Search method, device, electronic device and storage medium based on large model

Abstract

The present disclosure provides a large-model-based search method, apparatus, electronic device, and storage medium, relating to the fields of artificial intelligence technology, particularly large-model, computer vision, and deep learning. The large-model-based search method includes: determining, in response to a received search request, a search intent indicated by the search request; performing a content search based on the search intent using the large-model to obtain a plurality of output data, at least one of the plurality of output data being described in a specified language; and rendering the plurality of output data to obtain a first search result, wherein the rendering effect of the output data described in the specified language is different from the rendering effect of the remaining output data in the plurality of output data.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, in particular to technical fields such as large models, computer vision and deep learning, and specifically to search methods, devices, electronic devices, intelligent agents and storage media based on large models. Background Art

[0002] With the advancement of computer and network technologies, neural network models have been widely applied in various scenarios. For example, search engine products can use large language models (LLMs) to output search results that meet users' diverse search needs. However, the output data of large models currently has a single format, making it difficult to clearly and accurately present search results to users. Summary of the Invention

[0003] The present disclosure provides a large model-based search method, device, intelligent agent, electronic device and storage medium.

[0004] According to one aspect of the present disclosure, a search method based on a big model is provided, including: determining the search intent indicated by the search request in response to a received search request; performing a content search based on the search intent using the big model to obtain a plurality of output data, at least one of the plurality of output data being described in a specified language; and rendering the plurality of output data to obtain a first search result, wherein a rendering effect of the output data described in the specified language is different from a rendering effect of the remaining output data in the plurality of output data.

[0005] According to another aspect of the present disclosure, a data display method is provided, comprising: sending a search request in response to receiving a data query request; and displaying a first search result generated in response to the search request in response to obtaining the first search result; wherein the first search result is generated using the search method provided by an embodiment of the present disclosure.

[0006] According to another aspect of the present disclosure, a search device based on a big model is provided, including: a first determination module, for determining the search intent indicated by a search request in response to a received search request; a search module, for performing content search for the search intent using the big model to obtain multiple output data, at least one of the multiple output data being described in a specified language; and a rendering module, for rendering the multiple output data to obtain a first search result, wherein the rendering effect of the output data described in the specified language is different from the rendering effect of the remaining output data in the multiple output data.

[0007] According to another aspect of the present disclosure, a data display device is provided, including: a sending module for sending a search request in response to receiving a data query request; and a display module for displaying a first search result in response to obtaining a first search result generated in response to the search request; wherein the first search result is generated using the search device provided by an embodiment of the present disclosure.

[0008] According to another aspect of the present disclosure, an artificial intelligence agent is provided, which is configured to execute the method provided by the embodiment of the present disclosure.

[0009] According to another aspect of the present disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method described above.

[0010] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to execute the method described above.

[0011] According to another aspect of the present disclosure, a computer program product is provided, comprising a computer program, wherein the computer program implements the method described above when executed by a processor.

[0012] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The accompanying drawings are provided to facilitate a better understanding of the present invention and do not constitute a limitation of the present disclosure.

[0014] Figure 1 The following schematically illustrates an application scenario in which the large model-based search method and device according to an embodiment of the present disclosure can be applied;

[0015] Figure 2 The flowchart of the large model-based search method according to an embodiment of the present disclosure is schematically shown;

[0016] Figure 3 A schematic diagram schematically illustrates a first search result according to an embodiment of the present disclosure;

[0017] Figure 4 The following schematically illustrates a principle diagram of a large model-based search method according to an embodiment of the present disclosure;

[0018] Figure 5A The schematic diagram schematically shows the principle of updating the first search result according to an embodiment of the present disclosure;

[0019] Figure 5B Schematically illustrates a schematic diagram of the principle of updating the first search result according to another embodiment of the present disclosure;

[0020] Figure 5C Schematically illustrates a schematic diagram of the principle of updating the first search result according to another embodiment of the present disclosure;

[0021] Figure 6 The following schematically shows a flow chart of a data display method according to an embodiment of the present disclosure;

[0022] Figure 7 The following schematically shows a structural block diagram of a large model-based search device according to an embodiment of the present disclosure;

[0023] Figure 8 The following schematically shows a structural block diagram of a data display device according to an embodiment of the present disclosure;

[0024] Figure 9 A block diagram schematically illustrates the structure of an artificial intelligence agent according to an embodiment of the present disclosure; and

[0025] Figure 10 A block diagram of an electronic device according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0026] The following description of exemplary embodiments of the present disclosure is made in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding. These details should be considered as merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0027] Figure 1 The following schematically illustrates an application scenario in which the large model-based search method and device according to an embodiment of the present disclosure can be applied.

[0028] It should be noted that Figure 1 The examples shown are merely examples of system architectures to which the embodiments of the present disclosure may be applied, to help those skilled in the art understand the technical content of the present disclosure, but do not imply that the embodiments of the present disclosure may not be applied to other devices, systems, environments, or scenarios. For example, in another embodiment, an exemplary system architecture to which the large model-based search method and apparatus may be applied may include a terminal device, but the terminal device may implement the large model-based search method and apparatus provided by the embodiments of the present disclosure without interacting with a server.

[0029] like Figure 1 As shown, the application scenario 100 according to this embodiment may include a terminal device 101 and a server 102 .

[0030] Terminal device 101 can be connected to server 102 via a network. A user can send a data query request 103 through an interactive interface provided by terminal device 101. Terminal device 101 can convert data query request 103 into a search request 104 and send search request 104 to server 102, causing server 102 to call the large model and output a first search result 105. The server sends first search result 105 to terminal device 101, causing terminal device 101 to display first search result 105 to the user.

[0031] Various communication client applications may be installed on the terminal device 101, such as intelligent assistant applications, knowledge reading applications, web browser applications, search applications, instant messaging tools, email clients, and / or social platform software (for example only). A user may enter a data query request 103 in the interactive interface of these client applications, and these client applications will display the generated first search results 105 to the user.

[0032] In one embodiment, the server 102 may use a large language model to generate multimodal data and icon data, and call a rendering tool to generate a first search result 105 in a multimodal form, so as to display the first search result 105 in a multimodal form on the terminal device 101.

[0033] The terminal device 101 can be configured as various electronic devices with a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, etc.

[0034] Server 102 can be a server that provides various services, such as a backend management server (for example only) that supports the content browsed by users through the interactive interface of terminal device 101. The backend management server can invoke a large model to search data in response to received user instructions, and then feed the search results back to terminal device 101 and display them through the interactive interface. Server 102 can also be a cloud server, also known as a cloud computing server or cloud host. This is a host product within the cloud computing service system that addresses the management difficulties and poor scalability of traditional physical hosts and VPS services ("Virtual Private Servers" or "VPS"). Server 1302 can also be a server for a distributed system or a server integrated with blockchain.

[0035] It should be noted that the large model-based search method provided in the embodiment of the present disclosure can generally be executed by the server 102. Accordingly, the large model-based search device provided in the embodiment of the present disclosure can also be set in the server 102. The large model-based search method provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 102 and can communicate with the terminal device 101 and / or the server 102. Accordingly, the large model-based search device provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 102 and can communicate with the terminal device 101 and / or the server 102.

[0036] Alternatively, the large model-based search method provided in the embodiment of the present disclosure may also be generally executed by the terminal device 101. Accordingly, the large model-based search apparatus provided in the embodiment of the present disclosure may generally be provided in the terminal device 101.

[0037] It should be understood that Figure 1 The number of terminal devices and servers in the embodiment is merely illustrative. Any number of terminal devices and servers may be used as required.

[0038] In the technical solution disclosed herein, the collection, storage, use, processing, transmission, provision, disclosure and application of user personal information involved comply with the provisions of relevant laws and regulations, take necessary confidentiality measures, and do not violate public order and good morals.

[0039] In the technical solution disclosed herein, the user's authorization or consent is obtained before obtaining or collecting the user's personal information.

[0040] In light of this, the disclosed embodiments provide a search method based on a large model. By utilizing the large model to output search results in a multimodal format tailored to user search requirements, this method enriches the presentation of search results and enhances their expressiveness. Furthermore, multimodal presentation can clearly and concisely display complex search results, optimizing the visualization and readability of search results and improving the user experience.

[0041] The following will be combined Figures 2 to 4 、 Figure 5A 、 Figure 5B 、 Figure 5C The large model-based search method provided by the present disclosure is described in detail.

[0042] Figure 2 The flowchart of the large model-based search method according to an embodiment of the present disclosure is schematically shown.

[0043] like Figure 2 As shown, the method 200 includes operations S210 to S220.

[0044] In operation S210 , in response to a received search request, a search intent indicated by the search request is determined.

[0045] According to embodiments of the present disclosure, search intent can represent how a user expresses their search purpose and search results. For example, a search purpose can be a search object, such as the weather in a certain location, the business hours of a certain park, or the life story of a certain celebrity. The expression of search results can be the way the search results are presented to the user, such as in at least one of text, audio, images, video, and chart formats.

[0046] For example, the search intent may be obtained by parsing the search request using an intent recognition model, where the intent recognition model identifies keywords in the search request to determine the search intent.

[0047] For example, a search request can be in a format such as text, audio, image, video, or chart. The intent recognition model parses and identifies search requests in these formats, obtains the keywords included in the search request, and outputs the search intent. The search intent can also be expressed in at least one of these formats.

[0048] In operation S220 , a content search is performed based on the search intent using the large model to obtain a plurality of output data.

[0049] According to embodiments of the present disclosure, a large model may be an artificial intelligence (AI) large model, which is a machine learning model with extremely large parameters and complex computational structures. AI large models can process massive amounts of data and complete various complex tasks, such as natural language processing, image recognition, and computer vision. AI large models may include language large models, vision large models, and multimodal large models.

[0050] According to an embodiment of the present disclosure, at least one of the multiple output data is described in a specified language. The multiple output data may be described in multiple languages, and the output data described in different languages ​​can be presented to the user in different expressions on the front-end terminal device.

[0051] According to an embodiment of the present disclosure, the specified language can be a language used to describe the visual effects presented by the search results, and the specified language can be a code language used to describe non-text data. For example, the specified language can be a code language used to describe audio, pictures, videos, and charts.

[0052] For example, the designated language could be a code language used to describe charts. Since search results are presented to users in a single format, users cannot clearly and quickly access the useful information in the search results. Therefore, the large model needs to output data described in a graphical language so that front-end devices can display the search results to users in an image format, improving the readability of the search results.

[0053] In operation S230 , the plurality of output data are rendered to obtain a first search result.

[0054] According to an embodiment of the present disclosure, the rendering effect of output data described in a specified language is different from the rendering effect of the remaining output data in the plurality of output data. For example, the specified language describes chart data, and the remaining output data is described in a language that describes text data. The chart data described in the specified language is rendered as a chart, and the remaining output data is rendered as text.

[0055] For example, a display object rendered from output data described in a specified language can be editable. A display object is the content displayed on a terminal device, and can include images, audio, charts, and videos. For example, the style of a rendered chart can be editable, allowing for modification of its color, size, and other aspects.

[0056] For example, the specified language can be Mermaid. Output data described in Mermaid can be rendered to present the first search result to the user in a flowchart, mind map, timeline diagram, user journey map, block diagram, C4 model diagram, XY chart, etc. This allows the output data of large models to support the display requirements of different scenarios.

[0057] According to an embodiment of the present disclosure, the big model can perform content search and content generation based on the search intent, and directly output multiple output data described in multiple languages, so that the terminal device can display the first search result to the user in a centralized expression.

[0058] According to an embodiment of the present disclosure, the large model can also perform content search and content generation based on the search intent to obtain a code language describing the text data. The large model can then analyze the data content of the text data and convert the text data into image data, chart data, audio data, etc., thereby outputting multiple output data described in multiple languages. The large model can also perform content search and content generation based on the data content of the text data to obtain image data, chart data, audio data, etc. related to the text data, thereby outputting multiple output data described in multiple languages.

[0059] For example, the large model can analyze the obtained text data and convert it into data in other formats based on the data features included in the text data, so that the data content can be more accurately and clearly reflected in the other formats. If it is determined that the text data is difficult to understand or cannot accurately express the actual meaning, the large model can also conduct content search and content generation based on the data content of the text data again to obtain data in other formats related to the text data, so that the data in other formats can be used to illustrate and supplement the text data.

[0060] According to the embodiments of the present disclosure, by utilizing a large model to output data in different languages ​​based on user search requirements, the output data can be presented to users on the front end in different ways, enriching the presentation of search results and improving their expressiveness. For complex search results, displaying them in different ways can clearly and concisely convey the effective information in the search results, optimizing the visualization and readability of the search results and improving the user experience.

[0061] The following will be combined Figure 3 A schematic description is given of the first search result displayed by the terminal device. Figure 3 A schematic diagram schematically illustrates a first search result according to an embodiment of the present disclosure.

[0062] like Figure 3 As shown, in the display page 300 of the terminal device, the expression of the first search result includes text object 301, picture object 302, chart object 302, table object 340, audio object 305, video object 306, and link object 307.

[0063] According to an embodiment of the present disclosure, the text object 301 , the picture object 302 , the chart object 302 , the table object 340 , the audio object 305 , the video object 306 , and the link object 307 may be rendered from output data described in different languages.

[0064] For example, a user enters a search request such as "Search for Athlete A's career." Text object 301 may describe Athlete A's profile, image object 302 may contain photos of Athlete A, chart object 303 may describe the types of events Athlete A has participated in, and table object 304 may describe the awards Athlete A has won. Audio object 304 may contain an audio interview with Athlete A. Clicking on audio object 304 allows the user to listen to the interview. Video object 305 may contain a video of Athlete A's game. Clicking on video object 305 allows the user to watch the game. Link object 307 may contain Athlete A's personal homepage on a social networking site. Clicking on link object 307 allows the user to view Athlete A's personal homepage.

[0065] It should be noted that the data sources of the various display objects shown on the display page 300 are all in compliance with relevant regulations and are legal, and the display methods and display contents can comply with relevant regulations.

[0066] According to an embodiment of the present disclosure, search results are displayed to users in a variety of expressions according to their search needs, thereby increasing the readability of the search results.

[0067] The following will be combined Figure 4 The principle of the large model-based search method is schematically described.

[0068] Figure 4 The schematic diagram schematically shows the principle of the large model-based search method according to an embodiment of the present disclosure.

[0069] like Figure 4 As shown, the intent recognition model 410 performs intent recognition on the search request 410 to obtain the search intent. For example, the intent recognition model 410 can be a deep learning model running in the server.

[0070] According to an embodiment of the present disclosure, the search intent output by the intent recognition model 410 is input to the big model 420. The search intent may indicate a description language of the output data output by the big model 420 to indicate that the output data output by the big model 420 may be displayed in a variety of expressions.

[0071] For example, the intent recognition model 410 can determine whether the terminal device needs to display the first search result 406 in multiple expressions based on the search request 401. For example, the intent recognition model 410 determines whether expressions such as charts, text, audio, video, and pictures are needed.

[0072] For example, the search request may be "sportsman A's game video," and the intent recognition model 410 may recognize that the first search result 406 needs to be displayed as a video. For example, the search request may be "weather changes displayed in a chart," and the intent recognition model 410 may recognize that the first search result 406 needs to be displayed as a chart. For example, the search request may be "a certain tourist attraction introduced in a variety of forms," ​​and the intent recognition model 410 may recognize that the first search result 406 needs to be displayed in a variety of different expressions.

[0073] When the search request does not explicitly indicate that the first search result 406 should be displayed in only a single expression, the intent recognition model 410 may instruct the large model 420 to preferentially display the first search result 406 in multiple different expressions.

[0074] When the server calls the big model 420 , the server may obtain prompt information 402 related to the search request 401 , concatenate the prompt information 402 with the search request 401 , and input the concatenated information into the big model 420 .

[0075] According to an embodiment of the present disclosure, the prompt information 402 may be pre-set. For example, the prompt information 402 may be obtained during the training process of the large model 420. The prompt information 402 may describe the role, workflow, and precautions of the large model 420.

[0076] For example, the prompt information 402 may be an instruction sent to the large model 420 , or may be a text description, or may be a parameter description in a certain format.

[0077] For example, during the training process of large model 420, the user can adjust the large model by editing prompt information 402 and pre-store the edited prompt information 402 in the server's memory. During the use of large model 420, the server can call prompt information 402 from the memory and splice it with the search request 401 to input into large model 420.

[0078] For example, the user may also adjust the prompt information 402 according to actual operation requirements and operation results during the use of the large model 420 to improve the flexibility of the large model.

[0079] For example, the prompt information 402 may describe the role of the large model 420. For example, the large model role is a model that can write mermaid code. The role profile is that a diagram with a simple structure and clear process can be drawn using mermaid's diagram language according to the user's search request.

[0080] For example, if prompt 402 instructs large model 420 to generate a flowchart, prompt 402 describing the workflow of large model 420 may include the following: 1. Please describe the relevant conceptual information and knowledge involved in the search request in detail, comprehensively, and accurately. 2. Clarify the various prerequisites, possible accidents, potential actions, and processes for the occurrence of the event. 3. Use a diamond symbol for each decision point and clearly define the branching paths for each decision. 4. The logical structure must be clear, the wording must be concise, and the amount of information must be appropriate. 5. The examples provide various grammatical formats and organizational structures for flowcharts. Please refer to them but do not be bound by them. 6. Output mermaid code.

[0081] For example, the precautions described in the prompt information 402 may include requirements on the format of the mermaid code output by the large model 420 and suggestions on writing the mermaid code.

[0082] The large model 420 performs content search based on the prompt information 420 and the search intent, and outputs a plurality of output data. For example, at least one of the plurality of output data may be described in mermaid code to be rendered as a chart in a terminal device.

[0083] In some embodiments, the large model 402 may acquire real-time data related to the search intent and process the real-time data to obtain multiple output data. For example, the large model 420 performs data search and data generation based on the data resource 403 and outputs multiple output data.

[0084] For example, data resources 403 may include pre-built databases and resource libraries, as well as currently available, real-time resources across the entire web. The large model 420 may search the entire web for real-time data related to the search intent to obtain the latest information related to the search intent, thereby ensuring that the output data is factual, timely, and accurate.

[0085] For example, the large model 420 can query the real-time data for data that meets the search intent and search request, and output the data that meets the requirements. The large model 420 can also generate data that meets the search intent and search request based on the real-time data and output it.

[0086] According to an embodiment of the present disclosure, the output data output by the large model 420 includes chart data 404 and multimodal data 405. The chart data 404 may be described in mermaid code. The multimodal data 405 may include text data, image data, audio data, video data, web page link data, presentation files (PowerPoint PPT data), etc.

[0087] According to an embodiment of the present disclosure, the renderer 430 is used to render the chart data 404 and the multimodal data 405 so that the multiple output data are displayed on the front end in corresponding expressions.

[0088] In some embodiments, the renderer 430 may determine a display type of at least one output data described in a specified language, and render the at least one output data based on a presentation control of the display type to obtain the first search result 406 .

[0089] For example, based on the display type of the output data, a display control of that display type is called, and the output data is inserted into the display control. The renderer 420 renders the display space after the output data is inserted, so that the front end displays it using that display type. The display type can be the way the output data is presented. For example, display types can include text, images, audio, video, web links, PowerPoint presentations, and so on. The display control is used to display a specific type of output data. The output data's style can be edited and modified within the display control.

[0090] For example, the renderer 430 may insert the chart data 404 described in mermaid code into a presentation control for presenting a chart, and render the presentation control to display the chart data 404 in a chart format on the front end.

[0091] In some embodiments, the large model 420 may also output layout information and style information based on the search intent. The search intent indicates the user's desired display method and content. The large model 420 may analyze the search intent and determine the layout and style information of the multiple output data in the first search result 406 based on the relevance between the displayed content.

[0092] For example, the layout information may include positional information of multiple output data items within a display page, such as that described above. For example, the layout information may indicate that text data and image data should be positioned adjacent to each other, so that the text data can describe the image data. The layout information may also indicate that video data and chart data should be positioned adjacent to each other, so that the chart data can intuitively represent the content of the video data.

[0093] For example, the style information may include size information and color information of each of the plurality of output data. For example, the style information may indicate the font size and font of the text data. The style information may also indicate the resolution and size of the image data.

[0094] Renderer 430 renders the multiple output data based on the layout information and style information to obtain first search result 460. This allows the multiple output data in first search result 406 to be displayed in a corresponding layout and style, intelligently and selectively shuffling the content of first search result 406 in a clear and appropriate manner, and improving the readability and effectiveness of the output content.

[0095] According to an embodiment of the present disclosure, first search result 406 is sent to a front-end terminal device for display to the user. The user can view first search result 406 on the display screen of the terminal device. The user can send an update request 407 for first search result 406, causing the large model 402 to update first search result 406 and obtain a second search result.

[0096] In some embodiments, the large model 402 may determine the update intent of the update request 407 received for the first search result, and update the first search result 406 based on the update intent to obtain a second search result.

[0097] According to an embodiment of the present disclosure, after browsing the first search result 406 , the user may make another request for the content of the first search result 406 , thereby satisfying the user's extended request for the first search result 406 .

[0098] The following will be combined Figure 5A 、 Figure 5B and Figure 5C The principle of updating the first search result is schematically described.

[0099] Figure 5A The schematic diagram schematically shows the principle of updating the first search result according to an embodiment of the present disclosure.

[0100] like Figure 5A As shown, on the display page 500a of the terminal device, the user can input a data query request through the search bar and click the "Search" button. The terminal device converts the data query request into a search request and sends it to the server. The server calls the large model to output the first search result 502. The first search result 502 is displayed on the display page 500a.

[0101] The display page 500a may also display the information source 503 of the first search result 502. The user may modify the search scope of the first search result 502 by clicking a "directional search" button 504 or a "information source preference setting" button 505.

[0102] According to an embodiment of the present disclosure, if a user believes that the first search result 502 is not accurate enough, they can enter an update intent for the first search result 502 to narrow the search scope. For example, the target object indicated by the update intent in the first search result 502 is determined, and updated output data for the target object is output using a large model. The updated output data is rendered using a renderer to obtain an updated object, and the updated object is used to update the target object in the first search result 502 to obtain a second search result.

[0103] For example, a user clicks "Directed Search" button 504, and window 506 is displayed on page 500a. The user can enter the desired search scope and search content in window 506. Based on the search content, the macro model can determine a target object in first search result 502. The target object is the content in first search result 502 that the user wants to update. For example, the macro model can determine that content in first search result 502 that is relevant to or highly similar to the search content is the target object.

[0104] The search scope and search content entered by the user constitute an update request. Based on the search scope and search content, the large model determines the update intent as re-searching or re-generating content within the specified search scope for the target object in the first search result 502, and then renders the updated output data. The rendered updated object may replace the target object in the first search result 502 to obtain the second search result.

[0105] The representation form of the updated object is not limited, as long as the representation form of the updated object is compatible with the representation form of the target object.

[0106] For example, the user can modify the information source of the first search result 502 by clicking the "Information Source Preferences" button 505. For example, the user can set the web page from which the macro model obtains resources to modify the information source.

[0107] Figure 5B The following schematically shows a principle diagram of updating the first search result according to another embodiment of the present disclosure.

[0108] like Figure 5B As shown, on a display page 500b of a terminal device, first search result 502 includes a text object 521 and a chart object 522. Target object 5211 is text content that a user can select on display page 500b. After the user selects target object 5211, window 509 may be displayed on display page 500b. The user may enter an update request for target object 5211 in window 509.

[0109] The large model can perform content search and / or content generation based on the update request input by the user and the target object 5211 selected by the user, and output associated output data. For example, the associated output data is feedback to the user's update request, and the associated output data is rendered to obtain an associated object. The associated object can serve as supplementary content to the target object, and the first search result 502 is merged with the target object to obtain a second search result. For example, the target object 5211 describes relevant content about a rental contract. The user can enter an update request in window 509: "Is this item valid for shop rental?" The associated object can describe whether the target object 5211 is applicable to shop rentals, as well as the relevant content of the shop rental contract.

[0110] The representation form of the associated object is not limited, as long as the representation form of the associated object is compatible with the representation form of the target object.

[0111] For example, the user can click "Adjust Expression" button 507 to display window 508 on page 500b. The user can enter expression requirements for first search result 502 in window 508. For example, the user can enter expression requirements for chart object 522 in window 508 to modify the style of the chart object.

[0112] For example, the user may directly input a color update request and a size update request, so that the macro model outputs new icon data that satisfies the color update request and the size update request, and replaces the chart object 522 with the rendered new icon object to display the second search result.

[0113] For example, the user can also enter a modification suggestion for chart object 522. For example, in window 508, the user enters an update request, "I'm a junior high school student. Please explain this chart to me in a way I can understand." Based on the update request, the large model identifies chart object 522 as the target object and can generate text data based on the update request, using the rendered text object to describe chart object 522. At this point, first search result 502 and the text object are merged into the second search result.

[0114] The large model may also simplify the content of the chart object 522 based on the update request, or split the content described by the chart object 522 into multiple chart objects, and replace the image object 522 with a new image object to obtain a second search result.

[0115] Figure 5C The following schematically shows a principle diagram of updating the first search result according to another embodiment of the present disclosure.

[0116] like Figure 5C As shown, the display page 500c of the terminal device displays the first search result 502 and evaluation information 510. The evaluation information 510 can be the confidence score output by the large model regarding the first search result 502. The evaluation information 510 can also include an explanation of the confidence score and the information source of the first search result 502.

[0117] The user clicks "Information Verification" button 511, and window 512 appears on page 500c. In window 512, the user can enter an update request: the content to be verified. Based on the user's input, a target object is identified in the first search result 502. The target object is the content to be verified by the user. The large model is used to perform a confidence check on the target object to obtain a target confidence level. This target confidence level is then combined with the first search result and output as a second search result.

[0118] For example, the confidence level can be determined by the large model's scoring of information sources, for example, based on the reputation of the information source and whether the information provided by multiple information sources is consistent.

[0119] According to an embodiment of the present disclosure, during the update process of the first search result 502, the large model may only update part of the content in the first search result 502, thereby achieving a partial update of the first search result 502 to reduce the amount of calculation of the large model during the update process.

[0120] Based on the large model-based search method provided by the present disclosure, the present disclosure also provides a data display method. Figure 6 The data display method is described in detail.

[0121] like Figure 6 As shown, the data display method 600 of this embodiment may include operations S610 to S620.

[0122] In operation S610 , in response to receiving a data query request, a search request is sent.

[0123] According to an embodiment of the present disclosure, a data query request may be inputted by a user into a terminal device through an interactive interface provided by the terminal device. After receiving the inputted data query request, the terminal device may generate the search request and send the search request to the server.

[0124] For example, the terminal device can convert the data query request into a language that can be recognized by the large model, or add prompt information to the data query request to generate a search request.

[0125] For example, the prompt information added by the terminal device can be used to instruct the image recognition model to perform intent recognition.

[0126] In operation S620 , in response to obtaining a first search result generated in response to the search request, the first search result is displayed.

[0127] According to an embodiment of the present disclosure, the first search result is generated by the large model-based search method described above.

[0128] For example, after receiving the search request, the server may output the first search result using the large model-based search method described above and feed the first search result back to the terminal device. The terminal device may then display the first search result after obtaining the first search result.

[0129] Through the data display method provided by the present disclosure, the first search result can be displayed to the user in multiple ways based solely on the user's data query request, thereby improving the expressiveness and readability of the search results output by the large model and improving the user experience.

[0130] Based on the large model-based search method provided by the present disclosure, the present disclosure also provides a large model-based search device. Figure 7 The device is described in detail.

[0131] Figure 7 A block diagram of a large model-based search device according to an embodiment of the present disclosure is schematically shown.

[0132] like Figure 7 As shown, the large model-based search device 700 may include: a first determination module 710 , a search module 720 and a rendering module 730 .

[0133] The first determination module 710 is configured to determine, in response to a received search request, the search intent indicated by the search request.

[0134] The search module 720 is used to perform content search based on the search intent using the large model to obtain multiple output data, at least one of the multiple output data is described in a specified language.

[0135] The rendering module 730 is configured to render the plurality of output data to obtain a first search result, wherein the rendering effect of the output data described in the specified language is different from the rendering effect of the remaining output data in the plurality of output data.

[0136] According to an embodiment of the present disclosure, the large model-based search apparatus 700 further includes a second determination module and an update module. The second determination module is configured to, in response to a received update request for a first search result, determine the update intent of the update request. The update module is configured to update the first search result based on the update intent using the large model to obtain a second search result.

[0137] According to an embodiment of the present disclosure, the update module includes: a first determination submodule, used to determine the target object indicated by the update intention in the first search result; a first output submodule, used to output the updated output data of the target object using a large model; a first rendering submodule, used to render the updated output data to obtain an updated object; and an update submodule, used to update the target object in the first search result using the updated object to obtain a second search result.

[0138] According to an embodiment of the present disclosure, the update module includes: a first determination submodule, used to determine the target object indicated by the update intention in the first search result; a first search submodule, used to use a large model to perform content search for the update intention, and obtain associated output data associated with the target object; a second rendering submodule, used to render the associated output data and obtain an associated object; and a second output submodule, used to output a second search result based on the associated object and the first search result.

[0139] According to an embodiment of the present disclosure, the update module includes: a first determination submodule, used to determine the target object indicated by the update intention in the first search result; a verification submodule, used to use a large model to perform confidence verification on the target object to obtain a target confidence; and a third output submodule, used to output a second search result based on the first search result and the target confidence.

[0140] According to an embodiment of the present disclosure, the rendering module 730 includes: a second determination submodule, used to determine the display type of at least one output data described in a specified language; and a third rendering submodule, used to render at least one output data based on a display control of the display type to obtain a first search result.

[0141] According to an embodiment of the present disclosure, the search module 720 includes: a first acquisition submodule, which is used to use a large model to acquire real-time data related to the search intention; and a processing submodule, which is used to process the real-time data to obtain multiple output data.

[0142] According to an embodiment of the present disclosure, the search module 720 includes: a second acquisition sub-module, used to obtain prompt information related to the search intent, which is pre-set; and a second search sub-module, used to use a large model to perform content search based on the prompt information and search intent to obtain multiple output data.

[0143] According to an embodiment of the present disclosure, the rendering module 730 includes: a third determination submodule, used to determine the layout information and style information indicated by the search intention; and a fourth rendering submodule, used to render multiple output data based on the layout information and style information to obtain a first search result.

[0144] Based on the data display method provided by the present disclosure, the present disclosure also provides a data display device. Figure 8 The device is described in detail.

[0145] Figure 8 It is a structural block diagram of a data display device according to an embodiment of the present disclosure.

[0146] like Figure 8 As shown, the data display device 800 of this embodiment may include an information sending module 810 and a display module 820.

[0147] The sending module 810 is configured to send a search request in response to receiving a data query request. In one embodiment, the sending module 810 may be configured to execute the operation S610 described above, which will not be described in detail herein.

[0148] The display module 820 is configured to display the first search result in response to obtaining the first search result generated in response to the search request.

[0149] The first search result is generated by the large model-based search device described above. In one embodiment, the display module 820 can be used to perform the operation S620 described above, which will not be described in detail here.

[0150] Figure 9 The structural block diagram of an artificial intelligence agent according to an embodiment of the present disclosure is schematically shown.

[0151] In the embodiments of the present disclosure, inspired by the von Neumann structure in modern computer theory, such as Figure 9As shown, the AI ​​agent 900 may include five core modules: an input module 910 , a control module 920 , a storage module 930 , a calculation module 940 and an output module 950 .

[0152] Input module 910 is responsible for receiving or perceiving information such as queries, requests, instructions, signals, or data from the outside world (e.g., users or the external environment) and converting it into a format that AI agent 900 can understand and process. Input module 910 is the primary link for AI agent 900 to interact with the outside world. It enables AI agent 900 to efficiently and accurately obtain necessary "sensory" information from the outside world and respond to this information.

[0153] In an example, the input module 910 may input the search request and update request described above.

[0154] In this example, the control module 920 is the core support for the AI ​​agent 900 to handle complex tasks. The control module 920 can execute the large model-based search method described above.

[0155] In the example, the control module 920 will continuously interact with the storage module 930, the computing module 940, and / or the output module 950 during operation. However, it should be noted that in the embodiment of the present disclosure, the control module 920 acts as a single initiator to initiate communication with the storage module 930, the computing module 940, and / or the output module 950, and there is no communication coupling between the storage module 930, the computing module 940, and the output module 950.

[0156] In this example, the performance of control module 920 may be closely related to the large model underlying AI agent 900. To fully leverage the capabilities of the large language model, the internal structure of control module 920 may be designed to be highly configurable and extensible to handle a variety of different tasks and requirements in real-world scenarios.

[0157] The storage module 930 may be responsible for memorizing information such as historical conversations, event streams, etc. The aforementioned prompt information and data resources may be included in the storage module 930 .

[0158] In this example, after receiving a search request, AI agent 900 can use an intent recognition model to determine the search intent from the search request. The search intent can be stored in storage module 930. AI agent 900 can retrieve relevant search intent, prompt information, and data resources from storage module 930 and feed them back to control module 920. Control module 920 can then use the feedback prompt information and data resources to obtain a first search result corresponding to the search request and pass the first search result to output module 950.

[0159] The operation module 940 can be regarded as a predefined tool library. The renderer and display controls mentioned above can be included in the operation module 940.

[0160] In the example, when the AI ​​agent 900 needs to render multiple output data, it can call the relevant renderer and display control from the operation module 940 and feed it back to the control module 920. Then, the control module 920 can use the feedback renderer and display control to render the first search result and pass the first search result to the output module 950. It can be understood that although the large language model has excellent language understanding and generation capabilities, it is the same as a human being. Without the help of any tools, the tasks that can be solved are very limited. When the AI ​​agent 900 is given the ability to call tools, it can achieve tasks such as completing mathematical operations with the help of a calculator, completing data analysis with the help of python, and completing weather forecasts with the help of a search engine.

[0161] In an example, the output module 950 may output the first search result and the second search result described above.

[0162] The AI ​​agent 900 according to the embodiment of the present disclosure can simply and effectively improve the level of intelligence, and enhance flexibility and versatility.

[0163] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0164] According to an embodiment of the present disclosure, an electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method described above.

[0165] According to an embodiment of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause a computer to execute the method described above.

[0166] According to an embodiment of the present disclosure, a computer program product includes a computer program, and when the computer program is executed by a processor, the computer program implements the method described above.

[0167] Figure 10A schematic block diagram of an example electronic device 1000 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0168] like Figure 10 As shown, electronic device 1000 includes a computing unit 1001, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 1002 or a computer program loaded from a storage unit 1008 into a random access memory (RAM) 1003. Various programs and data required for the operation of device 800 may also be stored in RAM 1003. Computing unit 1001, ROM 1002, and RAM 1003 are connected to each other via bus 804. An input / output (I / O) interface 1005 is also connected to bus 1004.

[0169] Multiple components in the electronic device 1000 are connected to the I / O interface 1005, including an input unit 1006, such as a keyboard, a mouse, etc.; an output unit 1007, such as various types of displays, speakers, etc.; a storage unit 1008, such as a magnetic disk, an optical disk, etc.; and a communication unit 1009, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 1009 allows the electronic device 1000 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0170] The computing unit 1001 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the computing unit 1001 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 801 performs the various methods and processes described above, such as the large-model-based search method. For example, in some embodiments, the large-model-based search method can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 808. In some embodiments, part or all of the computer program can be loaded and / or installed onto the device 800 via the ROM 802 and / or the communication unit 809. When the computer program is loaded into the RAM 803 and executed by the computing unit 801, one or more steps of the large-model-based search method described above can be performed. Alternatively, in other embodiments, the computing unit 801 may be configured to execute the large model-based search method in any other appropriate manner (eg, by means of firmware).

[0171] Various embodiments of the systems and techniques described above can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-a-chip systems (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0172] The program code for implementing the method of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device so that when the program code is executed by the processor or controller, the functions / operations specified in the flow chart and / or block diagram are implemented. The program code can be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0173] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of machine-readable storage media may include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), optical fibers, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0174] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0175] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0176] A computer system may include a client and a server. The client and server are generally remote from each other and typically interact through a communication network. The client-server relationship arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server may be a cloud server, a server in a distributed system, or a server integrated with a blockchain.

[0177] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this disclosure can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved. This is not a limitation herein.

[0178] The above specific embodiments do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure shall be included within the scope of protection of this disclosure.

Claims

1. A search method based on a large model, comprising: In response to a received search request, determining a search intent indicated by the search request; Performing a content search based on the search intent using the large model to obtain a plurality of output data, at least one of the plurality of output data being described in a specified language; as well as Rendering the plurality of output data to obtain a first search result, wherein a rendering effect of the output data described in the specified language is different from a rendering effect of the remaining output data in the plurality of output data; The method further includes: in response to a received update request for the first search result, determining an update intent of the update request; and using the large model to update the first search result based on the update intent to obtain a second search result; The updating of the first search result based on the update intention using the large model to obtain the second search result includes: determining a target object indicated by the update intention in the first search result; performing a confidence check on the target object using the large model to obtain a target confidence; and outputting the second search result based on the first search result and the target confidence. The target confidence is determined by the large model based on the information source of the target object.

2. The method according to claim 1, wherein The updating of the first search result based on the update intention by using the large model to obtain a second search result includes: Determining a target object of the update intention indication in the first search result; outputting updated output data of the target object using the large model; Rendering the updated output data to obtain an updated object; and The target object in the first search result is updated using the update object to obtain the second search result.

3. The method according to claim 1, wherein The updating of the first search result based on the update intention by using the large model to obtain a second search result includes: Determining a target object of the update intention indication in the first search result; Performing content search for the update intention using the large model to obtain associated output data associated with the target object; Rendering the associated output data to obtain an associated object; and Based on the associated object and the first search result, the second search result is output.

4. The method according to claim 1, wherein Rendering the plurality of output data to obtain a first search result includes: determining a display type of at least one output data described in the specified language; and Based on the display control of the display type, the at least one output data is rendered to obtain the first search result.

5. The method according to claim 1, wherein The large model is used to perform content search based on the search intent, and multiple output data are obtained, including: Using the large model to obtain real-time data related to the search intent; and The real-time data is processed to obtain the plurality of output data.

6. The method according to claim 1, wherein The large model is used to perform content search based on the search intent, and multiple output data are obtained, including: Obtaining prompt information related to the search request, where the prompt information is preset; and The large model is used to perform content search based on the prompt information and the search intent to obtain the multiple output data.

7. The method according to claim 1, wherein Rendering the plurality of output data to obtain a first search result includes: Determining layout information and style information indicated by the search intent; and The plurality of output data are rendered based on the layout information and the style information to obtain the first search result.

8. A data presentation method, comprising: In response to receiving the data query request, sending the search request; as well as In response to obtaining the first search result generated in response to the search request, displaying the first search result; The first search result is generated by using the search method according to any one of claims 1 to 7.

9. A search device based on a large model, comprising: A first determining module, configured to determine, in response to a received search request, a search intent indicated by the search request; A search module, configured to use the large model to perform a content search based on the search intent, and obtain a plurality of output data, at least one of the plurality of output data being described in a specified language; as well as a rendering module, configured to render the plurality of output data to obtain a first search result, wherein a rendering effect of the output data described in the specified language is different from a rendering effect of the remaining output data in the plurality of output data; The device further comprises: a second determining module, configured to determine, in response to the received update request for the first search result, an update intention of the update request; and An updating module, configured to update the first search result based on the update intention using the large model to obtain a second search result; The update module includes: A first determining submodule, configured to determine a target object indicated by the update intention in the first search result; A verification submodule, configured to perform confidence verification on the target object using the large model to obtain target confidence; and a third output submodule, configured to output the second search result based on the first search result and the target confidence; The target confidence is determined by the large model based on the information source of the target object.

10. The device according to claim 9, wherein The update module includes: A first determining submodule, configured to determine a target object indicated by the update intention in the first search result; A first output submodule, configured to output updated output data of the target object using the large model; A first rendering submodule is configured to render the updated output data to obtain an updated object; and An updating submodule is configured to update the target object in the first search result using the update object to obtain the second search result.

11. The device according to claim 9, wherein The update module includes: A first determining submodule, configured to determine a target object indicated by the update intention in the first search result; A first search submodule is configured to perform a content search for the update intention using the large model to obtain associated output data associated with the target object; A second rendering submodule is configured to render the associated output data to obtain an associated object; and The second output submodule is configured to output the second search result based on the associated object and the first search result.

12. The apparatus according to claim 9, wherein the rendering module comprises: a second determining submodule, configured to determine a display type of at least one output data described in the specified language; as well as The third rendering submodule is configured to render the at least one output data based on the display control of the display type to obtain the first search result.

13. The device according to claim 9, wherein The search module includes: A first acquisition submodule is configured to acquire real-time data related to the search intent using the large model; and The processing submodule is used to process the real-time data to obtain the multiple output data.

14. The device according to claim 9, wherein The search module includes: A second acquisition submodule is configured to acquire prompt information related to the search request, where the prompt information is preset; and The second search submodule is used to use the large model to perform content search based on prompt information and search intent to obtain multiple output data.

15. The device according to claim 9, wherein The rendering module includes: A third determining submodule is configured to determine layout information and style information indicated by the search intent; and A fourth rendering submodule is configured to render the plurality of output data based on the layout information and the style information to obtain the first search result.

16. A data display device, comprising: A sending module, configured to send the search request in response to receiving the data query request; as well as a display module, configured to display the first search result in response to obtaining the first search result generated in response to the search request; The first search result is generated by using the search device according to any one of claims 9 to 15.

17. An artificial intelligence agent configured to execute the method according to any one of claims 1 to 8.

18. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 8.

19. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to enable the computer to execute the method according to any one of claims 1 to 8.

20. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the method according to any one of claims 1 to 8.

Images