Information interaction method and device, computer equipment and storage medium

Abstract

The invention relates to an information interaction method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: obtaining query content; displaying query related nodes; the query related node is used for representing a task generated in the process of obtaining the query result of the query content; executing the task represented by the query related node, and displaying an interactive question when the user needs to participate in interaction in the task execution process; the interactive questioning module is used for carrying out at least one of information confirmation and information inquiry aiming at the task; and in response to a reply operation for the interactive question, displaying reply content, the reply content being used for continuing to execute the task to obtain a task execution result. By adopting the method, the accuracy of the content generated by the intelligent agent can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computers, and in particular to an information interaction method and device, computer equipment and a storage medium. BACKGROUND

[0002] With the development of artificial intelligence technology, an intelligent dialogue system based on an artificial intelligence agent (AI Agent) has appeared. The artificial intelligence agent is an agent built on a large language model (LLM, Large Language Model) and has the ability to autonomously understand, perceive, plan, remember and use tools.

[0003] Currently, in the intelligent dialogue system, the randomness of the LLM can cause the artificial intelligence agent to make autonomous random decisions. For example, when the user inputs unclear query content, the artificial intelligence agent will randomly diverge to think about the problem, resulting in generating content that seems reasonable but does not meet the user's expectations, thereby reducing the accuracy of the content generated by the artificial intelligence agent. SUMMARY

[0004] Therefore, it is necessary to provide an information interaction method and device, computer equipment, computer readable storage medium and computer program product that can improve the accuracy of the generated content to solve the above technical problems.

[0005] In a first aspect, the present application provides an information interaction method, which comprises:

[0006] obtaining query content;

[0007] displaying a query-related node; the query-related node is used to represent a task generated in a query result process of obtaining the query content;

[0008] executing the task represented by the query-related node, and when user participation is required in the task execution process, displaying an interactive question; the interactive question is used for at least one of information confirmation and information follow-up for the task;

[0009] in response to a reply operation to the interactive question, displaying a reply content, which is used to continue executing the task to obtain a task execution result.

[0010] In a second aspect, the present application further provides an information interaction device, which comprises:

[0011] a query content obtaining module, configured to obtain query content;

[0012] a node display module configured to display a query-related node; the query-related node is configured to represent a task generated in a process of obtaining a query result of the query content;

[0013] a question interaction module configured to execute the task represented by the query-related node, and display an interactive question when user interaction is required in the process of executing the task; the interactive question is configured to at least one of the following: information confirmation and information follow-up question for the task; in response to a reply operation for the interactive question, display a reply content, the reply content is configured to continue executing the task to obtain a task execution result.

[0014] In one of the embodiments, the query-related node includes a task planning node; the node display module is configured to display the task planning node; the interactive question includes a planning confirmation question; the question interaction module is configured to execute the planning task represented by the task planning node, obtain a task planning result, and display a planning confirmation question for confirming the task planning result; the reply content includes a confirmation reply content; the question interaction module is configured to display the confirmation reply content for the task planning result in response to a reply operation for the planning confirmation question.

[0015] In one of the embodiments, the interactive question further includes a first follow-up question; the reply content further includes a first follow-up reply content; the question interaction module is further configured to display a first follow-up question for the confirmation reply content when the confirmation reply content represents a denial of the task planning result; display a first follow-up reply content for the first follow-up question in response to a reply operation for the first follow-up question; display a task planning result regenerated according to the first follow-up reply content, and display a planning confirmation question for the regenerated task planning result.

[0016] In one of the embodiments, the query-related node is a node in a mind map; the query-related node includes a task node in the mind map and a sub-node of the task node; the node display module is configured to display at least one task node; after displaying the at least one task node, display at least one sub-node of the task node for each task node;

[0017] the interactive question includes a second follow-up question; the question interaction module is configured to execute the task represented by the sub-node, and display a second follow-up question for information follow-up when information required for task execution is missing; the reply content further includes a second follow-up reply content; the question interaction module is configured to display a second follow-up reply content for the second follow-up question in response to a reply operation for the second follow-up question.

[0018] In one of the embodiments, the task status and execution time of the corresponding task are also displayed in the sub-nodes of the task node; the information interaction device further comprises a state time updating module, which is configured to continue executing the task based on the second follow-up reply content and dynamically display the execution time of the task in the node; when the task is executed, the task status of the task is changed from execution to execution end in the sub-nodes of the task node, and the dynamic updating of the execution time is stopped.

[0019] In one of the embodiments, the query-related node is displayed in the content display page; the asking interaction module is further configured to display an information interaction pop-up window in the content display page; and the interactive asking for the task is displayed in the information interaction pop-up window.

[0020] In one of the embodiments, the asking interaction module is further configured to, when multiple tasks represented by the query-related nodes are executed in parallel, for each task in the multiple tasks executed in parallel, when the user needs to participate in the interaction in the execution process of the current task, display the query-related node corresponding to the current task in the multiple query-related nodes, and display an information interaction pop-up window for the query-related node in the content display page.

[0021] In one of the embodiments, the content display page further displays an information interaction element; the asking interaction module is further configured to, in the execution process of the current task, display an information interaction pop-up window for the query-related node corresponding to the current task in response to a triggering operation on the information interaction element; and display input information in response to an information input operation in the information interaction pop-up window; the input information is used to supplement the information of the execution of the current task.

[0022] In one of the embodiments, the information interaction device further comprises a countdown module, which is configured to display a reply countdown when the interactive asking for the task is displayed; and if no reply content for the interactive asking is obtained until the reply countdown is zero, the interactive asking is skipped and the task is continued to be executed.

[0023] In one of the embodiments, the information interaction device further comprises a skipping module, which is configured to display an asking skipping element when the interactive asking for the task is displayed; and the interactive asking is skipped and the task is continued to be executed in response to a triggering operation on the asking skipping element.

[0024] In one of the embodiments, the interactive question further comprises a third follow-up question; the question interaction module is further configured to: when a task execution result of the task is obtained, and based on the task execution result, it is determined that the task execution fails, display a third follow-up question determined according to a reason for the task execution failure; in response to a reply operation to the third follow-up question, display a third follow-up reply content; the third follow-up reply content is used to re-execute the task.

[0025] In a third aspect, the present application further provides a computer device, comprising a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0026] obtaining query content;

[0027] displaying a query related node; the query related node is used to represent a task generated in a query result process of the query content;

[0028] executing the task represented by the query related node, and when user interaction is required in the task execution process, displaying an interactive question; the interactive question is used for at least one of information confirmation and information follow-up for the task;

[0029] in response to a reply operation to the interactive question, displaying a reply content, the reply content is used to continue executing the task to obtain a task execution result.

[0030] In a fourth aspect, the present application further provides a computer readable storage medium, the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the following steps:

[0031] obtaining query content;

[0032] displaying a query related node; the query related node is used to represent a task generated in a query result process of the query content;

[0033] executing the task represented by the query related node, and when user interaction is required in the task execution process, displaying an interactive question; the interactive question is used for at least one of information confirmation and information follow-up for the task;

[0034] in response to a reply operation to the interactive question, displaying a reply content, the reply content is used to continue executing the task to obtain a task execution result.

[0035] In a fifth aspect, the present application also provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the following steps:

[0036] obtaining query content;

[0037] displaying a query-related node; the query-related node is used to represent a task generated in a query result process of obtaining the query content;

[0038] executing the task represented by the query-related node, and when user interaction is required in the task execution process, displaying an interactive question; the interactive question is used for at least one of information confirmation and information follow-up for the task;

[0039] in response to a reply operation for the interactive question, displaying reply content, the reply content is used to continue executing the task to obtain a task execution result.

[0040] The above information interaction method, device, computer equipment, storage medium and computer program product, by displaying a query-related node, and the query-related node is used to represent a task generated in a query result process of obtaining query content, can facilitate understanding of the task executed by the intelligent agent through the query-related node. In the process of displaying the query-related node and executing the task represented by the query-related node, by displaying the interactive question, information confirmation and information follow-up can be performed based on the interactive question, thereby helping the intelligent agent to make correct decisions based on the reply content of the interactive question, and thus improving the accuracy of the content generated by the intelligent agent.

[0041] In a first aspect, the present application provides an information interaction method, which comprises:

[0042] receiving query content sent by a terminal;

[0043] generating a query-related node, and instructing the terminal to display the query-related node; the query-related node is used to represent a task generated in a query result process of obtaining the query content;

[0044] executing the task represented by the query-related node, and when user interaction is required in the task execution process, generating an interactive question; the interactive question is used for at least one of information confirmation and information follow-up for the task;

[0045] returning the interactive question for the task to instruct the terminal to display the interactive question;

[0046] obtaining reply content for the interactive question, and continuing to execute the task based on the reply content.

[0047] In a second aspect, the present application provides an information interaction device, which comprises:

[0048] a node generation module, configured to receive query content sent by a terminal, generate a query-related node, and instruct the terminal to display the query-related node; the query-related node is used to represent a task generated in a process of obtaining a query result of the query content;

[0049] a question generation module, configured to execute the task represented by the query-related node, and generate an interactive question when user interaction is required in the process of executing the task; the interactive question is used to at least one of the following: information confirmation and information follow-up inquiry for the task;

[0050] a reply acquisition module, configured to return the interactive question for the task to instruct the terminal to display the interactive question, acquire reply content for the interactive question, and continue to execute the task based on the reply content.

[0051] In one of the embodiments, the reply content submitted by the terminal is stored in a preset cache space; the reply acquisition module is configured to determine a question identifier of the interactive question; the preset cache space is searched for a key-value pair corresponding to the question identifier every preset time length; and the value in the key-value pair is extracted as the reply content of the interactive question when the value in the key-value pair is not empty.

[0052] In one of the embodiments, the information interaction device further comprises a waiting time length generation module, configured to, when the interactive question is generated, determine a complexity of the interactive question, predict a reply waiting time length of the interactive question according to the complexity, and trigger the terminal to display a reply countdown according to the reply waiting time length.

[0053] In one of the embodiments, the information interaction device further comprises a reanalysis module, configured to, when the reply content for the interactive question is not acquired until the reply countdown is zero, reanalyze the task, and continue to execute the task based on a result of the reanalysis.

[0054] In a third aspect, the present application provides a computer device, which comprises a memory and a processor; the memory stores a computer program; and the processor implements the following steps when executing the computer program:

[0055] receiving query content sent by a terminal;

[0056] generating a query-related node, and instructing the terminal to display the query-related node; the query-related node is used to represent a task generated in a process of obtaining a query result of the query content;

[0057] execute the task represented by the query-related node, and when user interaction is required in the process of executing the task, generate an interactive question; the interactive question is used for at least one of information confirmation and information follow-up for the task;

[0058] return the interactive question for the task to instruct the terminal to display the interactive question;

[0059] obtain reply content for the interactive question, and continue to execute the task based on the reply content.

[0060] In a fourth aspect, the present application further provides a computer readable storage medium, the computer readable storage medium has a computer program stored thereon, and the computer program is executed by a processor to implement the following steps:

[0061] receive query content sent by a terminal;

[0062] generate a query-related node, and instruct the terminal to display the query-related node; the query-related node is used to represent a task generated in a query result process of obtaining the query content;

[0063] execute the task represented by the query-related node, and when user interaction is required in the process of executing the task, generate an interactive question; the interactive question is used for at least one of information confirmation and information follow-up for the task;

[0064] return the interactive question for the task to instruct the terminal to display the interactive question;

[0065] obtain reply content for the interactive question, and continue to execute the task based on the reply content.

[0066] In a fifth aspect, the present application further provides a computer program product, the computer program product comprises a computer program, and the computer program is executed by a processor to implement the following steps:

[0067] receive query content sent by a terminal;

[0068] generate a query-related node, and instruct the terminal to display the query-related node; the query-related node is used to represent a task generated in a query result process of obtaining the query content;

[0069] execute the task represented by the query-related node, and when user interaction is required in the process of executing the task, generate an interactive question; the interactive question is used for at least one of information confirmation and information follow-up for the task;

[0070] returning an interactive question for the task to instruct the terminal to display the interactive question;

[0071] obtaining reply content for the interactive question, and continuing to execute the task based on the reply content.

[0072] The information interaction method, device, computer device, storage medium, and computer program product described above can trigger the terminal to display the query-related node by generating the query-related node. Since the query-related node is used to represent a task generated in a query result process of obtaining query content, the query-related node can facilitate understanding of the task executed by the intelligent agent. In the process of displaying the query-related node and executing the task represented by the query-related node, the interactive question can be generated to trigger the terminal to display the interactive question, and information confirmation and information follow-up questions can be performed based on the interactive question. Therefore, the intelligent agent can make correct decisions based on reply content of the interactive question, and the accuracy of content generated by the intelligent agent is improved. BRIEF DESCRIPTION OF DRAWINGS

[0073] Figure 1 An application environment diagram of an information interaction method in an embodiment;

[0074] Figure 2 A flowchart of an information interaction method in an embodiment;

[0075] Figure 3 A schematic diagram of a query content input page in an embodiment;

[0076] Figure 4 A schematic diagram of an interactive page in an embodiment;

[0077] Figure 5 A schematic diagram of an interactive question in an embodiment;

[0078] Figure 6 A schematic diagram of information interaction in an embodiment;

[0079] Figure 7 A schematic diagram of information interaction in another embodiment;

[0080] Figure 8 A schematic diagram of an interactive page in another embodiment;

[0081] Figure 9 A display diagram of a task state and execution time in an embodiment;

[0082] Figure 10 A display diagram of an information interaction pop-up window in an embodiment;

[0083] Figure 11 A flowchart of an information interaction method in another embodiment;

[0084] Figure 12 a schematic diagram of a mind map in one embodiment;

[0085] Figure 13 a schematic diagram of obtaining reply content in one embodiment;

[0086] Figure 14 a schematic diagram of an overall architecture of information interaction in one embodiment;

[0087] Figure 15 a structural block diagram of an information interaction device in one embodiment;

[0088] Figure 16 a structural block diagram of an information interaction device in another embodiment;

[0089] Figure 17 an internal structure diagram of a computer device in one embodiment. DETAILED DESCRIPTION

[0090] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not used to limit the present application.

[0091] The information interaction method provided by the embodiments of the present application can be applied to, for example, Figure 1The application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data required by the server 104 to process. The data storage system can be integrated on the server 104, or placed on the cloud or other servers. The terminal 102 can be used alone to execute the information interaction method provided in the embodiments of the application. The terminal 102 and the server 104 can also be used together to execute the information interaction method provided in the embodiments of the application. Taking the terminal 102 and the server 104 that can be used together to execute the information interaction method provided in the embodiments of the application as an example, the terminal 102 can send the query content to the server 104 in response to the input event of the query content. The server 104 generates at least one task corresponding to the query content based on the intelligent agent and the query content, and returns at least one query related node corresponding to the at least one task. The terminal 102 displays at least one query related node. The server 104 can execute the task corresponding to the node, and when the user needs to participate in the interaction in the task execution process, the server 104 generates an interaction question for the task. The terminal 102 can display the interaction question, so that the user can input reply content for the interaction question, and then the server 104 can continue to execute the task corresponding to the node based on the reply content input by the user. Among them, the terminal 102 can be, but is not limited to, various desktop computers, notebook computers, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things device can be a smart speaker, a smart television, a smart air conditioner, a smart vehicle device, etc. The portable wearable device can be a smart watch, a smart bracelet, a head-mounted device, etc. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers. The server 104 can also be a cloud server.

[0092] It should be noted that the "first", "second" and similar words used in the present application do not represent any order, quantity or importance, but are only used to distinguish different components. Unless the context clearly indicates otherwise, the singular form "one", "an" or "the" and the like similar words do not represent a quantity limitation, but represent the existence of at least one. The number of "multiple" or "multiple" mentioned in the embodiments of the present application refers to "at least two", for example, "multiple" refers to "at least two", and "multiple" refers to "at least two".

[0093] In order to clearly describe the technical solutions of the present application and facilitate the understanding of the technical solutions of the present application, the key concepts involved in the present application are explained as follows.

[0094] The query content refers to the content expected to be queried. The query content can include query content in any field expected to be answered, such as medical treatment, education, office work and shopping, etc. For example, "extract PPT video screenshot and identify title document".

[0095] The query-related node is used to represent a task generated in the process of obtaining a query result of the query content. For example, when multiple tasks need to be performed in the process of obtaining the query result, multiple query-related nodes can be generated, and each query-related node is used to represent a task.

[0096] The query-related node can be displayed in the form of a text stream, such as a task title, task description information, and task execution result of a task represented by a text description. Alternatively, the query-related node can be displayed in the form of a node in a mind map. The mind map, also known as a brain map, is a diagram that can show the ideas and processes of solving a problem. One query-related node represents one task, and each query-related node has its own task to complete. The connection relationship between the query-related nodes can represent at least one of the execution order of the tasks and the task inclusion relationship. For example, if the A query-related node is the parent node of the B query-related node, the task represented by the A query-related node can be executed before the B query-related node, or the task represented by the B query-related node can be obtained by splitting the task represented by the A query-related node. After all the tasks corresponding to the nodes in the mind map are executed, the query result of the query content is obtained. The query-related node can include a query content analysis node, a task planning node, a task node, and a sub-node. The tasks include a query content analysis task, a planning task, a basic task, a first-level sub-task of the basic task, and a second-level sub-task of the basic task.

[0097] The task represented by the query content analysis node is a query content analysis task, which is used to represent that the intelligent agent performs query content analysis to obtain a content analysis result. For example, when the query content is “I like hiking, and now I want to live in the mountains. I need to buy a tent. What purchase recommendations do you have?”, the content analysis result obtained by performing the query content analysis task can be “The user likes hiking and plans to go camping in the mountains. Please recommend a tent suitable for hiking and give purchase recommendations”.

[0098] The task represented by the task planning node is a planning task, which is used to represent that the intelligent agent performs task planning on the query content based on the content analysis result to obtain at least one basic task. For example, for the content analysis result described above, the at least one basic task planned by the intelligent agent includes a “popular hiking tent recommendation and comparison” basic task, a “user demand adaptability analysis” basic task, and a “integrate the comparison of goods and adaptability analysis content to generate a Markdown report” basic task.

[0099] The sub-tasks include a first-level sub-task, a second-level sub-task, a third-level sub-task, and the like. The sub-nodes include a first-level sub-node representing the first-level sub-task, a second-level sub-node representing the second-level sub-task, a third-level sub-node representing the third-level sub-task, and the like. The first-level sub-task is a sub-task generated by task decomposition of the base task. For example, for the base task of "popular tent recommendation and comparison for hiking", the base task can be decomposed into a first-level sub-task of "tent recommendation suitable for hiking" and a first-level sub-task of "tent comparison recommendation". The second-level sub-task refers to a sub-task generated by task decomposition of the first-level sub-task. For example, the first-level sub-task of "tent recommendation suitable for hiking" is decomposed into a sub-task. The third-level sub-task refers to a sub-task generated by task decomposition of the second-level sub-task.

[0100] The user interaction required in the task execution process refers to a situation that the agent encounters in the process of executing the task, and the situation needs to be provided with information, confirmed with selection, or authorized with operation by the user to continue. Whether the user interaction is required is specifically, for example, the following three scenarios: information missing type: the key data required by the agent to execute the task is not pre-installed and cannot be automatically obtained. For example, when executing the task of "checking business trip ticket", the "departure date", "destination city", or "passenger identity information" is missing. Decision selection type: there are multiple feasible branches or schemes for the task, and the next step needs to be determined according to the user's preference. For example, when executing the task of "generating a report", the agent provides three formats of "concise version", "detailed version", and "visualization version", and the user needs to select one of them; or when executing the planning task, the agent provides the task planning result for the user to confirm. Authorization operation type: the task involves external system calling, permission use, or actual impact, and the user needs to confirm the authorization. For example, when executing the task of "sending marketing email", the user needs to confirm "whether to send to 1000 customers" to avoid misoperation.

[0101] In one embodiment, as shown in Figure 2 , an information interaction method is provided. Taking the terminal in Figure 1 as an example, the method includes the following steps:

[0102] Step 202, acquiring query content.

[0103] Specifically, the user can input the query content into the intelligent dialogue system, for example, input "extract PPT video screenshot and identify title document".

[0104] In one of the embodiments, the query content can be input in the form of text. In another exemplary embodiment, the query content can also be input in the form of voice. Referring to Figure 3As shown, the query content input page 301 displays a query content input box 302, in which the query content "Recommendation of fragrance-free small package eyeglasses disinfection wet wipes" is input. Figure 3 A schematic diagram of a query content input page in an embodiment is shown.

[0105] In step 204, the query related nodes are displayed. The query related nodes are used to represent the tasks generated in the process of obtaining the query result of the query content.

[0106] Specifically, the terminal can display at least one query related node. In an embodiment, the terminal can display a query related node for representing a task each time a task is executed. For example, the terminal can display a query related node for representing a task in the form of a text stream each time a task is executed. For example, the terminal can display the task title, task description information, and task execution result of the task described by text in a node box in a page, and the task title, task description information, and task execution result of the task described by text in the node box constitute the query related node for representing the task.

[0107] Alternatively, the terminal can display a query related node for representing a task that has been executed and a query related node for representing a task that is being executed each time a task is executed. Alternatively, the terminal can display a query related node for representing a task that is being executed and a query related node for representing another task related to the task that is being executed each time a task is executed. Alternatively, the terminal can display all the query related nodes for representing the tasks that have been generated. For example, the terminal can display the query related nodes in the order of the execution of the tasks.

[0108] In an embodiment, the query related nodes can be displayed in the form of nodes in a mind map. For example, refer to Figure 4 The terminal can display a query content analysis node 401 and a task planning node 402 in the interactive page. Figure 4 A schematic diagram of an interactive page in an embodiment is shown.

[0109] In an embodiment, the query related nodes can include a task title and a task description. The task description is used to represent the summary content of the task content.

[0110] In an embodiment, the order of the query related nodes corresponds to the execution order of the tasks represented by the query related nodes. For example, the task planning node is displayed after the query content analysis node, so the content analysis task represented by the query content analysis node is executed first, and then the task planning node is displayed after the query content analysis node.

[0111] Step 206: Execute the task represented by the query node, and display interactive questions when user interaction is required during task execution; interactive questions are used for at least one of information confirmation and follow-up questions regarding the task.

[0112] Specifically, the terminal can invoke an intelligent agent to execute tasks represented by querying relevant nodes. For example, the task represented by querying relevant nodes can be input into the intelligent agent, which will then execute the task using models and tools to obtain the result. When the intelligent agent determines that user interaction is required during task execution, it can generate interactive questions for that task, thereby confirming or pursuing information based on these questions.

[0113] In one embodiment, the query-related nodes may also display task statuses, including at least one of "in execution," "execution completed," and "awaiting execution." The terminal can execute the task corresponding to the query-related node whose task status is "awaiting execution." For example, in the above example, when a query content parsing node and a task planning node are displayed, since the query content parsing node is in an execution completed state and the task planning node is in an "awaiting execution" state, the terminal can execute the planning task represented by the task planning node.

[0114] In one embodiment, during task execution, the intelligent agent can determine whether user interaction is required. For example, if user confirmation of information is required, then user interaction is determined, and the intelligent agent can generate an interactive question so that the terminal displays the interactive question. For example, refer to... Figure 4 When the agent executes the planning task represented by the task planning node and generates the task planning result, in order to confirm whether the generated task planning result meets the user's expectations, the agent can generate an interactive question 403 for information confirmation: "I will solve the problem according to the following approach. If you wish to adjust the plan, please let me know. 1. Based on user needs, search and recommend several disinfectant wipes that meet the criteria... Reject simulated or fabricated information. 2. Organize the filtered and recommended disinfectant wipe product information into Markdown format for easy reading and subsequent purchase by the user."

[0115] In one embodiment, during task execution, the intelligent agent can determine whether user interaction is required. For example, if user input is needed to supplement information, then user interaction is determined, and the intelligent agent can generate an interactive question so that the terminal displays the interactive question. For example, refer to... Figure 5When the agent executes the first-level subtask "retrieving product shopping links" represented by the first-level child node, since the product name and product model are missing, the agent can generate an interactive question 501 for further information: "Please provide the product name or model you want to query". Figure 5 A schematic diagram of interactive questioning in one embodiment is shown.

[0116] Step 208: In response to the reply to the interactive question, display the reply content. The reply content is used to continue executing the task to obtain the task execution result.

[0117] Specifically, when an interactive question is displayed, the user can reply to it, and the terminal can then display the reply. For example, see [reference]. Figure 4 When an interactive question is used to confirm information, the user's input response 404, "Confirm," can be displayed. For example, when an interactive question is used to follow up on information, the user's input response 502, "Product A, Model B," can be displayed.

[0118] When the terminal receives a response, it can input that response into the agent, which can then continue executing the task based on the response to obtain the task execution result. For example, in a task planning node, when an interactive question is used to confirm the task planning result, and the user inputs "confirm," the agent can determine that the currently generated task planning result meets the user's needs and thus use it as the final task planning result. If the user inputs "task planning result is incorrect," the agent can determine that the currently generated task planning result does not meet the user's needs and then regenerate the task planning result. As another example, in a first-level child node, when an interactive question is used to follow up on the name or model of a product, and the user inputs "Product A, Model B," the agent can continue searching for product shopping links based on the user's input of "Product A, Model B."

[0119] In a straightforward manner, at any stage of determining query results based on the query content, the agent can ask the user questions to request secondary confirmation or additional information, thereby helping the agent decide on the correct subsequent execution direction. For example, referencing... Figure 6 For task planning node 601, during the execution of the planning task represented by the task planning node, an interactive question 602 for confirming the task planning result can be displayed. After the user confirms the task planning result, for example, when the user enters "confirm", the intelligent agent can generate at least one basic task based on the task planning result, and then the terminal can display a task node 603 representing the basic task after the task planning node 601. Figure 6 A schematic diagram of information interaction in one embodiment is shown.

[0120] Furthermore, the agent can perform basic tasks. During the execution of these tasks, when the agent determines that information confirmation or follow-up questions are needed, it can generate interactive questions for such confirmation or follow-up. After the user inputs a response, the agent can continue performing the basic task based on that response. Upon completion of the basic task, the agent will reflect on whether the task has been resolved. If it determines that the task is unresolved, it will decompose the basic task to generate first-level subtasks, which will then be referenced... Figure 7 The terminal can display a first-level sub-node 702 representing the first-level sub-task after task node 701. The agent can execute the first-level sub-task. During the execution of the first-level sub-task, when the agent determines that information confirmation or follow-up inquiry is needed, it can generate an interactive question for information confirmation or follow-up inquiry, which the terminal can then display as interactive question 703. After obtaining the user's input response, the first-level sub-task can continue to be executed based on the response. This process continues until all query-related nodes in the mind map have been executed, at which point the query result is obtained. Figure 7 A schematic diagram of information interaction in another embodiment is shown.

[0121] In the aforementioned information interaction method, by displaying query-related nodes, which represent tasks generated during the process of obtaining query results, it is easy to understand the tasks performed by the agent through these nodes. During the process of displaying and executing the tasks represented by these nodes, interactive questions can be displayed to confirm and follow up on information. This helps the agent make correct decisions based on the responses to the interactive questions, thereby improving the accuracy of the content generated by the agent.

[0122] In one embodiment, querying related nodes includes task planning nodes; displaying query related nodes includes: displaying task planning nodes; interactive questions include planning confirmation questions; when the task represented by the execution node is required for user interaction during task execution, interactive questions for the task are displayed, including: executing the planning task represented by the task planning node, obtaining the task planning result, and displaying a planning confirmation question to confirm the task planning result; response content includes confirmation response content; in response to the response operation for the interactive question, displaying response content includes: in response to the response operation for the planning confirmation question, displaying confirmation response content for the task planning result.

[0123] Specifically, for ease of description, the interactive questions used to confirm information will be referred to as planning confirmation questions, and the responses to these questions will be called confirmation response content. The terminal can display task planning nodes and trigger the agent to execute the planning task represented by those nodes. Upon obtaining the task planning result, the agent can generate a planning confirmation question and trigger the terminal to display it. The planning confirmation question asks the user to confirm whether the task planning result meets their expectations. The user can respond to the planning confirmation question, for example, by entering "confirm" or "I think this plan is not accurate enough." The agent can then perform subsequent processing based on the user's response.

[0124] In this embodiment, by displaying a planning confirmation question for confirming information, the intelligent agent can obtain information for secondary confirmation from the user based on the planning confirmation question, and then make more accurate subsequent decisions based on the information for secondary confirmation.

[0125] In one embodiment, the method further includes: when the confirmation response content represents a denial of the task planning result, displaying a first follow-up question on the confirmation response content; in response to a reply operation on the first follow-up question, displaying a first follow-up response content on the first follow-up question; displaying a task planning result regenerated based on the first follow-up response content, and displaying a planning confirmation question on the regenerated task planning result.

[0126] Specifically, the intelligent agent can interact with the user in multiple rounds. For example, when the user inputs a confirmation response that represents a rejection of the task planning result, such as "I think this plan is not accurate enough," "This plan needs to be adjusted," or "This plan does not meet my expectations," the intelligent agent can generate an interactive question based on the confirmation response. This interactive question is called the first follow-up question, and the terminal can then display this first follow-up question. For example, the terminal can display "What parts of this planning result do you think need to be adjusted?" Furthermore, the user can reply to the first follow-up question, and the reply is called the first follow-up response. The intelligent agent can then regenerate the task planning result based on the first follow-up response, and the terminal can continue to display a planning confirmation question based on the regenerated task planning result.

[0127] In one embodiment, the terminal may display query-related nodes in a first area of ​​the interactive page and display the interaction content between the agent and the user in a second area of ​​the interactive page. For example, see reference... Figure 8 , Figure 8A schematic diagram of an interactive page in another embodiment is shown. The terminal can display query-related nodes in a first area 801 and display the interaction content between the agent and the user for at least one round in a second area 802. By displaying query-related nodes and interaction content between the agent and the user on the same page, it is easier for the user to associate the interaction content with the query-related nodes.

[0128] In one embodiment, the second area may also display a historical interaction content query button. When the user clicks the historical interaction content query button, the terminal can display the interaction content between the user and the agent within the historical time period. For example, the interaction content can be archived according to the query related nodes. Then, when the user clicks the historical interaction content query button, the terminal can display an archive table in the second area. The archive table may display mind map node markers. The user can then view the questions and answers between the agent and the user during the execution of the task represented by the query related node by querying the node markers. By archiving the interaction content, users can view the interaction content at historical moments through the archived results, thereby quickly recalling key information based on historical interaction content and assisting users in self-reflection and decision-making.

[0129] In the above embodiments, when a user expresses disagreement with the task planning result, a first follow-up question can be displayed. Based on this first follow-up question, further questions can be posed to the user, and a more accurate task planning result can be regenerated based on the response to the first follow-up question. This significantly improves the accuracy of the final generated task planning result.

[0130] In one embodiment, when the confirmation response indicates confirmation of the task planning result, the agent can take the currently generated task planning result as the final task planning result and generate at least one basic task based on the final task planning result. Then, the terminal can display the task nodes corresponding to each basic task.

[0131] In one embodiment, the query related node is a node in a mind map, which includes task nodes and child nodes of the task nodes in the mind map; displaying the query related node includes: displaying at least one task node; after displaying at least one task node, displaying at least one child node of the task node for each task node; executing the task represented by the query related node, and when user interaction is required during task execution, displaying an interactive question for the task, including: executing the task represented by each child node, and when the information required for task execution is missing, displaying a second follow-up question for further information; and displaying the response content in response to the response to the interactive question, including: displaying the second follow-up response content in response to the response to the second follow-up question.

[0132] Specifically, after obtaining the task planning results, at least one task node can be displayed based on the task planning results, for example, referring to... Figure 6 If the task planning result is "1. Based on user needs, retrieve and recommend several qualified disinfectant wipes... reject simulated or fabricated information; 2. Organize the filtered and recommended disinfectant wipe product information into Markdown format for easy user reading and subsequent purchase," then the task nodes generated based on this task planning result may include task node 603 corresponding to "1. Based on user needs, retrieve and recommend several qualified disinfectant wipes... reject simulated or fabricated information," and task node 603 corresponding to "2. Organize the filtered and recommended disinfectant wipe product information into Markdown format for easy user reading and subsequent purchase." Furthermore, the basic task represented by the task node can be executed, and at least one child node of the task node can be displayed after the task node.

[0133] For example, basic tasks can be performed by calling tools. These tools can be external functions, programs, APIs, or services called through predefined interfaces, with the aim of extending the core capabilities of the intelligent agent. Tools may include document conversion tools, web search tools, webpage reading tools, report generation tools, image generation tools, chart generation tools, webpage generation tools, etc. In this embodiment, when performing basic tasks, the process of calling a tool can be treated as a subtask. For example, if the basic task A represented by the task node involves a web search tool, the web search process can be treated as a first-level subtask of basic task A, and the corresponding first-level sub-nodes can be displayed after the task node. Another example is that basic tasks can also be performed by calling a large language model. For example, the complex basic task can be decomposed into easily executable first-level subtasks. For example, refer to... Figure 7For task node 701, the basic task corresponding to task node 701 is decomposed into first-level subtasks, and the first-level sub-node 702 corresponding to the first-level subtasks is displayed. It is easy to understand that the execution of the first-level subtask may also involve tool calls or task decomposition. The execution process of calling the tool can be regarded as the second-level subtask of the corresponding first-level subtask, or the subtasks obtained from the decomposition can be regarded as the second-level subtasks of the corresponding first-level subtasks.

[0134] Furthermore, for each child node, the task represented by the child node can be executed. For example, for the first-level child node 702, the first-level subtask represented by the first-level child node 702 can be executed. When information is missing during task execution, the terminal can display a second follow-up question for further information inquiry. For example, when executing the first-level subtask represented by the first-level child node 702, if the information required for the execution of the first-level subtask is missing, a second follow-up question can be displayed for further information inquiry.

[0135] Furthermore, the user can trigger a response to the second follow-up question, and the terminal can respond to the response to the second follow-up question by displaying the response content for the second follow-up question.

[0136] In this embodiment, since query-related nodes represent tasks generated during the process of obtaining query results, the overall process of the agent generating query results can be displayed in the form of a mind map, facilitating understanding of the agent's entire work arrangement and real-time progress. Furthermore, during task execution, when information required for task execution is lacking, interactive questions can be displayed to elicit further information. Based on these interactive questions, the necessary information can be obtained, allowing the task to continue execution more accurately.

[0137] In one embodiment, the child node also displays the task status and execution time of the corresponding task. The method further includes: continuing to execute the task based on the response to the second follow-up question, and dynamically displaying the execution time of the task in the node; when the task is completed, changing the task status of the task from "in execution" to "execution completed" in the child node, and stopping the dynamic updating of the execution time.

[0138] The task status is used to characterize the execution status of various tasks, including but not limited to at least one basic task for the query content, first-level subtasks, second-level subtasks, summary tasks corresponding to the summary process, and reflection tasks corresponding to the reflection process. The task status includes at least one of the following: executing, completed, and awaiting execution.

[0139] Execution time is used to indicate the actual time it takes for a task to be executed.

[0140] Specifically, child nodes can display task status and execution time. For example, see reference. Figure 9 The task status corresponding to the first-level child node 901 is "in execution," and the first-level child node 901 displays the "in execution" status indicator 902 and the execution time 903. Furthermore, when the task is completed, the task status in the child node is changed from "in execution" to "execution completed," and dynamic updates to the execution time are stopped. For example, when the first-level subtask represented by the first-level child node 904 is completed based on the second follow-up question's response, the "in execution" status indicator in the first-level child node 904 is changed to the "execution completed" status indicator 905, and dynamic updates to the execution time 906 are stopped. Figure 9 A schematic diagram showing the task status and execution time in one embodiment is shown.

[0141] In the above embodiments, displaying the task status in the query related nodes allows for an intuitive determination of the task's execution progress, facilitating the display of more detailed execution processes of the agent. Displaying the execution time in the query related nodes allows for an intuitive determination of the time spent executing the task.

[0142] In one embodiment, querying relevant nodes and displaying them on the content display page; displaying interactive questions for the task includes: displaying an information interaction pop-up on the content display page; and displaying interactive questions for the task in the information interaction pop-up.

[0143] Specifically, when user interaction is required during task execution, the terminal can display an information interaction pop-up window, such as a reference. Figure 4 An information interaction pop-up window (405) can be displayed on the content display page, and then interactive questions can be displayed in the information interaction pop-up window.

[0144] In one embodiment, when a response to an interactive question is received, the terminal can automatically close the pop-up window to prevent the information interaction pop-up window from obscuring the query-related nodes on the content display page.

[0145] In one embodiment, the user can drag the information interaction pop-up window to change its position on the content display page, and can also close the pop-up window as needed. The content display page can also display a shortcut button for the information interaction pop-up window; after the user closes the pop-up window, they can use the shortcut button to trigger the terminal to redisplay the pop-up window.

[0146] In the above embodiments, displaying interactive questions as pop-ups creates a visual focus by floating directly above the page content, guiding users to notice the questions and thus improving information reach. Furthermore, pop-ups can encourage immediate user interaction, thereby enhancing interaction efficiency.

[0147] In one embodiment, displaying an information interaction pop-up on the content display page includes: when multiple query-related nodes represent tasks in parallel, for each of the multiple tasks being executed in parallel, when user interaction is required during the execution of the current task, the query-related node corresponding to the current task is displayed differently among the multiple query-related nodes, and an information interaction pop-up for the distinguished query-related nodes is displayed on the content display page.

[0148] Specifically, multiple tasks represented by query-related nodes can be executed in parallel. For example, the tasks represented by each query-related node can be distributed to servers in a distributed system for execution. For each of the multiple tasks executed in parallel, when it is determined that user interaction is required during the execution of the current task, the terminal can display an interactive information pop-up window for the current task on the content display page. For example, for the task represented by query-related node A, when user interaction is required during the execution of the task represented by query-related node A, an interactive information pop-up window can be displayed above query-related node A. Furthermore, the terminal can also differentiate the display of query-related nodes corresponding to the current task. For example, in the above example, the text and boxes on query-related node A can be displayed in blue. Alternatively, query-related node A can be displayed with dynamic effects, including but not limited to flashing, scrolling, and color gradients.

[0149] In this embodiment, when multiple query-related nodes are executed in parallel, the query-related nodes that require user interaction are displayed in a different manner among the multiple query-related nodes, and information interaction pop-ups for the different displayed query-related nodes are displayed. This makes it easier for users to distinguish the interactive questions displayed in the information interaction pop-ups and confirm the query-related node to which the interactive question is directed.

[0150] In one embodiment, the content display page also displays information interaction elements; the method further includes: during the execution of the current task, in response to a trigger operation on the information interaction elements, displaying an information interaction pop-up window for querying relevant nodes corresponding to the current task; in response to an information input operation performed in the information interaction pop-up window, displaying input information; the input information is used to supplement the execution of the current task.

[0151] Specifically, during the execution of the current task, in addition to the terminal actively displaying information interaction pop-ups for information interaction, the user can also actively initiate information interaction. For example, the content display page also displays information interaction elements. When the user wishes to actively interact with the intelligent agent, the user can initiate a trigger operation on the information interaction element, and the terminal can then respond to the trigger operation by displaying the information interaction pop-up. For example, the information interaction element can be a shortcut display button for the information interaction pop-up. When the user clicks this shortcut display button, the terminal can display the information interaction pop-up. The user can then input information in the displayed information interaction pop-up, and the terminal can display the user's input information in the pop-up. For example, if the current task is a product link collection task, and the user finds that no product model has been specified, the user can actively initiate information interaction and enter the product model in the information interaction pop-up. When the terminal receives the user's supplementary input information, the terminal can send the input information to the intelligent agent, so that the intelligent agent can supplement the execution of the current task based on the user's supplementary input information.

[0152] In this embodiment, during the execution of a task, the user can actively supplement information through an information interaction pop-up window, thereby enabling the intelligent agent to make more accurate decisions based on the information supplemented by the user.

[0153] In one embodiment, the method further includes: displaying a response countdown when an interactive question for the task is displayed; if no response is received for the interactive question before the response countdown reaches zero, the interactive question is skipped and the task continues to be executed.

[0154] Specifically, when displaying interactive questions related to a task, a response countdown can also be shown, for example, in the reference section. Figure 10 The system can display a reply countdown of 1001 in the information interaction pop-up window, thus reminding the user to respond within the corresponding time period. Furthermore, if no response is received for the interactive question before the reply countdown reaches zero, the agent skips the interactive question and continues to execute the task. Figure 10 A schematic diagram of an information interaction pop-up window is shown in one embodiment.

[0155] Correspondingly, if a response is received before the countdown ends, the agent continues to perform the task based on that response.

[0156] In this embodiment, by displaying a reply countdown, users can be urged to reply as soon as possible, thereby improving the efficiency of obtaining reply content.

[0157] In one embodiment, the method further includes: displaying a question skipping element when an interactive question for the task is displayed; and skipping the interactive question and continuing to execute the task in response to a triggering operation on the question skipping element.

[0158] Specifically, when displaying interactive questions, you can also display question skipping elements, such as references. Figure 10 It can also display a skip button 1002. Users can skip the interactive question by asking a question, meaning they can choose not to answer the interactive question. When a user triggers an action on the question-skip element, such as clicking the skip button 1002, the agent can continue performing the task.

[0159] In this embodiment, by displaying a question skipping element, users can be given the right to choose, the interaction flexibility can be improved, users can avoid resistance due to being forced to answer questions, and the system can adapt to diverse usage scenarios.

[0160] In one embodiment, the interactive questioning further includes a third follow-up question; the method further includes: when the task execution result of the task is obtained, and it is determined that the task execution failed based on the task execution result, displaying a third follow-up question determined according to the reason for the task execution failure; in response to the reply operation to the third follow-up question, displaying the reply content of the third follow-up question; the reply content of the third follow-up question is used to re-execute the task.

[0161] Specifically, upon receiving the task execution result, the agent determines whether the problem has been solved. If the problem remains unsolved, the task is deemed a failure. For example, in the "product link collection" task, if no product links are collected, the problem of product link collection is considered unsolved, and the task is deemed a failure. The agent can then determine the reason for the failure and generate a third follow-up question. For instance, if the agent determines the failure is due to missing information, it can generate a third follow-up question to address the missing information. Similarly, if the agent determines the failure is due to a missing product name, it can generate a third follow-up question asking "Please output the product name," which is then displayed on the terminal. The user can respond to this third follow-up question, and the terminal displays the response. The agent can then re-execute the task based on the response; for example, it can re-collect product links based on the user-input product name.

[0162] In this embodiment, when a task fails to execute, a third follow-up question is displayed, and the response to the third follow-up question can be obtained based on the response. The task can then be re-executed based on the response, thereby improving the success rate of task execution.

[0163] In one embodiment, such as Figure 11 As shown, an information interaction method is provided, which can be applied to... Figure 1 Taking the server in the example, the following steps are included:

[0164] Step 1102: Receive the query content sent by the terminal.

[0165] Specifically, the server can receive query content sent by the terminal. This query content can include queries from any domain where the intelligent agent is expected to provide an answer, such as those related to healthcare, education, office work, and resource interaction. For example, a query could be "Extract screenshots from a PPT video and identify the title document."

[0166] Step 1104: Generate query-related nodes and instruct the terminal to display the query-related nodes; query-related nodes are used to represent the tasks generated during the process of obtaining the query results of the query content.

[0167] Specifically, after obtaining the query content, the server can generate query-related nodes. For example, it can generate query content parsing nodes, task planning nodes, and, after the user confirms the task planning results, task nodes. The server can then return the generated query-related nodes to the terminal to instruct the terminal to display them.

[0168] In one embodiment, since the query-related nodes are nodes in a mind map, and the mind map has a tree structure in the code, it has a root node, child nodes, and parent nodes. Furthermore, since each query-related node in the mind map represents a task, the structure of the query-related node (Task) generated by the server may include the task identifier (RootTaskId) corresponding to the root node, its own task identifier (TaskId), a list of dependent preceding task IDs (WaitTasks), a task title (Title), a task description (Description), a task creation time (CreatedAt), the task identifier (ParentId) corresponding to the parent node, the agent / executor handling the task (AgentInputs), the task's input parameters (AgentInputs), the task's dynamic type (DynamicType), the edit mode (EditMode), the subtasks contained in the current task (Children[]*Task), and the task status information (Status). If the task is completed, the task execution result can be stored in Status. Here, the task identifier corresponding to the node is the identifier of the task represented by the node.

[0169] In one embodiment, the server may also define a structure named TaskStatusInfo to record and describe the execution status information of a task. This structure works in conjunction with the preceding Task structure to specifically carry the task's status data throughout its lifecycle (such as whether it has started, completed, and its progress), facilitating the determination of the task's execution process. The TaskStatusInfo structure may include the task identifier (RootTaskId) corresponding to the root node, its own task identifier (TaskId), task status information (Status), the actual start time of the task (BeginAt), the task completion time (FinishedAt), the task progress (Progress), and the task execution result (Answer).

[0170] In simple terms, a tree (a mind map is a tree in code) can be constructed using its own task identifier (TaskId) and the task identifier of the root node (RootTaskId). This tree has a root node, child nodes, and a parent node. Based on the RootTaskId field, we can identify the root node of the tree; all nodes are assigned to tasks using the RootTaskId. (Tasks with the same RootTaskId belong to the same mind map). The root node's RootTaskId = TaskId, and it has no parent node. In short: by abstracting any step in the process of generating query results into a Task object, a query-related node can be displayed in the front-end interaction. Nodes are the smallest unit of subsequent user interaction.

[0171] In one embodiment, reference Figure 12 Mind maps can be represented as follows: Figure 11 The tree structure shown. Figure 12 A schematic diagram of a mind map in one embodiment is shown. Task 1 is the root node, Task 2 and Task 3 are child nodes of the root node, and Task 4 and Task 5 are parent nodes of Task 2.

[0172] Step 1106: Execute the task represented by the query related node, and generate interactive questions for the task when user interaction is required during task execution; interactive questions are used for at least one of information confirmation and information follow-up for the task.

[0173] Specifically, an intelligent agent can be deployed on the server. The intelligent agent will execute the task represented by the query-related nodes based on the large language model and tools. When user interaction is required during the task execution, the agent will generate interactive questions for the task. For example, it can generate planning confirmation questions to confirm the task planning results, or generate follow-up questions to ask for further information.

[0174] Step 1108: Return to the interactive prompt for the task to instruct the terminal to display the interactive prompt.

[0175] Step 1110: Obtain the response content to the interactive question, and continue to execute the task based on the response content.

[0176] Specifically, the server can return the interactive questions generated by the agent to the terminal, where the terminal can display the interactive questions and the user can then reply to them. The terminal can return the reply to the server, which can then input the received reply into the agent so that the agent can continue to perform its task.

[0177] In one embodiment, when any query-related node encounters a problem requiring user decision-making, follow-up questions can be triggered, and the interactive questions can be displayed above the query-related node to prompt the user to answer. Based on the user's feedback, the agent further adjusts its subsequent execution plan and continues to execute the task.

[0178] In one embodiment, each agent generates an execution result. After the agent completes its execution, there is a reflection node. The reflection node determines whether the task has solved the problem. If not, it may be due to insufficient information or the failure of some key steps. Further problem decomposition is needed, which may involve questioning the user for more specific information or breaking down the task into smaller sub-problems for resolving. Finally, a result-generating agent, such as a report generator or webpage generator, produces the final result.

[0179] In the aforementioned information interaction method, generating query-related nodes can trigger the terminal to display them. Since these nodes represent tasks generated during the process of obtaining query results, it is easy to understand the tasks performed by the agent through them. During the display and execution of the tasks represented by these nodes, generating interactive questions can trigger the terminal to display them. Information confirmation and follow-up questions can then be based on these interactive questions, helping the agent make correct decisions based on the responses to the interactive questions, thereby improving the accuracy of the content generated by the agent.

[0180] In one embodiment, the query-related nodes can be nodes in a mind map. Therefore, by generating query-related nodes, the overall process of the agent generating query results can be displayed in the form of a mind map, which makes it easier to understand the agent's entire work arrangement and real-time work progress.

[0181] In one embodiment, the response content submitted by the terminal is stored in a preset cache space; obtaining the response content for the interactive question includes: determining the question identifier of the interactive question; searching for the key-value pair corresponding to the question identifier in the preset cache space at preset intervals; and extracting the value in the key-value pair as the response content of the question when the value in the key-value pair is not empty.

[0182] Specifically, when the terminal receives a response from the user, it can send the response to the server, allowing the server to store the received response in a preset cache. For example, the response can be stored as key-value pairs, where the key is the question identifier and the value is the response. The agent can periodically search the cache for the corresponding key-value pair based on the question identifier, and if the value in the found key-value pair is not empty, it retrieves the value as the response.

[0183] In one embodiment, reference Figure 13 Specifically, the server can be a distributed server. After receiving the response, the terminal can send it to the storage server, which will then store it in the cache. The node server where the agent executing the task resides can poll the data in the cache to find the corresponding key-value pair based on the query identifier. If the value in the found key-value pair is not empty, it will extract the value as the response to the interactive query and continue executing the task based on the response. Figure 13 A schematic diagram illustrating the retrieval of response content in one embodiment is shown.

[0184] In the above embodiments, by storing the response content in the form of key-value pairs, the intelligent agent can accurately and quickly extract the corresponding response content from the cache space through the question identifier of the interactive question, thereby improving the efficiency and accuracy of the response content acquisition.

[0185] In one embodiment, the method further includes: determining the complexity of the interactive question when it is generated; predicting the response waiting time based on the complexity; and triggering the terminal to display a response countdown based on the response waiting time.

[0186] Specifically, after generating an interactive question, the agent can extract its textual features, such as the question's length, whether it contains technical terms, sentence structure complexity, and whether it's an open-ended question. Based on these textual features, the agent can determine the complexity of the interactive question, classifying it into simple, medium, and complex levels. Further, the agent determines the response waiting time based on the complexity level. For example, it can match different response waiting times to different complexity levels according to a pre-defined rule. Alternatively, the agent can input the complexity level into a duration prediction model trained on historical data to determine the response waiting time. When the agent obtains the response waiting time, it can send it to the terminal, allowing the terminal to start a real-time countdown based on the estimated time, dynamically displaying the remaining waiting time. If the response is completed ahead of schedule, the countdown is stopped and a notification is displayed.

[0187] In one embodiment, basic textual features of interactive questions can be extracted, such as the number of words and sentences in the question (generally, the longer the question, the more complex it is likely to be); secondly, the semantic depth of the interactive questions can be analyzed, including whether they contain professional terms, domain-specific vocabulary, and whether they involve multi-step reasoning or multi-layered logical connections; the sentence structure of the interactive questions can also be analyzed to determine whether they are nested interrogative sentences, multi-conditional compound sentences, or open-ended questions (not closed-ended questions like "yes / no"); finally, by combining these features, the questions are divided into simple, medium, and complex levels to quantify the complexity of the interactive questions.

[0188] In the above embodiments, this solution first determines the complexity of the interactive question, then predicts the waiting time for the user's response based on this, and triggers a countdown display on the terminal. This provides more precise rhythm control for machine interaction: on the one hand, the countdown can clarify the reasonable waiting boundary for the user's response, avoiding premature judgment of "user no response" or excessive waiting that consumes resources, thus optimizing the efficiency of the interaction process; on the other hand, the real-time countdown displayed on the terminal can provide users with clear time prompts, especially for complex questions (such as questions that require multiple steps of thought), helping users perceive the response rhythm, allocate thinking time reasonably, and reduce hasty or forgotten responses due to lack of time reference. Ultimately, this makes the interaction between the intelligent agent and the user more orderly and more in line with the expectations of both parties, balancing user experience and system resource utilization efficiency.

[0189] In one embodiment, the method further includes: if no response is received for the interactive question before the response countdown reaches zero, the task is re-analyzed, and the task continues to be executed based on the results of the re-analysis.

[0190] Specifically, when the terminal's response countdown reaches zero and no response has been received for the interactive question, the intelligent agent will automatically trigger a task re-analysis process: first, it will review the context and historical information of the interactive question, and then re-analyze the task based on factors such as the complexity of the original question and the user's past interaction habits. Based on the analysis results obtained from the re-analysis, it will adjust the task's execution strategy and then continue to advance the task execution according to the new strategy, ensuring that the interactive process can still flexibly adapt and continue to operate even if the user does not respond in time.

[0191] In the above embodiments, it is possible to effectively avoid the task from stalling due to the user's failure to respond in time, which reduces the ineffective use of system resources and makes the interaction process more flexible and fault-tolerant. At the same time, re-analyzing the task can make the subsequent execution strategy more in line with the actual situation, which not only ensures the continuity of task progress, but also reduces the interaction interruption caused by simply terminating the task, further improving the user experience and task completion rate.

[0192] In one embodiment, reference Figure 14 , Figure 14 This diagram illustrates the overall architecture of information interaction in one embodiment. The large language model understands the user's question and, using its task planning capabilities, proposes a solution. At this point, a human-computer interaction occurs through the front-end interface to confirm with the user whether the large language model's understanding of the question is flawed and whether the task planning results need adjustment. If the user confirms, the task execution begins. Each task can be solved by calling an intelligent agent. Internally, the intelligent agent is implemented based on the large language model and tools. Each task creates a query-related node on a mind map. Mind map code construction explanation: A tree (the mind map in the code) can be constructed using its own task identifier (TaskId) and the task identifier (ParentId) of its parent node. This tree has a root node, child nodes, and a parent node. Based on the task identifier (RootTaskId) field corresponding to the root node, it can be determined which node is the root of the tree; all nodes are assigned tasks through RootTaskId. The root node's RootTaskId = TaskId, and it has no ParentId. In summary: by abstracting any step into a task object, a query-related node can be displayed in the front-end interaction. The query-related node is the smallest unit of subsequent user interaction.

[0193] During task execution, each agent generates an interactive question when user interaction is required, triggering a display on the terminal and obtaining the user's response. The agent then continues task execution based on the response. Each agent generates a task execution result. After completion, a reflection node determines whether the task solved the problem. If not, it might be due to missing information or the failure of some key steps. Further problem decomposition is needed, which may involve questioning the user for more specific information or breaking down the problem into smaller sub-problems, generating next-level query nodes, and resolving the smaller problems. Finally, a result-generating agent, such as a report generator or webpage generator, produces the final result.

[0194] In a straightforward manner, when any query node encounters a problem requiring user decision-making, it can trigger follow-up questions and display the interactive question above the node, prompting the user to answer. Based on the user's feedback, the agent further adjusts its subsequent execution plan and continues task execution. Technically, a WebSocket connection is used to push the interactive question requiring user confirmation to the user's browser. After the user inputs the information, the response is temporarily stored in a cache (Redis). The task execution node (the server node where the agent executing the task resides) detects changes in the corresponding Redis key, obtains the user's response, and then the agent executes further based on the response. The interaction time between different nodes can be determined by the task's complexity, deciding how long to wait for the user. If the user does not respond within the timeout period, the agent will re-analyze the problem, attempt to skip the user intervention step, and make a decision independently.

[0195] In this embodiment, human-computer interaction is introduced to improve the reliability and accuracy of the final delivery results of the large language model multi-agent task system. The black-box steps of the agents are abstracted, and a mind map based on task nodes is constructed to facilitate user interaction. In cases where the user's task is unclear, or the large language model exhibits ambiguous understanding or illusion, the large language model can ask the user for more information. Users can continuously participate during task execution, adjusting the direction of task execution in a timely manner, thereby improving the accuracy of task execution.

[0196] Furthermore, compared to the situation where errors are discovered after obtaining the query results, requiring the query results to be regenerated, the ability for users to continuously participate during task execution and adjust the direction of task execution in a timely manner effectively solves the problem of the output results deviating from user expectations after a large amount of model computation, thereby saving computer and other resources consumed by the model's extensive computation.

[0197] In one embodiment, the aforementioned information interaction method includes a task planning section and a task execution section. Specifically, for the task planning section: the server obtains the query content and generates a task planning result, triggering the terminal to display a task planning node; the server executes the planning task represented by the task planning node, obtains the task planning result, and triggers the terminal to display a planning confirmation question to confirm the task planning result; the terminal responds to the reply operation to the planning confirmation question by displaying a confirmation reply content for the task planning result. When the confirmation reply content indicates a denial of the task planning result, the server generates a first follow-up question to the confirmation reply content and triggers the terminal to display the first follow-up question; the terminal responds to the reply operation to the first follow-up question by displaying a first follow-up reply content for the first follow-up question; the server regenerates the task planning result based on the first follow-up reply content, triggers the terminal to display the task planning result, and displays a planning confirmation question for the regenerated task planning result.

[0198] For the task execution portion: When the confirmation response indicates confirmation of the task planning result, the server generates at least one task node and instructs the terminal to display it. Subsequently, the server generates child nodes for each task node and triggers the terminal to display at least one child node for each task node after displaying the first task node. The server executes the task represented by the child node and, if the information required for task execution is missing, generates a second follow-up question for further information. The server determines the complexity of the interactive question; predicts the response waiting time based on the complexity; and triggers the terminal to display the second follow-up question and a response countdown based on the response waiting time. The terminal responds to the response to the second follow-up question by displaying the second follow-up response content. The server continues task execution based on the second follow-up response content, or, if no response content is obtained before the response countdown reaches zero, it skips the interactive question and continues task execution. The terminal dynamically displays the task execution time in the child nodes. When the task is completed, the task status is changed from "in execution" to "execution completed" in the child nodes of the task node, and dynamic updates to the execution time cease. When the task execution result is obtained, and the task execution failure is determined based on the task execution result, the server triggers the terminal to display a third follow-up question determined according to the reason for the task execution failure; the terminal responds to the reply operation for the third follow-up question and displays the reply content for the third follow-up question; the server uses the reply content for the third follow-up question to re-execute the task.

[0199] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0200] This application also provides an application scenario in which the above-described information interaction method is applied. Specifically, the application of the information interaction method in this scenario is as follows:

[0201] When a user requests recommendations for hiking gear, they can input the query "Please recommend hiking gear." The terminal then sends the query to the server, which generates relevant nodes. The terminal displays these nodes, and when the server executes the tasks corresponding to these nodes using the model, it can generate interactive questions if user interaction is required. These questions are then displayed on the terminal. The user can reply to the interactive questions, and when the terminal receives the user's reply, it sends the reply to the server, allowing the model on the server to continue executing tasks based on the reply.

[0202] The above application scenarios are merely illustrative. It is understood that the application of the information interaction methods provided in the embodiments of this application is not limited to the above scenarios.

[0203] Based on the same inventive concept, this application also provides an information interaction device for implementing the information interaction method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more information interaction device embodiments provided below can be found in the limitations of the information interaction method described above, and will not be repeated here.

[0204] In one embodiment, such as Figure 15 As shown, an information interaction device is provided, including: a query content acquisition module 1502, a node display module 1504, and a question interaction module 1506;

[0205] The query content retrieval module 1502 is used to retrieve query content;

[0206] The node display module 1504 is used to display query-related nodes; the query-related nodes represent the tasks generated during the process of obtaining the query results of the query content.

[0207] The question-and-answer interaction module 1506 is used to execute the task represented by the query-related node, and to display an interactive question when user interaction is required during the execution of the task; the interactive question is used to confirm information and ask follow-up questions about the task; in response to the reply operation to the interactive question, the reply content is displayed, and the reply content is used to continue the execution of the task to obtain the task execution result.

[0208] The aforementioned information interaction device, by displaying query-related nodes—which represent tasks generated during the process of obtaining query results—can show the entire process of the agent generating query results in the form of a mind map. This facilitates understanding of the agent's overall work arrangement and real-time progress. During the display of query-related nodes and the execution of the tasks represented by those nodes, interactive questions can be displayed to confirm and follow up on information. This helps the agent make correct decisions based on the responses to the interactive questions, thereby improving the accuracy of the content generated by the agent.

[0209] In one embodiment, the query related nodes include task planning nodes; the node display module 1504 is used to display the task planning nodes; the interactive question includes a planning confirmation question, and the question interaction module 1506 is used to execute the planning task represented by the task planning node, obtain the task planning result, and display a planning confirmation question to confirm the task planning result; the reply content includes a confirmation reply content, and the question interaction module 1506 is also used to respond to the reply operation for the planning confirmation question by displaying the confirmation reply content for the task planning result.

[0210] In one embodiment, the interactive questioning further includes a first follow-up question; the response content further includes the first follow-up response content; the questioning interaction module 1506 is further configured to display a first follow-up question regarding the confirmed response content when the confirmed response content represents a denial of the task planning result; in response to the response operation to the first follow-up question, display the first follow-up response content to the first follow-up question; display the task planning result regenerated based on the first follow-up response content, and display a planning confirmation question regarding the regenerated task planning result.

[0211] In one embodiment, the relevant nodes to be queried are nodes in a mind map, including task nodes and child nodes of the task nodes; the node display module 1504 is used to display at least one task node; after displaying at least one task node, for each task node, at least one child node of that task node is displayed; the interactive questioning includes a second follow-up question; the questioning interaction module 1506 is used to execute the task represented by the child node, and when the information required for task execution is missing, displays a second follow-up question for further information inquiry; the response content also includes the response content to the second follow-up question; the questioning interaction module 1506 is also used to respond to the response operation to the second follow-up question by displaying the response content to the second follow-up question.

[0212] In one embodiment, the child nodes of the task node also display the task status and execution time of the corresponding task; the information interaction device further includes a status time update module, which is used to continue executing the task based on the content of the second follow-up question reply, and dynamically display the execution time of the task in the node; when the task is completed, the task status of the task is changed from executing to executing in the child nodes of the task node, and the dynamic update of the execution time is stopped.

[0213] In one embodiment, the relevant nodes are displayed on the content display page; the question interaction module 1506 is also used to display an information interaction pop-up window on the content display page; and to display interactive questions for the task in the information interaction pop-up window.

[0214] In one embodiment, the question interaction module 1506 is further configured to, when multiple query-related nodes represent tasks in parallel execution, for each of the multiple tasks being executed in parallel, if user interaction is required during the execution of the current task, distinguish and display the query-related node corresponding to the current task among the multiple query-related nodes, and display an information interaction pop-up window for the distinguished query-related node on the content display page.

[0215] In one embodiment, the content display page also displays information interaction elements; the question interaction module 1506 is further configured to, during the execution of the current task, respond to the trigger operation on the information interaction elements and display an information interaction pop-up window for querying relevant nodes corresponding to the current task; respond to the information input operation performed in the information interaction pop-up window and display input information; the input information is used to supplement the execution of the current task.

[0216] In one embodiment, the information interaction device further includes a countdown module, which displays a response countdown when an interactive question for a task is displayed; if no response is received for the interactive question before the response countdown reaches zero, the interactive question is skipped and the task continues to be executed.

[0217] In one embodiment, the information interaction device further includes a skip module for displaying a question skip element when an interactive question for a task is displayed; and skipping the interactive question and continuing to perform the task in response to a triggering operation on the question skip element.

[0218] In one embodiment, the interactive questioning also includes a third follow-up question; the questioning interaction module 1506 is further configured to display a third follow-up question determined based on the reason for the task execution failure when the task execution result of the task is obtained and the task execution failure is determined based on the task execution result; in response to the reply operation for the third follow-up question, display the reply content of the third follow-up question; the reply content of the third follow-up question is used to re-execute the task.

[0219] In one embodiment, such as Figure 16 As shown, an information interaction device is provided, including: a node generation module 1602, a question generation module 1604, and a response acquisition module 1606;

[0220] The node generation module 1602 is used to receive the query content sent by the terminal; generate query-related nodes and instruct the terminal to display the query-related nodes; the query-related nodes are used to represent the tasks generated in the process of obtaining the query results of the query content;

[0221] The question generation module 1604 is used to execute the task represented by the query related node, and generate interactive questions when user interaction is required during the task execution process; interactive questions are used to confirm information and ask follow-up questions about the task.

[0222] The response acquisition module 1606 is used to return the interactive question for the task, so as to instruct the terminal to display the interactive question; to acquire the response content for the interactive question, and to continue to execute the task based on the response content.

[0223] The aforementioned information interaction device, by generating query-related nodes, can trigger the terminal to display these nodes. Since these nodes represent tasks generated during the process of obtaining query results, the overall process of the agent generating query results can be displayed in the form of a mind map, facilitating understanding of the agent's entire work arrangement and real-time progress. During the display and execution of the tasks represented by these nodes, generating interactive questions can trigger the terminal to display them. Information confirmation and follow-up questions can then be based on these interactive questions, helping the agent make correct decisions based on the responses to the interactive questions, thereby improving the accuracy of the content generated by the agent.

[0224] In one embodiment, the response content submitted by the terminal is stored in a preset cache space; the response acquisition module 1606 is used to determine the question identifier of the interactive question; every preset time interval, it searches for the key-value pair corresponding to the question identifier in the preset cache space; when the value in the key-value pair is not empty, it extracts the value in the key-value pair as the response content of the interactive question.

[0225] In one embodiment, the information interaction device further includes a waiting time generation module, which is used to determine the complexity of the interactive question when an interactive question is generated; predict the response waiting time of the interactive question based on the complexity; and trigger the terminal to display a response countdown based on the response waiting time.

[0226] In one embodiment, the information interaction device further includes a reanalysis module, which reanalyzes the task if no response is received for the interactive question before the response countdown reaches zero, and continues to execute the task based on the results of the reanalysis.

[0227] Each module in the aforementioned information interaction device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0228] In one embodiment, a computer device, which may be a server, is provided. The computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is connected to the system bus via the I / O interfaces. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device stores information exchange data. The I / O interfaces of the computer device are used for exchanging information between the processor and external devices. The communication interface of the computer device is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it implements an information exchange method.

[0229] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 17As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements an information interaction method. The display unit of the computer device is used to form a visually visible image. It can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0230] Those skilled in the art will understand that Figure 17 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0231] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0232] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0233] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the steps in the above method embodiments.

[0234] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0235] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, etc., and are not limited to these.

[0236] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0237] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. An information exchange method, characterized in that, The method includes: Retrieve query results; Display query-related nodes; the query-related nodes are used to represent the tasks generated during the process of obtaining the query results of the query content; The task represented by the query-related node is executed, and when user interaction is required during the execution of the task, an interactive question is displayed; the interactive question is used for at least one of information confirmation and information follow-up questions regarding the task. In response to the reply to the interactive question, the reply content is displayed, which is used to continue executing the task to obtain the task execution result.

2. The method according to claim 1, characterized in that, The query-related nodes include task planning nodes; the display query-related nodes include: Display task planning nodes; The interactive questions include planning confirmation questions. When executing the task represented by the node, and user interaction is required during task execution, interactive questions are displayed, including: Execute the planning task represented by the task planning node, obtain the task planning result, and display a planning confirmation question to confirm the task planning result; The response content includes a confirmed response. The response operation in response to the interactive question, displaying the response content, includes: In response to a question asking for confirmation of the plan, the confirmation response for the task planning result is displayed.

3. The method according to claim 2, characterized in that, The interactive questioning also includes a first follow-up question; the response content also includes the response content to the first follow-up question; the method further includes: When the confirmation response indicates a denial of the task planning result, a first follow-up question is displayed regarding the confirmation response. In response to the reply operation to the first follow-up question, the reply content of the first follow-up question is displayed; The system displays the task planning results regenerated based on the response to the first follow-up question, and displays a planning confirmation question for the regenerated task planning results.

4. The method according to claim 1, characterized in that, The query-related nodes are nodes in the mind map, including task nodes and child nodes of the task nodes in the mind map. The display query related nodes include: Display at least one task node; After displaying the at least one task node, for each task node, display at least one child node of that task node; The interactive questioning includes a second follow-up question; when the task represented by the query-related node is executed, and user interaction is required during the task execution, the interactive questioning is displayed, including: Execute the task represented by the child node, and if the information required for task execution is missing, display a second follow-up question for further information inquiry; The response content also includes a second follow-up question response; the response operation in response to the interactive question, displaying the response content, includes: In response to the reply to the second follow-up question, the reply to the second follow-up question is displayed.

5. The method according to claim 4, characterized in that, The child nodes of the task node also display the task status and execution time of the corresponding task; the method further includes: The task continues to be executed based on the response to the second follow-up question, and the execution time of the task is dynamically displayed in the sub-node; When the task is completed, in the child nodes of the task node, the task status is changed from "in execution" to "execution completed", and the dynamic updating of the execution time is stopped.

6. The method according to claim 1, characterized in that, The query-related nodes are displayed on the content display page; the display of interactive questions includes: An information interaction pop-up window is displayed on the content display page; The information interaction pop-up window displays interactive questions related to the task.

7. The method according to claim 6, characterized in that, The display of the information interaction pop-up window on the content display page includes: When multiple query-related nodes represent tasks in parallel, for each of the multiple tasks being executed in parallel, if user interaction is required during the execution of the current task, the query-related node corresponding to the current task is displayed separately among the multiple query-related nodes, and an information interaction pop-up window for the separately displayed query-related node is displayed on the content display page.

8. The method according to claim 7, characterized in that, The content display page also displays information interaction elements; the method further includes: During the execution of the current task, in response to the trigger operation on the information interaction element, an information interaction pop-up window for querying the relevant nodes corresponding to the current task is displayed; In response to the information input operation performed in the information interaction pop-up window, the input information is displayed; the input information is used to supplement the execution of the current task.

9. The method according to claim 1, characterized in that, The method further includes: When an interactive question is displayed for the task, a response countdown is shown; If no response is received for the interactive question before the response countdown reaches zero, the interactive question is skipped and the task continues.

10. The method according to claim 1, characterized in that, The method further includes: When displaying interactive questions for the task, a question skipping element is shown; In response to the triggered operation of the question skipping element, the interactive question is skipped, and the task continues to be executed.

11. The method according to claim 1, characterized in that, The interactive questioning also includes a third follow-up question; the method also includes: When the task execution result is obtained, and it is determined that the task execution failed based on the task execution result, a third follow-up question determined according to the reason for the task execution failure is displayed; In response to the reply to the third follow-up question, the reply to the third follow-up question is displayed; the reply to the third follow-up question is used to re-execute the task.

12. An information exchange method, characterized in that, Applied to a server, the method includes: The query content sent by the receiving terminal; Generate query-related nodes and instruct the terminal to display the query-related nodes; the query-related nodes are used to represent tasks generated during the process of obtaining query results for the query content; The task represented by the query-related node is executed, and when user interaction is required during the execution of the task, an interactive question is generated; the interactive question is used to confirm information and ask follow-up questions about the task. Returning to the interactive prompt instructs the terminal to display the interactive prompt; Obtain the response to the interactive question, and continue to execute the task based on the response.

13. The method according to claim 12, characterized in that, The response content submitted by the terminal is stored in a preset cache space; The step of obtaining the response content to the interactive question includes: Determine the question identifier of the interactive question; Every preset time interval, the key-value pair corresponding to the query identifier is retrieved from the preset cache space; When the value in the key-value pair is not empty, the value in the key-value pair is extracted as the response content to the interactive question.

14. The method according to claim 12, characterized in that, The method further includes: When generating the interactive question, determine the complexity of the interactive question; Predict the response time for the interactive question based on the complexity level; The terminal will display a countdown timer based on the specified response waiting time.

15. The method according to claim 14, characterized in that, The method further includes: If no response is received for the interactive question before the response countdown reaches zero, the task is re-analyzed, and the task continues to be executed based on the results of the re-analysis.

16. An information interaction device, characterized in that, The device includes: The query content retrieval module is used to retrieve query content; The node display module is used to display query-related nodes; the query-related nodes represent the tasks generated during the process of obtaining the query results of the query content. The question-and-answer interaction module is used to execute the task represented by the query-related node, and to display an interactive question when user interaction is required during the execution of the task; the interactive question is used to confirm information and ask follow-up questions about the task; in response to the reply operation to the interactive question, the reply content is displayed, and the reply content is used to continue the execution of the task to obtain the task execution result.

17. An information interaction device, characterized in that, The device includes: A node generation module is used to receive query content sent by a terminal; generate query-related nodes and instruct the terminal to display the query-related nodes; the query-related nodes are used to represent tasks generated during the process of obtaining query results for the query content; The question generation module is used to execute the task represented by the query-related node, and generate interactive questions when user interaction is required during the execution of the task; the interactive questions are used to confirm information and ask follow-up questions about the task. The response acquisition module is used to return the interactive question for the task, so as to instruct the terminal to display the interactive question; acquire the response content for the interactive question, and continue to execute the task based on the response content.

18. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 15.

19. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 15.

20. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 15.

Images