding to at least one decomposition task that is a child task of the parent task is displayed, in response to the second operation on the expand / collapse control, the decomposition task information corresponding to at least one decomposition task that is a child task of the parent task is undisplayed below the decomposition task information corresponding to the decomposition task of the parent task.

7. The method according to any one of claims 1 to 6, characterized in that, The status information corresponding to the decomposed task includes at least one of the following: task node icon, status text label; The task node icon is used to mark the decomposed task, and the display style of the task node icon is used to indicate the execution status of the decomposed task; The status text label is used to indicate the execution status of the decomposed task in text form.

8. The method according to claim 7, characterized in that, The color of the task node icon matches the execution status indicated by the status text label; When the task node icon corresponding to the decomposed task is the first color, the status text label corresponding to the decomposed task is used to indicate that the execution status of the decomposed task is completed. When the task node icon corresponding to the decomposed task is the second color, the status text label corresponding to the decomposed task is used to indicate that the execution status of the decomposed task is being executed. When the task node icon corresponding to the decomposed task is in the third color, the status text label corresponding to the decomposed task is used to indicate that the execution status of the decomposed task is not started.

9. The method according to claim 7 or 8, characterized in that, The status information corresponding to the decomposed task also includes: progress information, which is used to indicate the execution progress of the decomposed task.

10. A task execution method, characterized in that, The method includes: The system obtains input query information, which is used to generate a query task. The query task is a task that outputs the query results corresponding to the query information. Based on the query information, the system uses an artificial intelligence (AI) model to generate a list of decomposed tasks corresponding to the query information. The list of decomposed tasks includes at least two decomposed tasks, which are used to complete the query task. The AI ​​model executes each of the decomposition tasks in the decomposition task list to obtain the execution results corresponding to the at least two decomposition tasks respectively. The AI ​​model generates query results corresponding to the query information based on the execution results of the at least two decomposed tasks.

11. The method according to claim 10, characterized in that, The step of generating a list of decomposed tasks corresponding to the query information using the AI ​​model includes: The AI ​​model generates new query information based on the query information and historical query records, where the historical query records are those that precede the current query information during the current query process. The AI ​​model obtains an analysis result of the new query information based on the new query information. The analysis result is used to indicate whether the new query information depends on the retrieved data to generate the query result corresponding to the query information. When the analysis results indicate that the new query information depends on the retrieved data to generate the query results corresponding to the query information, the AI ​​model generates a list of decomposed tasks corresponding to the query information based on the new query information.

12. The method according to claim 11, characterized in that, The method further includes: When the analysis results indicate that the new query information does not rely on the retrieved data to generate the query results corresponding to the query information, the AI ​​model generates the query results corresponding to the query information based on the new query information.

13. The method according to any one of claims 10 to 12, characterized in that, The list of decomposed tasks includes at least one of the following: The identification information corresponding to each of the at least two decomposition tasks; The names corresponding to the at least two decomposition tasks; The description text corresponding to each of the at least two decomposed tasks; The intelligent agents invoked respectively when the at least two decomposition tasks are executed; The input parameters when the at least two decomposition tasks call the agent respectively; The execution order corresponding to the at least two decomposed tasks.

14. The method according to any one of claims 10 to 13, characterized in that, The task decomposition list includes at least one level of decomposed task list, and the hierarchical structure of the task decomposition list is used to define the parent-child relationship between the decomposed tasks. The step of generating a list of decomposed tasks corresponding to the query information using the AI ​​model includes: The AI ​​model generates a first-level decomposed task list corresponding to the query information based on the query information. During the execution of the decomposition task that serves as the parent task, the AI ​​model generates a list of decomposition tasks corresponding to the child tasks of the parent task based on the decomposition task that serves as the parent task.

15. The method according to claim 14, characterized in that, The step of generating query results corresponding to the query information using the AI ​​model based on the execution results of the at least two decomposed tasks includes: The AI ​​model generates a summary result corresponding to the at least two decomposed tasks based on the execution results of the at least two decomposed tasks respectively. The AI ​​model generates reflection results for the at least two decomposition tasks based on the summary results corresponding to the at least two decomposition tasks. The reflection results are used to indicate which decomposition tasks among the at least two decomposition tasks need to be re-executed. The AI ​​model generates new execution results corresponding to the at least two decomposed tasks based on the reflection results of the at least two decomposed tasks. The AI ​​model generates query results corresponding to the query information based on the new execution results corresponding to the at least two decomposed tasks.

16. The method according to any one of claims 10 to 15, characterized in that, The method further includes: Based on the decomposed task list corresponding to the query information, a directed acyclic graph corresponding to the query information is constructed. The directed acyclic graph is used to indicate the dependency relationship between the at least two decomposed tasks. Each task node in the directed acyclic graph is used to indicate a decomposed task. The directed edges in the directed acyclic graph are used to indicate the dependency relationship between two decomposed tasks connected by the directed edges. Traverse each task node in the directed acyclic graph. After the decomposition tasks corresponding to at least one of the preceding dependent nodes of the task node have been completed, add the decomposition task corresponding to the task node to the ready queue. The decomposition task corresponding to the preceding dependent node is the preceding task of the decomposition task corresponding to the task node. Invoke the agent to synchronously execute at least one decomposition task in the ready queue.

17. The method according to any one of claims 10 to 16, characterized in that, The method further includes: Generate task status information corresponding to the at least two decomposed tasks respectively. The task status information of the decomposed tasks includes at least one of the following: the creation time of the decomposed task, the execution status of the decomposed task, and the identification information corresponding to the subtasks of the decomposed task. During the execution of the at least two decomposed tasks, the task status information corresponding to the at least two decomposed tasks is updated respectively.

18. A task execution device, characterized in that, The device includes: An input display module is used to display input query information, which is used to generate a query task, and the query task is a task that outputs the query results corresponding to the query information. The task display module is used to display at least two decomposed tasks generated based on the query information in the form of a timeline; wherein the at least two decomposed tasks are used to complete the query task; A status display module is used to display the status information corresponding to the at least two decomposition tasks respectively, and the status information corresponding to the decomposition tasks is used to indicate the execution status of the decomposition tasks; The results display module is used to display the query results corresponding to the query information after the query task is completed.

19. A task execution device, characterized in that, The device includes: The input acquisition module is used to acquire input query information, which is used to generate a query task, and the query task is a task that outputs the query results corresponding to the query information. The task generation module is used to generate a list of decomposed tasks corresponding to the query information based on the query information using the AI ​​model. The list of decomposed tasks includes at least two decomposed tasks, which are used to complete the query task. The task execution module is used to execute each of the decomposed tasks in the decomposed task list through the AI ​​model, and obtain the execution results corresponding to the at least two decomposed tasks respectively. The result generation module is used to generate query results corresponding to the query information based on the execution results corresponding to the at least two decomposed tasks using the AI ​​model.

20. A computer device, characterized in that, The computer device includes a processor and a memory, the memory storing a computer program, the computer program being loaded and executed by the processor to implement the task execution method as described in any one of claims 1 to 9, or to implement the task execution method as described in any one of claims 10 to 17.

21. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which is loaded and executed by a processor to implement the task execution method as described in any one of claims 1 to 9, or to implement the task execution method as described in any one of claims 10 to 17.

22. A computer program product, characterized in that, The computer program product includes a computer program that is loaded and executed by a processor to implement the task execution method as described in any one of claims 1 to 9, or to implement the task execution method as described in any one of claims 10 to 17.

Images