Terminal interface generation method and device based on AI and storage medium

Abstract

The invention provides an AI-based terminal interface generation method and device and a storage medium, relates to the technical field of artificial intelligence, and can improve the man-machine interaction experience of a user. The method comprises the following steps: acquiring multi-source heterogeneous data of a target user collected by a terminal; the multi-source heterogeneous data comprises a current behavior executed by the target user on the terminal; based on the multi-source heterogeneous data, determining a portrait of the target user; predicting the next behavior of the target user based on the intention library of the target user, the current behavior, the portrait of the target user and a user behavior prediction model; generating a target user interface of the terminal based on the next behavior and an interface generation model; the target user interface is used for providing service for the next behavior of the target user.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, and in particular to an AI-based terminal interface generation method and device and storage medium. BACKGROUND

[0002] With the rapid development of the field of artificial intelligence, the application of human-computer interaction is becoming more and more widespread. Among them, human-computer interaction plays a very important role in intelligent driving, financial risk control, industrial automation, intelligent customer service, personalized content recommendation, and medical image diagnosis in many fields. Therefore, a reasonable terminal interface generation method is very important for users to provide efficient human-computer interaction.

[0003] At present, the traditional AI-based terminal interface mainly dynamically adjusts the terminal interface by learning user behavior and user preferences, and interacts with the user by responding to user clicks. Because the traditional AI-based terminal interface has a cumbersome interaction process, the user often needs to complete multiple steps or operations to achieve the operation purpose, in addition, the traditional AI-based terminal interface often provides inaccurate or even incorrect interaction interfaces for the user due to the simple understanding of the user's intention, resulting in the user's inability to achieve the operation purpose, therefore, the traditional AI-based terminal interface has the problem of poor user human-computer interaction experience. SUMMARY

[0004] The present application provides an AI-based terminal interface generation method, device and storage medium, which can improve the user's human-computer interaction experience.

[0005] To achieve the above purpose, the present application adopts the following technical solutions: In a first aspect, the present application provides an AI-based terminal interface generation method, which comprises: acquiring multi-source heterogeneous data of a target user collected from a terminal; the multi-source heterogeneous data includes the current behavior of the target user executing on the terminal; determining the portrait of the target user based on the multi-source heterogeneous data; determining the intention library of the target user based on the portrait of the target user, the multi-source heterogeneous data, and the intention analysis model; the intention library of the target user is used to indicate the target interaction intention; predicting the next behavior of the target user based on the intention library of the target user, the current behavior, the portrait of the target user, and the user behavior prediction model; generating the target user interface of the terminal based on the next behavior and the interface generation model; the target user interface is used to provide services for the next behavior of the target user.

[0006] The technical scheme has at least the following beneficial effects: The application obtains multi-source heterogeneous data of a user of a terminal, constructs a portrait of the target user, and predicts a next behavior of the target user based on the portrait of the target user and behavior information of the target user. The portrait of the user, the behavior information of the target user, the user intent library, the intent analysis model, and the user behavior prediction model can accurately predict the next behavior of the user, and a user interface with high matching degree to the next behavior of the target user is generated. Therefore, the application can provide an accurate user interface for the user by analyzing the user intent and combining the multi-modal interface generation model, and the accuracy of the interaction between the user and the user interface is improved. In addition, the user interface generated by the user interface generation model and the predicted next behavior of the target user has good interaction logic, which improves the efficiency and accuracy of the interaction between the user and the user interface, and further improves the human-computer interaction experience of the user.

[0007] In a possible implementation, the portrait of the target user includes a short-term portrait of the target user and a long-term portrait of the target user, and the portrait of the target user is constructed based on the multi-source heterogeneous data, including: performing feature recognition on the multi-source heterogeneous data to determine a behavior feature of the target user and a preference feature of the target user; inputting the behavior feature of the target user into a feature vector generation model to determine a behavior feature vector of the target user, and inputting the preference feature of the target user into the feature vector generation model to determine a preference feature vector of the target user; constructing the short-term portrait of the target user based on the behavior feature vector of the target user, and constructing the long-term portrait of the target user based on the preference feature vector of the target user.

[0008] In a possible implementation, the next behavior of the target user is predicted based on the intent library of the target user, the current behavior, the portrait of the target user, and the user behavior prediction model, including: determining a similar behavior feature of the target user based on the portrait of the target user and a user portrait library; the user portrait library includes portraits of multiple users and behavior features of the multiple users; determining a feature vector set of the target user based on the portrait of the target user, the similar behavior feature of the target user, the intent library of the target user, and the current behavior; and predicting the next behavior of the target user based on the feature vector set and the user behavior prediction model.

[0009] In a possible implementation, the similar behavior feature of the target user is determined based on the portrait of the target user and the user portrait library, including: determining a behavior mode of the target user based on the portrait of the target user; determining behavior modes of multiple users based on the user portrait library; in a case where a similarity between a behavior mode of a first user and the behavior mode of the target user is greater than a first threshold value, determining the first user as a similar user, the first user being one of the multiple users; and determining the similar behavior feature of the target user based on behavior features of the similar users in the user portrait library.

[0010] In a possible implementation, the determining of the feature vector set of the target user based on the portrait of the target user, the similar behavior feature of the target user, the intention library of the target user, and the current behavior includes: determining the interaction intention of the target user based on the current behavior and the intention analysis model; determining the interaction intention vector of the target user, the portrait vector of the target user, and the similar behavior feature vector of the target user based on the interaction intention of the target user, the portrait of the target user, the similar behavior feature of the target user, and the feature vector generation model; and determining the feature vector set of the target user based on the interaction intention vector of the target user, the portrait vector of the target user, and the similar behavior feature vector of the target user.

[0011] In a possible implementation, the predicting of the next behavior of the target user based on the feature vector set and the user behavior prediction model includes: determining a plurality of candidate behaviors and a prediction probability of each candidate behavior in the plurality of candidate behaviors based on the feature vector set and the user behavior prediction model; and determining a candidate behavior with the highest prediction probability in the plurality of candidate behaviors as the next behavior.

[0012] In a possible implementation, the interface generation model includes an image generation model and a layout generation model; and the generating of the target user interface of the terminal based on the next behavior and the interface generation model includes: inputting the next behavior into the image generation model to determine a candidate user interface of the terminal; and adjusting an interaction component of the candidate user interface and an interface constraint of the candidate user interface based on the next behavior and the layout generation model to determine the target user interface of the terminal.

[0013] In a second aspect, the present application provides an AI-based terminal interface generation device, which includes: a communication unit and a processing unit; the communication unit is configured to acquire multi-source heterogeneous data of a target user collected from a terminal; the multi-source heterogeneous data includes a current behavior of the target user performed on the terminal; the processing unit is configured to determine a portrait of the target user based on the multi-source heterogeneous data; the processing unit is further configured to determine an intention library of the target user based on the portrait of the target user, the multi-source heterogeneous data, and an intention analysis model; the intention library of the target user is used to indicate an interaction intention; the processing unit is further configured to predict a next behavior of the target user based on the intention library of the target user, the current behavior, the portrait of the target user, and a user behavior prediction model; and the processing unit is further configured to generate a target user interface of the terminal based on the next behavior and an interface generation model; the target user interface is used to provide a service for the next behavior of the target user.

[0014] In a possible implementation, the processing unit is specifically configured to: perform feature recognition on the multi-source heterogeneous data, and determine a behavior feature of the target user and a preference feature of the target user; input the behavior feature of the target user into a feature vector generation model, determine a behavior feature vector of the target user, and input the preference feature of the target user into the feature vector generation model, determine a preference feature vector of the target user; construct a short-term portrait of the target user based on the behavior feature vector of the target user, and construct a long-term portrait of the target user based on the preference feature vector of the target user.

[0015] In a possible implementation, the processing unit is specifically configured to: determine a similar behavior feature of the target user based on the portrait of the target user and a user portrait library; the user portrait library includes portraits of multiple users and behavior features of the multiple users; determine a feature vector set of the target user based on the portrait of the target user, the similar behavior feature of the target user, an intention library of the target user, and a current behavior; and predict a next behavior of the target user based on the feature vector set and a user behavior prediction model.

[0016] In a possible implementation, the processing unit is specifically configured to: determine a behavior mode of the target user based on the portrait of the target user; determine behavior modes of multiple users based on a user portrait library; in a case where a similarity between a behavior mode of a first user and the behavior mode of the target user is greater than a first threshold value, determine the first user as a similar user, the first user being one of the multiple users; and determine a similar behavior feature of the target user based on behavior features of the similar users in the user portrait library.

[0017] In a possible implementation, the processing unit is specifically configured to: determine an interaction intention vector of the target user, a portrait vector of the target user, and a similar behavior feature vector of the target user based on an interaction intention library of the target user, the portrait of the target user, the similar behavior feature of the target user, and a feature vector generation model; and determine a feature vector set of the target user based on the interaction intention vector of the target user, the portrait vector of the target user, and the similar behavior feature vector of the target user.

[0018] In a possible implementation, the processing unit is specifically configured to: determine multiple candidate behaviors and a prediction probability of each candidate behavior in the multiple candidate behaviors based on the feature vector set and a user behavior prediction model; and determine a candidate behavior with the highest prediction probability in the multiple candidate behaviors as the next behavior.

[0019] In a possible implementation, the processing unit is specifically configured to: input the next behavior into an image generation model, and determine a candidate user interface of the terminal; and adjust an interaction component of the candidate user interface and an interface constraint of the candidate user interface based on the next behavior and a layout generation model, and determine a target user interface of the terminal.

[0020] In a third aspect, the present application provides an AI-based terminal interface generation apparatus, which comprises a processor and a communication interface; the communication interface is coupled to the processor, and the processor is configured to run computer programs or instructions to implement the AI-based terminal interface generation method as described in the first aspect and any possible implementation manner of the first aspect.

[0021] In a fourth aspect, the present application provides a computer readable storage medium, which stores instructions, and when the instructions are run on a terminal, the terminal performs the AI-based terminal interface generation method as described in the first aspect and any possible implementation manner of the first aspect.

[0022] In a fifth aspect, the present application provides a computer program product comprising instructions, and when the computer program product is run on an AI-based terminal interface generation apparatus, the AI-based terminal interface generation apparatus performs the AI-based terminal interface generation method as described in the first aspect and any possible implementation manner of the first aspect.

[0023] In a sixth aspect, the present application provides a chip, which comprises a processor and a communication interface; the communication interface is coupled to the processor, and the processor is configured to run computer programs or instructions to implement the AI-based terminal interface generation method as described in the first aspect and any possible implementation manner of the first aspect.

[0024] Specifically, the chip provided in the present application further comprises a memory for storing computer programs or instructions. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 A structural schematic diagram of an AI-based terminal interface generation system provided for an embodiment of the present application; Figure 2 A component schematic diagram of an AI-based terminal interface generation apparatus provided for an embodiment of the present application; Figure 3 A flowchart of an AI-based terminal interface generation method provided for an embodiment of the present application; Figure 4 A method flowchart for predicting a next behavior of a target user provided for an embodiment of the present application; Figure 5 A data interaction flowchart of an AI-based terminal interface generation method provided for an embodiment of the present application; Figure 6 A structural schematic diagram of an AI-based terminal interface generation apparatus provided for an embodiment of the present application. DETAILED DESCRIPTION

[0026] The AI-based terminal interface generation method and device provided by the embodiments of the present application and the storage medium are described in detail below with reference to the accompanying drawings.

[0027] The term "and / or" in this document merely describes an association relationship of associated objects, and indicates that three relationships can exist, for example, A and / or B can represent three cases of existence of A alone, existence of A and B together, and existence of B alone.

[0028] The terms "first" and "second" and the like in the description of the present application and the accompanying drawings are used to distinguish different objects or different processing of the same object, and are not used to describe a specific order of the objects.

[0029] In addition, the terms "include" and "have" and any variations thereof mentioned in the description of the present application are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but can optionally include other steps or units not listed, or can optionally include other steps or units inherent to the process, method, product or device.

[0030] It should be noted that in the embodiments of the present application, the words "exemplary" or "for example" are used to mean serving as an example, instance, or illustration. Any embodiment or design solution described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of "exemplary" or "for example" is intended to present relevant concepts in a specific manner.

[0031] In the description of the present application, unless otherwise specified, the meaning of "a plurality of" is two or more.

[0032] With the rapid development of the field of artificial intelligence technology, the application of human-computer interaction is becoming more and more widespread. Among them, human-computer interaction plays a very important role in intelligent driving, financial risk control, industrial automation, intelligent customer service, personalized content recommendation, and medical image diagnosis in many fields. Therefore, a reasonable AI-based terminal interface generation method is very important to provide efficient human-computer interaction for users.

[0033] Currently, AI is rapidly developing under the influence of technologies such as machine learning, deep learning, natural language processing (NLP), computer vision, and reinforcement learning. AI has the ability to autonomously learn, analyze, and make decisions from massive amounts of data. AI applications have penetrated into many fields such as intelligent driving, medical image diagnosis, financial risk control, industrial automation, intelligent customer service, personalized content recommendation, and smart city management. AI, through its characteristics of automation, intelligence, and data-driven, not only affects traditional industries, but also continuously promotes the transformation of social production and life towards high-efficiency and intelligence.

[0034] With the rapid development of artificial intelligence technology, some technologies propose AI-based user interface technology, which marks the shift of human-computer interaction paradigm from function-oriented to context and emotion fusion-oriented. AI-based user interface technology, through the integration of multi-modal perception, natural language processing, computer vision, and generative artificial intelligence, restructures the interaction logic between users and digital environments, upgrading the experience from passive reception to active dialogue and co-creation.

[0035] Currently, traditional AI-based user interface methods adjust the layout of the user interface or the contrast of the user interface or the speech speed by learning user behavior and preferences in real time. This AI-based user interface method builds a strong cognitive support for users through technologies such as eye tracking, gesture recognition, and intelligent voice assistants, achieving information simplification, intent prediction, and operation guidance. AI-based user interface is not only a product of technology integration, but also a deep response to human cognitive habits and behavior patterns, laying the foundation for the interaction paradigm of future scenarios such as the metaverse and intelligent space, and has profound scientific and social significance.

[0036] As described above about the traditional AI-based user interface method, the traditional AI-based terminal interface mainly dynamically adjusts the terminal interface by learning user behavior and user preferences, and interacts with the user by responding to user clicks. Due to the complex interaction process of the traditional AI-based terminal interface, users often need to complete multiple steps or operations to achieve the operation purpose. In addition, the traditional AI-based terminal interface often provides inaccurate or even incorrect interaction interfaces due to simple understanding of user intent, resulting in users being unable to achieve the operation purpose. Therefore, the traditional AI-based terminal interface has the problem of poor human-computer interaction experience of users.

[0037] In view of this, the application provides an AI-based terminal interface generation method. The method obtains multi-source heterogeneous data of a user of a terminal, constructs a portrait of the target user, and predicts the next behavior of the target user based on the portrait of the target user and behavior information of the target user. The portrait of the user provided by the application can accurately predict the next behavior of the user in combination with the behavior information of the target user and the user intent library, and then generate a user interface with high matching degree for the next behavior of the target user. Therefore, the application can provide a more accurate user interface for the user by analyzing the user intent and combining an artificial intelligence model, thereby improving the accuracy of user and user interface interaction. In addition, the user interface generated by the user interface generation model and the predicted next behavior of the target user has good interaction logic, which improves the efficiency of user and user interface interaction, and thus improves the human-computer interaction experience of the user.

[0038] The technical scheme provided by the embodiments of the application can be applied to various communication systems, for example, a New Radio (NR) communication system using a 5th generation mobile communication technology (5G), a future evolution system, or a multi-communication fusion system.

[0039] Exemplarily, Figure 1 FIG. 1 shows a structure schematic diagram of an AI-based terminal interface generation system provided by an embodiment of the application. The AI-based terminal interface generation system can include at least one terminal 101 and at least one AI intelligent terminal platform 102.

[0040] In a possible implementation, the terminal 101 is configured to collect multi-source heterogeneous data of a target user. The multi-source heterogeneous data can include face recognition data, voice data, text data, sensory contact data, emotional expression data, and video data.

[0041] In a possible implementation, the terminal 101 can be a smart mobile terminal such as a smartphone or a tablet computer, a smart home terminal such as a smart speaker, a smart television, a smart projector, and a smart home robot, a smart vehicle terminal such as a smart cockpit system and a smart driving assistance system, a wearable terminal such as a smart watch, a bracelet, an augmented reality (AR) terminal, and a virtual reality (VR) head-mounted display, a public and commercial terminal such as a smart retail terminal, a service robot, and an interactive screen, a professional terminal such as a medical terminal, an educational terminal, and an industrial terminal. The embodiments of the application do not limit this.

[0042] In a possible implementation, the AI intelligent terminal platform 102 is configured to receive multi-source heterogeneous data from the terminal 101, determine a portrait of a target user based on the multi-source heterogeneous data, determine an intent library of the target user based on the portrait of the target user, the multi-source heterogeneous data, and an intent analysis model, indicate an interaction intent of the target user based on the intent library of the target user, a current behavior, the portrait of the target user, and a user behavior prediction model, predict a next behavior of the target user based on the next behavior and an interface generation model, and generate a target user interface of the terminal 101 based on the next behavior and the interface generation model, to provide a service for the next behavior of the target user.

[0043] In a possible implementation, the terminal 101 can be a device with a wireless transceiving function, and the present application does not make any limitation in this regard. For example, the terminal 101 can be a mobile terminal device such as a mobile phone, a tablet computer, or a VR glasses, and the like.

[0044] In a possible implementation, the AI intelligent terminal platform 102 can be a software and hardware integrated system integrating an artificial intelligence algorithm and a multi-modal interaction capability, and the present application does not make any limitation in this regard. For example, the AI intelligent terminal platform 102 can be a cloud AI intelligent terminal platform, or can also be a local AI intelligent terminal platform.

[0045] It should be noted that, Figure 1 is only an exemplary framework, Figure 1 The number of nodes included in the system architecture and the names of various devices are not limited, and in addition to the functional nodes shown in Figure 1 , the AI-based terminal interface generation system can further include other nodes such as a core network device, and the present application does not make any limitation in this regard.

[0046] The application scenarios of the embodiments of the present application are not limited. The system architecture and business scenarios described in the embodiments of the present application are for more clearly explaining the technical solutions provided by the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. It can be known by those skilled in the art that, with the evolution of network architecture and the appearance of new business scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0047] In specific implementation, Figure 1 The devices in Figure 2 may adopt the component structures shown in Figure 2 , or include the components shown in Figure 2A component diagram of an AI-based terminal interface generation apparatus 20 provided for an embodiment of the present application is shown in FIG. 2. The AI-based terminal interface generation apparatus 20 can be a terminal 101 or a chip or system on chip in the terminal 101. Alternatively, the AI-based terminal interface generation apparatus 20 can be an AI intelligent terminal platform 102 or a chip or system on chip in the AI intelligent terminal platform 102. As shown in FIG. 2, the AI-based terminal interface generation apparatus 20 can include a processor 201, a bus 202, a communication interface 203, and a memory 204. Figure 2

[0048] The processor 201, the memory 204, and the communication interface 203 can be connected through the bus 202.

[0049] The processor 201 can be a central processing unit (CPU), a general processor, a network processor (NP), a digital signal processing (DSP), a microprocessor, a microcontroller, a programmable logic device (PLD), or any combination thereof. The processor 201 can also be other devices with processing functions, such as a circuit, a device, or a software module, without limitation.

[0050] The bus 202 is used to transmit information between the components included in the AI-based terminal interface generation apparatus 20.

[0051] The communication interface 203 is used to communicate with other devices or other communication networks. The other communication networks can be an Ethernet, a radio access network (RAN), a wireless local area network (WLAN), or the like. The communication interface 203 can be a module, a circuit, a communication interface, or any device capable of communication.

[0052] The memory 204 is used to store instructions. The instructions can be a computer program.

[0053] ​The memory 204 can be a read-only memory (ROM) or other type of static storage device that can store static information and / or instructions, a random access memory (RAM) or other type of dynamic storage device that can store information and / or instructions, an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disk storage, a magneto-optical disk, a magnetic disk storage or other magnetic storage devices, etc., without limitation.

[0054] It should be noted that the memory 204 can exist independently of the processor 201 or can be integrated with the processor 201. The memory 204 can be used to store instructions or program codes or some data, etc. The memory 204 can be located in the AI-based terminal interface generation apparatus 20 or outside the AI-based terminal interface generation apparatus 20, without limitation.

[0055] In an example, the processor 201 can include one or more CPUs.

[0056] As an optional implementation, the AI-based terminal interface generation apparatus 20 includes multiple processors.

[0057] As an optional implementation, the AI-based terminal interface generation apparatus 20 can further include an output device and an input device. Exemplarily, the input device is a keyboard, a mouse, a microphone, or a joystick, etc., and the output device is a display screen, a speaker, etc.

[0058] It should be noted that the AI-based terminal interface generation apparatus 20 can be a desktop computer, a laptop computer, a network server, a mobile phone, a tablet computer, a wireless terminal, an embedded device, a chip system, or a device with a similar structure. In addition, Figure 1 the constituent structures shown in the foregoing do not constitute a limitation on each device in the foregoing, and in addition to the components shown in the foregoing, Figure 2 the constituent structures shown in the foregoing do not constitute a limitation on each device in the foregoing, and in addition to the components shown in the foregoing, Figure 1 and Figure 2 each device in the foregoing can include more or fewer components than shown, or combine certain components, or different component arrangements. Figure 2 Figure 1 and Figure 2 In the foregoing, each device can include more or fewer components than shown, or combine certain components, or different component arrangements.

[0059] ​In the embodiments of the present application, the chip system can be composed of a chip, or can include a chip and other discrete devices.

[0060] In addition, the actions, terms, etc. involved among the embodiments of the present application can be mutually referenced and are not limited. The message name or parameter name in the message between the devices in the embodiments of the present application is only an example, and other names can also be used in the specific implementation, which is not limited.

[0061] The AI-based terminal interface generation system shown in the following Figure 2 The AI-based terminal interface generation method provided by the embodiments of the present application is described. Among the embodiments of the present application, the actions, terms, etc. involved among the embodiments of the present application can be mutually referenced and are not limited. The message name or parameter name in the message between the devices in the embodiments of the present application is only an example, and other names can also be used in the specific implementation, which is not limited. The actions involved in the embodiments of the present application are only an example, and other names can also be used in the specific implementation, such as: "including" in the embodiments of the present application can be replaced by "carrying" or "carrying" and the like.

[0062] In order to solve the problems existing in the prior art, such as Figure 3 As shown in the present application, the embodiments of the present application propose an AI-based terminal interface generation method, which can improve the user's human-computer interaction experience. The method comprises: S301, the AI intelligent terminal platform acquires the multi-source heterogeneous data of the target user collected from the terminal.

[0063] Among them, the multi-source heterogeneous data target user's current behavior In a possible implementation, the AI intelligent terminal platform can establish a communication connection with the terminal and receive the multi-source heterogeneous data of the target user collected from the terminal.

[0064] In a possible implementation, the multi-source heterogeneous data can include: multi-modal data, target user interaction data, sensor data, terminal system data.

[0065] In a possible implementation, the multi-modal data can be audio data collected through a microphone, image and audio data collected through a camera, text data, and emotional expression data, and the like. The target user interaction data can be real-time interaction data of the target user (for example, mouse movement, click sequence, scroll speed, dwell time, and text input), behavior pattern of the target user, user attributes (for example, age, occupation, and preference), terminal device information (for example, device type, location data, interaction time, and network status), and component library data. The sensor data can be target user state data collected through a gyroscope, an accelerometer, a light sensor, a global positioning system (GPS), and the like, or surrounding environment data collected through the above sensors. The terminal system data can be device model, hardware configuration (for example, CPU or memory), operating system version, boot-up, shut-down, standby time point and frequency, network connection quality (for example, wireless fidelity (Wi-Fi) signal strength, network bandwidth, delay), and the like.

[0066] In S302, the AI intelligent terminal platform determines a portrait of the target user based on the multi-source heterogeneous data.

[0067] In a possible implementation, the AI intelligent terminal platform can fuse the collected multi-source heterogeneous data through a unified data model, and perform real-time alignment on the data of different sources and different frequencies in a message queue to generate an aligned multi-source data stream.

[0068] In a possible implementation, the unified data model can be an event model, a stream data model, or other models, which are not limited in the embodiments of the application.

[0069] In a possible implementation, the AI intelligent terminal platform can perform preprocessing on the multi-source data stream. The preprocessing process can include processing missing values and removing abnormal values. In addition, the preprocessing process can further include converting data of different sources (for example, application (App) logs, web page logs, and database records) into a unified format and protocol, and dividing continuous user activities into independent sessions according to certain rules (for example, an interval of more than 5 minutes).

[0070] Through preprocessing on the multi-source data stream, the multi-source data stream can be converted into a data stream suitable for model processing, improving the model processing efficiency and further improving the user interface generation efficiency.

[0071] In a possible implementation, the portrait of the target user includes a short-term portrait of the target user and a long-term portrait of the target user, and the process in which the AI intelligent terminal platform constructs the portrait of the target user based on the multi-source heterogeneous data is as follows: the AI intelligent terminal platform performs feature recognition on the multi-source heterogeneous data, and determines the behavior features of the target user and the preference features of the target user. The AI intelligent terminal platform inputs the behavior features of the target user into a feature vector generation model, determines a behavior feature vector of the target user, and inputs the preference features of the target user into the feature vector generation model, determines a preference feature vector of the target user. The AI intelligent terminal platform constructs a short-term portrait of the target user based on the behavior feature vector of the target user, and constructs a long-term portrait of the target user based on the preference feature vector of the target user.

[0072] For example, the AI intelligent terminal platform can classify the features in the multi-source data stream into user features, behavior features, content features, and context features through feature extraction and feature classification.

[0073] In a possible implementation, the user features can include user attribute features (for example, age, gender, and region) and aggregated statistical features of the user (for example, viewing time length in a preset time, film and television type, day and night viewing proportion, and interest).

[0074] In a possible implementation, the AI intelligent terminal platform can further convert the classified features into semantic vectors that are easy for the model to understand.

[0075] For example, the data stream of the shopping session of the target user is classified by features, and the elements in the features are sorted according to time to generate a session sequence. For example, session 1: [terminal home page ID, search box ID, search button ID, product A ID, add to shopping cart ID]. Session 2: [terminal home page ID, promotion activity B ID, buy now ID,...]. The session sequence is analyzed by a feature vector generation model to determine the semantic vector of each element in the session sequence. For example, the semantic vector of the search button ID can be [0.23, -0.45, 0.89,...], and the semantic vector of the buy now ID can be [0.25, -0.41, 0.91,...].

[0076] Further, the exemplary feature vector generation model can be a graph neural network (GNN) model, a recurrent neural network (RNN), a long short-term memory (LSTM), or another model.

[0077] In a possible implementation, the AI intelligent terminal platform can construct a short-term portrait and a long-term portrait of the target user based on the semantic vector. The short-term portrait includes interaction events (e.g., clicks, plays, searches, etc.), contexts of the interaction events (e.g., timestamps, item identifiers), and recent conversation data, etc. The long-term portrait includes preferences of the target user for product categories, product brands, and product prices, etc. In a possible implementation, the AI intelligent terminal platform can generate a user portrait vector and an interest label distribution based on the short-term portrait and the long-term portrait of the target user, and input the user portrait vector and the interest label distribution into a user portrait feature library. The user portrait feature library can include user portraits, interest label distributions, and behavior features of multiple users.

[0078] In a possible implementation, the AI intelligent terminal platform can generate a user portrait vector and an interest label distribution based on the short-term portrait and the long-term portrait of the target user, and input the user portrait vector and the interest label distribution into a user portrait feature library. The user portrait feature library can include user portraits, interest label distributions, and behavior features of multiple users.

[0079] In a possible implementation, the AI intelligent terminal platform can analyze the behavior trajectory of the target user by combining the interaction data of the target user on the terminal with a large language model (LLM), predict the interaction intent of the target user through user interface (UI) elements (e.g., buttons or pictures, etc.) and operation types (clicks, hovers, etc.), and generate a structured user intent library from the user portrait vector and the interest label in the generated user portrait library.

[0080] For example, the target user searches for electronic product information on the terminal interface. The AI intelligent terminal platform generates a structured intent description by receiving the current user behavior sequence, context features, and user portrait and label, which can be as follows: intent_type: such as Play, Search, Compare. slots: such as {content_id: "123", genre: "***", actor: "****"}. confidence: intent confidence.

[0081] context: Relevant contextual information (e.g., UI information).

[0082] personalized_context: Associated user profile tags (such as "triggered_interest:****") provide richer context for downstream tasks.

[0083] {"intent": "", "entities": {"products": ["phone_A", "phone_B"]}} {"intent": "edit_image", "parameters": {"action": "crop", "ratio": "16:9"}}.

[0084] S304, the AI ​​intelligent terminal platform predicts the next behavior of the target user based on the target user's intent library, current behavior, target user profile, and user behavior prediction model.

[0085] In one possible implementation, the process by which the AI-powered intelligent terminal platform predicts the next action of a target user based on the target user's intent database, current behavior, user profile, and user behavior prediction model can be described as follows: The AI-powered intelligent terminal platform determines similar behavioral characteristics of the target user based on the target user's profile and a user profile database. The user profile database includes profiles and behavioral characteristics of multiple users. The AI-powered intelligent terminal platform determines the target user's feature vector set based on the target user's profile, similar behavioral characteristics, and current behavior. The AI-powered intelligent terminal platform predicts the target user's next action based on the feature vector set and the user behavior prediction model.

[0086] In one possible implementation, the process by which the AI ​​intelligent terminal platform determines similar behavioral characteristics of a target user based on the target user's profile and a user profile database is as follows: The AI ​​intelligent terminal platform determines the target user's behavioral patterns based on the target user's profile. The AI ​​intelligent terminal platform determines the behavioral patterns of multiple users based on the user profile database. If the similarity between the behavioral pattern of the first user and the behavioral pattern of the target user is greater than a first threshold, the AI ​​intelligent terminal platform identifies the first user as a similar user, and the first user is one of multiple users. The AI ​​intelligent terminal platform determines the similar behavioral characteristics of the target user based on the behavioral characteristics of similar users in the user profile database.

[0087] Exemplarily, the AI intelligent terminal platform can retrieve a historical user most similar to the behavior mode of the target user from the user portrait feature library through the current behavior of the target user, i.e., a similar user, determine the behavior of the similar user as the user portrait of the target user, and determine the historical behavior of the historical similar user and the subsequent behavior sequence of the historical behavior. In addition, the AI intelligent terminal platform can combine the user's own behavior prompt word, the historical behavior of the historical similar user, and the subsequent behavior sequence of the historical behavior of the historical similar user into a final prompt word library to provide a quick index for subsequent quick retrieval of similar users.

[0088] In a possible implementation, the process in which the AI intelligent terminal platform determines the feature vector set of the target user based on the portrait of the target user, the similar behavior features of the target user, the intent library of the target user, and the current behavior is as follows: the AI intelligent terminal platform determines the interaction intent vector of the target user, the portrait vector of the target user, and the similar behavior feature vector of the target user based on the intent library of the target user, the portrait of the target user, the similar behavior features of the target user, and the feature vector generation model. The AI intelligent terminal platform determines the feature vector set of the target user based on the interaction intent vector of the target user, the portrait vector of the target user, and the similar behavior feature vector of the target user.

[0089] In a possible implementation, the AI intelligent terminal platform can construct the user feature vector set through the target user portrait vector set, the user intent vector set, the user interest vector set, the current behavior vector set, and the historical behavior vector set.

[0090] Exemplarily, the comprehensive user feature vector set can be expressed in the following form: { [user portrait vector set] + [user intent vector set] + [user interest vector set] + [current behavior vector set] + [historical behavior vector set]} or can also be expressed as {V_profile, V_intent, V_interest, V_current, V_history}. Wherein, the AI intelligent terminal platform can increase the vector set in the comprehensive user feature vector set or reduce the vector set in the comprehensive user feature vector set based on the actual business needs.

[0091] In a possible implementation, the process in which the AI intelligent terminal platform predicts the next behavior of the target user based on the feature vector set and the user behavior prediction model can be as follows: the AI intelligent terminal platform determines a plurality of candidate behaviors and a prediction probability of each candidate behavior in the plurality of candidate behaviors based on the feature vector set and the user behavior prediction model, and determines a candidate behavior with the highest prediction probability in the plurality of candidate behaviors as the next behavior.

[0092] In one possible implementation, the AI ​​smart terminal platform can input a set of user feature vectors into a hybrid neural model to determine multiple high-level vectors of user features. These high-level vectors of user features may include: a long-term user state vector, a sequence context vector, and a current user context vector.

[0093] For example, a user's long-term state vector can be expressed as: U_long = MLP_static(Concatenate(V_profile, V_interest)). A sequence context vector can be expressed as: C_seq = Transformer_Encoder(V_history). A user's immediate context vector can be expressed as: C_immediate = MLP_immediate(Concatenate(V_current, V_intent)).

[0094] In one possible implementation, the AI ​​intelligent terminal platform fuses multiple high-level vectors into a single high-level vector, and inputs this single high-level vector into a multilayer perceptron model, i.e., a user behavior prediction model, outputting the model's prediction result. The model's prediction result includes: the predicted candidate behavior, the number of predicted candidate behaviors, and the prediction probability of each candidate behavior. The AI ​​intelligent terminal platform then determines the candidate behavior with the highest prediction probability as the target user's next behavior.

[0095] For example, a single high-level vector can be represented as: `Fused_Vector = Concatenate(U_long, C_seq, C_immediate)`. A multilayer perceptron model can perform a fully connected linear transformation on the single high-level vector. The computational expression for the first layer is: `Hidden = ReLU(Dense_1(Fused_Vector))`. The computational expression for the second layer is: `Hidden = ReLU(Dense_2(Hidden))`. The expression for the number of predicted candidate actions is: `Output = Softmax(Dense_final(Hidden))`. Here, `Softmax` is the activation function.

[0096] For example, the prediction result for the next action could be: {Predicted action: "electronic product", product type: "smartphone", model: "a certain brand", color: "black", probability: 0.87}.

[0097] The S305 AI smart terminal platform generates the target user interface for the terminal based on the next behavior and interface generation model.

[0098] The target user interface is used to provide services for the next behavior of the target user.

[0099] In a possible implementation, the AI intelligent terminal platform can generate the target user interface of the terminal based on the next behavior and the interface generation model. The process can be as follows: the AI intelligent terminal platform inputs the next behavior into the image generation model to determine the candidate user interface of the terminal, and adjusts the interaction components of the candidate user interface and the interface constraints of the candidate user interface based on the next behavior and the layout generation model to determine the target user interface of the terminal.

[0100] In a possible implementation, the AI intelligent terminal platform can construct a training data set based on the next behavior and the user interface of the existing terminal. The training data set can include: an image data set, a text data set, a voice data set, a multi-modal data set, a context data set, and a conditional label data set.

[0101] In a possible implementation, the image data set includes: a vector label of a screenshot of the user interface of the current terminal, a vector label of a background image of the user interface of the current terminal, and a vector label of an icon of the user interface of the current terminal, etc. The text data set includes: a vector label of a natural language prompt word describing an image. The voice data set includes: a vector label of a natural language prompt word describing a voice. The multi-modal data set includes: a vector label of data describing emotional expression and sensory contact, etc. The context data set includes: a vector label of interaction time and interaction type. The conditional label data set includes: a structured label related to user behavior prediction.

[0102] In a possible implementation, the AI intelligent terminal platform can convert the predicted next behavior into a next behavior prediction result that is easy for the user interface generation model to understand.

[0103] For example, the predicted next behavior can be: {“predicted behavior”: “listen to relaxing music”, “emotional state”: “calm”, “music type”: “environmental music”, “time”: “night”}. The converted next behavior prediction result can include: a text prompt: “quiet, abstract, deep blue, flowing, background image, minimalism, no text”. A conditional prompt: [PRED: play_environmental_music] [EMOTION: calm] [TIME: night]; In one possible implementation, the AI-powered smart terminal platform can decompose the next behavior prediction result into prediction result prompts. These prompts are then input into a user interface generation model to generate an image of the terminal's user interface. Prediction result prompts could be: Predicted behavior = Play ambient music AND Emotion = Calm THEN; Generated prompts = "tranquil, abstract, dark blue, fluid, background image, minimalist".

[0104] In one possible implementation, the AI-powered smart terminal platform can also adjust the component layout of the terminal's user interface based on the predicted next action and the training dataset. The adjustment process includes: generating new components, adjusting component layout, adjusting component size, and adjusting component hierarchy.

[0105] For example, the AI-powered smart terminal platform determines prediction labels and interface constraints based on the predicted next behavior and the training dataset. For instance: {“Predicted Behavior”: “Explore News”, “Content Priority”: “High”, “Operation Intent”: “Quick Browsing”, “Screen Size”: “1080x2340”}. Based on the predicted labels and interface constraints, the AI-powered smart terminal platform adjusts the component layout of the terminal's user interface.

[0106] In one possible implementation, the AI ​​smart terminal platform can combine the component layout of the terminal's user interface with the image of the terminal's user interface to generate the terminal's user interface.

[0107] For example, the AI-powered smart terminal platform fuses predicted labels with image features, and then allows the model to generate layout descriptions or directly predict the bounding boxes of components, such as: Container(vertical) [Text(title, “NowPlaying”), Image(cover, size=large), Button(play, center))]. The AI-powered smart terminal platform inputs the layout descriptions or directly predicted bounding boxes of components into the terminal's user interface to generate the terminal's user interface.

[0108] In one possible implementation, the AI-powered smart terminal platform can also render images and layout content. After rendering, the AI-powered smart terminal platform can initiate network requests in advance to cache data and render the entire component or page in the background. Dynamic interface adjustments can be made, moving high-probability UI elements (e.g., buttons, menu items) to the most prominent positions, highlighting predicted buttons with pulsating animations to gently guide the user. Furthermore, a list of predicted content can be displayed in advance before the user searches for information.

[0109] In one possible implementation, the AI intelligent terminal platform can present the rendered user interface of the terminal to the multimedia terminal user UI interface, providing interaction for the user.

[0110] As shown in Figure 4 Fig. 1 is a flowchart of a method for predicting the next behavior of a target user provided by an embodiment of the present application. The AI intelligent terminal platform pre-processes multi-source heterogeneous data. The AI intelligent terminal platform formats and extracts features from the pre-processed multi-source heterogeneous data, and determines a feature value vector of the target user. The AI intelligent terminal platform constructs a portrait of the target user based on the feature vector of the target user. The AI intelligent terminal platform determines similar users based on the current user behavior and the user portrait of the target user. The AI intelligent terminal platform determines prompt words based on the portraits of the similar users and the target user. The AI intelligent terminal platform determines the intention of the target user based on the behavior of the target user, and constructs a user intention library based on the intention of the target user. The AI intelligent terminal platform predicts the next behavior of the user based on the user intention library and a prediction large model.

[0111] As shown in Figure 5 Fig. 2 is a data interaction flowchart of an AI-based terminal interface generation method provided by an embodiment of the present application. The AI intelligent terminal platform obtains a multimedia data stream. The AI intelligent terminal platform generates a user interface of the AI intelligent terminal based on the multimedia data stream, and sends the user interface to the AI intelligent terminal. The sources of the multimedia data stream include face recognition, sound capture, text corpus, sensory contact, emotional expression, and video data. The AI intelligent terminal platform includes an application layer, an algorithm model layer, and a data perception layer. The application layer is used for user interface generation, interface detection and optimization, user interface personalized adaptation, emotional interaction support, multi-modal interaction, and adaptive learning. The algorithm model layer is used for data stream pre-processing, feature value extraction and quantization, fine-tuning of a large model, and user perception and prediction. The data perception layer is used to store the following data: multi-modal data, user interaction data, sensor data, terminal system and performance data.

[0112] To solve the problems in the prior art, an AI-based terminal interface generation method is provided in an embodiment of the present application. The method obtains multi-source heterogeneous data of a user of a terminal, constructs a portrait of a target user, and predicts the next behavior of the target user based on the portrait of the target user and behavior information of the target user. The portrait of the user provided by the present application, combined with the behavior information of the target user, can accurately predict the next behavior of the user, and then generate a user interface that has a high matching degree with the next behavior of the target user. In addition, the user interface generated by the user interface generation model and the predicted next behavior of the target user has good interaction logic, thereby improving the human-computer interaction experience of the user.

[0113] It is understood that the aforementioned AI-based terminal interface generation method can be implemented by an AI-based terminal interface generation device. To achieve the above functions, the AI-based terminal interface generation device includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, based on the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein, the embodiments disclosed in this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in a hardware or software-driven manner depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the embodiments disclosed in this application.

[0114] The embodiments disclosed in this application can divide the functional modules according to the AI-based terminal interface generation device generated by the above method examples. For example, each function can be divided into its own functional module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module. It should be noted that the module division in the embodiments disclosed in this application is illustrative and is only a logical functional division. In actual implementation, there may be other division methods.

[0115] Figure 6 This is a schematic diagram of an AI-based terminal interface generation device provided in an embodiment of the present invention. Figure 6 As shown, the AI-based terminal interface generation device 60 can be used to perform... Figure 3 The illustrated method is an AI-based terminal interface generation method. The AI-based terminal interface generation apparatus 60 includes a communication unit 601 and a processing unit 602.

[0116] Communication unit 601 is used to acquire multi-source heterogeneous data of the target user collected from the terminal; the multi-source heterogeneous data includes the current behavior of the target user performed on the terminal; processing unit 602 is used to determine the profile of the target user based on the multi-source heterogeneous data; processing unit 602 is also used to determine the intent library of the target user based on the target user profile, the multi-source heterogeneous data, and the intent parsing model; the intent library of the target user is used to indicate the interaction intent of the target; processing unit 602 is also used to predict the next behavior of the target user based on the intent library of the target user, the current behavior, the profile of the target user, and the user behavior prediction model; processing unit 602 is also used to generate the target user interface of the terminal based on the next behavior and the interface generation model; the target user interface is used to provide services for the next behavior of the target user.

[0117] In a possible implementation, the processing unit 602 is specifically configured to: perform feature recognition on the multi-source heterogeneous data, and determine a behavior feature of the target user and a preference feature of the target user; input the behavior feature of the target user into a feature vector generation model, determine a behavior feature vector of the target user, and input the preference feature of the target user into the feature vector generation model, determine a preference feature vector of the target user; construct a short-term portrait of the target user based on the behavior feature vector of the target user, and construct a long-term portrait of the target user based on the preference feature vector of the target user.

[0118] In a possible implementation, the processing unit 602 is specifically configured to: determine a similar behavior feature of the target user based on the portrait of the target user and a user portrait library; the user portrait library includes portraits of a plurality of users and behavior features of the plurality of users; determine a feature vector set of the target user based on the portrait of the target user, the similar behavior feature of the target user, an intention library of the target user, and a current behavior; and predict a next behavior of the target user based on the feature vector set and a user behavior prediction model.

[0119] In a possible implementation, the processing unit 602 is specifically configured to: determine a behavior mode of the target user based on the portrait of the target user; determine behavior modes of a plurality of users based on a user portrait library; in a case where a similarity between a behavior mode of a first user and the behavior mode of the target user is greater than a first threshold value, determine the first user as a similar user, the first user being one of the plurality of users; and determine a similar behavior feature of the target user based on behavior features of the similar users in the user portrait library.

[0120] In a possible implementation, the processing unit 602 is specifically configured to: determine an interaction intention vector of the target user, a portrait vector of the target user, and a similar behavior feature vector of the target user based on the interaction intention library of the target user, the portrait of the target user, the similar behavior feature of the target user, and a feature vector generation model; and determine a feature vector set of the target user based on the interaction intention vector of the target user, the portrait vector of the target user, and the similar behavior feature vector of the target user.

[0121] In a possible implementation, the processing unit 602 is specifically configured to: determine a plurality of candidate behaviors and a prediction probability of each candidate behavior in the plurality of candidate behaviors based on the feature vector set and a user behavior prediction model; and determine a candidate behavior with a highest prediction probability in the plurality of candidate behaviors as the next behavior.

[0122] In a possible implementation, the processing unit 602 is specifically configured to: input a next behavior into a next behavior image generation model to determine a candidate user interface of the terminal; and adjust an interaction component of the candidate user interface and an interface constraint of the candidate user interface based on a next behavior and layout generation model to determine a target user interface of the terminal.

[0123] Those skilled in the art can clearly understand the technical solutions and the technical effects of the present application from the above description of the embodiments. For the convenience and brevity of description, only the division of the above functional modules is taken as an example in the above description. In actual applications, the above functions can be completed by different functional modules according to requirements, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0124] The present disclosure further provides a computer-readable storage medium having instructions stored thereon, which, when executed by a processor of an electronic device, enable the electronic device to perform the AI-based terminal interface generation method provided in the embodiments of the present disclosure.

[0125] The embodiments of the present disclosure further provide a computer program product containing instructions, which, when running on an electronic device, enable the electronic device to perform the AI-based terminal interface generation method provided in the embodiments of the present disclosure.

[0126] The computer readable storage medium, for example, can be, 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 (a non-exhaustive list) of the computer readable storage medium include: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), registers, a hard disk, an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing, or any other medium from which a computer can read instructions. An exemplary storage medium is coupled to the processor such that the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium can be integral to the processor. The processor and the storage medium can reside in an application-specific integrated circuit (ASIC). Alternatively, the processor and the storage medium can exist as discrete components. In some implementations, the processor and the storage medium can be located in a single ASIC. The steps of a method or algorithm can be embodied in a program of instructions, either directly or indirectly, stored in a storage medium such as the storage medium described above. In other implementations, the methods or algorithms can be embodied in computer readable instructions transmitted over a communication link. Those skilled in the art should appreciate that the functions described herein, including the functions of the software program, can be implemented using software, hardware or any combination of these techniques to implement the specified functions. The software program can be implemented as a program of instructions stored in a storage medium such as a storage medium described above, that can be read and executed by the processor. As such, the disclosure is not limited to any particular programming language or type of code. The software can be practiced using any programming language (e.g., C, C++, Java, JavaScript, etc.) and any type of code (e.g., object-oriented, procedural, etc.).

[0127] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited to this. Any change or replacement within the technical scope disclosed in the present application should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An AI-based terminal interface generation method, characterized by, The method comprises: acquiring multi-source heterogeneous data of a target user collected by a terminal; the multi-source heterogeneous data comprises a current behavior of the target user performed on the terminal; determining a profile of the target user based on the multi-source heterogeneous data; determining an intent library of the target user based on the profile of the target user, the multi-source heterogeneous data, and an intent analysis model; the intent library of the target user is used to indicate an interaction intent of the target; predicting a next behavior of the target user based on the intent library of the target user, the current behavior, the profile of the target user, and a user behavior prediction model; generating a target user interface of the terminal based on the next behavior and an interface generation model; the target user interface is used to provide a service for the next behavior of the target user, and the interface generation model is a multi-modal model.

2. The method of claim 1, wherein, The profile of the target user comprises a short-term profile of the target user and a long-term profile of the target user, and the profile of the target user is constructed based on the multi-source heterogeneous data, comprising: performing feature recognition on the multi-source heterogeneous data to determine a behavior feature of the target user and a preference feature of the target user; inputting the behavior feature of the target user into a feature vector generation model to determine a behavior feature vector of the target user, and inputting the preference feature of the target user into the feature vector generation model to determine a preference feature vector of the target user; constructing the short-term profile of the target user based on the behavior feature vector of the target user, and constructing the long-term profile of the target user based on the preference feature vector of the target user.

3. The method of claim 1, wherein, The next behavior of the target user is predicted based on the intent library of the target user, the current behavior, the profile of the target user, and a user behavior prediction model, comprising: determining a similar behavior feature of the target user based on the profile of the target user and a user profile library; the user profile library comprises profiles of multiple users and behavior features of the multiple users; determining a feature vector set of the target user based on the profile of the target user, the similar behavior feature of the target user, the intent library of the target user, and the current behavior; predicting the next behavior of the target user based on the feature vector set and a user behavior prediction model.

4. The method of claim 3, wherein, The similar behavior feature of the target user is determined based on the profile of the target user and a user profile library, comprising: determining a behavior mode of the target user based on the profile of the target user; determining behavior modes of the multiple users based on the user profile library; in a case where a similarity between a behavior mode of a first user and the behavior mode of the target user is greater than a first threshold value, determining the first user as a similar user, the first user being one of the multiple users; determining the similar behavior feature of the target user based on a behavior feature of the similar user in the user profile library.

5. The method of claim 3, wherein, The determining the feature vector set of the target user based on the portrait of the target user, the similar behavior feature of the target user, the intention library of the target user, and the current behavior includes: The determining the interaction intention vector of the target user, the portrait vector of the target user, and the similar behavior feature vector of the target user based on the intention library of the target user, the portrait of the target user, the similar behavior feature of the target user, and the feature vector generation model; The determining the feature vector set of the target user based on the interaction intention vector of the target user, the portrait vector of the target user, and the similar behavior feature vector of the target user.

6. The method of claim 3, wherein, The predicting the next behavior of the target user based on the feature vector set and the user behavior prediction model includes: The determining a plurality of candidate behaviors and a prediction probability of each candidate behavior in the plurality of candidate behaviors based on the feature vector set and the user behavior prediction model; The determining the candidate behavior with the highest prediction probability in the plurality of candidate behaviors as the next behavior.

7. The method of claim 1, wherein, The interface generation model includes an image generation model and a layout generation model; and the generating the target user interface of the terminal based on the next behavior and the interface generation model includes: The inputting the next behavior into the image generation model to determine a candidate user interface of the terminal; The adjusting an interaction component of the candidate user interface and an interface constraint of the candidate user interface based on the next behavior and the layout generation model to determine the target user interface of the terminal. 8.An AI-based terminal interface generation device, characterized by comprising: The AI-based terminal interface generation apparatus includes a communication unit and a processing unit. The communication unit is configured to acquire multi-source heterogeneous data of a target user collected from a terminal; the multi-source heterogeneous data includes a current behavior of the target user performed on the terminal. The processing unit is configured to determine a portrait of the target user based on the multi-source heterogeneous data. The processing unit is further configured to determine an intention library of the target user based on the portrait of the target user, the multi-source heterogeneous data, and an intention analysis model; the intention library of the target user is used to indicate an interaction intention of the target user. The processing unit is further configured to predict a next behavior of the target user based on the intention library of the target user, the current behavior, the portrait of the target user, and a user behavior prediction model. The processing unit is further configured to generate a target user interface of the terminal based on the next behavior and an interface generation model; the target user interface is used to provide a service for the next behavior of the target user. 9.A terminal interface generating apparatus based on AI, comprising: The apparatus includes: a processor and a communication interface; the communication interface and the processor are coupled, and the processor is configured to run a computer program or instructions to implement the AI-based terminal interface generation method in any one of claims 1-7.

10. A computer-readable storage medium having stored therein instructions, the computer-readable storage medium comprising: When a computer executes the instructions, the computer executes the AI-based terminal interface generation method in any one of claims 1-7.

Images