Interactive card generation method and electronic device

Abstract

The application discloses an interactive card generation method and an electronic device; wherein the method comprises: obtaining a user intention; wherein the user intention is used at least for reasoning a user task; determining a target card structure according to the user intention, and determining target content corresponding to at least one placeholder module in the target card structure; the target card structure is used for realizing a corresponding artificial intelligence interaction function; generating an artificial intelligence interaction card corresponding to the user task according to the target content and the target card structure.

Description

Technical Field

[0001] This application relates to, but is not limited to, the field of computer technology, and in particular to an interactive card generation method and an electronic device. Background Technology

[0002] In information-rich scenarios, users need to spend a significant amount of time reading information and extracting valuable insights from the raw data. For example, in a meeting, users need to spend considerable time piecing together key points from fragmented information scattered across multiple documents and organizing tasks that require follow-up processing; in everyday life, users need to peruse numerous old photos and objects to find clues to past memories; and so on.

[0003] Therefore, how to help users quickly extract useful information from a large amount of data and reduce their workload has become an urgent technical problem to be solved. Summary of the Invention

[0004] In view of this, this application provides at least one method for generating interactive cards and an electronic device.

[0005] The technical solution of this application is implemented as follows: On the one hand, this application provides an interactive card generation method, the method comprising: Obtain the user intent; wherein the user intent is used at least to infer the user task; Based on user intent, determine the target card structure and the target content corresponding to at least one placeholder module in the target card structure; the target card structure is used to implement the corresponding artificial intelligence interactive function. Based on the target content and target card structure, generate AI interactive cards corresponding to the user task.

[0006] Furthermore, this application also provides an electronic device, including: a memory and at least one processor; The memory stores a computer program; at least one processor is used to execute the computer program to implement the steps in any of the above methods.

[0007] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and are not intended to limit the technical solutions of this application. Attached Figure Description

[0008] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the technical solutions of this application.

[0009] Figure 1A schematic diagram illustrating the implementation process of an interactive card generation method provided in this application; Figure 2 A schematic diagram of an AI interactive card in one embodiment provided in this application; Figure 3 A schematic diagram of an AI interactive card in one embodiment provided in this application; Figure 4 A schematic diagram of an AI interactive card in one embodiment provided in this application; Figure 5 A schematic diagram of an AI interactive card in one embodiment provided in this application; Figure 6 A schematic diagram of an AI interactive card in one embodiment provided in this application; Figure 7 A schematic diagram of an AI interactive card in one embodiment provided in this application; Figure 8 A schematic diagram illustrating the implementation process of one embodiment provided in this application; Figure 9 This is a schematic diagram of the hardware entity of an electronic device provided in this application. Detailed Implementation

[0010] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application are further described in detail below with reference to the accompanying drawings and embodiments. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0011] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0012] The terms “first / second / third” are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that “first / second / third” may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0013] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used herein is for descriptive purposes only and is not intended to limit the scope of this application.

[0014] This application provides an interactive card generation method, which can be executed by an electronic device. The electronic device can be various types of terminals such as laptops, tablets, desktop computers, set-top boxes, and mobile devices (e.g., mobile phones, portable music players, personal digital assistants, dedicated messaging devices, portable gaming devices), or it can be implemented as a server. The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0015] The technical solutions in the embodiments of this application will now be clearly and completely described with reference to the accompanying drawings.

[0016] Figure 1 A schematic diagram illustrating the implementation process of an interactive card generation method provided in this application is shown below. Figure 1 As shown, the method includes the following steps S11 to S13: Step S11: Obtain user intent; wherein, the user intent is at least used to infer user task. Step S12: Based on the user intent, determine the target card structure and the target content corresponding to at least one placeholder module in the target card structure; the target card structure is used to implement the corresponding artificial intelligence interactive function. Step S13: Generate an AI interactive card corresponding to the user task based on the target content and the target card structure.

[0017] Here, user intent refers to the goal or need that a user expects to achieve. Based on this user intent, the corresponding user task can be obtained through model reasoning, identification, decomposition, or mapping.

[0018] In some implementations, user intent can be obtained in any way.

[0019] In some implementations, user intent can be determined and obtained through information actively input by the user. For example, explicit user intent can be determined and obtained through semantic understanding of the text or speech input by the user.

[0020] In some implementations, user behavior can be proactively monitored to predict potential future needs, which can then be interpreted as user intent. For example, if a user searches for attractions in City A, it can be predicted that the user may want to travel to City A during a recent holiday, thus generating a travel guide for City A as an implicit user intent.

[0021] In some implementations, user intent can be determined by actively monitoring user-related data. For example, by monitoring to-do list information in a calendar app and determining that a user will attend a friend's birthday party in three days, the purchase of a birthday gift can be considered an implicit user intent.

[0022] A card, or Artificial Intelligence (AI) interactive card, refers to a user interface that carries AI services or interactive functions in the form of a card.

[0023] Card structure refers to the card layout corresponding to AI interactive cards. A card structure may include at least one placeholder module.

[0024] Placeholder modules refer to blank or temporary areas reserved in a card structure, which can be used to represent visual elements that will be loaded later. In some implementations, placeholder modules can be categorized in various ways. In some implementations, they can be classified according to their function in the card structure, such as placeholder modules for main titles, placeholder modules for subtitles, and / or placeholder modules for detailed information. In some implementations, they can be categorized according to the content type they correspond to, such as placeholder modules for text content, placeholder modules for image content, placeholder modules for audio content, placeholder modules for video content, placeholder modules for multimodal data, and / or placeholder modules for interactive components.

[0025] Different card structures include different types of placeholder modules, thereby enabling different AI interaction functions. For example, card structures with placeholder modules corresponding to different content types can achieve different AI interaction functions; for instance, a card structure with placeholder modules corresponding to image content can achieve image interaction functions, and a card structure with placeholder modules corresponding to audio content can achieve audio interaction functions. As another example, a card structure with placeholder modules corresponding to operation components can provide users with operation functions; while a card structure without placeholder modules corresponding to operation components can display brief information to the user.

[0026] In this way, the target card structure that can fulfill the user's intent can be determined based on the user's intent. For example, when the user's intent is to generate a meeting summary, the card structure containing placeholder modules for the main title, subtitle, and detailed information can be determined as the target card structure; when the user's intent is to order lunch, the card structure containing placeholder modules for the operation components can be determined as the target card structure; and so on.

[0027] In some implementations, multiple card structures can be pre-stored. This allows the target card structure to be determined from the stored card structures after the user's intent is obtained.

[0028] In some implementations, the card structure generation logic can be pre-stored. This way, after obtaining the user's intent, the target card structure can be generated based on the user intent and the card structure generation logic. For example, the card structure generation logic may include selecting a target placeholder module from multiple placeholder modules that can fulfill the user's intent, and then generating the target card structure based on the selected target placeholder module.

[0029] After determining the target card structure, based on the role of each placeholder module within the card structure and the content type of the corresponding target content, target content is generated for each placeholder module according to the user's intent. For example, for a placeholder module corresponding to the main title in the target card structure, the core theme corresponding to the user's intent can be determined, and this core theme can be used as the target content for that placeholder module. As another example, for a placeholder module corresponding to an operation component in the target card structure, the application or tool that can fulfill the user's intent can be identified. Based on the application information of that application or the tool information of that tool, an operation component can be created, and this operation component can then be used as the target content for that placeholder module.

[0030] In this way, after generating the target content, firstly, the target content is bound or associated with the corresponding placeholder module; then, using the browser or rendering engine, layout calculations and pixel rendering are performed to generate the AI ​​interactive card corresponding to the user task. This AI interactive card corresponding to the user task is a dynamically generated intelligent interactive card based on the user's intent.

[0031] The interactive card generation method provided in this application first obtains the user intent, wherein the user intent is at least used to infer the user task; then, based on the user intent, a target card structure is determined, and the target content corresponding to at least one placeholder module in the target card structure is determined, wherein the target card structure is used to implement the corresponding artificial intelligence interactive function; finally, based on the target content and the target card structure, an artificial intelligence interactive card corresponding to the user task is generated. The method provided in this application dynamically generates AI interactive cards based on user intent, which can help users quickly extract useful information from a large amount of data, reduce the user's operational burden, and improve the user's AI interactive experience.

[0032] In some implementations, obtaining user intent, i.e., step S11 above, can be implemented as the following steps S111 to S112: Step S111: Obtain intent information and target knowledge; wherein, the intent information represents input data related to the user task, or trigger data related to the user task in the current system; the target knowledge represents knowledge related to the intent information determined from the user's historical knowledge.

[0033] Here, intent information refers to the information used to initiate user tasks.

[0034] Intent information can be input data related to the user's task. For example, intent information can be text data, image data, voice data, video data, or multimodal data entered by the user through a designated information input component.

[0035] Intent information can be trigger data related to the user task in the current system; that is, data actively monitored from the current system that can trigger the user task. Here, "current system" refers to any system running on the user device used to initiate the user task, such as the operating system or data storage system.

[0036] In some implementations, it can be determined whether there are any to-do items that meet the triggering conditions by monitoring the user's to-do items and the user's current status (e.g., current time information, geographical location, interactive objects, etc.); if there are to-do items that meet the triggering conditions, the to-do items can be used as trigger data.

[0037] In some implementations, the presence of a task to be performed can be determined by combining the user's current state (e.g., user actions, current time information, geographical location, interacting objects, etc.) with the user profile. For example, if a user habitually eats dinner between 6 and 7 pm every day, and the current time is 6 pm, it can be determined that dinner needs to be booked, and the current time and booking dinner for the user can then be used as trigger data.

[0038] User historical knowledge refers to any historical information related to the user.

[0039] In some implementations, historical knowledge can be original historical data related to the user, such as original documents, original images, original videos stored by the user, or emails, chat logs, etc. obtained from the application.

[0040] In some implementations, historical knowledge can be structured knowledge obtained by extracting information from historical data related to the user. For example, summarizing and knowledge structure-summarizing the original document can yield the historical knowledge corresponding to that original document. Another example is extracting features from original images, videos, or audio to determine the subjects, relationships, visual styles, compositions, emotions, etc., thereby obtaining structured historical knowledge based on the extracted information.

[0041] In some implementations, users' historical knowledge can be stored using document databases, relational databases, or graph databases, based on the data format of historical knowledge.

[0042] In this way, after obtaining the intent information, the user's historical knowledge can be retrieved based on the intent information, thereby determining the target knowledge related to the intent information.

[0043] In some implementations, at least one keyword can be extracted from the intent information, thereby retrieving the user's historical knowledge based on the at least one keyword to obtain the target knowledge.

[0044] In some implementations, at least one entity and / or at least one relationship can be extracted from the intent information, thereby retrieving the user's historical knowledge based on the at least one entity and / or at least one relationship to obtain the target knowledge.

[0045] In some implementations, semantic extraction can be performed on the intent information to obtain high-dimensional semantic information of the intent information, and then the user's historical knowledge can be retrieved based on the high-dimensional semantic information to obtain the target knowledge.

[0046] Step S112: Determine the user intent based on the intent information and target knowledge.

[0047] Here, after determining the intent information and target knowledge, intent recognition is performed to obtain the user intent. In some implementations, intent recognition can be performed using a large language model based on the intent information and target knowledge to obtain the user intent.

[0048] In some implementations, personalized data of the user can be retrieved based on intent information to obtain personalized data related to the intent information. Then, the user's intent can be determined based on the intent information, target knowledge, and the retrieved personalized data. Personalized data may include user preference data, personal profile data, etc.

[0049] In the embodiments provided in this application, user intent is determined by combining intent information and target knowledge related to the intent information. This makes the process of recognizing user intent not only based on a superficial understanding of intent information, but also includes a deep understanding of relevant background knowledge or contextual data, thereby improving the accuracy of intent recognition. In addition, when intent information contains semantically unclear content, using target knowledge to supplement the unclear content in intent information can reduce the burden on users to clarify intent information, improve interaction efficiency and convenience, and thus improve the user's interactive experience.

[0050] In some embodiments, the target card structure includes one of the following: a first card structure, a second card structure, and a third card structure; wherein, The first card structure is used to guide the user to perform an operation and includes at least one operation component; The second card structure is used to present structured information; The third card structure is used to present enhanced information corresponding to the user's historical knowledge.

[0051] The interactive card generation method provided in this application includes at least a first card structure, a second card structure, and a third card structure among the preset multiple card structures.

[0052] The first card structure is used to guide users to perform operations; that is, the first card structure contains explicit action guidance content. Therefore, the first card structure includes at least one operation component to guide users to perform the corresponding operation. It can be understood that the first card structure including at least one operation component before generating the corresponding AI interactive card means that the first card structure includes at least one placeholder module corresponding to the operation component.

[0053] In some implementations, the first card structure also includes a placeholder module for displaying brief information; wherein, the brief information may be text information, image information, or video information generated after performing a user task.

[0054] The second card structure is used to present structured information. In some implementations, the structured information can be obtained by organizing the information to be processed in a structured manner in any suitable way.

[0055] In some implementations, structured information can be information obtained by organizing the information to be processed in a structured manner according to time sequence. For example, structured information can be information such as a user's schedule or educational background generated in chronological order.

[0056] In some implementations, structured information can be information obtained by organizing the information to be processed in a structured manner according to geographical location. For example, structured information can be information generated according to geographical location, such as a user's travel records or movement routes.

[0057] In some implementations, structured information can be information obtained by organizing the information to be processed in a structured manner according to the logic of event development. For example, structured information can be event-related information generated according to the cause, process, climax, and result of an event.

[0058] The third card structure is used to present enhanced information corresponding to the user's historical knowledge.

[0059] Enhanced information refers to information obtained by enhancing a user's historical knowledge. Enhancement processing, in particular, refers to the process of generating new and diverse data based on a user's historical knowledge and according to specified goals. For example, enhancement processing could be extracting keyframes of interest and significance from historical videos to generate highlight videos of those videos. Another example is extracting key information from a specified document to generate a document summary.

[0060] Since augmented information can be of any data type, the output format of the third card structure is more flexible, supporting multimodal combinations such as images, text, video, and audio, and placing greater emphasis on context, emotion, and narrative expression.

[0061] In the embodiments provided in this application, the pending information or historical knowledge of the user's task can be transformed into an actionable, understandable structure or content that can be remembered for a long time, based on the user's intent and by utilizing three types of card structures. This significantly reduces the user's cognitive burden and enhances the lasting value of the pending information in actual use.

[0062] In some implementations, determining the target card structure based on the user intent in step S12 above can be implemented as the following steps S121 to S123: Step S121: If the user intent includes an operation intent, the first card structure is determined as the target card structure.

[0063] Here, "operational intent" refers to the intention to operate on a specified tool. The tool can be an application, a mathematical calculation engine, an information retrieval and acquisition tool, an application programming interface (API), etc.

[0064] In some implementations, intent information can be used to determine whether a user intends to operate a specific tool. For example, if a user enters the information "book a flight to City A", it can be directly determined from this input information that the user wants to operate the ticket booking software to perform the booking action.

[0065] In some implementations, during the execution of a user task, if the tool data corresponding to the user task includes operable data, the user intent is updated, and it is determined that the updated user intent includes an operable intent. Wherein: The tool data corresponding to a user task refers to the tool data that needs to be queried during the execution of the user task. For example, in the above-mentioned ticket booking task, it is necessary to query the user's calendar application to obtain the user's schedule information and to query the ticket booking application to obtain flight information, etc. Actionable data refers to data that has a time limit. Therefore, users need to decide whether to perform the corresponding action within the data's validity period.

[0066] Thus, when the user intent determined based on user intent information and target knowledge does not include operational intent, but the tool data includes actionable data, the user intent is updated to include operational intent in order to maximize user benefit. For example, if a user inputs "Please check flight tickets to City A," this input does not include the user's intent to open the booking software and perform the booking action immediately. However, if the searched booking information includes a discount valid for the day, the user intent is updated to "If there is a discount, ask if you want to book immediately." In this way, based on the updated user intent, it can be determined that the user has an operational intent.

[0067] When it is determined that the user's intent includes an operation intent, the first card structure is determined as the target card structure, so as to guide the user to perform the corresponding operation through the first card structure and open the tool for the user's expected operation using at least one operation component in the first card structure.

[0068] For example, when a user inputs "order lunch," and based on historical user data it's determined that the user frequently orders pizzas from the vendor "Mom's Kitchen Pizza," the user intent is identified as an ordering intent, which includes an operational intent: "currently perform the ordering operation." Thus, based on this operational intent, the first card structure is determined as the target card structure, and a corresponding AI interaction card is generated, such as... Figure 2 As shown. In Figure 2 The AI ​​interactive card includes an image 21 indicating that the order type is pizza, text 22 displaying brief information about the merchant, and an operation component 23. Clicking on the operation component 23 will take you to the ordering application.

[0069] Step S122: If the user intent represents an intent to enhance the user's historical knowledge, the third card structure is determined as the target card structure.

[0070] Here, when the user expects to enhance the specified historical knowledge based on the user's intent, the third card structure is determined as the target card structure so as to display the enhanced information corresponding to the specified historical knowledge using the third card structure.

[0071] In some implementations, user intents can include any type of enhanced processing intent. For example, user intents can include summarizing specified historical knowledge, extracting key inspirations, refining sentiment curves, summarizing new experiences, extracting highlight videos, generating to-do lists, etc.

[0072] For example, in one embodiment of this application, when the user inputs "generate a highlight video of 2025 based on life records from 2025", firstly, based on the user input, the user's historical knowledge is retrieved to obtain text records, image records, voice records, and / or video records related to "life records from 2025"; then, based on the user input and the retrieved relevant records, it is determined that the user's intent is to generate a highlight video based on the user's life records from 2025 (i.e., historical knowledge) (i.e., to perform enhancement processing); subsequently, since the user's intent is to enhance historical knowledge, the third card structure is determined as the target card structure; finally, based on the target card structure and the user's intent, an AI interactive card is generated, such as... Figure 3 As shown.

[0073] exist Figure 3In the device 30, an AI interactive card 31 and a prompt message 35 generated for the user are displayed. The AI ​​interactive card 31 includes video content 32, a main title 33, and a subtitle 34. Video content 32 is a highlight video generated based on retrieved user life records from 2025; the main title 33 summarizes the core theme of the highlight video, namely, "Time Capsule 2025"; the subtitle 34 summarizes the important events in the highlight video, namely, "Graduation Trip, Summer Trip to Italy, New Pet Arrival". The prompt message 35 displays the text "AI-Generated Content" to inform the user that the highlight video in the AI ​​interactive card 31 is an AI-generated video and not a continuous sequence of events.

[0074] In some implementations, when the third card structure is determined as the target card structure, corresponding AI interactive cards can be generated in the form of AI comics based on the third card structure.

[0075] For example, in one embodiment provided in this application, when the user inputs the intent information "write yesterday's AI diary", firstly, based on the intent information, data from the previous day related to the user is retrieved to obtain target knowledge; then, based on the intent information and target knowledge, it is determined that the user's intent is to enhance yesterday's data related to the user, therefore the third card structure is determined as the target card structure; subsequently, for at least one piece of target knowledge, a corresponding AI comic is generated; finally, based on the target card structure, at least one generated AI comic, the user intent, and the target knowledge, an AI interactive card is generated, such as... Figure 4 As shown.

[0076] exist Figure 4 In the AI ​​interactive card 40, the following information is included: a main title 41, which indicates the core information of the current AI interactive card, namely, "Comic Diary", and is accompanied by an image 411 for identifying the comic diary type; current time information 42; an image 43 of a comic effect generated using an image generation model or a multimodal model; and a concise image description 44 in the image 43, namely "Reading stories with your baby"; information 45, which records the number of steps taken the previous day; and information 46, which records the dietary information of the previous day, including the time the user drank coffee the previous day.

[0077] In some implementations, when generating AI comics, the user's personalized data can be retrieved to determine the user's preferred comic style, and then the generated AI comic can be limited according to the user's preferred style. For example, the user may prefer a simple style, a fresh style, or a romantic style.

[0078] Step S123: If the user intent does not include the operation intent and the intent to enhance the user's historical knowledge, the second card structure is determined as the target card structure; the second card structure is used to present structured content, steps, or status information to the user.

[0079] Here, when the user intent includes operational intent but does not include intent to enhance the user's historical knowledge, the second card structure is determined as the target card result to display the processing result for the user task using structured information.

[0080] It is evident that in scenarios where the user's goal is clear but the execution path is not yet clear, structured information can be extracted from the user's input information or the user's historical knowledge, and presented in the form of steps, processes, or key points using a second card structure.

[0081] For example, when users are unfamiliar with how to use a specific tool, a structured learning path or operation steps can be provided using a second card structure to enable users to quickly learn how to use the tool, thereby reducing the difficulty of getting started.

[0082] For example, when a user views their to-do list for the day, a second card structure can be used to display the to-do list in chronological order.

[0083] In some implementations, when determining the target card structure based on user intent, firstly, it is determined whether the user intent includes an intent to enhance the user's historical knowledge; if it includes an intent to enhance the knowledge, then the third card structure is determined as the target card structure; if it does not include an intent to enhance the knowledge, then it is determined whether the user intent includes an operational intent; if it includes an operational intent, then the first card structure is determined as the target card structure; if it does not include an operational intent, then the second card structure is determined as the target card structure.

[0084] In some implementations, when determining the target card structure based on user intent, firstly, it is determined whether the user intent includes an operational intent; if it includes an operational intent, the first card structure is determined as the target card structure; if it does not include an operational intent, it is determined whether the user intent includes an intent to enhance the user's historical knowledge; if it includes an intent to enhance the user's historical knowledge, the third card structure is determined as the target card structure; if it does not include an intent to enhance the user's historical knowledge, the second card structure is determined as the target card structure. For example, in one embodiment provided by this application, when the intent information is a daily plan generation task set by the user (i.e., a scheduled task), firstly, based on the scheduled task, the user's to-do list information for the day and other to-do lists nearing their expiration date are obtained from the user's calendar application, while simultaneously retrieving the user's historical knowledge to determine the user's lifestyle habits; then, based on the scheduled task and the to-do list information, the user intent is determined to be generating a plan for the day; subsequently, since the user intent does not include an operational intent and is not to enhance historical knowledge, the second card structure is determined as the target card structure; finally, based on the target card structure and the obtained to-do list information, an AI interactive card is generated, such as... Figure 5 As shown.

[0085] exist Figure 5 In the AI ​​interactive card 50, there are: a main title 51, which indicates the core theme of the current card, namely "Today's Schedule"; to-do reminders 52, which are reminders generated based on the user's upcoming to-do items and lifestyle habits, such as reminding the user to buy a birthday gift and supplement with vitamin E; a schedule generated in chronological order 53, which displays the items that need to be handled at three time points of the day (i.e., 7:56, 10:00, and 12:30); and daily statistics 54, which displays the user's water intake statistics and step count statistics obtained based on the analysis of the user's data for the day.

[0086] In the embodiments provided in this application, the target card structure is dynamically determined according to the user's intent. The target card structure is highly matched with the current user intent, so that the AI ​​interactive card generated based on the target card structure can more effectively display the processing results of the user's task. Furthermore, since the AI ​​interactive card is an independent card generated for the user task corresponding to the current user intent, it can be considered that the current AI interactive card only carries the core processing results of the corresponding user task, without including functions unrelated to the user task. This allows the user to quickly obtain task processing information for the user task from the AI ​​interactive card, thereby significantly reducing the user's cognitive burden and improving interaction efficiency and experience.

[0087] Furthermore, the embodiments provided in this application, by defining the logic for determining the target card structure, can eliminate the randomness of the generated AI interactive card structure, thereby providing users with a stable interactive experience. At the same time, the complex and open "design thinking" process is transformed into a rule-based card structure determination logic. In this way, when facing specific user intentions, it is not necessary to perform a complete "evaluation" and "generation" of all possible card structures every time. Instead, it is only necessary to determine the target card structure according to the predefined card structure determination logic, thereby reducing computational overhead and response latency.

[0088] In some implementations, where the target card structure is the first card structure and the operation intention represents an operation intention for a target tool, the AI ​​interactive card includes an operation component for indicating the entry point of the target tool.

[0089] Here, when it is determined that the user intent includes an operation intent, and that the operation intent is an operation intent for the target tool, the first card structure is used as the target card structure, and the target card structure contains a placeholder module corresponding to the operation component; then, the placeholder module corresponding to the operation component is bound to the target content used to control the jump to the target tool, so that the generated AI interaction card includes an operation component used to indicate the entry point of the target tool.

[0090] For example, when the target tool corresponding to the operation intent is a payment application, the redirection information used to control the jump to that payment application will be displayed in the form of an operation component in the AI ​​interaction card. Figure 6 As shown, the AI ​​interactive card 60 displays order information 61 and an operation component 62. Order information 61 displays the electronic bill amount; the operation component 62 displays the text "One-click payment" and controls the redirection to the corresponding payment application. Thus, in response to the user triggering the operation component 62, the user can be redirected to the corresponding payment application to continue the payment process.

[0091] In some implementations, the display effect of the operation component used to indicate the entry point of the target tool can be determined based on the tool type of the target tool. For example, the above-mentioned combination Figure 6 In the described embodiments, when the target tool is a payment application, the text "One-click payment" can be displayed in the operation component indicating the entry point of the payment application; when the target tool is a ticketing application, the text "Book Tickets" can be displayed in the operation component indicating the entry point of the ticketing application; when the target tool is a fitness application, the text "Start Fitness" can be displayed in the operation component indicating the entry point of the fitness application; and so on.

[0092] In the embodiments provided in this application, on the one hand, by displaying an operation component in the AI ​​interactive card to indicate the entry point of the target tool, the user can quickly jump to the corresponding target tool through the operation component to perform the corresponding operation; on the other hand, since the operation component is only used to trigger the jump to the target tool and not to indicate the internal operation process of the target tool, the amount of information that the card interaction method provided in this application needs to process can be reduced, lightweight operation is achieved, and the card generation efficiency can be improved.

[0093] As mentioned above, the enhanced information corresponding to the user's historical knowledge can all be presented using a third card structure. Therefore, the specific layout of the third card structure is quite flexible. In some implementations, the third card structure supports multimodal combinations such as text, images, audio, and video. Therefore, in specific implementations, the specific layout of the third card structure can be constrained based on the task data corresponding to the user's task.

[0094] Thus, in some embodiments, after determining the third card structure as the target card structure, that is, after step S122 above, the method further includes the following steps S124 to S125: Step S124: Perform feature extraction on multiple fragmented data to obtain multiple feature data; the multiple fragmented data represent the task data of the user task determined based on the intent information and the target knowledge.

[0095] Here, fragmented data refers to the task data corresponding to the user task, that is, the data of the processing object of the user task.

[0096] In some implementations, the fragmented data can be data contained within the intent information. For example, the fragmented data can be text to be processed, images, audio, or multimodal data within the intent information.

[0097] In some implementations, fragmented data can be target knowledge determined based on intent information. That is, the fragmented data can be task data retrieved from the user's historical knowledge based on intent information. Here, historical knowledge refers to knowledge obtained after structuring information related to the user. For example, when the user's intent is to enhance stored historical knowledge related to a specified product project to extract project experience, the fragmented data is the target knowledge retrieved from the user's historical knowledge.

[0098] In some implementations, fragmented data can be task data retrieved from a specified data storage location based on intent information and target knowledge. Here, firstly, the user intent is determined based on the intent information and target knowledge; then, the specified data storage location is retrieved according to the user intent to obtain the fragmented data. The specified data storage location can be a local storage location (e.g., internal hard drive, memory, external storage device, etc.), or a cloud storage location (e.g., a public cloud, private cloud, hybrid cloud, etc. deployed in the cloud); it can be the storage location corresponding to specified tool data, or the storage location corresponding to the user device's operation logs; and so on.

[0099] Thus, in some implementations, a large model can be used to perform feature extraction on multiple fragmented data sets to obtain multiple feature data sets. For example, a large language model can be used to perform feature extraction on fragmented text data; a multimodal model can be used to perform feature extraction on image, audio, video, or other multimodal data sets to obtain corresponding feature data sets.

[0100] In some implementations, the feature data can be the semantics of the corresponding fragment data, or it can be data obtained after enhancing the semantics of the fragment data by summarizing, simplifying or rewriting it.

[0101] Step S125: Determine the layout of the multiple feature data in the target card structure based on the relationship attributes between the multiple feature data, so as to optimize the target card structure.

[0102] The relationship attributes between multiple feature data refer to the nature or type of the relationship between them. In some implementations, the relationship attributes between multiple feature data may include subordinate relationships, process relationships, comparison relationships, temporal relationships, etc.

[0103] In some implementations, relation extraction models or preset relation judgment rules can be used to determine the relationship data between multiple feature data.

[0104] In this way, after determining the relationship attributes between multiple feature data, the layout of the multiple feature data in the target card structure is determined based on these relationship attributes.

[0105] For example, when there is a process relationship between multiple feature data, it can be determined that action-oriented strategies such as timelines, flowcharts, or event development axes can be used to lay out these multiple feature data in the target card structure.

[0106] For example, when there is a comparative relationship between multiple feature data, it can be determined that comparative layout strategies such as side-by-side arrangement, grid alignment, and icon juxtaposition should be used to arrange these multiple feature data in the target card structure.

[0107] For example, when multiple feature data are related in a temporal sequence, it can be determined whether to use a horizontal time axis, a vertical time axis, or a top-down progressive axis to arrange these multiple feature data in the target card structure.

[0108] In this way, after determining the layout of multiple feature data in the target card structure, the target card layout is optimized using the determined layout, that is, the specific layout of the target card structure is made to better fit the task data corresponding to the current user task.

[0109] In one embodiment provided in this application, such as Figure 7 As shown, when a process relationship is established between multiple fragmented data, feature data extracted from these fragmented data are linked together using the event development axis 70. Figure 7 In the diagram, data 71 represents the data obtained after feature extraction of data related to the start of the project, and the corresponding icon 711 is used to mark the starting point of the project; data 72 represents the key event data obtained after analyzing and summarizing the events that occur during the project, i.e., detecting the server from alarm information, and the corresponding icon 721 is used to mark the key event; data 73 represents the data obtained after feature extraction of data related to the end of the project, and the corresponding icon 731 is used to mark the success of the project.

[0110] In the embodiments provided in this application, by determining the relationship attributes between the feature data corresponding to multiple fragment data, the layout of the multiple feature data in the target card structure can be dynamically optimized. This can improve the adaptability of the optimized target card structure layout to the task data corresponding to the user task, improve the readability of the generated AI interactive card, and thus improve the intelligence of the interactive card generation method.

[0111] In some embodiments, the method further includes steps S126 to S127: Step S126: Determine target personalized data from the user's personalized data based on the user's intent.

[0112] Here, personalized user data can refer to user preference data or personal profile data. For example, this personalized data may include a user's image style preferences, text style preferences, overall card style preferences, reading habits, etc.

[0113] In this way, personalized data of users can be retrieved based on user intent in order to identify target personalized data that is relevant to that user intent.

[0114] Step S127: Optimize the target card structure based on the target personalized data.

[0115] Here, the target card structure is adjusted based on the target's personalized data to better meet the user's personalized needs.

[0116] For example, placeholder modules for image-type content or text-type content in the target card structure can be optimized based on the user's image style or text style preferences. This allows the image style preference or text style to be incorporated into the constraints of the corresponding placeholder modules, so that the generated images or text better meet the user's personalized requirements.

[0117] For example, the positions of each placeholder module within the target card structure can be updated based on the user's reading habits. For instance, when a user prefers reading images, the placeholder module corresponding to the image content can be moved to the center of the target card so that the user can quickly grasp the image information. Or, when a user prefers horizontal reading, the target card structure can be changed from a vertical to a horizontal structure, arranging the placeholder modules in the target card structure in a left-to-right order.

[0118] In the embodiments provided in this application, by optimizing the target card structure, the optimized target card structure can better meet the user's personalized requirements, thereby making the information layout, information priority, and interaction process in the generated AI interactive card highly matched with the user's cognitive habits, thereby reducing the user's cognitive burden, improving interaction efficiency and the user's interaction experience.

[0119] In some implementations, each of the placeholder modules has corresponding constraints.

[0120] Here, constraints refer to the conditions that must be met when generating the target content corresponding to the placeholder module.

[0121] In some implementations, constraints can be content-type constraints. For example, constraints may include the ability of corresponding placeholder modules to be bound to text data, image data, audio data, video data, or multimodal data.

[0122] In some implementations, the constraints can be constraints specific to the content generation style. For example, constraints on text style, image style, video style, etc.

[0123] In some implementations, constraints may include rendering condition limitations. For example, resolution, sampling parameters, lighting conditions, rendering precision, or color space.

[0124] In some implementations, constraints may include the deployment position of the corresponding placeholder module in the target card structure.

[0125] In some implementations, the constraints corresponding to the placeholder module can be determined in any way.

[0126] In some implementations, the constraints corresponding to the placeholder module can be determined based on the function of the placeholder module in the target card structure. For example, when the placeholder module corresponds to the main title in the target card structure, the constraints corresponding to the placeholder module include at least a word count constraint.

[0127] In some implementations, the constraints corresponding to the placeholder modules can be determined based on the user's personalized data (e.g., the target personalized information determined in step S126 above). For example, if the target personalized data indicates that the user's preferred font is SimSun, then the constraints corresponding to the text type placeholder modules in the target card structure will at least include restricting the font to SimSun.

[0128] Thus, determining the target content corresponding to at least one placeholder module in the target card structure in step S12 above can be implemented as the following step S128: Step S128: For each placeholder module, according to the constraints corresponding to the placeholder module and the user intent, call the target module to generate the target content corresponding to the placeholder module.

[0129] The target module can be any module capable of generating target content. In some implementations, the target module can be a model or an application module.

[0130] In some implementations, a target module capable of generating target content can be determined based on the constraints corresponding to the placeholder module. For example, when the constraints corresponding to the placeholder module include constraints on the content type, the target module can be determined based on that content type. For instance, if the content type defined by the constraints is text, a large language model capable of generating text content can be used as the target module; if the content type defined by the constraints is image, a multimodal model or image generation model capable of generating images can be used as the target module.

[0131] After identifying the target module, the user intent and the constraints corresponding to the placeholder modules are sent to the target module to generate the corresponding target content. For example, if the target module is a large language model, prompts can be generated based on constraints other than the user intent and the content type of the placeholder modules; these prompts are then sent to the large language model to guide it in generating the target content.

[0132] Here, the user intent sent to the target module includes at least the user intent related to the target content corresponding to the placeholder generation module. For example, the user intent sent to the target module may include intent information, semantic understanding information of the intent information, target knowledge required to generate the target content, and / or the user's personalized data, etc.

[0133] Thus, the step of generating the AI ​​interaction card corresponding to the user task based on the target content and the target card structure, i.e., step S13 above, can be implemented as the following step S131: Step S131: Update the corresponding placeholder module using the target content to obtain the AI ​​interactive card.

[0134] Here, firstly, after determining the target content, the target content is bound or associated with the corresponding placeholder module to update the placeholder module; then, the browser or rendering engine can be used to perform layout calculations and pixel rendering on the target card structure and its corresponding updated placeholder module, thereby generating the AI ​​interactive card corresponding to the user task.

[0135] In some implementations, the step of calling the target module based on the constraints corresponding to the placeholder module and the user intent, so as to generate the target content corresponding to the placeholder module using the target module, that is, the above step S128 can be implemented as the following steps S1281 to S1282: Step S1281: If the constraint includes a tool invocation condition for a specified tool, the component creation tool is invoked according to the tool invocation condition and the user intent to create the operation component of the tool corresponding to the placeholder module.

[0136] Here, when the constraints corresponding to the placeholder module include tool invocation conditions for a specified tool, the target content corresponding to the placeholder module is used to establish a jump association or trigger association with the specified tool, or to control the jump to the specified tool. Thus, in response to the user's trigger operation on the target content, the specified tool can be launched or the user can jump to the specified tool. In some implementations, launching or jumping to the specified tool may refer to launching or jumping to the main page of the specified tool; or it may refer to launching or jumping to a specified subpage of the specified tool, where the specified subpage is used to implement the function corresponding to the user task.

[0137] Thus, when the constraints include tool invocation conditions for a specific tool, a component creation tool capable of implementing the component creation function can be determined based on these conditions. Then, the component creation tool is invoked, and an operational component associated with the specified tool is created according to the tool invocation conditions and user intent. Therefore, in this embodiment, the target module is implemented as a component creation tool, and the target content is implemented as an operational component corresponding to the specified tool.

[0138] In some implementations, the component creation tool can be a visual user interface designer, an interface layout editor, a user interface component library, etc. Thus, by calling the aforementioned component creation tool, the corresponding operational components for the placeholder modules can be generated.

[0139] In some implementations, the component creation tool can be a generative large model. Based on the tool's invocation conditions and user intent, the generative large model can automatically generate the corresponding operational components (i.e., target content) for the placeholder modules.

[0140] Step S1282: If the constraints include content generation conditions, then according to the content generation conditions and the user intent, the generation model is invoked to generate the target content corresponding to the placeholder module.

[0141] When the constraints corresponding to a placeholder module include content generation conditions, the target content of that placeholder module is data content such as text, images, audio, or multimodal data. In some implementations, the content generation conditions may be constraints on content type, content generation style, rendering conditions, etc.

[0142] Thus, when generating target content, firstly, based on the content type constraints in the content generation conditions, a generation model capable of generating that content type is determined; then, based on the content generation conditions and user intent, the generation model is invoked to generate the corresponding target content. Therefore, in this embodiment, the target module is implemented as a generation model.

[0143] For example, when the content type is specified as text in the content generation conditions, the large language model can be determined as the corresponding generation model.

[0144] For example, when the content type is specified as image in the content generation conditions, the multimodal model can be determined as the corresponding generative model.

[0145] In the embodiments provided in this application, generating or creating corresponding target content based on the constraints corresponding to the placeholder module can improve the target content generation or creation speed, as well as the matching degree with the current user task, thereby improving the intelligence of the generated AI interactive card.

[0146] Below, in conjunction with Figure 8 An embodiment provided in this application will be described below. Figure 8 As shown, this embodiment includes the following steps S81 to S88: Step S81: Obtain user intent information; then, proceed to step S82. Here, user intent information can be text or audio information entered by the user through the information input box.

[0147] Step S82: Determine the semantics of the user intent information; then, proceed to step S83. Here, large language models or multimodal models can be used to perform semantic recognition of intent information.

[0148] Step S83: Based on the semantics of the user intent information, retrieve the user's knowledge graph and personalized data to obtain target knowledge and target personalized data; then, execute step S84. Here, knowledge graph can refer to the structured knowledge of users stored using graph databases.

[0149] In some implementations, a management unit can be used to perform retrieval on knowledge graphs and personalized data; wherein, the management unit can be implemented as a graph-based retrieval-augmented generation (GraphRAG) unit or an agent with graph retrieval capabilities.

[0150] Step S84: Generate user intent based on user intent information, semantics of intent information, target knowledge, and target personalized data; then, execute step S85. Here, the management unit can be used to call a model with semantic understanding capabilities, such as a large language model, to comprehensively identify and generate user intent information, semantics of intent information, target knowledge, and target personalized data, thereby obtaining user intent.

[0151] Step S85: Determine the card structure corresponding to the user intent based on the user intent and the card structure generation logic; then, execute step S86. In some implementations, the card structure generation logic can be stored in any database. For example, the card structure generation logic can be stored in a database based on Retrieval-Augmented Generation (RAG) technology, or other vector databases or text databases.

[0152] In this way, the management unit can be used to search the RAG-based database or other databases, and then determine the card structure corresponding to the user's intent, as well as at least one placeholder module in the card structure and the constraints corresponding to each placeholder module, based on the card generation logic therein.

[0153] In some implementations, the management unit represents the determined card structure with specified structured data to obtain the corresponding code skeleton.

[0154] Step S86: Based on the determined card structure, call the component creation tool and / or generate the model to create or generate the module content corresponding to at least one placeholder module in the card structure; then, execute step S87. Step S87: Establish the association between the placeholder modules and the corresponding module content in the card structure; then, execute step S88. Here, the management unit can be used to establish this relationship.

[0155] Step S88: Based on the card structure and the established relationships, use the rendering engine to perform rendering to obtain AI interactive cards.

[0156] Based on the foregoing embodiments, this application also provides an electronic device. For example... Figure 9 As shown, the electronic device includes a memory 910 and at least one processor 920; wherein, The memory stores a computer program; the at least one processor is used to execute the computer program to achieve: Obtain user intent; wherein, the user intent is at least used to infer user task; Based on the user intent, a target card structure is determined, and the target content corresponding to at least one placeholder module in the target card structure is determined; the target card structure is used to implement the corresponding artificial intelligence interactive function. Based on the target content and the target card structure, generate an AI interactive card corresponding to the user task.

[0157] In some embodiments, the at least one processor 920 is configured to: Acquire intent information and target knowledge; Wherein, the intent information represents input data related to the user task, or trigger data related to the user task in the current system; the target knowledge represents knowledge related to the intent information determined from the user's historical knowledge; The user intent is determined based on the intent information and target knowledge.

[0158] In some embodiments, the target card structure includes one of the following: a first card structure, a second card structure, and a third card structure; wherein, The first card structure is used to guide the user to perform an operation and includes at least one operation component; The second card structure is used to present structured information; The third card structure is used to present enhanced information corresponding to the user's historical knowledge.

[0159] In some embodiments, the at least one processor 920 is configured to: If the user intent includes an operational intent, the first card structure is determined as the target card structure. If the user intent represents an intent to enhance the user's historical knowledge, the third card structure is determined as the target card structure. If the user intent does not include the operation intent and the intent to enhance the user's historical knowledge, the second card structure is determined as the target card structure; the second card structure is used to present structured content, steps, or status information to the user.

[0160] In some implementations, where the target card structure is the first card structure and the operation intention represents an operation intention for a target tool, the AI ​​interactive card includes an operation component for indicating the entry point of the target tool.

[0161] In some embodiments, the at least one processor 920 is configured to: Feature extraction is performed on multiple fragmented data to obtain multiple feature data; the multiple fragmented data represent the task data of the user task determined based on the intent information and the target knowledge; Based on the relationship attributes between the multiple feature data, the layout of the multiple feature data in the target card structure is determined to optimize the target card structure.

[0162] In some embodiments, the at least one processor 920 is further configured to: Based on the user's intent, determine the target personalized data from the user's personalized data; Optimize the target card structure based on the target personalized data.

[0163] In some implementations, each of the placeholder modules has corresponding constraints. Determining the target content corresponding to at least one placeholder module in the target card structure includes: For each placeholder module, based on the constraints corresponding to the placeholder module and the user intent, the target module is invoked to generate the target content corresponding to the placeholder module. The corresponding placeholder module is updated using the target content to obtain the AI ​​interactive card.

[0164] In some embodiments, the at least one processor 920 is configured to: If the constraints include tool invocation conditions for a specified tool, the component creation tool is invoked according to the tool invocation conditions and the user intent, so as to create the operation component of the tool corresponding to the placeholder module using the component creation tool; When the constraints include content generation conditions, the generation model is invoked based on the content generation conditions and the user intent to generate the target content corresponding to the placeholder module.

[0165] It should be noted that the descriptions of the above device embodiments are similar to those of the above method embodiments, and have similar beneficial effects. In some embodiments, the functions or components / units included in the device provided in this application can be used to perform the methods described in the above method embodiments. For technical details not disclosed in the device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0166] If the technical solution of this application involves personal information, the product using this technical solution has clearly informed the user of the personal information processing rules and obtained the user's voluntary consent before processing the personal information. If the technical solution of this application involves sensitive personal information, the product using this technical solution has obtained the user's separate consent before processing the sensitive personal information, and also meets the requirement of "express consent". For example, at personal information collection devices such as cameras, clear and prominent signs are set up to inform users that they have entered the scope of personal information collection and that personal information will be collected. If an individual voluntarily enters the collection scope, it is deemed that they have agreed to the collection of their personal information; or on the personal information processing device, with clear signs / information informing users of the personal information processing rules, authorization is obtained from the user through pop-up information or by asking the user to upload their personal information; wherein, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the types of personal information processed.

[0167] It should be noted that, in the embodiments of this application, if the above-described interactive card generation method is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to the related technology, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, mobile hard drives, read-only memory (ROM), magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware, software, or firmware, or any combination of hardware, software, and firmware.

[0168] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above steps / processes do not imply a sequential order of execution; the execution order of each step / process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above embodiments of this application are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0169] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0170] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0171] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0172] In addition, each functional unit in the various embodiments of this application can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0173] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory (ROM), magnetic disks, or optical disks.

[0174] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, or the part that contributes to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, magnetic disks, or optical disks.

[0175] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for generating interactive cards, comprising: Obtain user intent; wherein, the user intent is at least used to infer user task; Based on the user intent, a target card structure is determined, and the target content corresponding to at least one placeholder module in the target card structure is determined; the target card structure is used to implement the corresponding artificial intelligence interactive function. Based on the target content and the target card structure, generate an AI interactive card corresponding to the user task.

2. The method according to claim 1, wherein obtaining user intent includes: Acquire intent information and target knowledge; Wherein, the intent information represents input data related to the user task, or trigger data related to the user task in the current system; the target knowledge represents knowledge related to the intent information determined from the user's historical knowledge; The user intent is determined based on the intent information and target knowledge.

3. The method according to claim 2, wherein the target card structure includes one of the following: a first card structure, a second card structure, and a third card structure; wherein, The first card structure is used to guide the user to perform an operation and includes at least one operation component; The second card structure is used to present structured information; The third card structure is used to present enhanced information corresponding to the user's historical knowledge.

4. The method according to claim 3, wherein determining the target card structure based on the user intent includes: If the user intent includes an operational intent, the first card structure is determined as the target card structure. If the user intent represents an intent to enhance the user's historical knowledge, the third card structure is determined as the target card structure. If the user intent does not include the operation intent and the intent to enhance the user's historical knowledge, the second card structure is determined as the target card structure; the second card structure is used to present structured content, steps, or status information to the user.

5. The method according to claim 4, wherein when the target card structure is the first card structure and the operation intention characterizes an operation intention for a target tool, the artificial intelligence interactive card includes an operation component for indicating the entry point of the target tool.

6. The method according to claim 4, further comprising, after determining the third card structure as the target card structure: Feature extraction is performed on multiple fragmented data points to obtain multiple feature data. The multiple fragmented data represent the task data of the user task determined based on the intent information and the target knowledge; Based on the relationship attributes between the multiple feature data, the layout of the multiple feature data in the target card structure is determined to optimize the target card structure.

7. The method according to any one of claims 2 to 6, further comprising: Based on the user's intent, determine the target personalized data from the user's personalized data; Optimize the target card structure based on the target personalized data.

8. The method according to claim 1, wherein each of the placeholder modules has corresponding constraints; Determining the target content corresponding to at least one placeholder module in the target card structure includes: For each placeholder module, based on the constraints corresponding to the placeholder module and the user intent, the target module is invoked to generate the target content corresponding to the placeholder module. The step of generating an AI interactive card corresponding to the user task based on the target content and the target card structure includes: The corresponding placeholder module is updated using the target content to obtain the AI ​​interactive card.

9. The method according to claim 8, wherein the step of invoking the target module based on the constraints corresponding to the placeholder module and the user intent, so as to generate the target content corresponding to the placeholder module using the target module, comprises: If the constraints include tool invocation conditions for a specified tool, the component creation tool is invoked according to the tool invocation conditions and the user intent, so as to create the operation component of the tool corresponding to the placeholder module using the component creation tool; When the constraints include content generation conditions, the generation model is invoked based on the content generation conditions and the user intent to generate the target content corresponding to the placeholder module.

10. An electronic device, comprising: Memory and at least one processor; The memory stores a computer program; the at least one processor is used to execute the computer program to implement the steps of the method according to any one of claims 1-9.

Images