Display device and display method

Abstract

The invention relates to the technical field of display equipment, and provides display equipment and a display method.The display equipment comprises a displayer and a controller, and the controller is configured to respond to a session text input by a user and drive an intention large model to recognize a user intention of the session text; under the condition that the user intention is the media asset retrieval intention, driving an analysis large model to analyze the user intention so as to obtain structured data for media asset retrieval; and retrieving the media asset data corresponding to the structured data in a media asset database, and driving a display to display the media asset data on a user interface. According to the technical scheme, when the session text input by the user is complex content, the intention large model can be used for accurately obtaining the user intention, and when the user intention is media asset retrieval, the structured data of the user media asset retrieval is obtained based on analysis of the large model, so that the media asset resources retrieved by using the structured data are ensured, and the user experience is improved. And the session text input by the user can be stably and reliably responded.

Description

Technical Field

[0001] The present application relates to the technical field of display devices, and in particular to a display device and a display method. Background Art

[0002] Media retrieval refers to the process of finding and locating media resources on a TV or other video playback device through specific search functions, such as finding movies, TV series, news clips and documentaries.

[0003] Existing media resource retrieval methods rely on the segmentation and labeling of various entity words in natural language. When receiving user statements, they are parsed according to the vocabulary and tag library in the existing language system. After obtaining business-related tags and segmentations, the corresponding media resources are retrieved in the database based on the tags and segmentations.

[0004] However, when faced with complex user statements, for example, when the user statement includes multiple intentions or there are many search restrictions, the output media results often fail to meet user expectations, resulting in the inability of existing media retrieval methods to respond to user statements stably and reliably. Summary of the invention

[0005] The present application provides a display device and a display method, so that the display device can stably and reliably respond to the user's media resource retrieval needs.

[0006] A first aspect provides a display device, comprising:

[0007] a display configured to display a user interface;

[0008] A controller is coupled to the display and is configured to:

[0009] In response to a conversation text input by a user, driving the intention big model to identify the user intention of the conversation text;

[0010] In the case where the user intention is a media resource retrieval intention, driving the large parsing model to parse the user intention to obtain structured data for media resource retrieval, and driving the display to display the structured data on the user interface;

[0011] The media asset data corresponding to the structured data is retrieved from a media asset database, and the display is driven to display the media asset data on a user interface.

[0012] In this way, the intention big model is used to identify the user intention of the conversation text, so that when the conversation text input by the user is complex, the user intention can be accurately obtained, and when the user intention is to retrieve media resources, the structured data of the user's media retrieval is obtained based on the parsing big model, thereby ensuring that the media resources retrieved using structured data can stably and reliably respond to the conversation text input by the user.

[0013] In an implementation of the first aspect, before retrieving the media asset data corresponding to the structured data in the media asset database, the controller is further configured to:

[0014] In response to an input channel of the conversation text, a media asset database corresponding to the structured data is determined.

[0015] In this way, it is possible to avoid a mismatch between the input channel and the media database, which results in the user being unable to view the retrieved media data, or the situation where the media data required by the user is not fully retrieved. That is, by determining the corresponding media database through the input channel of the conversation text, the retrieved media data can more easily meet the user's expectations.

[0016] In an implementation of the first aspect, if the acquired media asset data is multiple, in the step of driving the display to display the media asset data on the user interface, the controller is specifically configured as follows:

[0017] Based on a preset rule, determining a display order of the plurality of media asset data;

[0018] According to the display order of the plurality of media asset data, the display is driven to display the plurality of media asset data on the user interface.

[0019] In this way, through preset rules, multiple media data are displayed on the user interface according to the display ranking, so that users can determine the required target among the media data with the highest ranking as much as possible, avoiding the user having to view multiple media data before understanding or viewing the required media data, thereby improving the user experience.

[0020] In an implementation of the first aspect, in the step of driving the intention big model to identify the user intention of the conversation text in response to the conversation text input by the user, the controller is specifically configured to:

[0021] Constructing a first Prompt instruction for intent recognition according to the conversation text;

[0022] The first Prompt instruction is input into the intention big model to obtain the user intention.

[0023] In this way, by constructing the first Prompt instruction, the input information of the intention model is made more comprehensive, which in turn helps the intention model to improve the accuracy and standardization of intent recognition based on more comprehensive information.

[0024] In an implementation of the first aspect, in the step of driving the large parsing model to parse the conversation text to obtain structured data for media asset retrieval, the controller is specifically configured to:

[0025] Constructing a second prompt instruction for media resource retrieval according to the conversation text and the corresponding user intention;

[0026] The second Prompt instruction is input into the parsing large model to obtain the structured data.

[0027] In this way, by constructing the second Prompt instruction, the input information for parsing the large model is made more comprehensive, which in turn helps the large model to obtain more accurate and higher structured data based on more comprehensive information, and makes the structured data more standardized.

[0028] In an implementation of the first aspect, after the driving intention big model identifies the user intention of the conversation text, the controller is further configured to:

[0029] In the case where the user intention is a dialogue intention, the analysis model is driven to generate dialogue content corresponding to the conversation text, and the display is driven to display the dialogue content on the user interface.

[0030] In this way, when this solution is applied to a smart TV, it can not only meet the user's media resource retrieval needs, but also provide corresponding answers to the user's question-like conversation text.

[0031] In an implementation of the first aspect, before driving the intention model to identify the user intention of the conversation text in response to the conversation text input by the user, the controller is further configured to:

[0032] A historical conversation text input by a user is obtained, and based on the historical conversation text, the conversation text is adjusted.

[0033] In this way, the hidden information in the user conversation text can be accurately obtained during the conversation with the user, ensuring that the intent model can accurately identify the user's intent, thereby improving the user experience.

[0034] In an implementation of the first aspect, before driving the intention big model to identify the user intention of the conversation text, the controller is further configured to:

[0035] In a case where the content represented by the conversation text is not a control task, the conversation text is input into the intention big model.

[0036] In this way, by first determining whether the content represented by the conversation text is a control task, it is possible to avoid calling the large intention model to identify the user control task, thereby wasting the computing power resources of the controller.

[0037] In an implementation of the first aspect, the controller is further configured to:

[0038] Obtaining historical conversation texts of media resource retrieval and dialogue services, and marking user intentions corresponding to the historical conversation texts;

[0039] Constructing a first training set using the historical conversation texts and the user intentions corresponding to the historical conversation texts;

[0040] A first initial model is constructed, and the first training set is used to train the first initial model to obtain the intention large model.

[0041] In this way, the first initial model is trained through the first training set to obtain the intention big model. On this basis, the first training set can also be used to enhance the user intention analysis capability of the intention big model.

[0042] In an implementation of the first aspect, before the controller drives the large parsing model to parse the user intention, the controller is further configured to:

[0043] Acquire historical conversation texts retrieved by media resources, and drive the intention big model to identify user intentions corresponding to the historical conversation texts retrieved by media resources;

[0044] Annotating the corresponding structured data of the media resource retrieval;

[0045] Using the user intention and structured data corresponding to the historical conversation text, construct a second training set;

[0046] A second initial model is constructed, and the second initial model is trained using the second training set to obtain the analytical large model.

[0047] In this way, the second initial model is trained by the second training set to obtain the analytical large model. On this basis, the second training set can also be used to enhance the analytical ability of the analytical large model.

[0048] A second aspect provides a display method, which is applied to a display device, wherein the display device includes a display capable of displaying a user interface, and the display method includes:

[0049] In response to a conversation text input by a user, driving the intention big model to identify the user intention of the conversation text;

[0050] In the case where the user intention is a media resource retrieval intention, driving the parsing large model to parse the conversation text to obtain structured data for media resource retrieval, and driving the display to display the structured data on the user interface;

[0051] The media asset data corresponding to the structured data is retrieved from a media asset database, and the display is driven to display the media asset data on a user interface.

[0052] The display method provided by the present application uses the intention big model to identify the user intention of the conversation text, so that when facing complex conversation text, the user intention can be accurately obtained, and when the user intention is media resource retrieval, the structured data of the user media resource retrieval is obtained based on the parsing big model, so as to ensure that the media resource retrieved using the structured data can stably and reliably respond to the conversation text input by the user. The display device and display method provided by the embodiment of the present application, the display device includes a display and a controller, the controller is coupled to the display, and the controller is configured to: in response to the conversation text input by the user, drive the intention big model to identify the user intention of the conversation text; when the user intention is the media resource retrieval intention, drive the parsing big model to parse the user intention to obtain the structured data for media resource retrieval, and drive the display to display the structured data on the user interface; retrieve the media resource data corresponding to the structured data in the media resource database, and drive the display to display the media resource data on the user interface.

[0053] By applying the technical solution provided in the present application, the intention big model is used to identify the user intention of the conversation text, so that when the conversation text input by the user is complex, the user intention can be accurately obtained, and when the user intention is to retrieve media resources, the structured data of the user's media retrieval is obtained based on the parsing big model, thereby ensuring that the media resources retrieved using structured data can stably and reliably respond to the conversation text input by the user. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0055] Figure 1 A schematic diagram of an operation scenario between a display device and a control device provided in some embodiments of the present application;

[0056] Figure 2 A schematic diagram of the hardware configuration of a display device provided in some embodiments of the present application;

[0057] Figure 3 A schematic diagram of software configuration of a display device provided in some embodiments of the present application;

[0058] Figure 4 Schematic diagram of the configuration process of the controller of the display device provided in some embodiments of the present application Figure 1 ;

[0059] Figure 5 Schematic diagram of a conversation text input interface provided in some embodiments of the present application Figure 1 ;

[0060] Figure 6 Schematic diagram of a conversation text input interface provided in some embodiments of the present application Figure 2 ;

[0061] Figure 7 Schematic diagram of a conversation text input interface provided in some embodiments of the present application Figure 3 ;

[0062] Figure 8 Schematic diagram of a conversation text input interface provided in some embodiments of the present application Figure 4 ;

[0063] Fig. 9 A schematic diagram of a user interface for displaying structured data provided in some embodiments of the present application;

[0064] Fig.10 Schematic diagram of a user interface for displaying media asset data provided in some embodiments of the present application Figure 1 ;

[0065] Fig.11 Schematic diagram of a user interface for displaying media asset data provided in some embodiments of the present application Figure 2 ;

[0066] Fig.12 Schematic diagram of a user interface for displaying media asset data provided in some embodiments of the present application Figure 3 ;

[0067] Fig.13 Schematic diagram of the configuration process of the controller of the display device provided in some embodiments of the present application Figure 2 ;

[0068] Fig.14 Schematic diagram of the configuration process of the controller of the display device provided in some embodiments of the present application Figure 3 ;

[0069] Fig.15A timing diagram of a terminal interaction system of a display device provided in some embodiments of the present application. DETAILED DESCRIPTION

[0070] The following embodiments are described in detail, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following embodiments do not represent all implementations consistent with the present application. They are only examples of systems and methods consistent with some aspects of the present application as detailed in the claims.

[0071] It should be noted that the brief description of terms in this application is only for the convenience of understanding the embodiments described below, and is not intended to limit the embodiments of this application. Unless otherwise specified, these terms should be understood according to their common and usual meanings.

[0072] The terms "first", "second", "third", etc. in the specification and claims of this application and the above drawings are used to distinguish similar or similar objects or entities, and do not necessarily mean to limit a specific order or sequence, unless otherwise noted. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances.

[0073] The terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover but not exclude inclusion, for example, a product or device comprising a list of components is not necessarily limited to all the components expressly listed but may include other components not expressly listed or inherent to such product or device.

[0074] The term "module" refers to any known or later developed hardware, software, firmware, artificial intelligence, fuzzy logic or combination of hardware and / or software code that is capable of performing the functions associated with that element.

[0075] In the embodiment of the present application, the display device 200 generally refers to a device with image display, search capability and data processing capability. For example, the display device 200 includes but is not limited to a smart TV, a mobile terminal, a computer, a monitor, an advertising screen, a wearable device, a virtual reality device, an augmented reality device, etc.

[0076] Figure 1 This is a schematic diagram of an operation scenario between a display device and a control device provided in some embodiments of the present application. Figure 1 As shown in FIG. 1 , a user can operate the display device 200 through touch operation, the mobile terminal 300 and the control device 100. The control device 100 is used to receive operation instructions input by the user and convert the operation instructions into control instructions that can be recognized and responded to by the display device 200. For example, the control device 100 can be a remote controller, a stylus pen, a handle, etc.

[0077] The mobile terminal 300 can be used as a control device for performing human-computer interaction between the user and the display device 200. The mobile terminal 300 can also be used as a communication device for establishing a communication connection with the display device 200 and performing data interaction. In some embodiments, the mobile terminal 300 can install software applications with the display device 200, and achieve connection and communication through a network communication protocol to achieve the purpose of one-to-one control operation and data communication. The audio and video content displayed on the mobile terminal 300 can also be transmitted to the display device 200 to achieve a synchronous display function.

[0078] In some embodiments, the mobile terminal 300 or other electronic devices may also simulate the functions of the control device 100 by running an application program for controlling the display device 200 .

[0079] like Figure 1 As also shown in FIG. 4 , the display device 200 also communicates data with the server 400 through various communication methods. The display device 200 may be allowed to communicate and connect through a local area network (LAN), a wireless local area network (WLAN), and other networks.

[0080] The display device 200 may provide a broadcast receiving television function, and may also additionally provide an intelligent network television function with a computer support function, including but not limited to network television, smart television, Internet Protocol television (IPTV), and the like.

[0081] Figure 2 Some embodiments of the present application provide Figure 1 2 is a block diagram of the hardware configuration of the display device 200.

[0082] In some embodiments, the display device 200 may include at least one of a tuner 210, a communication device 220, a detector 230, a device interface 240, a controller 250, a display 260, an audio output device 270, a memory, a power supply, and a user input interface.

[0083] In some embodiments, the detector 230 is used to collect signals of the external environment or external interaction. For example, the detector 230 includes a light receiver, a sensor for collecting the intensity of ambient light; or, the detector 230 includes an image collector, such as a camera, which can be used to collect external environment scenes, user attributes or user interaction gestures; or, the detector 230 includes a sound collector, such as a microphone, etc., for receiving external sounds.

[0084] In some embodiments, the display 260 includes a display function component for presenting a picture, and a driving component for driving an image display. The display 260 is used to receive an image signal output from the controller 250 for display. For example, the display 260 can be used to display video content, image content, and components of a menu control interface and a user control UI interface.

[0085] In some embodiments, the communication device 220 is a component for communicating with an external device or server 400 according to various communication protocol types. The display device 200 may be provided with a plurality of communication devices 220 according to different supported communication modes. For example, when the display device 200 supports wireless network communication, the display device 200 may be provided with a communication device 220 including a WiFi function. When the display device 200 supports Bluetooth connection communication, the display device 200 needs to be provided with a communication device 220 including a Bluetooth function.

[0086] The communication device 220 can enable the display device 200 to communicate with the external device or server 400 by wireless or wired connection. Among them, the wired connection can connect the display device 200 with the external device through components such as data cables and interfaces. The wireless connection can connect the display device 200 with the external device through wireless signals or wireless networks. The display device 200 can establish a connection relationship with the external device directly, or indirectly establish a connection relationship through a gateway, a router, a connection device, etc.

[0087] In some embodiments, the controller 250 may include at least one of a central processing unit, a video processor, an audio processor, a graphics processor, and a power processor, and a first interface to an nth interface for input / output. The controller 250 controls the operation of the display device and responds to user operations through various software control programs stored in the memory. For example, there is a module for searching in the software architecture, so that after providing key data for retrieval, the corresponding results are retrieved from the database according to the key data. The controller 250 controls the overall operation of the display device 200.

[0088] In some embodiments, the controller 250 and the tuner-demodulator 210 may be located in different separate devices, that is, the tuner-demodulator 210 may also be located in an external device of the main device where the controller 250 is located, such as an external set-top box.

[0089] In some embodiments, the user may input a user command through a graphical user interface (GUI) displayed on the display 260 , and the user input interface receives the user input command through the graphical user interface (GUI).

[0090] In some embodiments, the audio output device 270 may be a local speaker of the display device 200, or may be an external audio output device of the display device 200. In particular, for the external audio output device of the display device 200, the display device 200 may also be provided with an external audio output terminal, and the audio output device may be connected to the display device 200 through the external audio output terminal to output the sound of the display device 200.

[0091] In some embodiments, the user input interface 280 may be used to receive instructions from a user.

[0092] In order to perform user interaction, in some embodiments, the display device 200 may run an operating system. The operating system is a computer program for managing and controlling hardware resources and software resources in the display device 200. The operating system can control the display device to provide a user interface. For example, the operating system can directly control the display device to provide a user interface, or provide a user interface by running an application program. The operating system also allows the user to interact with the display device 200.

[0093] It should be noted that the operating system can be a native operating system based on a specific operating platform, a third-party operating system deeply customized based on a specific operating platform, or an independent operating system specially developed for the display device.

[0094] The operating system can be divided into different modules or layers according to the functions implemented, such as Figure 3 As shown, in some embodiments, the system is divided into four layers, from top to bottom, namely, the application layer (Applications) layer (referred to as "application layer"), the application framework layer (Application Framework) layer (referred to as "framework layer"), the system library layer and the kernel layer.

[0095] In some embodiments, the application layer is used to provide services and interfaces for applications so that the display device 200 can run applications and interact with users based on the applications. At least one application can be run in the application layer, and these applications can be window programs, system settings programs, clock programs, etc. that come with the operating system; they can also be applications developed by third-party developers. In specific implementations, the application packages in the application layer are not limited to the above examples.

[0096] The framework layer provides application programming interfaces (APIs) and programming frameworks for applications. The application framework layer includes some predefined functions. The application framework layer is equivalent to a processing center that determines the actions that applications in the application layer take. Through the API interface, applications can access system resources and obtain system services during execution.

[0097] like Figure 3 As shown, in the embodiment of the present application, the application framework layer includes a view system, managers, content providers, etc., wherein the view system can design and implement the interface and interaction of the application, and the view system includes lists, grids, text boxes, buttons, etc. The manager includes at least one of the following modules: an activity manager for interacting with all activities running in the system; a location manager for providing system services or applications with access to system location services; a package manager for retrieving various information related to the application package currently installed on the device; a notification manager for controlling the display and clearing of notification messages; and a window manager for managing icons, windows, toolbars, wallpapers, and desktop widgets on the user interface.

[0098] In some embodiments, the activity manager is used to manage the life cycle of each application and the usual navigation back function, such as controlling the exit, opening, and back of the application. The window manager is used to manage all window programs, such as obtaining the display screen size, determining whether there is a status bar, locking the screen, capturing the screen, and controlling the display window changes, for example, reducing the display window, shaking the display, distorting the display, etc.

[0099] In some embodiments, the system runtime layer can provide support for the framework layer. When the framework layer is used, the operating system will run the instruction library contained in the system runtime layer, such as the C / C++ instruction library, to implement the functions to be implemented by the framework layer.

[0100] In some embodiments, the kernel layer is a functional layer between the hardware and software of the display device 200. The kernel layer can implement functions such as hardware abstraction, multitasking, and memory management. Figure 3As shown, the kernel layer may be configured with hardware drivers, and the drivers included in the kernel layer may be at least one of the following drivers: audio driver, display driver, Bluetooth driver, camera driver, WIFI driver, USB driver (Universal Serial Bus: USB), HD MI driver (High Definition Multimedia Interface: HDMI), sensor driver (such as fingerprint sensor, temperature sensor, pressure sensor, etc.), and power driver, etc.

[0101] It should be noted that the above example is only a simple division of the operating system functions and does not constitute a limitation on the specific operating system form of the display device 200 in the embodiment of the present application. Depending on factors such as the function of the display device and the type of operating system, the number of levels and specific level types contained in the operating system may be expressed in other forms.

[0102] The present application provides a display device, including: a display and a controller, the controller is coupled to the display, and the display is configured to display a user interface; Figure 4 As shown, the controller is configured to execute S410 to S430.

[0103] S410 , in response to a conversation text input by a user, driving the intention big model to identify the user intention of the conversation text.

[0104] The input method of the conversation text may be voice input or text input. For example, the user triggers the voice input control of the display device, and then the controller collects the user's voice through the sound collector and converts the user's voice into conversation text using the voice recognition model.

[0105] In some embodiments, Figure 5 As shown, the conversation text is: I want to watch a movie of actor A. The controller can drive the display to display the conversation text after voice conversion on the user interface. In this way, the user can view the converted text. At this time, if there is a voice recognition error, the user can find it in the first time, so that the user can correct the error in the conversation text.

[0106] For example, Figure 6 As shown, there is a text input bar on the user interface. After the user triggers the text input bar, the controller controls the display to display an electronic keyboard for inputting text on the user interface. The user types the conversation text by triggering the input control on the electronic keyboard.

[0107] For example, Figure 7There is a text input bar on the user interface. After the user triggers the text input bar, the controller controls the display to display a writing board for inputting text on the user interface. The user can input the conversation text by handwriting on the writing board, or use a mobile terminal to input the conversation text. The mobile terminal is a control device used to perform human-computer interaction between the user and the display device.

[0108] In some embodiments, when the user inputs the conversation text, he or she may also switch between the electronic keyboard and the writing board, and if necessary, he or she may continue to input the conversation text based on the voice input. Figure 8 As shown, after the conversation text is input by voice, the controller controls the display conversation text: I want to watch the movie of actor A. At this time, the trigger controller controls the display to display a continue input control on the user interface. After the user triggers the continue input control, the display can be controlled to display multiple input mode options on the user interface for the user to select the corresponding input mode.

[0109] After the user confirms the input conversation text, the controller uses the conversation text as the input of the intention big model and obtains the output of the intention big model, thereby obtaining the user intention corresponding to the conversation text.

[0110] For example, if the conversation text is: I want to watch movies starring actor A, then the output of the intent model is: [intent, video search; search content, movies starring actor A].

[0111] For example, if the conversation text is: I want to listen to the songs of singer B, then the output result of the intent model is: [intent, song search; search content, songs of singer B].

[0112] For example, if the conversation text is: What major news is there in region C today, then the output result of the intent model is: [intent, news search; search content, news in region A today].

[0113] For another example, if the conversation text is: Which team is the current champion of event D, then the output result of the intent model is: [intent, dialogue task].

[0114] The above user intentions include: movie search intention, music search intention, news search intention and conversation intention. In actual application, they may also include calendar search intention, weather search intention and TV control intention.

[0115] S420: When the user intention is to retrieve media resources, drive the large parsing model to parse the user intention to obtain structured data for media resource retrieval, and drive the display to display the structured data on the user interface.

[0116] In some embodiments, the controller can specifically determine whether the user's intention is to search for media resources based on the output result of the intention big model. For example, if the output result of the intention big model is: [Intent, song search; Search content, songs of singer B], or if the output result of the intention big model is: [Intent, video search; Search content, movies of actor A], then the user's intention is a media resource search intention.

[0117] For example, taking the conversation text: I want to watch actor A's movies as an example, the output result of the intention big model is: [Intent, video search; Search content, actor A's movies]. Then, the output result of the intention big model (video search, search content: actor A's movies) is input into the parsing big model, and the output result of the parsing big model can be obtained:

[0118] Intent: Video search;

[0119] Search content: Movies of actor A;

[0120] Search conditions: actor, A; singer, A; category, movie.

[0121] It should be noted that the search condition "singer" appears in the output of the big parsing model because actor A is also a singer. That is, the search condition that is not in the conversation text will be given in the output of the big parsing model to ensure the accuracy of the search. For example, in this example, the specific identity of A is also limited to the identity of a singer, that is, the structured data obtained in this example is: [Intent: video search; Search content: actor A's movies; Search conditions: actor, A; singer, A; category, movie].

[0122] In some embodiments, in order to ensure that the search results are as comprehensive as possible, an OR operation is used for the same type of search conditions. For example, in this example, the search condition for identity is: (actor, A) OR (singer, A), so as to avoid the situation where some movie details only mark A as one of the actor identity or the singer identity.

[0123] In some embodiments, Fig. 9 As shown, when the controller drives the display to display structured data on the user interface, a confirmation control, a cancel control or a modification control may be provided on the user interface. Thus, after the user triggers the controller to control the display to display the confirmation control on the user interface, the controller performs the operation of S430. After the user triggers the controller to control the display to display the cancel control on the user interface, the structured data is canceled, or the task returns to the operation of S410.

[0124] Alternatively, after the user triggers the controller to control the display to display a modification control on the user interface, the controller can open the editing authority of the structured data. On this basis, the user can freely edit the structured data.

[0125] For example, the structured data is: [intent, video search; search content, actor A's movies; search conditions: actor, A; singer, A; category, movie], which is edited and modified by the user to: [intent, video search; search content, actor A's movies; search conditions: actor, A; singer, A; category, movie; release date, 2023]. In this way, a new search condition is added by modification. Alternatively, it is edited and modified by the user to: [intent, video search; search content, actor A's movies; search conditions: actor, A; singer, A; category, action movie], so that the user's search needs can be adjusted by modifying the specific content of the existing search conditions.

[0126] S430, searching the media resource database for media resource data corresponding to the structured data, and driving the display to display the media resource data on the user interface.

[0127] In some embodiments, the retrieval of media asset data in the media asset database is implemented by the corresponding business microservice module. Specifically, the controller inputs the structured data into the business microservice module, and the business microservice module selects the corresponding media asset data in the media asset database according to the structured data. After selecting the corresponding media asset data, the controller drives the display to display the media asset data on the user interface.

[0128] In some embodiments, the business microservice module can output a search formula based on the output of media data, so that users can understand the search logic used by the retrieved media data, and then enable users to further understand whether the retrieved media data meets their own needs. For example: the structured data is: [intent, video search; search content, actor A's movies; search conditions: actor, A; singer, A; category, movie], then the output search formula is query = ((actor:\"A\")OR(singer:\"A\"))AND((category:\"movie\")). In layman's terms, the search results are movies related to actor A, or movies related to singer A, where actor is an actor, singer is a singer, and category is a category. In some embodiments, the controller can also drive the display to display the above search formula on the user interface.

[0129] The display device provided by the present application utilizes the intention big model to identify the user intention of the conversation text, so that when the conversation text input by the user is complex content, the user intention can be accurately obtained, and when the user intention is to retrieve media resources, the structured data of the user's media resource retrieval is obtained based on the parsing big model, thereby ensuring that the media resources retrieved using the structured data can stably and reliably respond to the conversation text input by the user.

[0130] In some implementations, before retrieving the media asset data corresponding to the structured data in the media asset database, the controller is further configured to: determine the media asset database corresponding to the structured data in response to the input channel of the conversation text.

[0131] The input channel refers to the data source of the controller session text.

[0132] Exemplarily, taking the display device as a smart TV as an example, the smart TV may include a local database 1. Video application 1 and video application 2 are installed on the smart TV, and video application 1 has a cloud database 1, and video application 2 has a cloud database 2.

[0133] When a user inputs "I want to watch a movie of actor A" from the homepage of the smart TV, the data source of the conversation text is the homepage of the smart TV, and the controller searches for media data from local database 1, cloud database 1, and cloud database 2. When driving the display to display media data on the user interface, the controller can drive the display to display the source mark on the display page of the media data. For example, Fig.10 As shown, the media asset data includes Movie 1, Movie 2, Movie 3, and Movie 4. In the detailed description of Movie 1 and Movie 2, the data source is marked as local, indicating that Movie 1 and Movie 2 are from local database 1. In the detailed description of Movie 3, the data source is marked as video application 1, indicating that Movie 3 is from the cloud database. In the detailed description of Movie 4, the data source is marked as video application 2, indicating that Movie 4 is from cloud database 2.

[0134] When the user inputs from the interface of video application 1: I want to watch the movie of actor A, and the data source of the conversation text is video application 1, the controller will search for media data in cloud database 1. At this time, when the controller drives the display to display the media data on the user interface, the data source of the media data may not be marked, and the media data is assumed to come from cloud database 1.

[0135] In some embodiments, if video application 1 and video application 2 are related applications, and databases are interoperable during use, when the user enters from the interface of video application 1: I want to watch a movie of actor A, the controller can also search for media data in cloud database 1 and cloud database 2. Fig.11 As shown, when the display is driven to display media data on the user interface, the data sources of movies 7 and 8 from the cloud database 2 can be marked, and the media data without marked data sources are assumed to come from the cloud database 1.

[0136] In this way, it is possible to avoid a mismatch between the input channel and the media database, which results in the user being unable to view the retrieved media data, or the situation where the media data required by the user is not fully retrieved. That is, by determining the corresponding media database through the input channel of the conversation text, the retrieved media data can more easily meet the user's expectations.

[0137] In some implementations, if multiple media data are obtained, in the step of driving the display to display the media data on the user interface, the controller is specifically configured to: first, determine the display positions of the multiple media data based on preset rules; and then drive the display to display the multiple media data on the user interface according to the display positions of the multiple media data.

[0138] For example, the preset rule is: the later the media data is released, the higher the display rank is. For another example, the preset rule is: the more times the media data is viewed, the higher the display rank is. For another example, the higher the matching degree between the media data and the conversation text, the higher the display rank is. Among them, the later the release time means that the release time is closer to the retrieval time; the higher the display rank means: the higher the priority display level and the front display position; the matching degree between the media data and the conversation text refers to the degree of correlation between the information of the conversation text and the media data.

[0139] In some embodiments, the user can select specific preset rules. Fig.12 As shown, when the controller drives the display to display multiple media data on the user interface, multiple sorting controls corresponding to preset rules are displayed on the user interface at the same time. After the user triggers the sorting control, the display is driven to adjust the display order of the media data according to the corresponding preset rules.

[0140] In this way, through preset rules, multiple media data are displayed on the user interface according to the display position, so that users can determine the required target among the media data with a higher display position as much as possible, avoiding the user having to view multiple media data before understanding or viewing the required media data, thereby improving the user experience.

[0141] In some implementations, in the step of driving the intention big model to recognize the user intent of the conversation text in response to the conversation text input by the user, the controller is specifically configured to: first construct a first Prompt instruction for intent recognition based on the conversation text; and then input the first Prompt instruction into the intention big model to obtain the user intent.

[0142] Specifically, in order to enable the intention big model to perform tasks better, it is necessary to input a first Prompt instruction with more comprehensive information into the intention big model.

[0143] For example, taking the conversation text: I want to watch a movie by actor A as an example, the first Prompt instruction example is as follows: "You are a user intent analysis AI assistant and need to perform intent analysis on the conversation text. Your output should follow the following format: {"intentInfo":[{"intent":"intent","query":"query"}]}, where intent is the user intent, and its enumeration value is movie_search (movie search), music_search (music search), calendar_search (calendar_search), news_search (news search), weather_search (weather search), or chat (conversation intent); query is the statement corresponding to each user intent. Next, my input is: {{I want to watch a movie by actor A}}, and your response is: "Output result", so that the intent model outputs according to the specified data format.

[0144] In this way, by constructing the first Prompt instruction, the input information of the intention model is made more comprehensive, which in turn helps the intention model to improve the accuracy and standardization of intent recognition based on more comprehensive information.

[0145] In some implementations, in the step of driving the parsing big model to parse the conversation text to obtain structured data for media resource retrieval, the controller is specifically configured to: first construct a second Prompt instruction for media resource retrieval based on the conversation text and the corresponding user intention; and then input the second Prompt instruction into the parsing big model to obtain structured data.

[0146] Specifically, when the output result of the intention big model is used as the input parameter of the parsing big model, in order to enable the parsing big model to better perform the user intention parsing task, it is necessary to input a second Prompt instruction with more comprehensive information into the parsing big model.

[0147] For example, taking the conversation text: I want to watch actor A's movies, the second Prompt instruction example is as follows: "You are a film and television retrieval expert and need to perform film and television information query and analysis on the conversation text. Your output should follow the following format: {"intent":"intent",query":"query","conditionInfos":[{""condition":{"title":["title1","title2"],"tag":["tag1","tag2"]}}]}, where intent is the user's intent; query is the conversation text corresponding to the user's intent; title is the media retrieval condition, and title1 and title2 are the retrieval condition values; tag is the media retrieval condition, and tag1 and tag2 are the retrieval condition values. Next, my input is: {{intent}}, {{query}}, and your response is: "Structured data", and the parsing model outputs according to the specified data format.

[0148] In this way, by constructing the second Prompt instruction, the input information for parsing the large model is made more comprehensive, which in turn helps the large model to obtain more accurate and higher structured data based on more comprehensive information, and makes the structured data more standardized.

[0149] In some implementations, after driving the intention big model to identify the user intention of the conversation text, the controller is also configured to: when the user intention is a conversation intention, drive the parsing big model to generate conversation content corresponding to the conversation text, and drive the display to display the conversation content on the user interface.

[0150] Specifically, taking the second Prompt instruction as an example, the second Prompt instruction example is as follows: "You are a dialogue task expert and need to give the dialogue content for the conversation text. Your output should follow the following format: {"intent":"intent",query":"qu ery","conditionInfos":[{""condition":{"title":["title1","title2"],"tag":["tag1","tag2"]}}],"answer":"answer"}, where intent is the user intent; query is the conversation text corresponding to the user intent; title is the conversation condition, title1 and title2 are the conversation condition values; tag is the answer condition, tag1 and tag2 are the retrieval condition values; answer is the answer given for the conversation text.

[0151] In this way, when this solution is applied to a smart TV, it can not only meet the user's media resource retrieval needs, but also provide corresponding answers to the user's question-like conversation text.

[0152] In some implementations, before driving the intention big model to identify the user intention of the conversation text in response to the conversation text input by the user, the controller is further configured to: obtain the historical conversation text input by the user, and adjust the conversation text based on the historical conversation text.

[0153] For example, a user has a continuous conversation on a smart TV. The first conversation text input by the user is: Please briefly evaluate the movie genres of actor A. After the control area gives an answer based on the intent big model and the parsing big model, the second conversation text input by the user is: I want to watch his movies. At this time, if only based on the conversation text input by the user for the second time, the intent big model can only conclude that the user wants to perform media resource retrieval, specifically retrieval of movies, but cannot determine the specific identity of "him".

[0154] In view of the above problems, in the technical solution provided by this application, the controller obtains the historical conversation text input by the user, which in this example is: Please briefly evaluate the movie types of actor A. Based on: "Please briefly evaluate the movie types of actor A", the controller adjusts the conversation text to obtain: "I want to watch the movie of actor A". In this way, the hidden information of the user's conversation text can be accurately obtained through the process of communicating with the user, ensuring that the intention model can accurately identify the user's intention, thereby improving the user experience.

[0155] In some implementations, before driving the intention big model to identify the user intention of the conversation text, the controller is further configured to: input the conversation text into the intention big model if the content represented by the conversation text is not a control task.

[0156] In actual application, taking the display device as a smart TV as an example, the smart TV will also provide users with some automatic controls based on voice or text. For example, the user inputs voice through the remote control: Please adjust the volume to 30%. At this time, the function the user wants to achieve is to adjust the TV volume, and this control function can be directly implemented by the controller based on voice recognition and control components. In actual application, the control tasks of the smart TV are internally marked, and the controller can easily identify whether the content represented by the conversation text is a control task for the smart TV. In this regard, the present application first determines whether the content represented by the conversation text is a control task. If the content represented by the conversation text is not a control task, the conversation text is input into the intention large model.

[0157] In this way, by first determining whether the content represented by the conversation text is a control task, it is possible to avoid calling the large intention model to identify the user control task, thereby wasting the computing power resources of the controller.

[0158] In some implementations, such as Fig.13 As shown, the controller responds to the conversation text input by the user and drives the intention big model to recognize the user intention of the conversation text. It is also configured to obtain the intention big model by adopting the training method of S1301 to S1303.

[0159] S1301, obtaining historical conversation texts of media resource retrieval and dialogue services, and marking user intentions corresponding to the historical conversation texts.

[0160] First, before obtaining the historical conversation text of the media resource retrieval and dialogue services, for the mainstream services of the existing voice terminal interaction system, an intent description is defined for each service. The intent description is a specified data format and contains user intent information.

[0161] Then, the conversation texts input by mainstream business users in the existing voice terminal interaction system are sorted out, and the user intentions of the conversation texts are annotated. For example, if the conversation text is: I want to watch a movie starring actor A, then the annotated user intention is: [Intent, video search; Search content, movie starring actor A].

[0162] S1302: Construct a first training set using historical conversation texts and user intentions corresponding to the historical conversation texts.

[0163] Specifically, a first training set for voice terminal interaction is constructed, in which the user intent corresponding to the conversation text is annotated in the first data set, that is, the conversation text data set includes user intent and conversation text. For example, if the conversation text is: I want to watch a movie starring actor A, then the first training set includes: [intent, video search; search content, movies starring actor A; I want to watch a movie starring actor A].

[0164] S1303, construct a first initial model, and use the first training set to train the first initial model to obtain a large intention model.

[0165] The first initial model of the present application is a generative large model. The present application uses the first training set to train the first initial model: on the basis of the pre-trained first initial model, by fine-tuning the data of the user intent recognition task, the first initial model can more accurately understand and generate text or dialogue based on user intent. The specific steps include: model pre-training, fine-tuning training, performance evaluation and optimization adjustment.

[0166] Among them, model pre-training: use large-scale text datasets for unsupervised learning to allow the first initial model to learn language structure and semantics.

[0167] Fine-tuning training: Based on the pre-trained first initial model, supervised fine-tuning training is performed using the first training set to enable the first initial model to understand and generate text that meets the specific user's intentions.

[0168] Performance evaluation: The performance of the fine-tuned first initial model on the user intent recognition task is tested through evaluation indicators (such as accuracy, F1 score, and BLEU, etc.). Among them, the F1 score is an indicator used in statistics to measure the accuracy of a binary classification model. The F1 score takes into account both the precision and recall of the classification model and can be regarded as the harmonic mean of the precision and recall. BLEU (Bilingual Evaluation Understudy) is a widely used indicator for automatically evaluating the quality of machine translation. The core idea of ​​BLEU scoring is to calculate the precision and adjust the shorter output results through a penalty term to encourage the completeness and richness of the output results.

[0169] Optimization and adjustment: According to the evaluation results, adjust the fine-tuning parameters, learning rate or data set to further optimize the performance of the first initial model. In the fine-tuning process, you can flexibly choose a variety of methods, such as full parameter fine-tuning (full) and efficient fine-tuning methods LoRA, Freeze or Galore. In view of the continuous iterative progress of large model fine-tuning technology, when choosing a fine-tuning method, you can keep an open attitude and not be subject to specific restrictions. You only need to combine your own hardware conditions with the currently feasible efficient fine-tuning method to make the best choice.

[0170] In this way, the first initial model is trained through the first training set to obtain the intention large model. On this basis, the first training set can also be used to enhance the user intention analysis capability of the intention large model.

[0171] In some implementations, such as Fig.14 As shown, before the controller drives the large parsing model to parse the user's intention, it is also configured to adopt the training method of S1401 to S1404 to obtain the adopted large parsing model.

[0172] S1401, obtaining historical conversation texts retrieved from media resources, and driving the intention big model to identify user intentions corresponding to the historical conversation texts retrieved from media resources.

[0173] S1402, marking structured data corresponding to the historical conversation text retrieved by the media resources.

[0174] In actual application, the intention big model can also be used to adjust the data of the first training set. For example, the user intention in the first training set can be adjusted according to the output of the intention big model to obtain the third data set, thereby avoiding the need to repeatedly obtain historical conversation texts. On this basis, the structured data of the conversation text is annotated in the third training set.

[0175] S1403: Construct a second training set using the user intention and structured data corresponding to the historical conversation text.

[0176] After the structured data is annotated using the third training set, the annotated structured data is added to the third training set to obtain the second training set. The second training set includes three parts: user intention, conversation text, and structured data for media asset retrieval.

[0177] S1404, constructing a second initial model, and using the second training set to train the second initial model to obtain a large analytical model.

[0178] Similarly, the second initial model of the present application is a generative large model. The present application uses the second training set to train the second initial model: on the basis of the pre-trained first initial model, by fine-tuning the data of the user intent parsing task, the second initial model can more accurately understand and generate text or dialogue based on structured data. The specific steps include: model pre-training, fine-tuning training, performance evaluation and optimization adjustment.

[0179] Among them, model pre-training: use large-scale text datasets for unsupervised learning to allow the second initial model to learn language structure and semantics.

[0180] Fine-tuning training: Based on the pre-trained model, supervised fine-tuning training is performed using the second training set to enable the second initial model to understand and generate specific structured data.

[0181] Performance evaluation: The performance of the fine-tuned second initial model on the user or intent parsing task is tested through evaluation indicators (such as accuracy, F1 score, and BLEU, etc.). Among them, the F1 score is an indicator used in statistics to measure the accuracy of a binary classification model. The F1 score takes into account both the precision and recall of the classification model and can be regarded as the harmonic mean of the precision and recall. BLEU (Bilingual Evaluation Understudy) is a widely used indicator for automatically evaluating the quality of machine translation. The core idea of ​​BLEU scoring is to calculate the precision and adjust the shorter output results through a penalty term to encourage the completeness and richness of the output results.

[0182] Optimization and adjustment: According to the evaluation results, adjust the fine-tuning parameters, learning rate or data set to further optimize the performance of the second initial model. In the fine-tuning process, you can flexibly choose a variety of methods, such as full parameter fine-tuning (full) and efficient fine-tuning methods LoRA, Freeze or Galore. In view of the continuous iterative progress of large model fine-tuning technology, when choosing a fine-tuning method, you can keep an open attitude and not be subject to specific restrictions. You only need to combine your own hardware conditions with the currently feasible efficient fine-tuning method to make the best choice.

[0183] In this way, the second initial model is trained by the second training set to obtain the analytical large model. On this basis, the second training set can also be used to enhance the analytical ability of the analytical large model.

[0184] In order to more clearly illustrate the technical solution of this application, see Fig.15 , the following introduces the technical solution of the present application from the perspective of a terminal interaction system timing diagram.

[0185] S1501, the user inputs the conversation text through the central control.

[0186] Among them, the central control is a central control system, and at least one controller is integrated in the central control to execute the user's control tasks. When multiple controllers are integrated, the controller of this application is one of the multiple controllers.

[0187] The input method of the conversation text can be voice input or text input. For example, the user triggers the voice input control of the display device, and then the controller collects the user's voice through the sound collector and converts the user's voice into conversation text using the voice recognition model.

[0188] S1502, the central control calls the intention big model to recognize the conversation text.

[0189] S1503, the intention big model returns the user intention corresponding to the conversation text to the central control.

[0190] In actual application, the central control can drive the display to display the user intention on the user interface so that the user can confirm whether the user intention is accurate. It is also possible not to provide feedback to the user and directly execute S1504 to speed up the data processing process. In particular, when the user uses language to input the conversation text, the controller drives the display to display the conversation text on the user interface, which can reduce the risk of incorrect recognition of the user intention.

[0191] In some embodiments, the central control can drive the display to display the user intention on the user interface, and even if the user has not confirmed whether the user intention is accurate, synchronously execute S1504, so that the processing flow can be accelerated as much as possible. If the user finds that the user intention is inaccurate, modifies the conversation text or cancels the task, the controller responds to the empty finger instruction to modify the conversation text or cancel the task, and stops driving the large model to parse the user intention.

[0192] S1504, the central control calls the large parsing model to parse the user's intention.

[0193] S1505, parsing the large model and returning structured data for media resource retrieval to the central control.

[0194] In some embodiments, the large parsing model is not limited to the case where the user's intention is to retrieve media resources. The large parsing model that can complete other tasks can also be trained according to the needs of the display device. The technical solution of this application is mainly used on display devices, especially smart TVs, so that the media resource retrieval capability of smart TVs can be improved.

[0195] S1506, the central control sends structured data for media resource retrieval to the business microservice module.

[0196] S1507, the business microservice module returns the media resource data to the central control big model.

[0197] In some embodiments, the business microservice module can be integrated into the display device. It can also be deployed in other devices. In the actual application process, the central control sends structured data for media resource retrieval to other devices, and the business module deployed in other devices sends the media resource data to the central control after retrieving the media resource data.

[0198] S1508, the central control displays the media data to the user through a display.

[0199] In this way, the intention big model is used to identify the user intention of the conversation text, so that when the conversation text input by the user is complex, the user intention can be accurately obtained, and when the user intention is to retrieve media resources, the structured data of the user's media retrieval is obtained based on the parsing big model, thereby ensuring that the media resources retrieved using structured data can stably and reliably respond to the conversation text input by the user.

[0200] An embodiment of the present application also provides a display method, which can be applied to the display device in the aforementioned embodiment, the display device includes a display that can display a user interface, and the display method may include: in response to a conversation text input by a user, driving the intention big model to identify the user intention of the conversation text; when the user intention is a media retrieval intention, driving the parsing big model to parse the conversation text to obtain structured data for media retrieval, and driving the display to display the structured data on the user interface; retrieving media data corresponding to the structured data in a media database, and driving the display to display the media data on the user interface.

[0201] The display method provided in the present application utilizes the intention big model to identify the user intention of the conversation text, so that when the conversation text input by the user is complex content, the user intention can be accurately obtained, and when the user intention is to retrieve media resources, the structured data of the user's media resource retrieval is obtained based on the parsing big model, thereby ensuring that the media resources retrieved using the structured data can stably and reliably respond to the conversation text input by the user.

[0202] The display device and display method provided by the embodiments of the present application include a display and a controller, wherein the display device includes a display and a controller, wherein the controller is coupled to the display, and the controller is configured to: in response to a conversation text input by a user, drive the intention big model to identify the user intention of the conversation text; when the user intention is a media retrieval intention, drive the parsing big model to parse the user intention to obtain structured data for media retrieval, and drive the display to display the structured data on a user interface; retrieve the media data corresponding to the structured data from a media database, and drive the display to display the media data on the user interface.

[0203] By applying the technical solution provided in the present application, the intention big model is used to identify the user intention of the conversation text, so that when the conversation text input by the user is complex, the user intention can be accurately obtained, and when the user intention is to retrieve media resources, the structured data of the user's media retrieval is obtained based on the parsing big model, thereby ensuring that the media resources retrieved using structured data can stably and reliably respond to the conversation text input by the user.

Claims

1. A display device, characterized in that: include: a display configured to display a user interface; A controller is coupled to the display and is configured to: In response to a conversation text input by a user, driving the intention big model to identify the user intention of the conversation text; In the case where the user intention is a media resource retrieval intention, driving the large parsing model to parse the user intention to obtain structured data for media resource retrieval, and driving the display to display the structured data on the user interface; The media asset data corresponding to the structured data is retrieved from a media asset database, and the display is driven to display the media asset data on a user interface.

2. The display device according to claim 1, characterized in that Before retrieving the media asset data corresponding to the structured data in the media asset database, the controller is further configured to: In response to an input channel of the conversation text, a media asset database corresponding to the structured data is determined.

3. The display device according to claim 1, characterized in that If the acquired media resource data is multiple, in the step of driving the display to display the media resource data on the user interface, the controller is specifically configured as follows: Based on a preset rule, determining a display order of the plurality of media asset data; According to the display order of the plurality of media asset data, the display is driven to display the plurality of media asset data on the user interface.

4. The display device according to claim 1, characterized in that In the step of driving the intention model to identify the user intention of the conversation text in response to the conversation text input by the user, the controller is specifically configured to: Constructing a first Prompt instruction for intent recognition according to the conversation text; The first Prompt instruction is input into the intention big model to obtain the user intention.

5. The display device according to claim 1, characterized in that In the step of driving the large parsing model to parse the conversation text to obtain structured data for media resource retrieval, the controller is specifically configured to: Constructing a second prompt instruction for media resource retrieval according to the conversation text and the corresponding user intention; The second Prompt instruction is input into the parsing large model to obtain the structured data.

6. The display device according to claim 1, characterized in that After the driving intention big model identifies the user intention of the conversation text, the controller is further configured to: In the case where the user intention is a dialogue intention, the analysis model is driven to generate dialogue content corresponding to the conversation text, and the display is driven to display the dialogue content on the user interface.

7. The display device according to claim 1, characterized in that Before driving the intention model to identify the user intention of the conversation text in response to the conversation text input by the user, the controller is further configured to: A historical conversation text input by a user is obtained, and based on the historical conversation text, the conversation text is adjusted.

8. The display device according to claim 1, characterized in that Before driving the intention big model to identify the user intention of the conversation text, the controller is further configured to: In a case where the content represented by the conversation text is not a control task, the conversation text is input into the intention big model.

9. The display device according to claim 1, characterized in that: The controller is further configured to: Obtaining historical conversation texts of media resource retrieval and dialogue services, and marking user intentions corresponding to the historical conversation texts; Constructing a first training set using the historical conversation texts and the user intentions corresponding to the historical conversation texts; A first initial model is constructed, and the first training set is used to train the first initial model to obtain the intention large model.

10. The display device according to claim 9, characterized in that Before the controller drives the large parsing model to parse the user intention, the controller is further configured to: Acquire historical conversation texts retrieved by media resources, and drive the intention big model to identify user intentions corresponding to the historical conversation texts retrieved by media resources; Annotating the corresponding structured data of the media resource retrieval; Using the user intention and structured data corresponding to the historical conversation text, construct a second training set; A second initial model is constructed, and the second initial model is trained using the second training set to obtain the analytical large model.

11. A display method, characterized in that: Applied to a display device, the display device includes a display capable of displaying a user interface, and the display method includes: In response to a conversation text input by a user, driving the intention big model to identify the user intention of the conversation text; In the case where the user intention is a media resource retrieval intention, driving the parsing large model to parse the conversation text to obtain structured data for media resource retrieval, and driving the display to display the structured data on the user interface; The media asset data corresponding to the structured data is retrieved from a media asset database, and the display is driven to display the media asset data on a user interface.

Images