Task interaction method and device based on intelligent agent, equipment and storage medium

Abstract

The embodiment of the invention provides an agent-based task interaction method and device, equipment and a storage medium. The method comprises the steps that in response to a received request for indicating a target task, an execution plan of the target task is determined through an intelligent agent, and the execution plan indicates at least one subtask; during execution of the target task according to the execution plan, presenting a content object representing an intermediate execution result of the target task, the content object having an at least partially structured visual pattern, the intermediate execution result comprising at least a portion of a corresponding execution result of the at least one sub-task; in response to receiving feedback for at least a portion of the content object, a response to the feedback is provided with the agent.

Description

Technical Field

[0001] The exemplary embodiments disclosed herein generally relate to the field of computers, and particularly to methods, apparatuses, devices, computer-readable storage media, and computer-executable instruction products for agent-based task interaction. Background Technology

[0002] With the development of information technology, an increasing number of intelligent systems possess the ability to autonomously analyze and generate content. These systems typically obtain user commands through natural language interaction and combine this with model reasoning capabilities and tool invocation capabilities to generate target content. However, in practical applications, most systems employ relatively simple generation methods. Usually, after the user inputs a request, the corresponding generated content is directly output. Users often have no choice but to wait for the final result after submitting their request. Summary of the Invention

[0003] In a first aspect of this disclosure, a method for task interaction based on an intelligent agent is provided. The method includes: in response to receiving a request instructing a target task, utilizing an intelligent agent to determine an execution plan for the target task, the execution plan instructing at least one subtask; during execution of the target task according to the execution plan, presenting a content object representing intermediate execution results of the target task, the content object having a visually structured style at least partially, the intermediate execution results including at least a portion of the corresponding execution results of at least one subtask; and in response to receiving feedback on at least a portion of the content object, utilizing the intelligent agent to provide a response to the feedback.

[0004] In a second aspect of this disclosure, an apparatus for agent-based task interaction is provided. The apparatus includes: a plan determination module configured to, in response to receiving a request instructing a target task, utilize an agent to determine an execution plan for the target task, the execution plan indicating at least one subtask; a presentation module configured to, during execution of the target task according to the execution plan, present a content object representing intermediate execution results of the target task, the content object having at least partially structured visual styles, the intermediate execution results including at least a portion of the corresponding execution results of at least one subtask; and a response module configured to, in response to receiving feedback on at least a portion of the content object, utilize an agent to provide a response to the feedback.

[0005] In a third aspect of this disclosure, an electronic device is provided. The device includes at least one processor; and at least one memory coupled to the at least one processor and storing instructions for execution by the at least one processor. When executed by the at least one processor, the instructions cause the device to perform the method of the first aspect.

[0006] In a fourth aspect of this disclosure, a computer-readable storage medium is provided. The computer-readable storage medium stores computer-executable instructions that can be executed by a processor to implement the method of the first aspect.

[0007] In a fifth aspect of this disclosure, a computer program product is provided, including computer-executable instructions, wherein when executed by a processor, the computer-executable instructions implement the method according to a first aspect of this disclosure.

[0008] It should be understood that the content described in this content section is not intended to limit the key or essential features of the embodiments of this disclosure, nor is it intended to restrict the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0009] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein:

[0010] Figure 1 A schematic diagram of an example environment in which embodiments of the present disclosure can be implemented is shown;

[0011] Figures 2A to 2F A schematic diagram of an interface for interacting with an intelligent agent according to some embodiments of the present disclosure is shown;

[0012] Figure 3 A flowchart of a method for task interaction according to some embodiments of the present disclosure is shown;

[0013] Figure 4 A schematic structural block diagram of a device for task interaction according to some embodiments of the present disclosure is shown; and

[0014] Figure 5 A block diagram of an electronic device in which one or more embodiments of the present disclosure may be implemented is shown. Detailed Implementation

[0015] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0016] In the description of embodiments of this disclosure, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The term "some embodiments" should be understood as "at least some embodiments". Other explicit and implicit definitions may also be included below.

[0017] In this document, unless explicitly stated otherwise, performing a step in response to A does not mean that the step is performed immediately after A, but may include one or more intermediate steps.

[0018] It is understood that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and related provisions.

[0019] It is understood that before using the technical solutions disclosed in the various embodiments of this disclosure, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this disclosure through appropriate means in accordance with relevant laws and regulations, and user authorization should be obtained.

[0020] For example, in response to receiving a user's active request, a prompt message is sent to the user to clearly inform the user that the requested operation will require the acquisition and use of the user's personal information, thereby enabling the user to choose whether to provide personal information to the software or hardware such as electronic devices, applications, servers or storage media that perform the operation of the technical solution disclosed herein, based on the prompt message.

[0021] As an optional but non-restrictive implementation, in response to a user's active request, a prompt message can be sent to the user, such as a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0022] It is understood that the above notification and user authorization process are merely illustrative and do not constitute a limitation on the implementation of this disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this disclosure.

[0023] As used in this paper, the term "model" refers to a model that learns the relationship between inputs and outputs from training data, enabling it to generate corresponding outputs for a given input after training. Model generation can be based on machine learning techniques. Deep learning is a machine learning algorithm that processes inputs and provides corresponding outputs using multiple layers of processing units. A neural network model is an example of a deep learning-based model. In this paper, "model" may also be referred to as a "machine learning model," "learning model," "machine learning network," or "learning network," and these terms are used interchangeably.

[0024] A neural network is a machine learning network based on deep learning. A neural network processes input and provides a corresponding output, typically consisting of an input layer, an output layer, and one or more hidden layers between the input and output layers. Neural networks used in deep learning applications often include many hidden layers, thus increasing the network's depth. The layers of a neural network are connected sequentially, so that the output of the previous layer is provided as the input to the next layer. The input layer receives the input to the neural network, while the output layer's output serves as the final output. Each layer of a neural network includes one or more nodes (also called processing nodes or neurons), each node processing the input from the layer above.

[0025] Machine learning typically comprises three phases: training, testing, and application (also known as inference). In the training phase, a given model is trained using a large amount of training data, iteratively updating its parameter values ​​until the model can consistently generate inferences that meet the expected goals from the training data. Through training, the model can be considered to have learned the relationship between inputs and outputs (also known as the input-output mapping) from the training data. The parameter values ​​of the trained model are determined. In the testing phase, test inputs are applied to the trained model to test whether it can provide the correct output, thus determining the model's performance. In the application phase, the model can be used to process actual inputs based on the trained parameter values ​​to determine the corresponding output.

[0026] Figure 1 A schematic diagram of an example environment 100 in which embodiments of the present disclosure can be implemented is shown. In this example environment 100, an application 125 is installed on a terminal device 110. A user 140 can interact with the application 125 via the terminal device 110 and / or an attached device of the terminal device 110.

[0027] In some embodiments, application 125 can be downloaded and installed on terminal device 110. In some embodiments, application 125 can also be accessed in other ways, such as through a web page. Figure 1In environment 100, in response to application 125 being launched, terminal device 110 can display the interface 150 of application 125.

[0028] In some embodiments, terminal device 110 communicates with server 130 to provide services to application 125. Terminal device 110 can be any type of mobile terminal, fixed terminal, or portable terminal, including mobile phones, desktop computers, laptop computers, notebook computers, netbook computers, tablet computers, media computers, multimedia tablets, personal communication system (PCS) devices, personal navigation devices, personal digital assistants (PDAs), audio / video players, digital cameras / camcorders, television receivers, radio receivers, e-book devices, gaming devices, or any combination thereof, including accessories and peripherals of these devices or any combination thereof. In some embodiments, terminal device 110 can also support any type of user-facing interface (such as "wearable" circuitry). Application 125 can be various types of computing systems / servers capable of providing computing power, including but not limited to mainframes, edge computing nodes, computing devices in cloud environments, etc.

[0029] In embodiments of this disclosure, application 125 can provide interaction functionality with an intelligent agent. Application 125 can be an application dedicated to providing intelligent agent services, or an application integrated with an intelligent agent (i.e., it can provide functions or services other than those of an intelligent agent). Although Figure 1 The image shows a single application, but in reality, multiple applications can be installed on the terminal device 110.

[0030] In this disclosure, agents 160-1, 160-2, ..., 160-N (collectively or individually referred to as agent 160) can be deployed locally on terminal device 110 or remotely. In the case of remote deployment, terminal device 110 can directly invoke the agent or invoke the agent via server 130.

[0031] In embodiments of this disclosure, agent 160 may possess intelligent dialogue and task processing capabilities. Terminal device 110 provides an interface 150 that can present interactions with agent 160. In interface 150, user 140 can initiate task requests to agent 160 by inputting natural language (text input or voice input). Alternatively or additionally, user 140 can upload online or offline file conversations to instruct agent 160 to assist in completing various tasks.

[0032] In embodiments of this disclosure, during interaction with user 140, agent 160 can respond to user 140's requests and handle tasks instructed by the user. In some embodiments, during task processing, agent 160 can invoke one or more tools 165-1, 165-2, ... 165-M (collectively or individually referred to as tools 165) to assist in task execution and the provision of task results, as needed for the task. These tools 165 can be any type of tool, such as text generation tools, file reading tools, information search tools, online or offline databases, image processing tools, chart generation tools, web page creation tools, etc.

[0033] In some embodiments, environment 100 may further include a management node for a plurality of agents 160-1, 160-2, ... 160-N, which can interact with the agents 160-1, 160-2, ... 160-N. In some examples, the management node may, in response to a task request from user 140, determine the task requirements corresponding to the task request. The management node may then, based on the task requirements, assign the task request to the agent 160 that matches the task requirements, requesting that agent 160 to execute the task. In other examples, the management node may also determine an execution plan for the task based on the task requirements. The execution plan may indicate one or more subtasks required to complete the task. The management node may assign the one or more subtasks to one or more agents 160, which will then execute their respective subtasks. Regarding the management node, in some examples, the management node may be implemented by one of the agents 160-1, 160-2, ... 160-N (in this case, agent 160 is also referred to as scheduling agent 160). In other examples, management nodes can be implemented using machine learning models, such as language models (LM).

[0034] In some embodiments, agent 160 may be constructed based on one or more machine learning models. In some embodiments, the machine learning model upon which agent 160 is based may include at least a language model (LM), such as a large language model (LLM). In some embodiments, the machine learning model upon which agent 160 is based may include a multimodal model capable of handling multiple modalities of input, such as text input, visual input (e.g., images, videos), audio input, etc. These machine learning models may include content-generating models capable of generating corresponding outputs based on model inputs. In some embodiments, the machine learning model may receive text-modal model inputs (e.g., natural language and / or machine language) and / or non-text-modal model inputs (e.g., images, speech, videos, etc.), and may obtain corresponding model outputs based on model inputs and prompts, thereby completing the task execution.

[0035] It should be understood that the structure and function of the various elements in environment 100 are described for illustrative purposes only and do not imply any limitation on the scope of this disclosure.

[0036] As mentioned above, with the development of information technology, more and more intelligent systems are capable of performing tasks based on user input. These systems typically obtain user instructions through natural language interaction and combine model reasoning and tool invocation capabilities to provide users with various task processing services, including content generation, question answering, recommendation, and search. For example, based on user input keywords and descriptive statements, they can output various generated results containing text, images, videos, or page configurations.

[0037] In practical applications, traditional intelligent systems generally adopt a direct-output generation process based on "request-response," where the system directly executes the generation operation and outputs the result after the user inputs a request. This generation method has several limitations. For example, on the one hand, the system relies on understanding the surface semantics of the user's initial input, making it difficult to promptly identify problems such as insufficient expression of needs or deviations in direction. On the other hand, the system's generation process is invisible to the user. While waiting, the user cannot understand the model's thought process or interim results, nor can they supplement, correct, or refine their needs during task execution. This type of interaction mode easily leads to deviations between the generated content and the user's actual needs, increasing the cost of repeated modifications and waiting. Therefore, how to ensure task execution efficiency while making the task results more in line with user needs has become a noteworthy issue.

[0038] In related technologies, some intelligent systems or tools offer functions similar to "deep search" or "deep thinking," displaying the agent's reasoning or thought processes in a certain form during the generation process. However, the presentation of such functions is mostly in the form of simple text descriptions, and no interactive or corrective mechanisms are provided during task execution. Therefore, this approach still cannot solve the aforementioned problems of user experience and result quality.

[0039] In view of this, embodiments of this disclosure propose an improved scheme for agent-based task interaction. In this improved scheme, firstly, in response to receiving a request indicating a target task, an agent is used to determine an execution plan for the target task, the execution plan indicating at least one subtask. During the execution of the target task according to the execution plan, a content object representing intermediate execution results of the target task can be presented, the content object having at least a partially structured visual style. The intermediate execution results may include at least a portion of the corresponding execution results of at least one subtask. Further, in response to receiving feedback on at least a portion of the content object, an agent can be used to provide a response to the feedback.

[0040] In the embodiments of this disclosure, structured content objects representing intermediate execution results are presented during task execution. Furthermore, users can provide feedback on specific parts of the content objects. The system can provide corresponding responses based on this feedback, thereby flexibly supplementing or correcting requirements during task execution. This approach avoids repeated modifications caused by insufficient expression of requirements in traditional direct output modes and improves the matching degree between the generated results and the user's actual needs. In addition, the structured presentation of content objects allows users to quickly locate, view, and adjust the task execution process within a visual interface. This further enhances the transparency of interaction and operational efficiency.

[0041] In this disclosure, "trigger" refers to one or more interactive actions performed by a user on a terminal device. Furthermore, these interactive actions may be triggered within the same user interface / pop-up window or within different user interfaces / pop-up windows. This disclosure is not limited in this respect.

[0042] For ease of understanding, the following text will refer to Figure 2A and Figure 2F Examples of interfaces for interacting with an intelligent agent in some embodiments of this disclosure are described below. These example interfaces may be presented by application 125 in terminal device 110. It should be understood that the interfaces shown in the figures are merely examples, and various interface designs may exist in practice. The various graphical elements in the interface may have different arrangements and different visual representations, one or more elements may be omitted or replaced, and one or more other elements may also be present. Embodiments of this disclosure are not limited in this respect.

[0043] exist Figure 2A The interface 200A shown can present an interactive interface 200A between user 140 and agent 160. This interactive interface 200A can receive requests for new tasks initiated by the user. For example... Figure 2A As shown, users can trigger a new task by activating the "New Task" control 211. In response to the activation of the "New Task" control 211, a task-specific interactive interface is presented in interface 200A, such as... Figure 2B The interactive interface 200B shown is used to initiate new tasks. In some embodiments, the interface 200A also provides an input area 215 for task input, in which the user can enter a request for the task to be triggered and initiate the task request via a "send" control 216. The input area 215 can support text input, such as entering text in a text input box, and can also support voice input, such as by triggering a voice control 217. In addition, the input area 215 can also provide an upload control 216 to support uploading attachments to indicate the user's task request. In some embodiments, the interactive interface 200A also provides a task list 212 initiated to the agent 160, the task list 212 indicating each initiated task and the execution status of the task (e.g., interrupted, in progress, completed, etc.).

[0044] In some embodiments, multiple operating modes for the user 140 and the agent 160 may be provided, and the user may flexibly switch between these operating modes, for example via... Figure 2A The mode switching control 218 is shown. When a certain mode is triggered, the corresponding interactive area is displayed to facilitate interaction between user 140 and agent 160. The interaction methods between user 140 and agent 160 are different in different operation modes, which can flexibly adapt to the interaction needs of different application scenarios.

[0045] In some embodiments, in response to receiving a request indicating a target task, terminal device 110 may utilize agent 160 to determine an execution plan for the target task. The execution plan indicates at least one subtask. For example, agent 160 may break down the subtasks required to complete the target task based on a natural language request input by the user through input area 215.

[0046] As an example, upon receiving a request instructing on a target task, terminal device 110 can present... Figure 2BThe interactive interface 200B is shown. In the interactive interface 200B, user-initiated requests can be presented via message 221. For example, a user's request could be, "I'm considering buying a bicycle. My needs are XXXX. Please help me with a comprehensive study, provide information on choosing a bicycle, recommend suitable models based on my actual situation, and advise me on what I should pay attention to when purchasing." Alternatively or additionally, the user can specify the scope of the task, preferences, output format, or publishing platform in the request.

[0047] In some examples, during task execution, agent 160 can analyze the user-initiated task request to determine task requirements, thereby generating an execution plan for the task. The task execution plan can indicate one or more subtasks required to complete the task. Examples include information retrieval, data analysis, and content generation. In response to the determination of the execution plan, agent 160 can automatically or in response to user confirmation execute the target task according to the execution plan.

[0048] In some examples, to help users better understand the agent's task execution process, after the execution plan is determined, descriptions of at least one subtask can be presented in the interactive interface 200B. For example, this could include the task's analysis approach, execution steps, prompt word construction, and tool call paths. The descriptions of at least one subtask can be presented in the interactive interface 200B in the form of a message list 222.

[0049] The following description continues with reference to the accompanying drawings, outlining some exemplary embodiments of this disclosure. The embodiments described herein can be implemented on terminal device 110. It should be noted that the operations performed by terminal device 110 may specifically be performed by relevant applications (e.g., application 120) installed on terminal device 110. Some operations described with reference to terminal device 110 may require the assistance of server 130 to complete.

[0050] In some embodiments, the terminal device 110 may present a content object representing intermediate execution results of the target task during execution of the target task according to the execution plan. The content object has a visually structured style that is at least partially structured. The intermediate execution results include at least a portion of the corresponding execution results of at least one subtask.

[0051] To enhance user perception of task progress and interactive experience, intermediate execution results of the target task can be dynamically presented during task execution. These intermediate results can include internal model inferences or raw, interim outputs from the agent 160 during subtask execution. For example, intermediate results may include searched information, task analysis processes, inference chains, calculated intermediate values, and so on.

[0052] Content objects representing intermediate execution results of a target task are content that is determined based on these intermediate results and is viewable or interactive for the user. Content objects can be understood as intermediate products of task execution, possessing at least a partially structured presentation. Presenting content objects representing intermediate execution results of the target task during its execution process helps users understand the task's progress more promptly and intuitively.

[0053] In some embodiments, agent 160 may execute at least one subtask based on a determined execution plan to obtain corresponding execution results for at least one subtask. Further, agent 160 may determine whether the execution results of one or more subtasks currently obtained meet predetermined presentation conditions. For example, agent 160 may determine whether these execution results include sufficient key information to support user understanding or decision-making, and whether they meet set completeness or credibility thresholds. As another example, agent 160 may determine whether these execution results require user confirmation or feedback.

[0054] In some embodiments, in response to the execution results of one or more subtasks in at least one subtask satisfying predetermined presentation conditions, the agent 160 can generate a content object based on the execution results of one or more subtasks. The content object can serve as an intermediate presentation unit during task execution. For example, the content object can be content suitable for user reading and understanding, formed by abstracting and structuring the phased results generated by the agent 160 during the execution of subtasks. During the generation of the content object, the agent 160 can filter, summarize, and format the intermediate execution results to highlight key information, logical points, and visual elements.

[0055] In some embodiments, the terminal device 110 may utilize the intelligent agent 160 to generate multiple rich text cards corresponding to the multiple subtasks, based on the corresponding execution results of the multiple subtasks in at least one subtask. The rich text cards can be constructed based on the execution results of their corresponding subtasks. A rich text card can be a specific implementation of a content object, containing text, icons, hyperlinks, tables, or other at least partially structured elements. By presenting rich text cards, intermediate execution results related to the subtask can be presented in a concise and to-the-point manner.

[0056] Furthermore, the terminal device 110 can present multiple rich text cards in the interface, using multiple subtasks as a dimension. Each rich text card can be used to present the portion of the intermediate execution result corresponding to that subtask. The presentation style of the rich text cards can include card-based layouts, section labels, or lists. In this way, the originally scattered subtask outputs can be integrated and transformed into intuitive and visual presentation units, thereby improving the readability and organization of information during task execution.

[0057] As an example, terminal device 110 can Figure 2C The interface 200C shown presents content objects representing intermediate execution results of the target task. For example, the terminal device 110 can present access points 224-1 and 224-2 for the thinking process interface and the final result interface. Access points 224-1 and 224-2 are used to present intermediate execution results and target execution results, respectively, so that users can switch between viewing the task execution process and the final result.

[0058] Continue to refer to Figure 2C Content object 230 may include one or more specific rich text cards. Each rich text card may correspond to the execution result of a different subtask. For example, card 230-1 may present the interim analysis result of "comparison of major vehicle models" obtained by executing a subtask. The content of card 230-1 may include key information such as recommended vehicle models, reasons for recommendation, and configurations. Card 230-2 may include "basic configuration knowledge" retrieved and organized by executing another subtask. The content of card 230-2 may include descriptions of the vehicle's basic structure, key components, and their functions. In this way, the interim results of different subtasks can be intuitively displayed in a card-based layout, making it easy for users to distinguish and understand the current task execution ideas and progress of the intelligent agent.

[0059] Alternatively or additionally, content objects may include other presentation formats. Content objects can present intermediate execution results as visual charts, interactive lists, collapsible panels, or knowledge graph nodes. For example, for subtasks involving numerical calculations or trend analysis, data distribution can be visually displayed through charts or statistical panels. For subtasks involving conceptual relationships or knowledge topology, the connections between information can be shown through knowledge graphs or relational network diagrams. These formats can be combined with rich text cards or presented independently to adapt to different types of subtasks.

[0060] In some examples, terminal device 110 can present a content list 231 in the interactive interface 200C to display the topics of various contents contained in the current content object. For example, content list 231 may include topic items such as "Bicycle Buying Guide," "New and Used Bicycle Selection Analysis," and "Storage and Anti-theft Solutions." Each item may correspond to a specific rich text card or other form of intermediate execution result presentation unit in the current content object. Users can quickly locate and view one or more rich text cards corresponding to a topic item by selecting it in content list 231.

[0061] In some embodiments, the request indicating the target task is received in a first interface area for a session with the agent. See, for example, [link to relevant documentation]. Figure 2C The first interface area can be the left-hand area of ​​interface 200C, used to receive and present dialogue messages between user 140 and agent 160. Content objects can be presented in the second interface area. The first interface area differs from the second interface area. See also [link to relevant documentation]. Figure 2C The second interface area can be the right-hand area of ​​interface 200C, used to present content objects representing the intermediate execution results of the target task. The first and second interface areas are independent of each other, ensuring both the continuity of the session interaction and the clear presentation of the intermediate execution results.

[0062] In some embodiments, in response to receiving supplementary description information for a task during the execution of a target task according to the execution plan, the terminal device 110 can update the execution plan based on the supplementary description information and the requirement information. (Continue to see...) Figure 2C The first interface area may include an input box 223. During the process of determining the execution plan or executing the target task according to the execution plan, the user 140 can provide supplementary descriptive information to the agent 160 via the input box 223. The supplementary descriptive information may indicate modifications, expansions, or restrictions on the direction of task execution. For example, the supplementary descriptive information may indicate adding new subtasks to the original execution plan, adjusting the analysis perspective of existing subtasks, or deleting subtasks that are no longer needed.

[0063] As an example, via input box 223, user 140 can input supplementary description information such as "Please add analysis for models priced below 1,000 yuan". Terminal device 110 can then transmit this supplementary information to agent 160. Agent 160 can determine new subtasks based on the original user requirements and supplementary description information, and insert the new subtasks into the existing task execution flow.

[0064] Furthermore, agent 160 can continue to execute the target task according to the updated execution plan. For example, in Figure 2CIn the interface 200C shown, the agent can add a new rich text card 230 based on the updated execution plan to display the interim results of "analysis of car models under 1000 yuan". In this way, users can not only see new task results being added to the original card sequence in real time, but also continuously iterate and improve their needs during execution, thereby obtaining target execution results that better meet actual needs.

[0065] In some embodiments, in response to the selection of descriptive information for a subtask in at least one subtask, the terminal device 110 may highlight a portion of the content object generated based on the execution result of the selected subtask. See also Figure 2C User 140 can select the corresponding subtask item from message list 222 to jump to the corresponding rich text card of the subtask displayed in the second interface area. Alternatively, the rich text card can be highlighted. In this way, a quick correspondence and linkage can be achieved between conversation interaction and result display, thereby improving the understandability and ease of operation of the task execution process.

[0066] In some embodiments, in response to receiving feedback on at least a portion of the content object, the terminal device 110 may utilize the intelligent agent 160 to provide a response to the feedback. To further enhance user engagement and flexibility during task execution, the terminal device 110 may support users providing interactive feedback on any part of the content object. In other words, users can not only view intermediate execution results during task execution but also provide feedback based on their understanding of or questions regarding these results.

[0067] As an example, in Figure 2D In the interface 200D shown, user 140 selects a portion of the content object presented on the interface to provide feedback. For example, user 140 can select the theme of the entire content object to provide overall feedback. Alternatively, user 140 can select a specific rich text card 230 or a specific part of a rich text card 230 (such as a recommendation reason or a parameter value) to provide feedback.

[0068] In some examples, in response to at least a portion of a content object being selected, terminal device 110 may present an action list 240 for the selected content in interface 200D. Action list 240 is used to trigger available feedback actions. For example, action list 240 may include query control 241. Query control 241 may be used to provide feedback for at least a portion of the content object. The feedback may indicate a query for at least a portion of the content object. For example, further questions or other action requests regarding the selected content. Alternatively or additionally, action list 240 may include delete control 242. Delete control 242 may be used to remove or mask the selected portion of the content from intermediate execution results and the final target execution result.

[0069] Continue to refer to Figure 2E In response to the triggering of query control 241, terminal device 110 can present input control 251 in interface 200E. Input control 251 can be used to obtain user input, including natural language. Terminal device 110 can provide this user input as feedback to agent 160. For example, in the first interface area for a conversation with agent 160, feedback determined based on user input and the specific content referenced by the feedback are provided to agent 160 together. In interactive interface 200E, the feedback and the specific content referenced by the feedback can be presented via message 242.

[0070] Furthermore, agent 160 can determine the action to take in response to user feedback based on the specific content being referenced, the task context, and user input. Subsequently, agent 160 can execute the corresponding action. For example, it can generate a new answer, clarify or expand existing content, or update the execution plan or intermediate execution results. Agent 160's response can be returned and displayed on the interface in real time, forming an immediate closed loop for user feedback. This approach allows users to dynamically participate during task execution, improving the controllability of the results and the interactive experience.

[0071] In some embodiments, terminal device 110 can utilize agent 160 to update intermediate execution results based on feedback. For example, agent 160 can parse user feedback and the specific content referenced in the feedback, and, in conjunction with the context information of the current task, determine whether the execution direction needs to be adjusted. If it is determined that the intermediate execution results need to be updated, agent 160 can generate new or modified subtask execution results and update the corresponding intermediate execution results. Terminal device 110 can present a content object representing the updated intermediate execution results, thereby enabling the user to view the latest version of the interim results.

[0072] In some embodiments, agent 160 can generate the target execution result of the task based on the updated intermediate execution results. The updated intermediate execution results can serve as new contextual information for agent 160 and participate in subsequent reasoning. Based on the updated intermediate execution results, agent 160 can fully incorporate user supplementary or corrective requests when generating the target execution result of the task. In this way, it can be ensured that the target execution result of the task remains consistent with the latest user requirements.

[0073] In some embodiments, terminal device 110 can utilize an agent to generate a response for at least a portion of the content object based on feedback. For example, when a user only needs to explain, respond to, or supplement a field in a rich text card, agent 160 can determine that there is no need to update the intermediate execution results as a whole, but instead generate an immediate response related to the selected content. The generated response can be for the selected content object as a whole or for a specific part thereof, and is returned to the user in the form of a message or prompt. In this way, the user can quickly obtain the information they need without waiting for the entire intermediate execution results to be refreshed.

[0074] As an example, query control 241 can be understood as a user-facing general query entry point. Users can use this entry point to submit various requests in natural language. Alternatively or additionally, terminal device 110 can provide preset shortcut options for common user feedback needs. For example, in Figure 2F In the interface 200F shown, in response to at least a portion of the content object being selected, the terminal device 110 can display a feedback control 261. The feedback control 261 is used to quickly select a preset feedback type. For example, the feedback control 261 can be used to trigger feedback operations such as general queries, explanations, in-depth analysis, and authenticity checks. Each operation can be bound to preset prompts. In this case, the user can either freely input a feedback request or quickly trigger feedback through shortcut options. This reduces input costs and unifies feedback semantics.

[0075] In some embodiments, feedback can instruct the explanation of at least a portion of the content object. This type of feedback can be used to instruct the explanation or clarification of the meaning or background of the selected content, thereby helping the user understand the interim results during task execution. For example, user 140 can select the "Transmission System" field in card 230-2 in interface 200F and trigger a corresponding feedback operation through feedback control 261. Terminal device 110 can determine the preset feedback prompt corresponding to this operation, such as, "Please explain what this means before performing the next action." Further, terminal device 110 can provide the feedback prompt and the selected specific content together to agent 160. Agent 160 can generate a targeted explanation based on the task context and provide the explanation result to the user in the form of a message in the first interface area.

[0076] In some embodiments, feedback can indicate detailed results for determining at least a portion of the content object. Such feedback can be used to instruct for expanding or refining the details of the selected content, or to allow for a more in-depth discussion of a particular analysis or viewpoint. For example, user 140 can select the "Model A - Most Recommended" field in card 230-1 in interface 200F and trigger a corresponding feedback operation via feedback control 261. Terminal device 110 can provide the corresponding preset prompt and selected content to agent 160. Agent 160 can generate more detailed comparison information or expanded data based on this feedback. This newly generated information can be determined by agent 160 to be presented in the form of messages or to update the corresponding intermediate execution results and content objects.

[0077] In some embodiments, feedback can instruct the verification of the authenticity of at least a portion of the content object. This type of feedback can be used to instruct the verification of the data source or authenticity of the selected content, preventing erroneous or outdated content from affecting task results. For example, user 140 can select a section of text involving statistical data in a card on interface 200F, triggering a corresponding feedback operation. Terminal device 110 can provide the agent 160 with a preset prompt word corresponding to the feedback operation and the selected content. Agent 160 can then perform a verification process, such as confirming the reliability of the content by searching for references, comparing data from different sources, or performing logical consistency verification. Furthermore, agent 160 can generate a response or update intermediate execution results based on the verification results.

[0078] In the embodiments of this disclosure, by supporting multiple feedback-based response modes, a highly flexible interactive experience can be provided while ensuring execution efficiency. On the one hand, when the task direction or interim results need to be adjusted, the intelligent agent 160 can promptly update intermediate execution results and optimize the final output accordingly. On the other hand, for cases where no overall adjustment is required, users can directly obtain quick responses to selected content, reducing waiting time and improving interaction efficiency. This approach enhances user participation and controllability during task execution, thereby improving the overall user experience of the intelligent system.

[0079] In some embodiments, feedback for at least a portion of the content object can be provided to the agent 160 through manipulation of graphical interface elements within the content object. In other words, the user 140 can adjust the content object and corresponding intermediate execution results via a graphical user interface (GUI) to further influence the task execution process and the final result. For example, the user 140 can reorganize the order of rich text cards corresponding to different subtasks through drag-and-drop operations. Alternatively, the user 140 can directly edit, modify, or delete certain fields or paragraphs on the cards. The user 140 can also add a completely new card or merge parts of existing card content through interface controls. Such direct GUI-based interactive operations can be provided to the agent 160 as feedback information. The agent 160 can then update intermediate execution results accordingly or apply the modifications as contextual information to the generation of the target execution result.

[0080] By employing the above methods, the system can simultaneously support both natural language and graphical interaction modes, thus providing more flexible and intuitive task execution control capabilities. This approach reduces the operational costs for users modifying or supplementing requirements, improves the consistency between intermediate task outputs and the final execution result, and enhances user satisfaction.

[0081] In summary, the embodiments disclosed herein significantly improve the transparency and flexibility of agent interaction by presenting intermediate execution results during task execution and supporting user feedback based on these results. On one hand, by visually presenting the intermediate execution results of subtasks in structured formats such as rich text cards, users can intuitively understand the agent's task execution logic and progress, thus promptly identifying deviations or shortcomings before execution is complete. On the other hand, users are allowed to provide feedback during task execution using natural language or graphical operations. Feedback can trigger immediate responses or drive dynamic updates to the execution plan or intermediate results. Through this two-way interactive mechanism, users no longer need to input all their requirements at once, but can continuously supplement, clarify, or correct them during execution. This reduces interaction costs and improves the consistency between the results and the user's actual needs.

[0082] Figure 3 A flowchart of a process 300 for task interaction according to some embodiments of the present disclosure is shown. Process 300 may be implemented or included in terminal device 110. Reference will be made below. Figure 1 Describe the process 300.

[0083] In block 310, in response to receiving a request indicating a target task, terminal device 110 uses an agent to determine an execution plan for the target task. The execution plan indicates at least one subtask.

[0084] In box 320, during the execution of a target task according to the execution plan, terminal device 110 presents a content object representing intermediate execution results of the target task. The content object has a visually structured style that is at least partially structured, and the intermediate execution results include at least a portion of the corresponding execution results of at least one subtask.

[0085] In box 330, terminal device 110, in response to receiving feedback on at least a portion of the content object, utilizes an intelligent agent to provide a response to the feedback.

[0086] In some embodiments, the content object is determined by: executing at least one subtask based on an execution plan to obtain the corresponding execution result of at least one subtask; and generating the content object based on the corresponding execution result of one or more subtasks in the at least one subtask in response to the fact that the corresponding execution result of one or more subtasks in the at least one subtask satisfies a predetermined presentation condition.

[0087] In some embodiments, presenting a content object representing intermediate execution results includes: generating multiple rich text cards corresponding to multiple subtasks based on the corresponding execution results of multiple subtasks in at least one subtask, wherein the rich text cards in the multiple rich text cards are based on the execution results of the corresponding subtasks; and presenting multiple rich text cards with multiple subtasks as a dimension.

[0088] In some embodiments, providing a response to feedback includes: using an agent to update intermediate execution results based on the feedback; and generating a target execution result for the task based on the updated intermediate execution results.

[0089] In some embodiments, providing a response to feedback includes: using an agent to generate a reply for at least a portion of the content object based on the feedback.

[0090] In some embodiments, the feedback indicates at least one of the following: for a query on at least a portion of the content object, to interpret at least a portion of the content object, to determine detailed results for at least a portion of the content object, or to verify the authenticity of at least a portion of the content object.

[0091] In some embodiments, feedback is determined by at least one of the following: user input in natural language, or actions performed on graphical interface elements in a content object.

[0092] In some embodiments, the request includes requirement information indicating the task, and process 300 further includes: in response to receiving supplemental description information for the task during the execution of the target task according to the execution plan, updating the execution plan based on the supplemental description information and the requirement information; and continuing to execute the target task according to the updated execution plan.

[0093] In some embodiments, process 300 further includes: presenting descriptive information of at least one subtask; and, in response to the selection of descriptive information of a subtask in the at least one subtask, highlighting a portion of the content object generated based on the execution result of the selected subtask.

[0094] In some embodiments, the request is received in a first interface area for a session with the agent, the content object is presented in a second interface area, and the first interface area is different from the second interface area.

[0095] Embodiments of this disclosure also provide corresponding apparatus for implementing the above methods or processes. Figure 4 A schematic structural block diagram of an example device 400 for task interaction according to certain embodiments of the present disclosure is shown. Device 400 may be implemented as or included in terminal device 110. Various modules / components in device 400 may be implemented by hardware, software, firmware, or any combination thereof.

[0096] like Figure 4 As shown, the apparatus 400 includes: a plan determination module 410 configured to determine an execution plan for the target task using an agent in response to receiving a request indicating a target task, the execution plan indicating at least one subtask; a presentation module 420 configured to present a content object representing intermediate execution results of the target task during execution of the target task according to the execution plan, the content object having a visually structured style at least partially, the intermediate execution results including at least a portion of the corresponding execution results of at least one subtask; and a response module 430 configured to provide a response to feedback using an agent in response to receiving feedback on at least a portion of the content object.

[0097] In some embodiments, the content object is determined by: executing at least one subtask based on an execution plan to obtain the corresponding execution result of at least one subtask; and generating the content object based on the corresponding execution result of one or more subtasks in the at least one subtask in response to the fact that the corresponding execution result of one or more subtasks in the at least one subtask satisfies a predetermined presentation condition.

[0098] In some embodiments, the presentation module 420 is further configured to: generate multiple rich text cards corresponding to the multiple subtasks based on the corresponding execution results of the multiple subtasks in at least one subtask, wherein the rich text cards in the multiple rich text cards are based on the execution results of the corresponding subtasks; and present the multiple rich text cards with the multiple subtasks as a dimension.

[0099] In some embodiments, the response module 430 is further configured to: utilize an agent to update intermediate execution results based on feedback; and generate a target execution result for the task based on the updated intermediate execution results.

[0100] In some embodiments, the response module 430 is further configured to: utilize an agent to generate a response for at least a portion of the content object based on feedback.

[0101] In some embodiments, the feedback indicates at least one of the following: for a query on at least a portion of the content object, to interpret at least a portion of the content object, to determine detailed results for at least a portion of the content object, or to verify the authenticity of at least a portion of the content object.

[0102] In some embodiments, feedback is determined by at least one of the following: user input in natural language, or actions performed on graphical interface elements in a content object.

[0103] In some embodiments, the request includes requirement information indicating the task, and the apparatus 400 further includes a plan update module configured to: update the execution plan based on the supplementary description information and requirement information in response to receiving supplementary description information for the task during the execution of the target task according to the execution plan; and continue to execute the target task according to the updated execution plan.

[0104] In some embodiments, the presentation module 420 is further configured to: present description information of at least one subtask; and, in response to the selection of description information of a subtask in the at least one subtask, highlight a portion of the content object generated based on the execution result of the selected subtask.

[0105] In some embodiments, the request is received in a first interface area for a session with the agent, the content object is presented in a second interface area, and the first interface area is different from the second interface area.

[0106] The units and / or modules included in device 400 can be implemented in various ways, including software, hardware, firmware, or any combination thereof. In some embodiments, one or more units and / or modules can be implemented using software and / or firmware, such as machine-executable instructions stored on a storage medium. In addition to or as an alternative to machine-executable instructions, some or all of the units and / or modules in device 400 can be implemented at least partially by one or more hardware logic components. By way of example and not limitation, exemplary types of hardware logic components that can be used include field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-chips (SoCs), complex programmable logic devices (CPLDs), and so on.

[0107] It should be understood that one or more steps in the above methods can be performed by suitable electronic devices or combinations of electronic devices. Such electronic devices or combinations of electronic devices may include, for example, […]. Figure 1 Terminal device 110 in the middle.

[0108] Figure 5 A block diagram of an electronic device 500 in which one or more embodiments of the present disclosure may be implemented is shown. It should be understood that... Figure 5 The electronic device 500 shown is merely exemplary and should not be construed as limiting the functionality and scope of the embodiments described herein. Figure 5 The illustrated electronic device 500 may include or be implemented as Figure 1 The terminal device 110, or Figure 4 Device 400.

[0109] like Figure 5 As shown, electronic device 500 is in the form of a general-purpose electronic device. Components of electronic device 500 may include, but are not limited to, one or more processors or processing units 510, memory 520, storage devices 530, one or more communication units 540, one or more input devices 550, and one or more output devices 550. Processing unit 510 may be a physical or virtual processor and is capable of performing various processes according to programs stored in memory 520. In a multiprocessor system, multiple processing units execute computer-executable instructions in parallel to improve the parallel processing capability of electronic device 500.

[0110] Electronic device 500 typically includes multiple computer storage media. Such media can be any accessible media that is accessible to electronic device 500, including but not limited to volatile and non-volatile media, removable and non-removable media. Memory 520 can be volatile memory (e.g., registers, cache, random access memory (RAM)), non-volatile memory (e.g., read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory), or some combination thereof. Storage device 530 can be removable or non-removable media and can include machine-readable media, such as flash drives, disks, or any other media that can be used to store information and / or data and can be accessed within electronic device 500.

[0111] Electronic device 500 may further include additional removable / non-removable, volatile / non-volatile storage media. Although not explicitly stated... Figure 5 As shown, disk drives for reading from or writing to removable, non-volatile disks (e.g., "floppy disks") and optical disk drives for reading from or writing to removable, non-volatile optical disks can be provided. In these cases, each drive can be connected to a bus (not shown) via one or more data media interfaces. Memory 520 may include computer program product 525 having one or more program modules configured to perform various methods or actions of various embodiments of this disclosure.

[0112] Communication unit 540 enables communication with other electronic devices via a communication medium. Additionally, the functionality of components of electronic device 500 can be implemented using a single computing cluster or multiple computing machines capable of communicating via communication connections. Therefore, electronic device 500 can operate in a networked environment using logical connections to one or more other servers, network personal computers (PCs), or another network node.

[0113] Input device 550 can be one or more input devices, such as a mouse, keyboard, trackball, etc. Output device 550 can be one or more output devices, such as a monitor, speaker, printer, etc. Electronic device 500 can also communicate with one or more external devices (not shown) via communication unit 540 as needed. External devices include storage devices, display devices, etc., and can communicate with one or more devices that enable user interaction with electronic device 500, or with any device that enables electronic device 500 to communicate with one or more other electronic devices (e.g., network card, modem, etc.). Such communication can be performed via input / output (I / O) interface (not shown).

[0114] According to an exemplary implementation of this disclosure, a computer-readable storage medium is provided that stores computer-executable instructions thereon, wherein the computer-executable instructions are executed by a processor to implement the methods described above. According to an exemplary implementation of this disclosure, a computer program product is also provided, which is tangibly stored on a non-transitory computer-readable medium and includes computer-executable instructions, which are executed by a processor to implement the methods described above.

[0115] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatuses, devices, and computer program products implemented according to this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0116] These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processing unit of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0117] Computer-readable program instructions can be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions that execute on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0118] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction, which contains one or more executable instructions for implementing the specified logical function. In some, as newer, implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0119] Various implementations of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed implementations. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described implementations. The terminology used herein is chosen to best explain the principles, practical applications, or improvements to technology in the market, or to enable others skilled in the art to understand the various implementations disclosed herein.

Claims

1. A task interaction method based on an intelligent agent, comprising: In response to receiving a request indicating a target task, the agent determines an execution plan for the target task, the execution plan indicating at least one subtask; During the execution of the target task according to the execution plan, a content object representing intermediate execution results of the target task is presented, the content object having at least a partially structured visual style, the intermediate execution results including at least a portion of the corresponding execution results of the at least one subtask; as well as In response to receiving feedback for at least a portion of the content object, the agent provides a response to the feedback.

2. The method according to claim 1, wherein the content object is determined in the following manner: Based on the execution plan, execute the at least one subtask to obtain the corresponding execution results of the at least one subtask; and In response to the fact that the execution results of one or more of the at least one subtask meet the predetermined presentation conditions, the content object is generated based on the execution results of the one or more subtasks.

3. The method according to claim 1, wherein presenting the content object representing the intermediate execution result includes: Based on the corresponding execution results of multiple subtasks in the at least one subtask, multiple rich text cards are generated, each corresponding to one of the multiple subtasks. The rich text cards in the multiple rich text cards are based on the execution results of the corresponding subtasks. as well as The multiple rich text cards are presented using the multiple sub-tasks as dimensions.

4. The method of claim 1, wherein providing a response to the feedback comprises: The intermediate execution result is updated based on the feedback using the intelligent agent. as well as Based on the updated intermediate execution results, the target execution result of the task is generated.

5. The method of claim 1, wherein providing a response to the feedback comprises: Using the intelligent agent, a response is generated for at least a portion of the content object based on the feedback.

6. The method of claim 1, wherein the feedback indicates at least one of the following: A query targeting at least a portion of the content object. At least a portion of the content object is interpreted. Detailed results determining at least a portion of the content object, or Verify the authenticity of at least a portion of the content object.

7. The method of claim 1, wherein the feedback is determined by at least one of the following: Including user input in natural language, or Operations on the graphical interface elements in the content object.

8. The method of claim 1, wherein the request includes requirement information indicative of the task, and the method further comprises: In response to receiving supplemental description information for the task during the execution of the target task according to the execution plan, the execution plan is updated based on the supplemental description information and the requirement information; and Continue executing the target task according to the updated execution plan.

9. The method according to claim 1, further comprising: Present descriptive information about the at least one subtask; as well as In response to the selection of descriptive information for a subtask in the at least one subtask, the portion of the content object generated based on the execution result of the selected subtask is highlighted.

10. The method of claim 1, wherein the request is received in a first interface area for a session with the agent, the content object is presented in a second interface area, and the first interface area is different from the second interface area.

11. A task interaction device based on an intelligent agent, comprising: A plan determination module is configured to, in response to receiving a request indicating a target task, use the agent to determine an execution plan for the target task, the execution plan indicating at least one subtask; A presentation module is configured to present a content object representing intermediate execution results of the target task during execution of the target task according to the execution plan. The content object has a visually structured style at least partially, and the intermediate execution results include at least a portion of the corresponding execution results of the at least one subtask. as well as A response module is configured to provide a response to the feedback received in response to at least a portion of the content object, using the agent.

12. An electronic device, comprising: At least one processor; as well as At least one memory coupled to the at least one processor and storing instructions for execution by the at least one processor, the instructions causing the electronic device to perform the method according to any one of claims 1 to 10 when executed by the at least one processor.

13. A computer-readable storage medium having stored thereon computer-executable instructions that can be executed by a processor to implement the method according to any one of claims 1 to 10.

14. A computer program product comprising computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, implement the method according to any one of claims 1 to 10.

Images