Interactive question-answering task processing method and related device based on multi-agent collaboration

Abstract

The present disclosure provides an interactive question and answer task processing method based on multi-agent cooperation and related devices, which relates to the fields of artificial intelligence technologies such as generative large language models and agents. The method comprises: determining a target task to be solved according to a natural language input of a user; decomposing the target task into a plurality of sub-target tasks containing at least an interactive question and answer task, and issuing each sub-target task to each sub-target agent containing at least an interactive question and answer agent, different sub-agents being used to process different types of sub-tasks; controlling the interactive question and answer agent to present an interactive graphical user interface corresponding to the interactive question and answer task to the user, and to present an answer result corresponding to the received feedback answer, the interactive graphical user interface containing a question and an interactive answer area; controlling the associated sub-target agent in the plurality of sub-target agents to generate a corresponding associated processing result based on the answer result output by the interactive question and answer agent, the associated sub-target agent being a sub-target agent that relies on the answer result as part of the task input information; and determining a target processing result corresponding to the target task based on the associated processing result. By applying this scheme, better comprehensive task processing results can be achieved at a lower comprehensive cost, especially for interactive question and answer tasks that require users to provide subjective answers to the questions raised and to verify the results based on the answers. The interactive question and answer agent provides an interactive graphical interface containing a question and an interactive answer area to the user, which makes the user's task requirements more quickly satisfied in the form of interactive answers, improving the task solving effect.

Description

Technical Field

[0001] This disclosure relates to the field of data processing technology, specifically to the fields of generative large language models, intelligent agents, and other artificial intelligence technologies, and particularly to a method, apparatus, electronic device, computer-readable storage medium, and computer program product for interactive question-and-answer task processing based on multi-agent collaboration. Background Technology

[0002] With the rapid development and iteration of generative large language models, they have a better ability to understand user input needs and provide corresponding results.

[0003] In order to make the output of generative large language models more in line with specific needs, intelligent agents that use generative large language models as base models and combine them with pre-defined role parameters have emerged.

[0004] How to use intelligent agents to solve complex problems raised by users remains a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] This disclosure presents an interactive question-and-answer task processing method, apparatus, electronic device, computer-readable storage medium, and computer program product based on multi-agent collaboration.

[0006] In a first aspect, embodiments of this disclosure propose an interactive question-and-answer task processing method based on multi-agent collaboration, comprising: determining a target task to be solved based on the user's natural language input; decomposing the target task into multiple sub-target tasks, each containing at least an interactive question-and-answer type task, and distributing each sub-target task to a corresponding sub-target agent containing at least an interactive question-and-answer agent; wherein different sub-agents are used to process different types of sub-tasks; controlling the interactive question-and-answer agents to present an interactive graphical user interface corresponding to the interactive question-and-answer type task to the user, and presenting an answer result corresponding to the received feedback answer; wherein the interactive graphical user interface includes a question and an interactive answer area; controlling the associated sub-target agents among the multiple sub-target agents to generate corresponding associated processing results based on the answer results output by the interactive question-and-answer agents; wherein the associated sub-target agents are sub-target agents that rely on the answer results as part of the task input information; and determining the target processing result corresponding to the target task based on the associated processing results.

[0007] Secondly, embodiments of this disclosure propose an interactive question-and-answer task processing device based on multi-agent collaboration, comprising: a target task determination unit configured to determine a target task to be solved based on a user's natural language input; a task decomposition and corresponding assignment unit configured to decompose the target task into multiple sub-target tasks, each including at least an interactive question-and-answer type task, and assign each sub-target task to each sub-target agent, each including at least an interactive question-and-answer agent; wherein different sub-agents are used to process different types of sub-tasks; an interactive interface control and presentation unit configured to control the interactive question-and-answer agents to present an interactive graphical user interface corresponding to the interactive question-and-answer type task to the user, and to present an answer result corresponding to the received feedback answer; wherein the interactive graphical user interface includes a question and an interactive answer area; an association processing result control and generation unit configured to control the association sub-target agents among the multiple sub-target agents to generate corresponding association processing results based on the answer results output by the interactive question-and-answer agents; wherein the association sub-target agents are sub-target agents that rely on the answer results as part of the task input information; and a target processing result determination unit configured to determine the target processing result corresponding to the target task based on the association processing result.

[0008] Thirdly, embodiments of this disclosure provide an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to implement the interactive question-answering task processing method based on multi-agent collaboration as described in the first aspect.

[0009] Fourthly, embodiments of this disclosure provide a non-transitory computer-readable storage medium storing computer instructions that enable a computer to implement the interactive question-answering task processing method based on multi-agent collaboration as described in the first aspect.

[0010] Fifthly, embodiments of this disclosure provide a computer program product including a computer program that, when executed by a processor, can implement the steps of the interactive question-answering task processing method based on multi-agent collaboration as described in the first aspect.

[0011] The interactive question-and-answer task processing scheme based on multi-agent collaboration provided in this disclosure involves a master agent determining the target task to be solved based on the user's natural language input, then decomposing the target task into multiple sub-target tasks, including at least interactive question-and-answer tasks. Each sub-target task is then distributed to a corresponding sub-target agent, including at least an interactive question-and-answer agent. Under the control of the master agent, the interactive question-and-answer agent presents an interactive graphical user interface corresponding to the interactive question-and-answer task, as well as answer results corresponding to the received feedback answers. The related sub-target agents, depending on the answer results, also generate corresponding related processing results based on the answer results under the control of the master agent. Finally, the master agent obtains the target processing result corresponding to the target task based on the related processing results.

[0012] This disclosure employs an agent cluster consisting of a pre-built master agent and multiple sub-agents to handle user-initiated task requests. The master agent is responsible for understanding the user's task requirements and decomposing them into multiple sub-tasks that can be executed by different sub-agents. Each sub-agent then processes its assigned sub-task according to the master agent's instructions. Through collaboration between the master agent and its sub-agents, each entity can effectively handle different components of a complex task. Since different sub-agents are pre-built specifically for handling different types of tasks, this collaborative approach of a master agent and multiple sub-agents offers better performance on single-type tasks compared to using a single, all-encompassing agent. Furthermore, the relatively smaller size of the sub-agents allows for flexible addition and modification of their functions, resulting in better overall task processing performance at a lower overall cost.

[0013] Especially for interactive question-and-answer tasks that require users to subjectively answer questions and verify the results, the interactive question-and-answer agent can provide users with an interactive graphical interface that includes question and interactive answer areas, so that users' task requirements can be met more quickly through interactive answering, thereby improving the task solving effect.

[0014] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0015] Other features, objects, and advantages of this disclosure will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0016] Figure 1This is an exemplary system architecture to which this disclosure can be applied;

[0017] Figure 2 A flowchart illustrating an interactive question-answering task processing method based on multi-agent collaboration, provided as an embodiment of this disclosure;

[0018] Figure 3 A flowchart illustrating a method for determining a target task based on natural language input, provided as an embodiment of this disclosure;

[0019] Figure 4 A flowchart illustrating a method for decomposing a target task into multiple sub-target tasks, including real-time information inquiry tasks, based on task elements, is provided in this embodiment of the disclosure.

[0020] Figure 5 A flowchart illustrating a method for creating a sub-agent provided in this disclosure embodiment;

[0021] Figure 6 A flowchart illustrating a method for controlling an interactive question-answering agent to output answer results, provided in an embodiment of this disclosure;

[0022] Figure 7 A flowchart illustrating a method for controlling an interactive question-answering agent, an associated sub-target agent to generate associated processing results, and obtaining target processing results, provided in an embodiment of this disclosure;

[0023] Figure 8 A two-branch schematic diagram of a control sub-target intelligent agent outputting corresponding sub-processing results, provided in an embodiment of this disclosure;

[0024] Figures 9-1 to 9-14 These are all example diagrams of an interactive question-answering task processing method based on multi-agent collaboration in an application scenario provided by embodiments of this disclosure;

[0025] Figure 10 A structural block diagram of an interactive question-answering task processing device based on multi-agent collaboration provided in an embodiment of this disclosure;

[0026] Figure 11 This is a schematic diagram of the structure of an electronic device suitable for performing an interactive question-answering task processing method based on multi-agent collaboration, provided in an embodiment of this disclosure. Detailed Implementation

[0027] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding; these should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description. It should be noted that, unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0028] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0029] Figure 1 An exemplary system architecture 100 is shown, in which embodiments of the interactive question-answering task processing method, apparatus, electronic device, and computer-readable storage medium based on multi-agent collaboration of the present disclosure can be applied.

[0030] like Figure 1 As shown, system architecture 100 may include terminal devices 101, 102, and 103, a network 104, and a server 105. Network 104 serves as the medium for providing communication links between terminal devices 101, 102, and 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0031] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Various applications for enabling information communication between the terminal devices 101, 102, and 103 and server 105 can be installed. These applications include complex task processing applications, browser applications, and instant messaging applications.

[0032] Terminal devices 101, 102, and 103 and server 105 can be either hardware or software. When terminal devices 101, 102, and 103 are hardware, they can be various electronic devices with displays, including but not limited to smartphones, tablets, laptops, and desktop computers. When terminal devices 101, 102, and 103 are software, they can be installed in the aforementioned electronic devices, and can be implemented as multiple software programs or software modules, or as a single software program or software module; no specific limitation is made here. When server 105 is hardware, it can be implemented as a distributed server cluster composed of multiple servers, or as a single server. When server 105 is software, it can be implemented as multiple software programs or software modules, or as a single software program or software module; no specific limitation is made here.

[0033] Server 105 can provide various services through its built-in applications. Taking a complex task processing application that provides one-click processing services for complex tasks as an example, when running this application, server 105 can achieve the following: First, it receives natural language input from the user via terminal devices 101, 102, and 103 through network 104; then, it determines the target task to be solved based on the natural language input. Next, it decomposes the target task into multiple sub-target tasks, each containing at least an interactive question-and-answer type task, and distributes each sub-target task to its corresponding sub-target agent, which contains at least an interactive question-and-answer agent. The agent is used to handle different types of subtasks. Next, the interactive question-and-answer agent is controlled to present the user with an interactive graphical user interface corresponding to the interactive question-and-answer task, and to present the answer result corresponding to the received feedback answer. This interactive graphical user interface includes a question and an interactive answer area. Furthermore, the associated sub-target agent among multiple sub-target agents is controlled to generate corresponding associated processing results based on the answer result output by the interactive question-and-answer agent. This associated sub-target agent is a sub-target agent that depends on the answer result as part of the task input information. Finally, the target processing result corresponding to the target task is determined based on the associated processing result.

[0034] Furthermore, the server 105 can also transmit the target processing result back to the terminal devices 101, 102, and 103 via the network 104, so that the terminal devices 101, 102, and 103 can display the received target processing result to the user.

[0035] It should be noted that, in addition to being obtained from terminal devices 101, 102, and 103 via network 104, natural language input can also be pre-stored locally on server 105 through various means. Therefore, when server 105 detects that this data is already stored locally (e.g., when starting to process previously reserved pending tasks), it can choose to directly retrieve this data from locally. In this case, the exemplary system architecture 100 may also exclude terminal devices 101, 102, and 103 and network 104.

[0036] Because processing complex tasks requires significant computing resources and power, the interactive question-and-answer task processing methods based on multi-agent collaboration provided in the subsequent embodiments of this disclosure are generally executed by a server 105 with strong computing power and abundant computing resources. Correspondingly, the interactive question-and-answer task processing device based on multi-agent collaboration is also generally located in the server 105. However, it should also be noted that when terminal devices 101, 102, and 103 also possess sufficient computing power and resources, they can also complete the aforementioned calculations performed by the server 105 through complex task processing applications installed on them, thereby outputting the same results as the server 105. Especially when multiple terminal devices with different computing capabilities exist simultaneously, but a complex task processing application determines that the terminal device it is on has strong computing power and sufficient remaining computing resources, the terminal device can be allowed to perform the aforementioned calculations, thereby appropriately reducing the computing pressure on server 105. Correspondingly, the interactive question-and-answer task processing device based on multi-agent collaboration can also be set in terminal devices 101, 102, and 103. In this case, the exemplary system architecture 100 may also exclude server 105 and network 104.

[0037] It should be noted that the main intelligent agent and each sub-target intelligent agent can be installed on server 105 at the same time. Each sub-target intelligent agent called and controlled by the main intelligent agent can also be installed on other servers or terminal devices different from server 105. No specific restrictions are made here.

[0038] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0039] Please refer to Figure 2 , Figure 2 A flowchart of an interactive question-answering task processing method based on multi-agent collaboration provided in this disclosure embodiment, wherein process 200 includes the following steps:

[0040] Step 201: Determine the target task to be solved based on the user's natural language input;

[0041] This step aims to transform the executor of an interactive question-answering task processing method based on multi-agent collaboration (e.g., Figure 1 The server 105 shown, which carries the main intelligent agent, fully understands and analyzes the natural language input received from the user (e.g., through...). Figure 1The network 104 shown receives natural language input from the user through the terminal devices 101, 102, and 103, and then determines the target task that the user wishes to solve, that is, the target task corresponds to a task requirement.

[0042] Specifically, this step is typically a multi-step integrated processing procedure involving complex intelligent reasoning processes such as language understanding, context analysis, and task modeling. Examples include input reception, preprocessing and normalization, semantic understanding and intent recognition, and task modeling and reasoning.

[0043] In this process, the main intelligent agent, acting as the execution subject, first receives the user's natural language input, then preprocesses and normalizes the received natural language input to ensure that it can be effectively understood and processed. Next, it understands the user's intention reflected in the preprocessed and normalized input information based on the context, thereby determining the target task.

[0044] The input of natural language can take the form of speech, text, images, video, or other forms of language data. To facilitate processing, conversion techniques can be used to convert non-textual natural language input into natural language text that is easy to recognize and process. Considering that natural language input can present request information in various forms (e.g., direct requests, indirect requests, and multi-step requests), preprocessing and normalization processes, including word segmentation and annotation, noise removal, spell correction, and synonym processing, can be used to effectively understand the user's actual task requirements. In the intent recognition stage, keyword extraction, entity recognition, and the combination of contextual information can be used to accurately identify the user's intent expressed through natural language input. To convert the identified intent into a specific task, task modeling operations, including task abstraction (mapping the user's needs to a system-recognizable task framework) and target task definition, can be used to finally obtain the target task corresponding to the natural language input.

[0045] Step 202: Decompose the target task into multiple sub-target tasks that include at least interactive question-and-answer tasks, and distribute each sub-target task to each sub-target agent that includes at least an interactive question-and-answer agent.

[0046] Building upon step 201, this step aims to decompose the target task into multiple sub-target tasks by the aforementioned executing entity, and to distribute each sub-target task to a corresponding sub-target agent. Among the multiple sub-target tasks decomposed in this step, at least one interactive question-and-answer task should be included. Correspondingly, each sub-target agent should include at least one interactive question-and-answer agent specifically designed for handling interactive question-and-answer tasks.

[0047] Different sub-agents are used to handle different types of sub-tasks. That is, sub-agents are created in advance according to different task types, each dedicated to a single type of task. There are many ways to classify task types, such as according to the task processing steps, the task processing method, the task complexity, the data types involved in the task, etc. No specific limitation is made here. Different sub-agents are applied to handle different atomic tasks (i.e., the smallest unit of task. Here, "smallest" is a relative concept, not an absolute concept, and refers to the smallest task unit that can be decomposed at present).

[0048] Interactive question-and-answer tasks are a type of interactive task that requires users to subjectively answer questions and evaluate their responses. Examples include knowledge-learning tasks (especially numerical knowledge, such as the chicken-and-rabbit problem) and question-and-answer game-like tasks. Taking the task "Want to learn the chicken-and-rabbit problem" as an example, it's clear that this task not only requires learning the definition of the problem but also, after understanding the definition, learning and mastering the knowledge point as much as possible by answering practice questions and tests. Therefore, the focus of this type of task is to provide interactive problem solutions through interactive questions, enabling users to solve their task requirements during the interactive process.

[0049] It should be noted that in this embodiment, the multiple sub-target tasks decomposed from the target task in this step should include at least interactive question-and-answer tasks. Apart from that, there is no limitation on the task type of the other sub-target tasks decomposed from the target task. They can be any type of task. Correspondingly, a sub-target intelligent agent with corresponding task processing capabilities is required.

[0050] Step 203: Control the interactive question and answer to present the user with an interactive graphical user interface corresponding to the interactive question and answer task, and present the answer result corresponding to the received feedback answer;

[0051] Building upon step 202, this step aims to present the user with an interactive graphical user interface (GUI) corresponding to the interactive question-and-answer task, and to display the answer result corresponding to the received feedback answer. This GUI includes a question and an interactive answer area. The user's feedback answer should be entered in this interactive answer area, and there are various input methods, such as selecting options, entering the answer in a blank area, or a specific graphic. The answer result corresponding to the feedback answer is a judgment of whether the feedback answer is correct, and may include, for example, an incorrect answer and a correct answer.

[0052] Step 204: Control the related sub-target agents among multiple sub-target agents to generate corresponding related processing results based on the answer results output by the interactive question-answering agent;

[0053] Based on step 203, this step aims to have the aforementioned executing entity control the associated sub-target intelligent agents among multiple sub-target intelligent agents to generate corresponding associated processing results based on the answer results output by the interactive question-and-answer intelligent agent. The associated sub-target intelligent agent is a sub-target intelligent agent that relies on the answer results as part of the task input information. Therefore, the execution of the corresponding sub-target task by the associated sub-target intelligent agent is actually arranged after the interactive question-and-answer intelligent agent outputs the answer results.

[0054] It should be understood that each sub-target agent, including the interactive question-answering agent and the associated sub-target agent, should actually output sub-processing results corresponding to the issued sub-target tasks under the invocation or arrangement of the main agent. Specifically, the main agent can simultaneously inform the sub-target agents of the execution order or execution trigger signal when issuing each sub-target task, so that each sub-target agent executes the corresponding sub-target task according to the informed execution order or executes the corresponding sub-target task when it is determined that the current state meets the requirements of the execution trigger signal. That is, each sub-target agent actively executes the corresponding sub-target task at an appropriate time. Alternatively, the main agent can actively invoke different sub-target agents to execute the corresponding sub-target tasks as required when it determines that different sub-target agents meet the execution time. That is, each sub-target agent only passively executes the corresponding sub-target task according to the invocation instruction.

[0055] Both the active and passive execution mechanisms mentioned above are different forms of collaboration between the main intelligent agent and multiple sub-intelligent agents to process tasks. The specific choice of which one to use depends on the actual needs of the application scenario, and no specific restrictions are made here.

[0056] Furthermore, during the process of each sub-target agent executing its corresponding sub-target task and obtaining its corresponding sub-processing result, the user's personalized preferences can be incorporated. Among these, multiple sub-processing results include at least the answer output by the interactive question-answering agent corresponding to its respective interactive question-answering task, and the associated processing result output by the related sub-target agent. The remaining sub-processing results can also be the results obtained by other agents when processing corresponding types of tasks. These sub-processing results may depend on the results of other sub-processing results, or they may be determined independently based on the task description information of the sub-target task without relying on other sub-processing results.

[0057] Step 205: Determine the target processing result corresponding to the target task based on the association processing result.

[0058] Building upon step 203, this step aims to determine the target processing result corresponding to the target task based on the correlation processing result.

[0059] Specifically, in the process of processing the associated processing results to obtain the target processing results, we can fully refer to the decomposition method of decomposing the target task into multiple sub-target tasks, and adopt a variety of processing methods, including deduplication, adjusting the expression, adjusting the sentence structure and enriching the expression, so that the final target processing results can not only meet the user's task requirements, but also highlight the key requirement result information as much as possible and improve the information recognition.

[0060] The interactive question-answering task processing method based on multi-agent collaboration provided in this disclosure employs an agent cluster consisting of a pre-built master agent and multiple sub-agents to process user-proposed task requests. The master agent is responsible for understanding the user's task requests and decomposing the overall task requirements into multiple sub-target tasks that can be executed by different sub-agents. Each sub-agent then processes its assigned sub-task according to the master agent's instructions. Through collaboration between the master agent and its sub-agents, each entity can effectively handle different components of a complex task. Since different sub-agents are pre-built specifically for handling different types of tasks, this scheme, which uses a master agent and multiple sub-agents to collaboratively process task requests, achieves better processing results for single-type tasks compared to using a single, all-encompassing agent. Furthermore, the relatively small number of sub-agents allows for flexible addition and modification of their functions, resulting in better overall task processing performance at a lower overall cost.

[0061] Especially for interactive question-and-answer tasks that require users to subjectively answer questions and verify the results, the interactive question-and-answer agent can provide users with an interactive graphical interface that includes question and interactive answer areas, so that users' task requirements can be met more quickly through interactive answering, thereby improving the task solving effect.

[0062] To deepen your understanding of how to determine the target task, please refer to [the relevant documentation / reference]. Figure 3 , Figure 3 A flowchart of a method for determining a target task based on natural language input is provided for embodiments of this disclosure, wherein process 300 includes the following steps:

[0063] Step 301: Convert the user's natural language input into natural language text;

[0064] This step aims to have the aforementioned implementing entity convert various forms of natural language input into natural language text that is easy to recognize and extract, that is, to unify various forms of input into text form. For example, speech-to-text technology can be used to convert natural speech signals in speech form into natural language text.

[0065] Step 302: Perform intent recognition on the natural language text to obtain intent recognition results that include the intent to learn and master knowledge points;

[0066] Building upon step 301, this step aims to have the aforementioned executing entity perform intent recognition on the natural language text, obtaining intent recognition results that include the intent to learn and master knowledge points. That is, the intent recognition results must at least contain the intent to learn and master knowledge points, and may also contain other intents, which are not specifically limited here.

[0067] One implementation method, including but not limited to, may include the following specific steps:

[0068] First, the aforementioned execution entity performs semantic understanding on the natural language text to obtain semantic understanding results. Then, the aforementioned execution entity determines the preliminary intent (or suspected intent) based on the semantic understanding results. In order to improve the accuracy of intent determination, the aforementioned execution entity can also use the user's personalized preferences to personalize and modify the preliminary intent, thereby obtaining intent recognition results that include the need verification intent combined with real-time information.

[0069] This approach first uses semantic understanding to determine the initial intent directly corresponding to the natural language text. Then, it uses the user's personalized preferences to personalize the initial intent, such as filtering, excluding, or supplementing implicit or missing information, so that the final intent recognition result is more in line with the user's actual expectations.

[0070] In addition to this implementation method, other implementation methods that determine the intent recognition result can also be constructed by using various technologies involved in the multiple processing steps mentioned in step 201, combined with the actual application scenario. These will not be listed one by one here.

[0071] Step 303: Determine the target task to be solved based on the intent recognition results.

[0072] Building upon step 302, this step aims to have the aforementioned executing entity determine the target task to be solved based on the intent recognition results. The target task includes tasks related to learning and mastering knowledge points.

[0073] Specifically, in order to convert the identified intent into a specific task, task modeling operations, including task abstraction (mapping user needs to a system-recognizable task framework) and target task definition, can be used to ultimately obtain the target task corresponding to the natural language input.

[0074] This embodiment provides a more specific implementation method for determining the target task based on natural language input through steps 301-303. It includes key processing steps such as transformation of the presentation form, intent recognition, and determination of the target task based on the intent recognition result, so as to provide a feasible and more accurate implementation scheme for the target task.

[0075] To deepen your understanding of how to break down a target task into multiple sub-target tasks, please refer to [link to relevant documentation]. Figure 4 , Figure 4 A flowchart of a method for decomposing a target task into multiple sub-target tasks, including tasks with explicit result acquisition, based on task elements provided in this disclosure is provided. The process 400 includes the following steps:

[0076] Step 401: Identify the multiple task elements that constitute the target task;

[0077] This step aims to have the aforementioned implementing entities identify multiple task elements that constitute the target task, with different task elements corresponding to different types of tasks.

[0078] In this context, task elements refer to the basic units or components that constitute the target task. Each target task may consist of several independent elements, which are the key steps, sub-tasks, or resources necessary to achieve the target task. Each task element may correspond to an independent task, and they work together to promote the completion of the target task.

[0079] Suppose the objective is "to develop a new software product". This objective can be broken down into multiple task elements, each corresponding to a different type of task:

[0080] 1) Requirements Analysis (Management Task): Determine user needs and market demands;

[0081] 2) System Design (Technical Tasks): Architecture design, database design, interface design;

[0082] 3) Coding Development (Technical Tasks): Programmers write code and implement functions;

[0083] 4) Testing and quality control (technical and managerial tasks): Unit testing, integration testing, quality assurance;

[0084] 5) Marketing and promotion (management and communication tasks): Develop marketing plans and promote products.

[0085] Step 402: Based on the specific task element that requires users to subjectively answer the questions raised and to verify whether the knowledge points to which the questions belong have been learned and mastered based on the answer results, the interactive question-and-answer tasks for learning and mastering knowledge points are decomposed from the target task.

[0086] This step aims to have the aforementioned implementing entity decompose interactive question-and-answer tasks from the target task. Specifically, it relies on individual task elements that require users to subjectively answer questions and verify whether the knowledge points related to the questions have been learned and mastered based on the answers.

[0087] Specifically, to ensure a thorough grasp of a particular knowledge point, the interactive question-and-answer agent can be designed in two parts: one is the knowledge point practice and answering agent corresponding to the knowledge point practice phase, and the other is the knowledge point mastery testing agent corresponding to the practice result verification phase. The difference is that one provides multiple practice questions, while the other only provides a small number of test questions.

[0088] Step 403: Decompose the target task into multiple sub-target tasks, each containing only a single task element.

[0089] This step aims to decompose the target task into multiple sub-target tasks, each containing only a single task element, by referring to the task element decomposition method in step 402 to obtain interactive question-and-answer tasks. That is, each sub-target task contains at least an interactive question-and-answer task.

[0090] This embodiment provides a task decomposition method through steps 401-403, which decomposes each task element into a sub-target task according to the task elements constituting the target task. Combined with the pre-definition of task elements (e.g., according to the way different sub-intelligent agents are constructed), the task decomposition can be flexibly completed according to the actual situation.

[0091] Furthermore, considering that even the same task element may have additional conditions, it is necessary to decompose it into sub-target tasks carrying these additional conditions during task decomposition. This is so that when the sub-target tasks are subsequently assigned to the sub-target agents, the appropriate sub-target agent can be selected as the sub-target agent based on these additional conditions.

[0092] One possible implementation, including but not limited to, is as follows: In response to the existence of additional conditions corresponding to task elements, the target task is decomposed into multiple sub-target tasks, each containing only a single task element and corresponding additional conditions. These additional conditions may include: a specified task processing method and / or a specified result presentation format. Different task processing methods are abstracted from different processing logics for the same type of task (e.g., the same type of task can have multiple different processing logics, but all can ultimately yield the correct answer); the result presentation format may include at least one of the following: text, table, image, video, and interactive card formats.

[0093] To deepen your understanding of how to pre-create different sub-agents, please also refer to [link to relevant documentation]. Figure 5 , Figure 5 A flowchart of a method for creating a sub-agent provided in this disclosure embodiment is included in process 500, which includes the following steps:

[0094] Step 501: Determine the set of task types to be used to create the sub-agent;

[0095] Step 502: For each type of task in the task type set, construct a sub-agent with the corresponding task processing capabilities.

[0096] This embodiment provides a scheme through steps 501-502 for constructing sub-agents with corresponding task processing capabilities for each of the identified multiple types of tasks. Furthermore, the number of sub-agents corresponding to each type of task can be only one or multiple. These multiple sub-agents may differ from each other or be completely identical. Whether differences are necessary depends on whether the task type typically includes further additional conditions.

[0097] Specifically, for a first-objective-type task with at least two task processing methods, a sub-agent corresponding to each task processing method can be constructed; for a second-objective-type task with at least two result presentation formats, a sub-agent corresponding to each result presentation format can be constructed.

[0098] To deepen your understanding of how interactive question-answering agents output their responses, please refer to [link to relevant documentation]. Figure 6 , Figure 6 A flowchart of a method for controlling an interactive question-answering agent to output answer results, provided in an embodiment of this disclosure, is included in the following steps:

[0099] Step 601: Control the interactive question-answering agent to determine the question and interactive answer area based on the received task input information;

[0100] The aforementioned executing entity controls the interactive question-and-answer agent to determine the question and interactive response area based on the received task input information. This received task input information is determined based on task description information related to interactive question-and-answer tasks, decomposed from the target task. Furthermore, the received task input information can also be determined jointly based on the task description information related to interactive question-and-answer tasks decomposed from the target task and the user's personalized preferences. That is, by additionally incorporating the user's personalized preferences, the degree of matching between the determined task input information and the user's actual task needs is maximized.

[0101] Specifically, the personalized preference may include at least one of the following:

[0102] Age, education level, areas of interest, and historical consumption preferences.

[0103] Taking age in personalized preferences as an example, when the aforementioned executing entity controls the interactive question-answering agent to determine the question based on the received task input information, it will specifically manifest as the aforementioned executing entity controlling the interactive question-answering agent to determine the question that matches the cognitive level corresponding to the age based on the received task input information, so that the generated question matches the cognitive level of the user's current age.

[0104] Step 602: Control the interactive question-and-answer agent to present an interactive graphical user interface containing question and interactive answer areas to the user;

[0105] Based on step 601, this step aims to have the aforementioned executing entity control the interactive question-and-answer agent to present an interactive graphical user interface containing questions and interactive answer areas to the user, so as to enable the user to answer in the interactive answer area by presenting an interactive graphical user interface containing questions and interactive answer areas, thereby facilitating the receipt of feedback answers.

[0106] Step 603: Control the interactive question-answering agent to present the answer result determined based on the degree of difference between the feedback answer received from the interactive answering area and the standard answer to the question;

[0107] This step aims to have the aforementioned implementing entity control the interactive question-answering agent to present a response result determined by the degree of difference between the feedback answer received from the interactive response area and the standard answer to the question. That is, the greater the difference between the feedback answer and the standard answer, the more likely the response will be incorrect; conversely, the smaller the difference, the closer the response will be to the correct answer.

[0108] This step, through steps 601-603, provides a specific implementation method for controlling an interactive question-and-answer agent to present an interactive graphical user interface, receive feedback answers, and analyze the feedback answers to obtain the answer results.

[0109] To further deepen your understanding of how the intelligent agent generates associated processing results based on the answer results and how it obtains the target processing results based on the associated processing results, please refer to [link to relevant documentation]. Figure 7 , Figure 7 A flowchart of a method for controlling an interactive question-answering agent, associating a sub-target agent to generate an association processing result, and obtaining a target processing result, provided in this embodiment of the disclosure, is included in the following steps:

[0110] Step 701: In response to the feedback answer being an incorrect answer that is different from the standard answer, control the interactive question-answering agent to regenerate a new question that is different from the question that was answered incorrectly the last time and an interactive graphical user interface that matches the new question;

[0111] This step addresses a situation where the interactive question-answering agent determines the current feedback answer is incorrect. Based on this, the aforementioned executing entity will control the interactive question-answering agent to regenerate a new question that differs from the one that appeared when the answer was incorrect, along with an interactive graphical user interface matching the new question's interactive answer area. To increase the accuracy of the new question's answer, it is further controlled to be generated based on the question that appeared when the user answered correctly last time. In other words, by making the new question as close as possible to the previously correctly answered question, yet with certain differences, the learning and mastery of the knowledge points related to the question are further reinforced.

[0112] Step 702: In response to obtaining a preset number of answers, control the interactive question-and-answer agent to output a prompt message indicating that the interactive question-and-answer process is complete;

[0113] This step addresses the situation where the interactive question-and-answer agent continuously initiates a preset number of questions for the user to answer, and then obtains a preset number of answers. In this case, the aforementioned executing entity also controls the interactive question-and-answer agent to output a prompt message indicating that the interactive question-and-answer process has been completed. That is, the prompt message indicates that the preset number of questions have been answered.

[0114] Step 703: In response to receiving the prompt message, the control accuracy calculation agent determines the answer accuracy rate based on a preset number of answer results;

[0115] Based on step 702, this step aims to have the execution entity control the accuracy calculation agent to determine the response accuracy based on a preset number of response results when it receives the prompt information. That is, the accuracy calculation agent in this step is a specific related sub-target agent.

[0116] Step 704: Control the learning suggestion agents in multiple sub-target agents to output matching learning suggestions based on the response accuracy;

[0117] Building upon step 703, this step aims to allow the aforementioned executing entity to control the learning suggestion agent to output matching learning suggestions based on the response accuracy.

[0118] Step 705: Summarize the number of responses, the accuracy rate of responses, and learning suggestions to obtain the target processing results corresponding to the target task.

[0119] Based on step 704, this step aims to have the aforementioned implementing entities summarize the number of answers, the accuracy rate of responses, and learning suggestions to obtain the target processing results corresponding to the target task, which are used to comprehensively describe the mastery of the knowledge points in this study.

[0120] Based on any of the above embodiments, considering that users may modify the content of some of the output sub-processing results or issue new restriction information at any time during the entire process of each sub-target intelligent agent executing its corresponding sub-target task and outputting sub-processing results, please also refer to... Figure 8 , Figure 8 A two-branch schematic diagram of controlling a sub-target intelligent agent to output corresponding sub-processing results, provided in an embodiment of this disclosure, includes the following steps in process 800:

[0121] Step 801: In response to the instruction input box being selected, control the currently executing sub-target agent to pause the output of the corresponding sub-processing result;

[0122] The instruction input box is used for user input of instructions. The instruction input box is in an unselected state during the process of the sub-target intelligent agent outputting the corresponding sub-processing result. That is, in the solutions provided in the various embodiments, once the instruction input box is selected, it indicates that the user needs to input some new instructions or needs to interrupt the original sub-processing result output process.

[0123] Step 802: In response to the fact that no new instruction is generated during the process of the instruction input box returning from the selected state to the unselected state, control the currently executing sub-target agent to continue to output the corresponding sub-processing result;

[0124] This step corresponds to a branch case where no new command is generated during the process of the command input box returning from the selected state to the unselected state. That is, the user did not input a new command, so the output result of the sub-processing that was interrupted by the selection of the command input box will continue.

[0125] Step 803: A new instruction is generated in response to the process of the instruction input box returning from a selected state to an unselected state, and correction information is extracted from the new instruction;

[0126] In this step, during the process of the instruction input box returning from a selected state to an unselected state, a new instruction is generated, which is another branch case. Therefore, the aforementioned execution entity needs to extract correction information from the new instruction. This correction information includes the modification of the original information and the addition of new information.

[0127] Step 804: Determine the sub-target agents affected by the correction information;

[0128] Step 805: Control the affected sub-target agent to re-output the corresponding sub-processing result based on the correction information.

[0129] Steps 804 and 805 involve the aforementioned executing entity first determining the sub-target intelligent agents affected by the correction information, and then the executing entity controlling the affected sub-target intelligent agents to re-output the corresponding sub-processing results based on the correction information. Specifically, if all sub-target intelligent agents are affected, the process of outputting the sub-processing results from the first executed sub-target intelligent agent can be restarted based on the correction information.

[0130] To enhance understanding, this disclosure also provides a specific implementation scheme that attempts to eliminate existing technical defects and overcome existing technical problems, taking into account the actual existing technical defects in specific application scenarios:

[0131] The current method of meeting user needs primarily relies on search engines, employing a multi-layered system funnel for recalling, ranking, and heterogeneous result mixing of relevant web pages. Each strategy funnel ranks web pages based on basic relevance, user feedback behavior, and authority, and then truncates the output before sending it to the next layer. The disadvantages of this system are that it can only match relevance at the content level, failing to understand and solve user problems at the task level. Furthermore, the web pages can only meet fixed user needs, making it impossible to achieve high-quality, consistent satisfaction when users express personalized needs or require multiple rounds of further fulfillment. Finally, the entire matching process is unexplainable to the user, leaving them only with final information filtering.

[0132] Related assistant-type products based on large models typically attempt to solve this problem by creating an all-in-one assistant based on a large model. However, a single large model is unlikely to achieve high-quality satisfaction across different open domains and tasks. In professional fields and special scenarios, a single assistant struggles to build trust among users and achieve the desired satisfaction.

[0133] In other words, the relevant technologies face the following challenges in completing tasks in open domains:

[0134] 1) How to efficiently understand and break down the user's expressed needs into key steps;

[0135] 2) How to find the best way to satisfy each key step;

[0136] 3) How to output the execution process and integration results in a user-friendly and complete manner.

[0137] To address the aforementioned challenges, this embodiment proposes a solution for user-demand interactive end-to-end task completion based on a multi-agent collaboration approach. The features of this solution are as follows:

[0138] Previously, searches were typically point-to-point, with a single query (query term, query statement, search terms, search statement) fulfilling one requirement. Each requirement required user input and definition, which was too costly for users. For complex tasks, users often didn't know how to break down the task or devise multiple queries for searching. The paradigm proposed in this embodiment aims to completely assist users in completing tasks. This task completion paradigm allows users to directly and efficiently complete the problem or task at hand. Compared to the single-query fulfillment of searches, the task fulfillment paradigm deeply understands and breaks down the user's complete task, better meeting user needs.

[0139] This embodiment leverages collaboration among multiple intelligent agents to break down and fulfill user needs in multiple steps, ultimately automatically integrating them into a complete solution that satisfies the entire task. This solution can perform the following functions:

[0140] 1) Understand the key steps of the requirements, build the main intelligent agent, understand the user's expressed requirements, and break them down into multiple sub-tasks that the intelligent agent can complete.

[0141] 2) Subtask agent generation: Based on the understanding of key steps and agent capabilities, the main agent is responsible for generating multiple subtask agent candidates needed to complete the current task.

[0142] 3) Interactive task completion: Multiple sub-task agents generate personalized interactive steps and user interfaces to meet the task requirements based on the input of task, key steps and user personalized preferences, allowing the user to complete the entire task through interaction with multiple agents.

[0143] 4) The integrated output of the complete solution: the main intelligent agent integrates the task decomposition process and generates results interactively with multiple intelligent agents, and the interactive presentation of the entire solution is completed through dialogic orchestration.

[0144] In this embodiment, a master agent is constructed to schedule multiple sub-task agents to cooperate. The entire task is completed through cooperation between the master agent and multiple sub-task agents, as well as among the multiple sub-task agents. Compared with search or intelligent assistant products, this method can use agents with better performance in specific fields to achieve better results.

[0145] Furthermore, by providing a scheduling and distribution mechanism based on end-to-end generative large models, the large models have stronger expressive power and can overcome the problems of insufficient performance consistency and global optimality of traditional multi-level ranking mechanisms, enabling task-oriented global optimal combination scheduling optimization. In terms of model capabilities, it can achieve end-to-end aligned training of models from the perspectives of agent understanding and task understanding.

[0146] It also offers a conversational orchestration approach to task completion. Through conversational orchestration, complex tasks are presented in a logical order according to the application scenario, breaking them down into multiple sub-tasks. This provides a very smooth user experience, rather than arbitrarily combining or presenting all the content at once, which can lead to overwhelming information. Furthermore, because the entire task completion process is personalized, from task breakdown to sub-task completion, everything is generated based on user input and preferences. Different solutions, sub-task agent completion schemes, and presentation effects are generated for different users, enhancing user satisfaction.

[0147] The implementation block diagram of this solution is as follows: Figure 9-1 As shown, the system's inputs include user-input requirements and a set of distributable agents (including the basic settings, attached plugins, workflows, etc. of these agents), and the output is multi-layered generated content to complete the task. The entire system is built on a multi-agent collaborative approach, mainly comprising two types of agents: 1) Main Agent: performs overall task understanding, decomposition, and connection, and provides the final result. 2) Task Agent: the Main Agent invokes and summons agents to collaboratively complete the tasks assigned by the Main Agent.

[0148] The solution provided in this embodiment can be widely applied to various information fulfillment applications or standalone products. The following example, using the scenario of the Baidu application, illustrates the specific application form of this embodiment:

[0149] While performing a traditional search, users can switch to the functionality provided in this embodiment with a single click (as shown by the "AI button" in the bottom bar of the image below). For example, if a primary school student enters the phrase "chicken and rabbit in the same cage problem," the results of the web search and the intelligent answer will be displayed (corresponding to...). Figure 9-2 Switch to the interactive multi-agent satisfaction page for the "chicken and rabbit in the same cage problem" via the AI ​​button, as shown below. Figure 9-3 The displayed page generates a graphical teaching method that suits the learning style of primary school students.

[0150] The task completion page displays an automatically generated interactive interface for multiple task agents, allowing users to complete the entire process of teaching, practicing, and testing the "chicken and rabbit in a cage" problem through interaction. The agent "Elementary School Math Teacher Zhang" is responsible for breaking down the entire problem into sub-tasks, including summoning the "Problem-Solving Method Teacher" agent to complete the teaching task, the "Socratic Teaching Assistant" agent to complete the practice task, and the "Chicken and Rabbit in a Cage Test" agent to complete the testing task. The results of each task completion agent are presented through a user-friendly, generative, and interactive UI to complete the sub-tasks, such as... Figures 9-4 to 9-7 The process of completing the interactive teaching task by the "Problem-Solving Method Little Teacher" intelligent agent was presented in sequence.

[0151] Once the user completes an interactive task with an AI agent, the host AI agent, "Elementary School Math Teacher Zhang," guides them to the next practice session. Similarly, the user interacts with the automatically generated UI to complete the "chicken and rabbit in the same cage" problem exercise, such as... Figures 9-8 to 9-9 Then, the main agent schedules the sub-task agent "Chicken and Rabbit in the Same Cage Test" to complete the test of this knowledge point, such as... Figures 9-10 to 9-12 .

[0152] Personalized effects are generated for different users. For example, for junior high school students aged 12-14, the same "chicken and rabbit in the same cage problem" will generate the following results: Figures 9-13 to 9-14 As the comparison shows, it is evident that Figure 9-14 It provides a matching scheme for solving equations based on junior high school knowledge points.

[0153] The above examples demonstrate more clearly that this embodiment, compared to the prior art, has the following improvements and technical effects:

[0154] 1) The task satisfaction paradigm elevates user requests from being treated as mere information retrieval to a comprehensive multi-agent task completion satisfaction model.

[0155] Current search engines can retrieve information by matching the relevance of requests, and intelligent assistants can break down a task into multiple steps and search for solutions step by step. However, task fulfillment through multi-agent collaboration to break down tasks and meet the complete user needs represents a completely new user needs fulfillment model. Furthermore, this paper proposes training that aligns task breakdown with multi-agent capabilities, achieving a high degree of consistency between task breakdown and task completion, thus significantly improving task performance.

[0156] 2) Multi-agent candidate generation based on multi-agent cooperation

[0157] Based on the task decomposition results, a target collaborative agent set is generated end-to-end through a large model. This is a novel approach, distinct from the traditional method of ranking agents through a retrieval and recall system. For multi-agent collaborative tasks, the goal is to combine multiple agents even when the task is incomplete. This requires the large model to deeply understand the task decomposition and the capabilities of each agent, meticulously characterizing the difficulty boundaries of the problem. Ultimately, the optimal agent set for the current objective is obtained through combinatorial optimization. The challenges include:

[0158] Challenge 1: A deep understanding of the capabilities of intelligent agents, including the boundaries of those capabilities and the tasks they excel at, and a detailed description of the subtle differences between their abilities, such as the different creative styles of painting agents and the subtle differences in their image editing capabilities.

[0159] Challenge 2: Understanding the matching relationship between tasks and agent capabilities is not semantic similarity matching, but deep matching in terms of task completion capabilities. This requires a deep characterization and understanding of the tasks, task types, and task boundaries. Matching these on the open set of agent capabilities and task requirements is a significant challenge.

[0160] Challenge 3: Given a multi-agent collaborative task, obtaining the optimal combination is a combinatorial optimization problem, which requires combinatorial optimization.

[0161] 3) The task completion process and results are satisfied through end-to-end generation and automated orchestration of intelligent agents. Currently, search engines struggle to display the retrieval process to users, while intelligent assistants display the thought process and single-step execution results, with the UI generally presented in a pre-defined style. Thanks to the task decomposition, combination, and multi-subtask completion modes of multi-agent collaboration, this invention can present results through end-to-end generation, including the steps, style, and results of each step. Furthermore, the results of each step and the subtasks completed by each intelligent agent are based on the user's personalized preferences, and the results can be further interacted with. This end-to-end generation experience is a comprehensive innovation. Specifically, the end-to-end generative experience includes several parts:

[0162] ① End-to-end generative model: The construction of the main task model goes from task decomposition to generating an intelligent agent set, and the tasks of the intelligent agents are completed by a unified model for training.

[0163] ② End-to-end generation of interactions: The agent's output is not generated from single-turn content, but from multi-turn interactive content and UI interfaces. Compared to the traditional methods of generating single content and pre-defined graphical user interfaces (GUIs) or natural language user interfaces (LUIs), the automated generation and planning of combined interactive interfaces significantly raises the ceiling of a single agent's capabilities. To achieve this, the agent needs to enhance its planning capabilities. In addition to planning basic logic, it needs to complete interaction planning (UIplanning). This planning capability is a major innovation in agents, enabling them to plan not only in terms of thinking tasks but also in terms of external interactive presentation, greatly improving the agent's ability to meet needs and its user-friendliness.

[0164] ③ Interactive orchestration-based interactive fulfillment: From task decomposition to automated orchestration and interactive presentation of task fulfillment, the effectiveness of task completion can be significantly improved. It's equivalent to automating the overall generation of task completion planning, the content of task steps, and the interactive design of step completion, greatly enhancing the multi-agent capability to fulfill task scenarios. This is especially true for complex tasks requiring multiple steps and intricate interactions. For example, learning knowledge points in primary and secondary school textbooks requires a complete process of teaching, practice, and testing for students to truly master them. Traditional educational applications can only solidify the content, failing to adopt different teaching methods for different knowledge points, and struggling to achieve hierarchical coordination of learning, practice, and testing for specific knowledge points. This invention's method can achieve end-to-end complete generation from planning and interaction design to content generation, resulting in a revolutionary improvement in user experience (see application examples for details).

[0165] ④ Incorporating personalization into the entire end-to-end generation process can further enhance the satisfaction effect. For the same task, there are significant differences between different users in terms of solutions, satisfaction methods, interaction design, and outcome planning. End-to-end generation that incorporates personalized information can make the entire effect completely tailored to the user, greatly raising the ceiling of effectiveness.

[0166] Further reference Figure 10 As an implementation of the methods shown in the above figures, this disclosure provides an embodiment of an interactive question-answering task processing device based on multi-agent collaboration. This device embodiment is similar to... Figure 2 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.

[0167] like Figure 10As shown, the interactive question-and-answer task processing device 1000 based on multi-agent collaboration in this embodiment may include: a target task determination unit 1001, a task decomposition and corresponding distribution unit 1002, an interactive interface control and presentation unit 1003, an associated processing result control and generation unit 1004, and a target processing result determination unit 1005. The system includes the following components: a target task determination unit 1001, configured to determine the target task to be solved based on the user's natural language input; a task decomposition and corresponding assignment unit 1002, configured to decompose the target task into multiple sub-target tasks, each containing at least an interactive question-and-answer type task, and assign each sub-target task to a corresponding sub-target agent containing at least an interactive question-and-answer agent; wherein different sub-target agents are used to handle different types of sub-tasks; an interactive interface control and presentation unit 1003, configured to control the interactive question-and-answer agent to present an interactive graphical user interface corresponding to the interactive question-and-answer type task to the user, and to present the answer result corresponding to the received feedback answer; wherein the interactive graphical user interface includes a question and an interactive answer area; an association processing result control and generation unit 1004, configured to control the association sub-target agents among the multiple sub-target agents to generate corresponding association processing results based on the answer results output by the interactive question-and-answer agent; wherein the association sub-target agent is a sub-target agent that relies on the answer result as part of the task input information; and a target processing result determination unit 1005, configured to determine the target processing result corresponding to the target task based on the association processing result.

[0168] In this embodiment, the specific processing and technical effects of the target task determination unit 1001, task decomposition and corresponding distribution unit 1002, interactive interface control and presentation unit 1003, associated processing result control and generation unit 1004, and target processing result determination unit 1005 in the multi-agent collaborative interactive question-and-answer task processing device 1000 can be referred to respectively. Figure 2 The relevant descriptions of steps 201-205 in the corresponding embodiments will not be repeated here.

[0169] In some optional implementations of this embodiment, the target task determination unit 1001 may include:

[0170] The conversion subunit is configured to convert the user's natural language input into natural language text;

[0171] The intent recognition subunit is configured to perform intent recognition on natural language text to obtain intent recognition results that include the intent to learn and master knowledge points.

[0172] The target task determination subunit is configured to determine the target task to be solved based on the intent recognition result; wherein, the target task includes tasks for learning and mastering knowledge points.

[0173] In some optional implementations of this embodiment, the intent recognition subunit can be further configured as follows:

[0174] Perform semantic understanding on natural language text to obtain semantic understanding results;

[0175] The initial intent is determined based on the semantic understanding results, and then personalized adjustments are made based on the user's individual preferences to obtain an intent recognition result that includes the intent to learn and master knowledge points.

[0176] In some optional implementations of this embodiment, the task decomposition and corresponding distribution unit 1002 may include a task decomposition subunit configured to decompose the target task into multiple sub-target tasks, including at least interactive question-and-answer type tasks. The task decomposition subunit may include:

[0177] The task element determination module is configured to determine multiple task elements that constitute the target task; different task elements correspond to different types of tasks.

[0178] The interactive question-and-answer task decomposition module is configured to decompose interactive question-and-answer tasks for learning and mastering knowledge points from the target task based on the individual task elements that require users to subjectively answer the questions raised and to verify whether the knowledge points to which the questions belong have been learned and mastered based on the answer results.

[0179] The sub-target task decomposition module is configured to decompose the target task into multiple sub-target tasks, each containing only a single task element; wherein, the sub-target tasks include at least interactive question-and-answer tasks.

[0180] In some optional implementations of this embodiment, the interactive question-answering agent includes a knowledge point practice answering agent corresponding to the knowledge point practice stage and a knowledge point mastery testing agent corresponding to the practice result verification stage.

[0181] In some optional implementations of this embodiment, the sub-target task decomposition module can be further configured as follows:

[0182] In response to the existence of additional conditions corresponding to task elements, the target task is decomposed into multiple sub-target tasks, each containing only a single task element and corresponding additional conditions; wherein, the additional conditions include: a specified task processing method and / or a specified result presentation format.

[0183] In some optional implementations of this embodiment, the interactive question-answering task processing device 1000 based on multi-agent collaboration may further include: a sub-agent building unit configured to construct different sub-agents, the sub-agent building unit may include:

[0184] The task type set determines the sub-unit, which is configured to determine the set of task types used to create the sub-agent;

[0185] Sub-units are created and configured to be task types in the task type set, and sub-agents with task processing capabilities for the corresponding task type are constructed.

[0186] In some optional implementations of this embodiment, the creation of sub-units can be further configured as follows:

[0187] In response to the fact that the first target type task has at least two task processing methods, a sub-agent corresponding to each task processing method is constructed for the first target type task; wherein, the different task processing methods are abstracted from different processing logics for the same type of task.

[0188] In some optional implementations of this embodiment, the creation of sub-units can be further configured as follows:

[0189] In response to the second objective type task having at least two result display formats, a sub-agent corresponding to each result display format is constructed for the second objective type task; wherein, the result display formats include at least one of the following: text, table, image, video, and interactive card.

[0190] In some optional implementations of this embodiment, the interactive interface control presentation unit 1003 may include:

[0191] The first control subunit is configured to control the interactive question-and-answer agent to determine the question and interactive answer area based on the received task input information; wherein, the received task input information is determined based on task description information related to interactive question-and-answer tasks decomposed from the target task;

[0192] The second control subunit is configured to control the interactive question-and-answer agent to present an interactive graphical user interface containing a question and answer area to the user.

[0193] The third control subunit is configured to control the interactive question-and-answer agent to present a response result determined based on the degree of difference between the feedback answer received from the interactive response area and the standard answer to the question.

[0194] In some optional implementations of this embodiment, the interactive question-answering task processing device 1000 based on multi-agent collaboration may further include:

[0195] The regeneration control unit is configured to respond to an incorrect answer that differs from the standard answer by controlling the interactive question-answering agent to regenerate a new question that is different from the question that was answered incorrectly the previous time, and an interactive graphical user interface that matches the new question; wherein the new question is generated based on the question that was answered correctly the previous time.

[0196] In some optional implementations of this embodiment, the interactive question-answering task processing device 1000 based on multi-agent collaboration may further include:

[0197] The prompt message output control unit is configured to control the interactive question-and-answer agent to output a prompt message indicating that the interactive question-and-answer process is complete in response to receiving a preset number of answer results;

[0198] Correspondingly, the association processing result control generation unit 1004 is further configured as follows:

[0199] In response to receiving a prompt, the control accuracy calculation agent determines the response accuracy based on a preset number of response results.

[0200] In some optional implementations of this embodiment, the target processing result determination unit 1005 can be further configured to:

[0201] The learning suggestion agent in a multi-sub-target intelligent agent outputs matching learning suggestions based on the correctness of the responses;

[0202] By summarizing the number of responses, the accuracy rate of responses, and learning suggestions, we can obtain the target processing results corresponding to the target task.

[0203] In some optional implementations of this embodiment, the received task input information is determined based on the task description information related to interactive question-and-answer tasks decomposed from the target task and the user's personalized preferences.

[0204] In some optional implementations of this embodiment, personalized preferences include at least one of the following:

[0205] Age, education level, areas of interest, and historical consumption preferences.

[0206] In some optional implementations of this embodiment, the first control subunit may include a question determination control module configured to control the interactive question-answering agent to determine the question based on the received task input information. The question determination control module may be further configured to:

[0207] In response to personalized preferences, including age, the interactive question-answering agent determines questions that match the cognitive level corresponding to the age based on the received task input information.

[0208] In some optional implementations of this embodiment, the interactive question-answering task processing device 1000 based on multi-agent collaboration may further include:

[0209] The pause output control unit is configured to control the currently executing sub-agent to pause the output of the corresponding sub-processing result in response to the command input box being selected; wherein, the command input box is used for user input of commands, and the command input box is in an unselected state during the process of the sub-agent outputting the corresponding sub-processing result;

[0210] The correction information extraction unit is configured to extract correction information from the new instruction generated during the process of the instruction input box returning from a selected state to an unselected state.

[0211] The affected sub-target agent determination unit is configured to determine the sub-target agents affected by the correction information;

[0212] The re-output control unit is configured to control the affected sub-target agent to re-output the corresponding sub-processing results based on the correction information.

[0213] This embodiment exists as a device embodiment corresponding to the above method embodiment. The interactive question-and-answer task processing device based on multi-agent collaboration provided in this embodiment adopts an agent cluster formed by a pre-built master agent and multiple sub-agents to process the task requirements proposed by the user. The master agent is responsible for understanding the user's task requirements and decomposing the overall task requirements into multiple sub-target tasks that can be executed by different sub-agents. Each sub-agent processes the sub-tasks that match its own according to the master agent's instructions. That is, through the collaboration between the master agent and each sub-agent, each can perform its own duties to process different components of a complex task. Since different sub-agents are pre-built to handle different types of tasks, the scheme of master agent and multiple sub-agents collaboratively processing task requirements adopted in this disclosure has better processing effect on a single type of task than using a single, all-around agent. Moreover, the relatively small scale of the sub-agents also makes it easier to flexibly add and modify the corresponding functions. This scheme can bring better overall task processing effect with lower overall cost.

[0214] Especially for interactive question-and-answer tasks that require users to subjectively answer questions and verify the results, the interactive question-and-answer agent can provide users with an interactive graphical interface that includes question and interactive answer areas, so that users' task requirements can be met more quickly through interactive answering, thereby improving the task solving effect.

[0215] According to embodiments of this disclosure, this disclosure also provides an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enable the at least one processor to implement the interactive question-answering task processing method based on multi-agent collaboration described in any of the above embodiments.

[0216] According to embodiments of this disclosure, this disclosure also provides a readable storage medium storing computer instructions that enable a computer to execute the interactive question-answering task processing method based on multi-agent collaboration described in any of the above embodiments.

[0217] According to embodiments of this disclosure, this disclosure also provides a computer program product that, when executed by a processor, can implement the interactive question-answering task processing method based on multi-agent collaboration described in any of the above embodiments.

[0218] Figure 11 A schematic block diagram of an example electronic device 1100 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0219] like Figure 11 As shown, device 1100 includes a computing unit 1101, which can perform various appropriate actions and processes according to a computer program stored in read-only memory (ROM) 1102 or a computer program loaded from storage unit 1108 into random access memory (RAM) 1103. The RAM 1103 may also store various programs and data required for the operation of device 1100. The computing unit 1101, ROM 1102, and RAM 1103 are interconnected via bus 1104. Input / output (I / O) interface 1105 is also connected to bus 1104.

[0220] Multiple components in device 1100 are connected to I / O interface 1105, including: input unit 1106, such as keyboard, mouse, etc.; output unit 1107, such as various types of monitors, speakers, etc.; storage unit 1108, such as disk, optical disk, etc.; and communication unit 1109, such as network card, modem, wireless transceiver, etc. Communication unit 1109 allows device 1100 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0221] The computing unit 1101 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 1101 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 1101 performs the various methods and processes described above, such as a multi-agent cooperative interactive question-answering task processing method. For example, in some embodiments, the multi-agent cooperative interactive question-answering task processing method can be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 1108. In some embodiments, part or all of the computer program can be loaded and / or installed on device 1100 via ROM 1102 and / or communication unit 1109. When the computer program is loaded into RAM 1103 and executed by the computing unit 1101, one or more steps of the multi-agent cooperative interactive question-answering task processing method described above can be performed. Alternatively, in other embodiments, computing unit 1101 may be configured by any other suitable means (e.g., by means of firmware) to perform an interactive question-and-answer task processing method based on multi-agent cooperation.

[0222] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0223] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0224] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0225] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0226] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0227] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and Virtual Private Server (VPS) services, such as high management difficulty and weak business scalability.

[0228] The technical solution of this disclosure employs an agent cluster consisting of a pre-built master agent and multiple sub-agents to handle user-initiated task requests. The master agent is responsible for understanding the user's task requirements and decomposing the overall task requirements into multiple sub-target tasks that can be executed by different sub-agents. Each sub-agent then processes its assigned sub-task according to the master agent's instructions. Through collaboration between the master agent and the sub-agents, each entity can effectively handle different components of a complex task. Since different sub-agents are pre-built specifically for handling different types of tasks, this scheme, which uses a master agent and multiple sub-agents to collaboratively process task requests, achieves better processing results for single-type tasks compared to using a single, all-encompassing agent. Furthermore, the relatively small number of sub-agents allows for flexible addition and modification of their functions, resulting in better overall task processing performance at a lower overall cost.

[0229] Especially for interactive question-and-answer tasks that require users to subjectively answer questions and verify the results, the interactive question-and-answer agent can provide users with an interactive graphical interface that includes question and interactive answer areas, so that users' task requirements can be met more quickly through interactive answering, thereby improving the task solving effect.

[0230] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0231] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for processing interactive question-answering tasks based on multi-agent collaboration, applied to the main agent, comprising: The target task to be solved is determined based on the user's natural language input; The target task is decomposed into multiple sub-target tasks, each of which includes at least an interactive question-and-answer task. Each sub-target task is then assigned to a sub-target agent that includes at least an interactive question-and-answer agent. Different sub-agents are used to handle different types of sub-tasks. The interactive question-answering agent is controlled to present an interactive graphical user interface (GUI) corresponding to the interactive question-answering task to the user, and to present the answer result corresponding to the received feedback answer; the interactive GUI includes a question and an interactive answer area; in response to the feedback answer being an incorrect answer that differs from the standard answer, the interactive question-answering agent is controlled to regenerate a new question that differs from the question corresponding to the previous incorrect answer and an interactive GUI matching the new question; the new question is generated based on the question corresponding to the previous correct answer; The associated sub-target intelligent agents among the multiple sub-target intelligent agents are controlled to generate corresponding associated processing results based on the answer results output by the interactive question-answering intelligent agent; The associated sub-target agent is a sub-target agent that relies on the answer result as part of the task input information; Based on the association processing results, determine the target processing result corresponding to the target task; In response to the command input box being selected, the currently executing sub-agent pauses the output of the corresponding sub-processing result; The instruction input box is used for the user to input instructions. The instruction input box is in an unselected state during the process of the sub-agent outputting the corresponding sub-processing result. A new instruction is generated in response to the process of the instruction input box returning from the selected state to the unselected state. Correction information is extracted from the new instruction. The sub-target agent affected by the correction information is determined. The affected sub-target agent is controlled to re-output the corresponding sub-processing result in combination with the correction information.

2. The method according to claim 1, wherein, The step of determining the target task to be solved based on the user's natural language input includes: Convert the user's natural language input into natural language text; Intent recognition is performed on the natural language text to obtain intent recognition results that include the intent to learn and master knowledge points; The target task to be solved is determined based on the intent recognition result; wherein, the target task includes a task for learning and mastering knowledge points.

3. The method according to claim 2, wherein, The process of performing intent recognition on the natural language text to obtain intent recognition results that include the intent to learn and master knowledge points includes: Perform semantic understanding on the natural language text to obtain the semantic understanding result; The initial intent is determined based on the semantic understanding results, and the initial intent is then personalized and modified using the user's personalized preferences to obtain an intent recognition result that includes the intent to learn and master knowledge points.

4. The method according to claim 2, wherein, The step of decomposing the target task into multiple sub-target tasks, including at least interactive question-and-answer tasks, includes: Identify multiple task elements that constitute the target task; wherein, different task elements correspond to different types of tasks; Based on the specific task element that requires users to subjectively answer the questions raised and to verify whether the knowledge points to which the questions belong have been learned and mastered based on the answer results, interactive question-and-answer tasks for learning and mastering knowledge points are decomposed from the target task. The target task is decomposed into multiple sub-target tasks, each containing only a single task element; wherein, each sub-target task contains at least the interactive question-and-answer task.

5. The method according to claim 4, wherein, The interactive question-and-answer agent includes a knowledge point practice and answering agent corresponding to the knowledge point practice stage and a knowledge point mastery testing agent corresponding to the practice result verification stage.

6. The method according to claim 4, wherein, The step of decomposing the target task into multiple sub-target tasks, each containing only a single task element, includes: In response to the existence of additional conditions corresponding to the task element, the target task is decomposed into multiple sub-target tasks, each containing only a single task element and a corresponding additional condition; wherein the additional conditions include: a specified task processing method and / or a specified result presentation format.

7. The method according to claim 1, wherein, The process of constructing different sub-agents includes: Determine the set of task types to be used to create the sub-agent; For each type of task in the set of task types, a sub-agent with the task processing capabilities of the corresponding type of task is constructed.

8. The method according to claim 7, wherein, For each type of task in the set of task types, a sub-agent with the corresponding task processing capabilities is constructed, including: In response to a first target type task having at least two task processing methods, a sub-agent corresponding to each of the task processing methods is constructed for the first target type task; wherein, the different task processing methods are abstracted from different processing logics for the same type of task.

9. The method according to claim 7, wherein, For each type of task in the set of task types, a sub-agent with the corresponding task processing capabilities is constructed, including: In response to a second target type task having at least two result display formats, a sub-agent is constructed for the second target type task, corresponding to each of the result display formats; wherein, the result display formats include at least one of: text, table, image, video, and interactive card.

10. The method according to claim 1, wherein, The control mechanism of the interactive question-and-answer agent presents an interactive graphical user interface corresponding to the interactive question-and-answer task to the user, and presents the answer result corresponding to the received feedback answer, including: The interactive question-answering agent is controlled to determine the question and the interactive answering area based on the received task input information; wherein, the received task input information is determined based on task description information related to interactive question-answering tasks decomposed from the target task; The interactive question-and-answer agent is controlled to present an interactive graphical user interface to the user, which includes the question and the interactive answer area. The interactive question-and-answer agent is controlled to present a response result determined based on the degree of difference between the feedback answer received from the interactive response area and the standard answer to the question.

11. The method according to claim 1, further comprising: In response to receiving a preset number of answers, the interactive question-and-answer agent is controlled to output a prompt message indicating that the interactive question-and-answer process is complete. Correspondingly, the controlling sub-target intelligent agent among the plurality of sub-target intelligent agents generates corresponding association processing results based on the answer results output by the interactive question-answering intelligent agent, including: In response to receiving the prompt information, the control accuracy calculation agent determines the response accuracy based on the preset number of response results.

12. The method according to claim 11, wherein, The step of determining the target processing result corresponding to the target task based on the association processing result includes: The learning suggestion agent among the multiple sub-target agents outputs matching learning suggestions based on the response accuracy. By summarizing the number of responses, the accuracy rate of the responses, and the learning suggestions, the target processing result corresponding to the target task is obtained.

13. The method according to claim 10, wherein, The received task input information is determined based on the task description information related to interactive question-and-answer tasks decomposed from the target task and the user's personalized preferences.

14. The method according to claim 13, wherein, The personalized preferences include at least one of the following: Age, education level, areas of interest, and historical consumption preferences.

15. The method according to claim 14, wherein, Controlling the interactive question-answering agent to determine the question based on the received task input information includes: In response to the personalized preferences, including the age, the interactive question-answering agent is controlled to determine a question that matches the cognitive level corresponding to the age based on the received task input information.

16. An interactive question-answering task processing device based on multi-agent collaboration, applied to a main agent, comprising: The target task determination unit is configured to determine the target task to be solved based on the user's natural language input. The task decomposition and corresponding distribution unit is configured to decompose the target task into multiple sub-target tasks, which at least include interactive question-and-answer tasks, and distribute each sub-target task to each sub-target agent, which at least includes an interactive question-and-answer agent; wherein, different sub-agents are used to handle different types of sub-tasks. An interactive interface control and presentation unit is configured to control the interactive question-and-answer agent to present an interactive graphical user interface corresponding to the interactive question-and-answer task to the user, and to present the answer result corresponding to the received feedback answer; wherein, the interactive graphical user interface includes a question and an interactive answer area; The regeneration control unit is configured to, in response to an incorrect answer that differs from the standard answer, control the interactive question-answering agent to regenerate a new question that differs from the question that appeared when the previous answer was incorrect, and an interactive graphical user interface that matches the new question; wherein the new question is generated based on the question that appeared when the previous answer was correct. The association processing result control generation unit is configured to control the association sub-target intelligent agent among the plurality of sub-target intelligent agents to generate corresponding association processing results based on the answer results output by the interactive question-answering intelligent agent; wherein, the association sub-target intelligent agent is a sub-target intelligent agent that depends on the answer results as part of the task input information; The target processing result determination unit is configured to determine the target processing result corresponding to the target task based on the associated processing result; The pause output control unit is configured to control the currently executing sub-agent to pause the output of the corresponding sub-processing result in response to the instruction input box being selected; the instruction input box is used for the user to input instructions, and the instruction input box is in an unselected state during the process of the sub-agent outputting the corresponding sub-processing result; The correction information extraction unit is configured to extract correction information from the new instruction when a new instruction is generated during the process of the instruction input box returning from the selected state to the unselected state. The affected sub-target agent determination unit is configured to determine the sub-target agents affected by the correction information; The re-output control unit is configured to control the affected sub-target agent to re-output the corresponding sub-processing result in conjunction with the correction information.

17. The apparatus according to claim 16, wherein, The target task determination unit includes: The conversion subunit is configured to convert the user's natural language input into natural language text; The intent recognition subunit is configured to perform intent recognition on the natural language text to obtain intent recognition results that include the intent to learn and master knowledge points. The target task determination subunit is configured to determine the target task to be solved based on the intent recognition result; wherein the target task includes a task for learning and mastering knowledge points.

18. The apparatus according to claim 17, wherein, The intent recognition subunit is further configured to: Perform semantic understanding on the natural language text to obtain the semantic understanding result; The initial intent is determined based on the semantic understanding results, and the initial intent is then personalized and modified using the user's personalized preferences to obtain an intent recognition result that includes the intent to learn and master knowledge points.

19. The apparatus according to claim 17, wherein, The task decomposition and corresponding distribution unit includes a task decomposition subunit configured to decompose the target task into multiple sub-target tasks, including at least interactive question-and-answer tasks. The task decomposition subunit includes: The task element determination module is configured to determine multiple task elements that constitute the target task; wherein, different task elements correspond to different types of tasks. The interactive question-and-answer task decomposition module is configured to decompose interactive question-and-answer tasks for learning and mastering knowledge points from the target task based on a single task element that requires users to subjectively answer questions and verify whether the knowledge points to which the questions belong have been learned and mastered based on the answer results. The sub-target task decomposition module is configured to decompose the target task into multiple sub-target tasks that each contain only a single task element; wherein, the sub-target task contains at least the interactive question-and-answer type task.

20. The apparatus according to claim 19, wherein, The interactive question-and-answer agent includes a knowledge point practice and answering agent corresponding to the knowledge point practice stage and a knowledge point mastery testing agent corresponding to the practice result verification stage.

21. The apparatus according to claim 19, wherein, The sub-target task decomposition module is further configured to: In response to the existence of additional conditions corresponding to the task element, the target task is decomposed into multiple sub-target tasks, each containing only a single task element and a corresponding additional condition; wherein the additional conditions include: a specified task processing method and / or a specified result presentation format.

22. The apparatus of claim 16, further comprising: Sub-agent building units are configured to construct different sub-agents, the sub-agent building units comprising: The task type set determines the sub-unit, which is configured to determine the set of task types used to create the sub-agent; Sub-units are created and configured to be tasks of each type in the set of task types, and sub-agents with task processing capabilities for the corresponding task type are constructed.

23. The apparatus according to claim 22, wherein, The creation of the subunit is further configured to: In response to a first target type task having at least two task processing methods, a sub-agent corresponding to each of the task processing methods is constructed for the first target type task; wherein, the different task processing methods are abstracted from different processing logics for the same type of task.

24. The apparatus according to claim 22, wherein, The creation of the subunit is further configured to: In response to a second target type task having at least two result display formats, a sub-agent is constructed for the second target type task, corresponding to each of the result display formats; wherein, the result display formats include at least one of: text, table, image, video, and interactive card.

25. The apparatus according to claim 16, wherein, The interactive interface control and presentation unit includes: The first control subunit is configured to control the interactive question-answering agent to determine the question and the interactive answering area based on the received task input information; wherein the received task input information is determined based on task description information related to interactive question-answering tasks decomposed from the target task; The second control subunit is configured to control the interactive question-and-answer agent to present an interactive graphical user interface containing the question and the interactive answer area to the user. The third control subunit is configured to control the interactive question-answering agent to present a response result determined based on the degree of difference between the feedback answer received from the interactive answering area and the standard answer to the question.

26. The apparatus of claim 16, further comprising: The prompt information output control unit is configured to control the interactive question-and-answer agent to output a prompt information indicating that the interactive question-and-answer is completed in response to receiving a preset number of answer results; Correspondingly, the association processing result control generation unit is further configured to: In response to receiving the prompt information, the control accuracy calculation agent determines the response accuracy based on the preset number of response results.

27. The apparatus according to claim 26, wherein, The target processing result determination unit is further configured to: The learning suggestion agent among the multiple sub-target agents outputs matching learning suggestions based on the response accuracy. By summarizing the number of responses, the accuracy rate of the responses, and the learning suggestions, the target processing result corresponding to the target task is obtained.

28. The apparatus according to claim 25, wherein, The received task input information is determined based on the task description information related to interactive question-and-answer tasks decomposed from the target task and the user's personalized preferences.

29. The apparatus according to claim 28, wherein, The personalized preferences include at least one of the following: Age, education level, areas of interest, and historical consumption preferences.

30. The apparatus according to claim 29, wherein, The first control subunit includes a question-determination control module configured to control the interactive question-answering agent to determine the question based on received task input information, the question-determination control module being further configured to: In response to the personalized preferences, including the age, the interactive question-answering agent is controlled to determine a question that matches the cognitive level corresponding to the age based on the received task input information.

31. An electronic device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the interactive question-answering task processing method based on multi-agent cooperation as described in any one of claims 1-15.

32. A non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the interactive question-answering task processing method based on multi-agent cooperation as described in any one of claims 1-15.

33. A computer program product comprising a computer program that, when executed by a processor, implements the steps of the interactive question-answering task processing method based on multi-agent collaboration according to any one of claims 1-15.

Images