Recommendation method and device

Abstract

The invention discloses a recommendation method and device, and belongs to the technical field of artificial intelligence. The method comprises the steps that under the condition that a first page is displayed, the type of a problem encountered when a user operates the electronic equipment is determined according to operation information of the user operating the electronic equipment and equipment information of the electronic equipment; n function setting items related to the first page are determined through a semantic model according to the question type, the operation information and the equipment information, the semantic model is obtained by training sample data of a large language model, and N is a positive integer; and displaying prompt information of the function setting item in the first page.

Description

Technical Field

[0001] The present application belongs to the field of artificial intelligence technology, and specifically relates to a recommendation method and device thereof. Background Art

[0002] With the continuous development of science and technology, electronic devices can provide the function of customizing and modifying device usage parameters through settings, such as changing ringtones, setting font size, automatically changing screen color over time, setting scheduled tasks, etc., to meet the needs of users to operate electronic devices. For example, older users need larger fonts or young users need to limit the usage time of electronic devices.

[0003] However, when users operate electronic devices, they are often unclear about or do not understand the names and functions of setting items, encounter many operational problems, do not know how to use the functions of setting items, and cannot find setting items through keyword searches, which reduces the efficiency of users in using setting items in electronic devices. Summary of the Invention

[0004] The purpose of the embodiments of the present application is to provide a recommendation method, device, electronic device, storage medium and program product, which can improve the user's efficiency in using setting items in an electronic device.

[0005] In a first aspect, an embodiment of the present application provides a recommendation method, comprising:

[0006] In the case of displaying the first page, determining the type of problem encountered by the user in operating the electronic device according to the operation information of the user operating the electronic device and the device information of the electronic device;

[0007] Determine N function setting items related to the first page based on the semantic model, question type, operation information, and device information, where N is a positive integer.

[0008] The prompt information of the function setting items is displayed on the first page.

[0009] In a second aspect, an embodiment of the present application provides a recommendation device, including:

[0010] a determination module, configured to determine, when the first page is displayed, a type of problem encountered by the user in operating the electronic device based on operation information of the user operating the electronic device and device information of the electronic device;

[0011] Determine N function setting items related to the first page based on the semantic model, question type, operation information, and device information, where N is a positive integer.

[0012] The display module is used to display prompt information of the function setting items on the first page.

[0013] In a third aspect, an embodiment of the present application provides an electronic device comprising a processor and a memory, wherein the memory stores programs or instructions that can be run on the processor, and when the programs or instructions are executed by the processor, the steps of the recommended method shown in the first aspect are implemented.

[0014] In a fourth aspect, an embodiment of the present application provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the recommended method shown in the first aspect are implemented.

[0015] In a fifth aspect, an embodiment of the present application provides a chip, which includes a processor and a display interface, the display interface and the processor are coupled, and the processor is used to run programs or instructions to implement the steps of the recommended method shown in the first aspect.

[0016] In a sixth aspect, an embodiment of the present application provides a computer program product, which is stored in a storage medium and executed by at least one processor to implement the steps of the recommendation method shown in the first aspect.

[0017] In an embodiment of the present application, when the first page is displayed, the type of problem encountered by the user in operating the electronic device can be determined based on the user's operation information and device information of the electronic device, and N function setting items related to the first page can be determined based on the problem type, operation information and device information through a semantic model, and prompt information of the function setting items can be displayed on the first page. In this way, when the user operates the electronic device, the problem encountered by the user in operating the electronic device can be determined based on the user's operation information and device information of the electronic device, thereby triggering the determination of N function setting items related to the first page, and displaying prompt information of the N function setting items related to the first page on the first page viewed by the user, so as to realize the function of triggering the function setting items through the prompt information, solve the problem encountered by the user in operating the electronic device, eliminate the need for the user to manually search for the setting items, reduce user operations, and avoid the user being unable to use the function of the setting items to solve the problem encountered by the user in operating the electronic device when the user is unclear or does not understand the name and function of the setting items, thereby improving the convenience of the user in using the setting items in the electronic device, and thereby improving the efficiency of the user in using the system setting items in the electronic device. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A flowchart of a recommended method provided in an embodiment of the present application;

[0019] FIG2( a ) is a schematic diagram of a page of a recommendation method provided in an embodiment of the present application;

[0020] FIG2( b ) is a schematic diagram of a page of a recommendation method provided in an embodiment of the present application;

[0021] FIG3( a ) is a schematic diagram of a page of a recommendation method provided in an embodiment of the present application;

[0022] FIG3( b ) is a schematic diagram of the logical dependency relationship of multiple reference function setting items of a recommendation method provided in an embodiment of the present application;

[0023] FIG3( c ) is a schematic diagram of a page of a recommendation method provided in an embodiment of the present application;

[0024] FIG3( d ) is a schematic diagram of a page of a recommendation method provided in an embodiment of the present application;

[0025] FIG4( a ) is a schematic diagram of a page of a recommendation method provided in an embodiment of the present application;

[0026] FIG4( b ) is a schematic diagram of a page of a recommendation method provided in an embodiment of the present application;

[0027] FIG4( c ) is a schematic diagram of a page of a recommendation method provided in an embodiment of the present application;

[0028] FIG4( d ) is a schematic diagram of a page of a recommendation method provided in an embodiment of the present application;

[0029] FIG4( e ) is a schematic diagram of a page of a recommendation method provided in an embodiment of the present application;

[0030] FIG4( f ) is a schematic diagram of a page of a recommendation method provided in an embodiment of the present application;

[0031] Figure 5 A schematic diagram of a page showing a daily report of settings items for a recommended method provided in an embodiment of the present application;

[0032] FIG6( a ) is a flow chart of a recommended method provided in an embodiment of the present application;

[0033] FIG6( b ) is a schematic diagram of a page of a recommendation method provided in an embodiment of the present application;

[0034] FIG6( c ) is a schematic diagram of a physical button according to a recommended method provided in an embodiment of the present application;

[0035] FIG6( d ) is a schematic diagram of a virtual key according to a recommended method provided in an embodiment of the present application;

[0036] FIG6( e ) is a schematic diagram of a virtual key according to a recommended method provided in an embodiment of the present application;

[0037] FIG6( f ) is a schematic diagram of a page of a recommendation method provided in an embodiment of the present application;

[0038] Figure 7A schematic diagram of the structure of a recommended device provided in an embodiment of the present application;

[0039] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application;

[0040] Figure 9 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0041] The following will be combined with the accompanying drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the embodiments described are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of this application.

[0042] The terms "first," "second," and the like in the specification and claims of this application are used to distinguish similar objects, and are not used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of this application can be implemented in an order other than that illustrated or described herein, and that the objects distinguished by "first," "second," and the like are generally of the same type, and do not limit the number of objects; for example, the first object can be one or more. In addition, the term "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0043] In order to solve the problems in the related art, the embodiments of the present application provide a recommended method, apparatus, device, storage medium and program product. Figures 1 to 9 , the recommended methods, devices, equipment, media and program products provided in the embodiments of this application are described in detail through specific embodiments and their application scenarios.

[0044] First, combine Figure 1 A recommended method provided in an embodiment of the present application is described in detail.

[0045] Figure 1 A flowchart of a recommended method provided in an embodiment of the present application.

[0046] like Figure 1 As shown, the recommendation method in the embodiment of the present application can be applied to electronic devices, and the recommendation method may include the following steps:

[0047] Step 110: When the first page is displayed, determine the type of problem encountered by the user in operating the electronic device based on the operation information of the user operating the electronic device and the device information of the electronic device;

[0048] Step 120 , determining N function setting items related to the first page based on the question type, operation information, and device information using a semantic model, where the semantic model is trained using sample data from a large language model, and N is a positive integer;

[0049] Step 130: Display prompt information of the function setting items on the first page.

[0050] For example, as shown in FIG2(a), if the first page is page 20 playing video A, the operation information of the user operating the electronic device includes the user's sliding information on the display screen of the electronic device, and the device information of the electronic device includes the device configuration information when the electronic device displays page 20, such as the current volume level, screen brightness, and font size of the electronic device, and the page information of the first page, such as the full-screen display of video A. Based on this, if when page 20 is displayed, based on the user's sliding information on the first area 21 of the display screen of the electronic device, the current volume level, screen brightness, font size of the electronic device, and the full-screen display of video A, the type of problem encountered by the user operating the electronic device can be determined, such as an abnormal click problem. The operation information includes the number of clicks on physical buttons and / or virtual buttons related to volume and / or brightness, such as 10 times, the click content is the same, such as 10 times, and the click duration exceeds 2 seconds. The device information includes screen brightness 20 and volume level 30. Based on this, a semantic model can be used to capture and understand the question type, operation information, and words, phrases, sentences, and the relationships between them in the device information to determine which function setting items in the electronic device these contents are suitable for. Here, two function setting items related to the first page, such as function setting item 1 for adjusting the screen brightness and function setting item 2 for adjusting the font size of the subtitles in video A, can be determined as the two function setting items related to the first page. Then, as shown in Figure 2(b), prompt information 1 for function setting item 1, such as "Do you need to lower the brightness?", and prompt information 2 for function setting item 2, such as "Click here to increase the subtitle font size," are displayed on page 20.

[0051] In this way, when a user operates an electronic device, prompt information of N function setting items related to the first page can be displayed on the first page viewed by the user based on the user's operation information of the electronic device and the device information of the electronic device, so as to realize the function of triggering the function setting item through the prompt information, solve the problem of operating the electronic device encountered by the user, and make the user do not need to manually search for the setting item. While reducing user operations, it avoids the situation where the user is not clear about or does not understand the name and function of the setting item and cannot use the function of the setting item to solve the situation of operating the electronic device faced by the user, improves the convenience of the user in using the setting items in the electronic device, and thereby improves the efficiency of the user in using the system setting items in the electronic device.

[0052] First, the technical terms appearing in the embodiments of the present application are described in detail below.

[0053] A large language model is a deep learning model trained using large amounts of text data, enabling it to generate natural language text or understand the meaning of text. This model is trained on massive datasets that provide in-depth knowledge and language production across a wide range of topics. The sheer size of the training datasets required requires powerful hardware, such as high-performance processors and large amounts of content.

[0054] A multimodal large model is a deep learning model capable of processing data from multiple modalities, such as text, images, and audio. This model can integrate information from different modalities to achieve cross-modal semantic understanding and generation, demonstrating powerful capabilities in various application scenarios.

[0055] A semantic model is an encoding model with language understanding capabilities. It captures and understands the meanings of words and their relationships in a high-dimensional vector space. For example, it captures and understands the relationships between words, phrases, and sentences. This model adds new data constructors and data processing primitives to the relational model, creating a new type of data model that expresses complex structures and rich semantics. This model can be used to improve language understanding in recommendation, information retrieval, and natural language processing tasks, with lower computational and hardware requirements than large language models.

[0056] Functional settings can include operating system and / or application settings within an electronic device. These are a set of options that users can access and modify to control and personalize the various functions and behaviors of the electronic device. These settings typically include, but are not limited to, network and connectivity, display and sound, application management, privacy, and security.

[0057] The decision tree model is a classification algorithm based on a tree structure. The decision tree model can include at least one of the following algorithms: greedy algorithm (ID3), C4.5 algorithm, and classification and regression tree algorithm (CART). The decision tree model has simple logic and low computational requirements.

[0058] The above steps are described in detail below.

[0059] First, involving step 110, in some embodiments of the present application, the operation information of the user operating the electronic device in the embodiments of the present application includes the user's click information on the input buttons of the electronic device or the user's sliding information on the display screen of the electronic device, and the device information includes at least one of the following: device configuration information when the electronic device displays the first page, and page information of the first page.

[0060] Specifically, the input buttons of the electronic device may include at least one of the following: a physical button of the electronic device, such as a volume button; a virtual button displayed on the electronic device, such as a control displayed on the electronic device, such as a function control in the embodiment of the present application. Information about user clicks on the input buttons of the electronic device may include, but is not limited to, at least one of the following: the number of clicks, the content of the clicks, and the duration of the clicks on the physical and / or virtual buttons.

[0061] The user's sliding information on the display screen of the electronic device may include but is not limited to at least one of the following: the sliding starting point, direction, distance, number of times, duration, and sliding content on the display screen.

[0062] The device configuration information when the electronic device displays the first page includes but is not limited to at least one of the following: the battery level of the electronic device, screen brightness, volume level, font size, and ringtone settings. The types of problems encountered by users operating electronic devices in the embodiments of the present application include but are not limited to at least one of the following: abnormal click problems, abnormal sliding problems, and abnormal device feature problems. Among them, abnormal click problems include but are not limited to at least one of the following: high-frequency invalid clicks, abnormal click duration, and misplaced click content. Abnormal sliding problems include but are not limited to at least one of the following: repeated sliding in a short time, frequent corrections during a single sliding, and abnormal sliding distance. Abnormal device feature problems include but are not limited to at least one of the following: low battery, low brightness, and abnormal changes in volume.

[0063] Therefore, based on the operation information and device information of the user's operation of the electronic device, it can be detected whether there is a problem with the user operation or room for improvement. If there is a problem with the user operation or room for improvement, the type of problem encountered by the user in operating the electronic device can be determined, so that through the semantic model, according to the problem type, operation information and device information, the functional setting items that solve the problem with the user operation or improve the space for the user to operate the electronic device can be determined. While reducing user operations, it avoids the situation where the user is unable to use the function of the setting item to solve the problem of operating the electronic device when the user is unclear or does not understand the name and function of the setting item, thereby improving the convenience of the user in using the setting items in the electronic device, and thereby improving the efficiency of the user in using the system setting items in the electronic device.

[0064] In addition, in an embodiment of the present application, the type of problem encountered by the user when operating the electronic device can also be determined based on the problem type, operation information, device information, and combined with user information.

[0065] User information may include at least one of the following: user individual information, user biometric information, and user emotional information. User individual information includes at least one of the following: user age and user gender. User biometric information includes, but is not limited to, at least one of the following: heart rate and blood oxygen level. User emotional information may include feelings of nervousness, anger, happiness, or sadness. Heart rate and blood oxygen level may be obtained via a wearable device that is communicatively connected to an electronic device, and user emotional information may be determined based on user biometric information such as heart rate and blood oxygen level.

[0066] Therefore, based on the operation information and device information of the user operating the electronic device and combined with the user information, different user states can be further analyzed, so that under the same operation information and device information, different function setting items can be determined for different user information, thereby improving the accuracy of determining the function setting items.

[0067] It should be noted that if the problem type includes at least two items, a comprehensive judgment can be made. For example, if the problem type includes abnormal click problems and abnormal device feature problems, the function setting items can be comprehensively determined through a semantic model based on the power and / or brightness of the electronic device combined with high-frequency invalid clicks.

[0068] Thus, the question type can be determined based on a variety of factors, and the semantic model can be used to determine the function settings related to the first page based on the question type. This eliminates the need for a high-performance processor and large-capacity memory, reduces response time, and reduces computational overhead, making the recommendation method suitable for use in scenarios with fast responses and limited computing resources. Furthermore, the semantic model in the embodiments of this application is trained using sample data from a large language model, which not only has the functionality of a large language model but also avoids the need for powerful hardware support for the large language model, reducing the difficulty of applying the recommendation method.

[0069] In some implementations, a decision tree model with a simple, low-power algorithm may be used to detect whether a user encounters a problem operating an electronic device and the type of the problem. Based on this, step 110 may specifically include steps 1101 and 1102.

[0070] Step 1101: Determine abnormal operation information of the user operating the electronic device through a decision tree model based on the operation information and the device information.

[0071] Exemplarily, through the if-then rules of the decision tree model, the feature space of the operation information and the device information is recursively divided into multiple non-overlapping sub-regions, and the samples in each sub-region have the same predicted abnormal operation information. Through these if-then rules, the operation information and the device information are divided into multiple subsets, thereby realizing classification or regression analysis of the abnormal behavior of the user operating the electronic device to determine the abnormal operation information of the user operating the electronic device, namely, click abnormal information, sliding abnormal information, and device feature abnormal information.

[0072] Step 1102: Determine the type of problem encountered by the user when operating the electronic device according to the type of abnormal operation information.

[0073] Among them, the type of click abnormality information, namely the click abnormality problem, the type of sliding abnormality information, namely the sliding abnormality problem, and the type of device feature abnormality information, namely the device feature abnormality problem, can be determined as the type of problems encountered by users when operating electronic devices.

[0074] Therefore, the decision tree model can be used to determine abnormal operation information of users operating electronic devices. It can not only process high-dimensional data, but also process the nonlinear relationship between operation information and device information. While improving the accuracy of determining abnormal operation information of users operating electronic devices, it can also reduce computing load and reduce computing resources.

[0075] It should be noted that the decision tree model in the embodiment of the present application can be determined in the following manner, as shown below.

[0076] Step 1: Construct decision tree training data. A large language model is used to generate decision tree training data using functional information of each reference function setting item in a plurality of reference function setting items, such as functional explanations and corresponding problems to be solved. The decision tree training data includes positive and negative samples of the decision tree. For example, a large amount of normal operation information of users operating electronic devices and abnormal operation information when encountering problems are generated using reinforcement learning-large language model-decision tree (RL-LLM-DT). The abnormal operation information may include click abnormality information, sliding abnormality information, device feature abnormality information, and user status abnormality information corresponding to user information. The user status abnormality information here may specifically include user biological abnormality information corresponding to user biological information and user emotional abnormality information corresponding to user emotional information.

[0077] In addition, the decision tree training data may also include data on the user's active triggering of search function settings as supplementary data, that is, collecting features before the user searches for a problem as supplementary data when the user encounters a problem and needs to be reminded immediately.

[0078] Step 2: Construct the problems and problem types encountered by users when operating electronic devices. The problems encountered by users can be simulated, and the problems may include abnormal click problems, abnormal sliding problems, abnormal device feature problems, and abnormal user status problems.

[0079] In the embodiment of the present application, the above steps 1 and 2 can be executed successively or in parallel. Of course, step 2 can be executed first and then step 1. At this time, for example, the "font size" function is to adjust the font size of the displayed page. Please simulate the problem of the user using this function setting item and the abnormal user status information corresponding to the problem. The characteristic results of the simulation may include abnormal click problems - the abnormal user status information corresponding to the problem may include vision degeneration of the middle-aged and elderly groups, the operation of pinching to enlarge the font, the discovery of layout disorder in some applications, and secondary settings caused by font reset after system update.

[0080] Step 3, training the decision tree model. Specifically, the initial decision tree model can be trained by using the content constructed in steps 1 and 2. For example, if the initial decision tree model includes the ID3 algorithm, feature selection and decision tree construction can be performed based on the classification information entropy reduction Gain(D,A)=H(D)-H(D|A), and the accuracy, F1-score, ROC-AUC and other indicators are used to evaluate the effect of the initial decision tree model. If the accuracy, F1-score, ROC-AUC and other indicators of the trained initial decision tree model meet the preset effects, the trained initial decision tree model can be determined as the decision tree model in the embodiment of the present application.

[0081] In some embodiments, the operation information and device information may be quantified to more accurately determine abnormal operation information. Based on this, the above step 1101 may specifically include:

[0082] Determine the rationality index value of the user's operation of the electronic device through a decision tree model based on the operation information and device information;

[0083] When the rationality index value is less than or equal to a preset threshold, the operation information is determined to be abnormal operation information of the user operating the electronic device.

[0084] Exemplarily, through a decision tree model, the rationality index value of the user's operation of the electronic device is determined based on the operation information and device information, so as to judge whether the user encounters a problem based on whether the user operation is reasonable. For example, the operation information and device information are scored separately, and a weighted sum calculation is performed based on the score of the operation information, the score of the device information, the reference weight value of the operation information, and the reference weight value of the device information to obtain the rationality index value of the user's operation of the electronic device.

[0085] When it is determined that the rationality index value is less than or equal to the preset threshold value, it is determined that the user's operation of the electronic device is unreasonable, indicating that the user encounters problems in operating the electronic device. At this time, the user's operation information of the electronic device, such as the user's click information on the input buttons of the electronic device, the user's sliding information on the display screen of the electronic device, etc., can be determined as abnormal operation information.

[0086] It should be noted that in the embodiments of the present application, the rationality index value of the user's operation of the electronic device can also be determined based on the operation information, device information, and user information. Among them, the user information can be scored. For example, if the heart rate is greater than 120bpm, or if the elderly user, i.e., a user greater than 65 years old, performs rapid continuous clicks or complex gestures, the score is set to a value higher than the reference threshold. In this way, the rationality index value of the user's operation of the electronic device is obtained by performing a weighted sum calculation based on the score of the operation information, the score of the device information, the score of the user information, the reference weight value of the operation information, the reference weight value of the device information, and the reference weight value of the user information.

[0087] In this way, multiple types of information such as operation information, device information, and user information can be quantified, which can reflect the differences in importance of various types of information and improve the accuracy of whether user operations are reasonable.

[0088] Regarding step 120, in some embodiments, the above step 120 may specifically include steps 1201 to 1203, as shown below.

[0089] Step 1201: Encode the question type, operation information, and device information through a semantic model to obtain encoded information.

[0090] Step 1202 : Determine correlation information between the coding information and each reference coding information according to the coding information and the reference coding information of each reference function setting item in at least two reference function setting items.

[0091] Step 1203 : Determine N function setting items based on the correlation information between the coding information and each reference coding information, where the correlation information between the reference coding information and the coding information of the function setting item is greater than or equal to a preset threshold.

[0092] Exemplarily, the above steps 1201 to 1203 are explained in detail through the following example. First, the operation information, question type and device information are encoded by the semantic model trained with the sample data of the large language model and converted into coded information. Here, the coded information can be represented by a coding vector. The coding vector is used to calculate the correlation with the reference coding information of each reference function setting item, that is, the reference coding vector, to obtain correlation information. Here, the correlation information can be represented by a correlation score. A target correlation score greater than or equal to a preset threshold, such as 0.65, is screened from the correlation scores of the coding information and at least two reference coding information, and the reference function setting item corresponding to the target correlation score is determined as the function setting item, so that the text description of its function setting item can be used as the prompt information of the function setting item.

[0093] It should be noted that in the embodiment of this application, the semantic model is used to determine N function setting items. Compared with the large language model, the computational requirements are small, the reasoning speed is fast, and there are multiple feasible deployment methods. Among them, the embodiment of this application provides the following two deployment methods, which are specifically shown below.

[0094] Deployment method 1 is deployed on the remote server corresponding to the electronic device. With the computing resources of a 16-core processor, it can provide the ability to process 260 requests per second. Through cache computing, it saves computing resources while providing higher processing power.

[0095] The second deployment method is to deploy on electronic devices, deploy semantic model reasoning on electronic devices, use the computing chips of electronic devices, complete semantic understanding reasoning and calculation on electronic devices, and use recommendation and search functions even when the user is offline.

[0096] Therefore, N function setting items are determined through the semantic model, which does not require a high-performance processor and a large-capacity memory, can reduce the response time and reduce the computing overhead, so that the recommendation method can be used in scenarios with fast response and limited computing resources.

[0097] It should be noted that the semantic model in the embodiment of the present application can be determined through the following steps 210 to 230.

[0098] In step 210, the recommended item characteristic information of multiple reference function setting items in the electronic device is determined through a multimodal large language model, wherein the recommended item characteristic information includes logical dependency information between the multiple reference function setting items, functional information of the reference function setting items, and reference adjustment information of the reference function setting items.

[0099] Exemplarily, if the reference function setting item is sound and touch, then the reference function setting item that has logical dependencies on sound and touch may be ringtones and reminders and system sounds and touch. Furthermore, the reference function setting item that has logical dependencies on the reference function setting item of ringtones and reminders may be ringtones of various applications. At this time, the logical dependency relationship between multiple reference function setting items, the name of each reference function setting item and other information can be determined as logical dependency information.

[0100] Taking the reference function setting item "sound and haptics" as an example, the function information of the reference function setting item is described below. The sound and haptics function information may refer to the function of an electronic device providing feedback through sound and haptic feedback, such as vibration, when receiving notifications, incoming calls, alarms, and other operations.

[0101] Still taking the reference function setting item of sound and touch as an example, the reference adjustment information of the reference function setting item is described. The reference adjustment information of sound and touch can be the ring tone type, ring tone volume, etc. of a phone ring tone.

[0102] Step 220: construct sample data of a large language model based on the characteristic information of the recommendation item and the cold start inference data corresponding to the characteristic information of the recommendation item, wherein the cold start inference data includes at least one of the following: reference operation information of the user operating the electronic device, reference question type of the user operating the electronic device, and reference device information of the electronic device.

[0103] Step 230 : Train the reference semantic model using the sample data of the large language model until the reference training conditions are met, thereby obtaining a semantic model.

[0104] Therefore, by training the semantic model with sample data from the large language model, the semantic model not only has the functions of the large language model, but also avoids the need for powerful hardware support for the large language model, reducing the difficulty of applying the recommendation method and enabling the recommendation method to be used in scenarios with fast response and limited computing resources.

[0105] The above steps 210 to 230 are described in detail below.

[0106] In some embodiments, multiple reference function setting items can be understood through a multimodal large language model. The purpose of step 210 is to independently understand multi-level and multi-type reference function setting items, and establish recommended item characteristic information for multiple reference function setting items in the electronic device, that is, a knowledge base of multiple reference function setting items. Based on this, the above step 210 can specifically include steps 2101 to 2104.

[0107] Step 2101 : traverse multiple reference function setting items in the electronic device through a multimodal large language model to construct a tree data structure of the multiple reference function setting items.

[0108] Exemplarily, as shown in FIG3(a), the multimodal large language model performs operations such as clicking, jumping, rewinding, and capturing screen information on multiple reference function setting items in the display page menu, such as flight mode, WLAN, Bluetooth, dual SIM and mobile network, notification and status bar, sound and touch, and reference function setting items at the next level of sound and touch, such as ringtones and reminders, and system sound and touch, through electronic device development tools in a reference order, such as a depth-first traversal, a breadth-first traversal, or the order of the database according to the display page of the reference function setting items, such as the display page 30 of the reference function setting items set in the operating system, to obtain a tree data structure of multiple reference function setting items, as shown in FIG3(b).

[0109] Therefore, through the multimodal large model, multi-level menus and multi-type setting items can be traversed and understood independently, and information can be extracted to form a knowledge base and training data, reducing dependence on manual labor.

[0110] Step 2102 : Determine logical dependency relationship information between the multiple reference function setting items according to the tree data structure of the multiple reference function setting items.

[0111] Exemplarily, according to the tree data structure shown in Figure 3(b), the logical dependencies between reference function setting items such as setting reference function setting items such as flight mode, WLAN, Bluetooth, dual SIM cards and mobile networks, notifications and status bar, and sound and touch can be determined, as well as the logical dependencies between reference function setting items such as sound and touch and reference function setting items such as ringtones and reminders, and system sound and touch. The aforementioned logical dependencies and information such as the name of each reference function setting item involved are determined as logical dependency information.

[0112] Step 2103 : Perform semantic analysis on the page content related to the reference function setting item of each branch in the tree data structure to obtain first function information of the reference function setting item.

[0113] For example, as shown in FIG3(c), the reference function setting item may be a healthy use device, and the page may be a page of healthy use device. The page content shown in FIG3(c) is understood through a multimodal large language model, and a text description is output. For example, the page content includes usage time, application time, available time and application limit.

[0114] Specifically, a semantic analysis is performed on the "usage time" in the "healthy use device" to obtain the first functional information of the healthy use device, which can be expressed through a text description, that is, the usage time is associated with the healthy use device, and the weekly and daily usage time statistics of the electronic device can be viewed. It can also include viewing the usage details to understand the usage of applications (apps) in the electronic device.

[0115] By performing semantic analysis on the "inactivity time", the first functional information of the healthy use device is obtained, which can be expressed through a text description, and the inactivity time can be associated with the healthy use device. The inactivity time can be set, and basic applications such as calls and text messages and applications that are always allowed can be used during the application time.

[0116] A semantic analysis is performed on the "available time" to obtain the first functional information of the healthy use device, which can be expressed through a text description, so that the available time can be associated with the healthy use device, and the daily available time of the electronic device can be set. After the daily available time is exceeded, the use of the electronic device will be prohibited. During the prohibition period, basic applications such as calls and text messages and applications that are always allowed can be used.

[0117] By performing semantic analysis on the "application limit", the first functional information of the healthy use device is obtained, which can be expressed through a text description, and the application limit can be associated with the healthy use device. The daily available time of the application can be set. After exceeding the daily available time, the use of the application will be prohibited and authorization is required to continue using it.

[0118] Step 2104: Determine the function information of the reference function setting item based on the first function information.

[0119] Exemplarily, the first functional information involved in the above step 2103 may be determined as functional information of the health use device.

[0120] Therefore, the use scenarios of multiple reference function setting items in electronic devices can be expanded through a multimodal large language model, and are not limited to the preset functions of the reference function setting items. By performing semantic analysis on the page content related to the reference function setting items of each branch, the complex operation behaviors that may occur by the user can be analyzed, and the user's operation behavior can be associated with the function of the reference function setting item obtained through semantic analysis, so as to more accurately determine the functional information of the reference function setting item.

[0121] It should be noted that for adjustable function setting items such as switches, buttons and sliders, the multimodal large language model dynamically changes them, compares the changes before and after, and understands the specific role of the reference function setting item. Based on this, before step 2104, the recommendation method may also include step 2105. When the reference function setting item is an adjustable function setting item, the second function information of the reference function setting item is determined according to the reference adjustment information of the adjustable function setting item.

[0122] Based on this, step 2104 may specifically include:

[0123] Function information of a reference function setting item is determined according to the first function information and the second function information.

[0124] Exemplarily, as shown in FIG3(d), the adjustable function setting items may be font size and thickness. In the manner shown in FIG3(c), the page of font size and thickness may be understood by a multimodal large language model for the page content shown in FIG3(d), and a text description may be output, such as the page content including font size and font thickness.

[0125] Perform semantic analysis on "font size" to obtain the first functional information of font size and weight. Comparing the changes before and after the page, it can be expressed through text description, that is, the font size and weight can be associated with the font size, that is, the font size has 7 levels of size adjustment options: "smaller-small-standard-large-larger-extra large-largest".

[0126] By performing semantic analysis on "font weight", the first functional information of font size and weight is obtained. By comparing the changes before and after the page, it can be expressed through text description, that is, the font size and weight are associated with the font weight, that is, the font weight can be adjusted steplessly.

[0127] Therefore, the use functions of adjustable function setting items in electronic devices can be expanded through a multimodal large language model, and are not limited to the preset functions of fixed reference function setting items. By performing semantic analysis on the page content related to each adjustable function setting item, the complex operation behaviors that may occur by the user can be analyzed, and the user's operation behavior can be associated with the function of the reference function setting item through semantic analysis, so as to more accurately determine the functional information of the reference function setting item.

[0128] In some embodiments, a multimodal large language model can be used to obtain a tree data structure, logical dependency information, and functional information according to the above steps 2101 to 2103 for information expansion, and a search statement related to the reference function setting item can be obtained to obtain the cold start inference data of the large language model in the reference function setting item. Based on this, an embodiment of the present application also provides a step for determining the cold start inference data. Before step 220, the recommendation method may also include steps 2401 and 2402.

[0129] Step 2401 : Using a multimodal large language model, based on the logical dependency information between multiple reference function setting items, the function information of the reference function setting items is expanded to obtain a search statement corresponding to the reference function setting items.

[0130] Among them, the information expansion processing methods in the embodiments of the present application include but are not limited to: title synonym conversion, problems solved by setting item functions, expected results that can be achieved, associated behaviors, etc.

[0131] For example, let's use the reference function setting item "Usage duration" as an example. Synonymous transformations could be "usage time," "device usage statistics," or "today's device usage." The problem being solved could be "how long did you use your phone today?" or "how long did you use short video apps?" The expected result could be "check today's device usage" or "this week's device usage." The associated behavior could be a user searching for "ringtone" after their alarm goes off. This user likely wants to change the alarm ringtone, not the phone or text message ringtone.

[0132] Step 2402: determine the operation information corresponding to the search statement as reference operation information for the user to operate the electronic device.

[0133] Therefore, through steps 210, 220, 2401 and 2402, the multimodal large language model can independently understand the reference function setting items in the electronic device, expand the search statements and operation information, and extract the reference adjustment information for training the model. In this way, knowledge can be automatically summarized and sample data for training the model can be generated, which can greatly reduce the dependence on manual intervention. Compared with the manual unified configuration solution, it has higher compatibility with electronic devices, can understand the different functions provided by different electronic device models, system versions and software versions, reduce the maintenance difficulties caused by changes in the actual functions of the functions due to upgrades of the operating system and application versions, and is more flexible.

[0134] In some embodiments, involving step 230, the data constructed in the above steps 210, 220, 2401 and 2402 can be used in combination with the knowledge base of the reference function setting items to perform supervised training on the reference semantic model until the training conditions are met to obtain a semantic model.

[0135] Then, involving step 130, in some embodiments of the present application, the prompt information includes at least one of the following: a touch point of the function setting item, a touch prompt text of the function setting item, and the first page includes at least one of the following display areas: a first display area, a second display area. Based on this, step 130 can specifically include step 1301 and step 1302.

[0136] Step 1301: When the prompt information includes a touch point of a function setting item, the touch point is displayed in a first display area.

[0137] For example, as shown in Figures 4(a) to 4(c), a touch point is displayed in the first display area to remind the user that there is a function setting item that can solve the current problem. For example, if the user repeatedly presses the volume up button, as shown in Figure 4(a), a touch point 41 may be displayed in the first display area 401, i.e., above the volume bar 40; if a video freeze is detected, as shown in Figure 4(b), a touch point 42 may be displayed in the first display area 401, i.e., in the middle of the screen; if there are too many messages in the notification center, as shown in Figure 4(c), a touch point 43 may be displayed in the upper right corner of the first display area 401, i.e., the message list.

[0138] Step 1302: When the prompt information includes a touch prompt text of the function setting item, the touch prompt text is displayed in the second display.

[0139] For example, as shown in Figures 4(d) to 4(f), a touch prompt text is displayed in the second display area to remind the user that there is a function setting item that can solve the current problem. For example, if the user repeatedly presses the volume up button, as shown in Figure 4(d), a touch prompt text 44 may be displayed in the second display area, such as at the top 402 of the screen: "Do you need to increase the volume?"; if a video freeze is detected, as shown in Figure 4(e), a touch prompt text 45 may be displayed at the top 402 of the screen: "Switch to clarity A to avoid freezes."; if there are too many messages in the notification center, as shown in Figure 4(f), a touch prompt text 46 may be displayed at the top 402 of the screen: "Find important information."

[0140] It should be noted that there are many ways to determine the first display area and / or the second display area in the embodiment of the present application, including but not limited to: a simple calculation method, a user-related method, and a content-related method. Specifically, the simple calculation method is to determine the page elements with which users frequently interact, such as the four corner positions and the four directions. The user-related method is to calculate the point where the user's line of sight falls, the position where the user's finger operates, etc. The content-related method is to calculate the attention heat map of the current content and the device screen, and the position where attention is high.

[0141] In this way, the interface can be kept simple through touch points, and the original content in the page can be blocked as little as possible to ensure that users can quickly identify the content. This can not only effectively convey information but also avoid interfering with the user experience, and can take into account both functionality and visual experience.

[0142] In some embodiments of the present application, the problem type, the difficulty level of the function setting item to solve the problem operation corresponding to the problem type, and the user information are used to determine the display urgency level of the function setting item, so as to determine whether to immediately trigger the display of the above-mentioned prompt information. Based on this, before step 130, the recommendation method can also include steps 1301 and 1302.

[0143] Step 1301 : Determine the display urgency level of the function setting item according to the problem type, the difficulty level of the operation of the function setting item to solve the problem corresponding to the problem type, and user information.

[0144] Among them, the difficulty level of the function setting item's operation to solve the problem corresponding to the problem type can be determined based on whether the user needs to understand and learn the function setting item and whether it involves the selection of multiple function setting items. Specifically, if the user needs to understand and learn the function setting item and it involves the selection of multiple function setting items, the difficulty level is high. If the user does not need to understand and learn the function setting item and it requires the selection of multiple function setting items, or if the user needs to understand and learn the function setting item and it does not require the selection of multiple function setting items, the difficulty level is medium. If the user does not need to understand and learn the function setting item and it involves the selection of multiple function setting items, the difficulty level is low.

[0145] Here, in addition to determining the display urgency level of the function setting item based on the problem type, the difficulty level of the operation corresponding to the problem type, and user information, the display urgency level of the function setting item can also be determined in combination with device information of the electronic device, such as device modes such as game mode, power saving mode, and airplane mode. Here, if the device mode of the electronic device is in a device mode such as game mode, power saving mode, and airplane mode, it indicates that it is not urgent. Conversely, if the device mode of the electronic device is not in a device mode such as game mode, power saving mode, and airplane mode, it indicates that it is urgent.

[0146] The urgency level can be displayed based on at least one of the following criteria: time, severity, and comprehensive. Time refers to issues with low difficulty and immediate resolution. Severity refers to issues with high impact, high priority, or severe physiological abnormalities. Comprehensive refers to a combination of multiple criteria, such as a user who frequently performs operations but fails to complete them, or is emotionally unstable and in need of urgent assistance.

[0147] Step 1302: When the displayed emergency level is greater than or equal to the reference displayed emergency level, prompt information of the function setting item is displayed on the first page.

[0148] On the contrary, when the display urgency level is lower than the reference display urgency level, it means that the prompt information of the function setting item does not need to be displayed immediately.

[0149] Therefore, by setting the display emergency level, non-critical or short-term anomalies can be filtered out, invalid prompts caused by user errors can be avoided, and redundant recommendations can be avoided, while saving recommendation resources and costs.

[0150] It should be noted that in the recommendation method provided in the embodiment of the present application, when a prompt message is displayed and the user does not click on the prompt message, on the one hand, the data related to the prompt message can be used as a condition for judging the display urgency level of the function setting item, that is, the user may not need the prompt message, and the display of such prompt messages can be reduced in the future. On the other hand, it may be because the way of displaying the prompt message does not meet the user's preferences. In this case, it can be used as a basis for adjusting the display form of the prompt message. Specifically, the user's operation behavior data, problem type and solution can be recorded and wait for the user to view, such as Figure 5 As shown, the daily report of the setting item records three types of problems that do not require immediate reminders.

[0151] Type 1: Requires user understanding and learning, such as learning "quick page turning skills". For example, its daily setting items may include: 10:20, it is detected that the desktop right swipe is repeated 10 times, do you need to: 1) set a quick opening of the newly installed application; 2) change the desktop page turning style; 3) learn the quick page turning skills.

[0152] Type 2: requires the user to select from multiple options, such as selecting an alarm ringtone. For example, the daily setting item may include: at 14:00, after the alarm rings, your heart rate increases instantly. Do you need to: 1) change to a softer alarm ringtone; 2) turn on the alarm advance reminder function.

[0153] Type 3: Involves user preferences, such as changing the input method. For example, at 3:10 PM, you receive a verification code and manually enter it. Do you want to: 1) have the input method automatically fill in the verification code? 2) Switch to the official input method.

[0154] In addition, the recommendation method provided in the embodiment of the present application may also include steps 1401 to 1403.

[0155] Step 1401: Receive a first input of prompt information from a user.

[0156] Step 1402: In response to the first input, display a function control corresponding to the function setting item.

[0157] Step 1403: Adjust the function corresponding to the function setting item according to the user's input to the function control.

[0158] For example, when it is detected that the user repeatedly presses the volume up button when the volume is already at maximum, the function control corresponding to the volume setting item is displayed, guiding the user to turn on the volume switch to 120% through the function control. Alternatively, when it is detected that the content the user is watching is stuck, combined with operating system data such as network connection status, processor temperature, memory usage, etc., the function control corresponding to the network is displayed to guide the user to adjust the network status, switch WLAM or traffic, clean up the background, clean up the memory, etc. When the user has too much information in the notification center, a pop-up window prompts "Find important messages" to guide the user to manage the information in the notification center and block notifications of non-important information.

[0159] Therefore, the recommendation method provided by the embodiment of the present application can effectively solve the problem of users not knowing which function can solve the current problem when they encounter an operation problem, and cannot retrieve the mobile phone setting items in a colloquial way. The method has the high-level semantic understanding ability of a large model, efficient reasoning speed and high practicality.

[0160] Based on this, the recommendation method provided in the embodiment of the present application can be applied to solve application scenarios where users encounter operational problems and do not know what function can solve the current problem, and are unable to verbally retrieve the system setting items and application settings of each application in the electronic device.

[0161] In order to better illustrate the recommended method provided in the embodiment of this application, the following Figure 6(a) to Figure 6(f) The recommended method provided in the embodiments of the present application is described in detail.

[0162] FIG6( a ) is a flow chart of a recommended method provided in an embodiment of the present application.

[0163] As shown in FIG6( a ), the recommendation method provided in the embodiment of the present application may include steps 101 to 111 , as specifically described below.

[0164] Step 101 provides function setting switches to meet the needs of different users. The function setting item settings page, shown in FIG6(b), provides switches for function setting item recommendations, function setting item daily reports, display of prompt information records, and quick assistance settings. Users can activate the recommended functions in the recommended function setting items by switching them on. When disabled, only the active search function in step 109 is available.

[0165] Step 102: Provide recommendation and search functions. The instance will continuously detect user operations and recommend function settings when conditions are met. It also provides a search function for users to actively trigger.

[0166] Step 103: Detect whether the user is experiencing an operational problem. A simple, low-power algorithm, such as a decision tree model, is used to detect whether the user is experiencing an operational problem with the electronic device. The detection features used may include user operation information, device information, and user information. This information can be used to determine the legitimacy of the user's operation of the electronic device and, therefore, whether a problem has occurred. This determination includes, but is not limited to, the following: click anomalies: frequent invalid clicks, abnormal click duration, and misplaced click content; sliding anomalies: repeated sliding within a short period of time, frequent corrections during a single slide, and abnormal sliding distance; user characteristic anomalies: abnormal heart rate / blood oxygen levels (e.g., heart rate >120 bpm) accompanied by a sudden increase in operation speed or an increased error rate; elderly users (>65 years old) performing rapid, continuous clicks or complex gestures; device characteristic anomalies: low battery, low brightness, and abnormal volume fluctuations; and combined anomalies: a combination of the above anomalies is used to determine the problem, for example, low device battery + frequent invalid clicks + elevated user heart rate; elderly users + insufficient sliding distance + small font size. If the system detects that the user is experiencing an issue, it proceeds to step 104; otherwise, it proceeds to step 111.

[0167] Step 104: Infer and calculate the semantic model of the function setting items to provide function setting item recommendations. The semantic model, trained with large language model data, encodes the user's electronic device operation information, page information, the type of operational problem encountered, and device information, converting it into an encoding vector. The encoding vector is then correlated with the setting functions provided by the function setting items to obtain a score. Setting functions or text descriptions of problems encountered with a correlation above a preset threshold, such as 0.65, are selected as recommendations or search results.

[0168] Here, it should be noted that if in step 104 after the active search function in step 109, the object of encoding can be the operation information, page information, type of operation problem encountered, device information and search query provided by the semantic model trained with large language model data for the user's operation of the electronic device, wherein the search query provided is used to represent the search type triggered by the user, such as global search, setting item search, etc.

[0169] Step 105, detect the type of problem encountered by the user, and determine whether to immediately pop up a window. The example uses a decision tree model to detect the type of problem encountered by the user and the solution, and determines whether an immediate pop-up window is needed. The features used include the features of step 103 and step 104, and the additional information is as follows: the difficulty level of the function setting item to solve the problem operation corresponding to the problem type, and the user status message. The conditions and judgment logic that require an immediate pop-up window include but are not limited to: Time conditions: the problem difficulty is low and can be solved immediately. Serious conditions: the problem has a large impact and high priority, and the user's physiological condition is seriously abnormal. Comprehensive judgment: A comprehensive judgment of multiple conditions, such as the user performs high-frequency operations but does not complete the operations, is in an abnormal emotional state, and needs urgent help.

[0170] When it is not necessary to remind immediately, the process proceeds to step 106. When it is determined that it is necessary to remind immediately, the process proceeds to step 108 for immediate reminder.

[0171] Step 106: Display the touch point to remind the user. In the selected display area, the touch point pops up to remind the user that there is a function setting item that can solve the current problem.

[0172] When the user clicks the touch point, the process proceeds to step 108. When the user ignores the touch point, the process proceeds to step 107.

[0173] Step 107: Record the user's actions and solutions. Specifically, record the user's action data, problem type, and solution, and wait for the user to review. For example, in the system, for example, the daily report of the function setting item records three types of problems that do not require immediate reminders.

[0174] In step 108, the electronic device displays a solution to the current problem. If an immediate reminder is determined, or if the user proactively clicks a touch point, the system displays the recommended results from step 104 via a pop-up window, notifying the user of the current problem. Upon clicking, the user executes the solution or is redirected to the corresponding function setting result page.

[0175] In step 109, the user actively triggers the recommendation and search functions. In the quick assistance settings of Figure 6(b), binding functions are provided for the physical button 60 shown in Figure 6(c), the sidebar virtual button 61 shown in Figure 6(d), and the bottom virtual button 62 shown in Figure 6(e). After binding, a detector is registered. When the user clicks the button, the instance analysis and recommendation functions are triggered. Features such as the user's current operation and system status are calculated, and the semantic model of step 104 is requested to propose the user's problem and solution, as shown in Figure 6(f). A spoken search function is provided. The user uses voice input or enters text in the search input box to describe the specific problem they encounter, and the semantic model of step 104 is requested to provide and display a solution. Here, the semantic model supports the spoken search requirements of function setting items, which can improve the user's search experience and efficiency. It also significantly reduces the amount of computation. The semantic model can be deployed on limited computing resources, making it highly practical.

[0176] Step 110: Data Feedback. After applying the semantic search capability for function settings, collect and analyze user backward behavior, such as exposure, click volume, and click-through rate of function settings, to generate better data for the large language model, achieving a positive cycle of increasing usability.

[0177] Step 111: No abnormal operation or problem is detected, and the service is in standby mode. If no abnormal user operation is detected, the service enters the standby mode and waits for the next detection result.

[0178] It should be noted that, in addition, the user's active triggered search in step 109 is used as supplementary data. The features before the user searches for a problem are collected as supplementary data when the user encounters a problem and needs to be reminded immediately.

[0179] Therefore, the embodiments of the present application can realize efficient semantic recommendation and search methods for function setting items, through the design of recommendation logic for mobile phone function setting items, the independent establishment of knowledge base, the migration of large language model reasoning capabilities, the deployment and application of semantic models, and the data collection, processing and feedback methods. It can detect problems encountered by users when using mobile phones with low power consumption, provide functional daily reports, low-intrusive touch points, and timely pop-up function setting item recommendations, and provide one-click trigger assistance and colloquial search functions. The invention integrates the recommendation and search capabilities of function setting items, with high integration and ideal and efficient reasoning speed, and has the ability to be deployed offline, which is highly practical. The invention can assist users who are not familiar with mobile phone functions to understand and learn, make full use of the functions on the mobile phone, greatly reduce the difficulty of using complex systems, improve the ease of use of the system, and enhance the competitiveness of the system.

[0180] The recommendation method provided in the embodiment of the present application can be executed by a recommendation device. In the embodiment of the present application, the recommendation device performing the recommendation is taken as an example to illustrate the device of the recommendation method provided in the embodiment of the present application.

[0181] Based on the same inventive concept, this application also provides a recommendation device. Figure 7 Provide detailed explanation.

[0182] Figure 7 A schematic structural diagram of a recommended device provided in an embodiment of the present application.

[0183] like Figure 7 As shown, the recommendation device 70 can be applied to an electronic device, and the recommendation device 70 can specifically include:

[0184] A determination module 701 is configured to determine, when the first page is displayed, the type of problem encountered by the user in operating the electronic device based on the operation information of the user operating the electronic device and the device information of the electronic device;

[0185] The determination module 701 may also be configured to determine N function setting items related to the first page based on the question type, operation information, and device information using a semantic model, where the semantic model is trained using sample data of a large language model, and N is a positive integer.

[0186] The display module 702 is used to display prompt information of the function setting items on the first page.

[0187] The recommendation device 70 in the embodiment of the present application is described in detail below, as shown below.

[0188] In some embodiments of the present application, the determination module 701 may be specifically configured to determine abnormal operation information of the user operating the electronic device through a decision tree model based on the operation information and the device information;

[0189] The type of problem encountered by the user in operating the electronic device is determined based on the type of abnormal operation information.

[0190] In some embodiments of the present application, the determination module 701 may be specifically configured to, when the operation information includes information of a user clicking an input button of the electronic device or information of a user sliding a display screen of the electronic device, and the device information includes at least one of the following: device configuration information when the electronic device displays a first page, and page information of the first page, determine, based on the operation information and the device information, a rationality index value of the user operating the electronic device through a decision tree model;

[0191] When the rationality index value is less than or equal to a preset threshold, the operation information is determined to be abnormal operation information of the user operating the electronic device.

[0192] In some embodiments of the present application, the recommendation device 70 in the embodiment of the present application may further include an encoding module for encoding the question type, operation information, and device information through a semantic model to obtain encoded information;

[0193] The determining module 701 may also be configured to determine, based on the coding information and the reference coding information of each reference function setting item in the at least two reference function setting items, correlation information between the coding information and each reference coding information;

[0194] The determination module 701 may also be configured to determine N function setting items based on correlation information between the coding information and each reference coding information, wherein the value of the correlation information between the reference coding information and the coding information of the function setting item is greater than or equal to a preset threshold.

[0195] In some embodiments of the present application, the determination module 701 may also be configured to determine, using a multimodal large language model, recommendation item characteristic information of multiple reference function setting items in the electronic device, wherein the recommendation item characteristic information includes logical dependency information between the multiple reference function setting items, function information of the reference function setting items, and reference adjustment information of the reference function setting items.

[0196] The recommendation device 70 in the embodiment of the present application may further include a construction module for constructing sample data of a large language model based on the characteristic information of the recommendation item and the cold-start inference data corresponding to the characteristic information of the recommendation item, wherein the cold-start inference data includes at least one of the following: reference operation information of the user operating the electronic device, reference question type of the user operating the electronic device, and reference device information of the electronic device;

[0197] The recommendation device 70 in the embodiment of the present application may further include a training module for training the reference semantic model using sample data of the large language model until the reference training conditions are met to obtain the semantic model.

[0198] In some embodiments of the present application, the recommendation device 70 in the embodiment of the present application may further include a traversal module for traversing multiple reference function setting items in the electronic device using a multimodal large language model to construct a tree data structure of the multiple reference function setting items;

[0199] The determining module 701 may also be configured to determine logical dependency relationship information between the plurality of reference function setting items according to the tree data structure of the plurality of reference function setting items;

[0200] The recommendation device 70 in the embodiment of the present application may further include an analysis module for performing semantic analysis on the page content related to the reference function setting item of each branch in the tree data structure to obtain the first function information of the reference function setting item;

[0201] The determining module 701 may also be configured to determine function information of a reference function setting item based on the first function information.

[0202] In some embodiments of the present application, the determining module 701 may also be configured to, when the reference function setting item is an adjustable function setting item, determine the second function information of the reference function setting item according to the reference adjustment information of the adjustable function setting item;

[0203] Function information of a reference function setting item is determined according to the first function information and the second function information.

[0204] In some embodiments of the present application, the recommendation device 70 in the embodiments of the present application may further include a display module, configured to display the touch point in the first display area when the prompt information includes at least one of the following: a touch point of a function setting item, a touch prompt text of the function setting item, and the first page includes at least one of the following display areas: a first display area and a second display area, and when the prompt information includes the touch point of the function setting item;

[0205] In a case where the prompt information includes a touch prompt text of the function setting item, the touch prompt text is displayed in the second display.

[0206] In some embodiments of the present application, the recommendation device 70 in the embodiment of the present application may further include a receiving module for receiving a first input of the user on the prompt information;

[0207] The recommendation device 70 in the embodiment of the present application may further include a display module for displaying a function control corresponding to the function setting item in response to the first input;

[0208] The recommendation device 70 in the embodiment of the present application may further include a recommendation module, which is used to adjust the function corresponding to the function setting item according to the user's input to the function control.

[0209] The recommendation device in the embodiments of the present application can be an electronic device or a component in an electronic device, such as an integrated circuit or chip. The electronic device can be a terminal or other device other than a terminal. For example, the electronic device can be a mobile phone, a tablet computer, a laptop computer, a PDA, an in-vehicle electronic device, a mobile Internet device (MID), an augmented reality (AR) / virtual reality (VR) device, a robot, a wearable device, an ultra-mobile personal computer (UMPC), a netbook or a personal digital assistant (PDA), etc. It can also be a server, a network attached storage (NAS), a personal computer (PC), a television (TV), a teller machine or a self-service machine, etc., and the embodiments of the present application do not specifically limit it.

[0210] The recommendation device in the embodiment of the present application may be a device having an operating system. The operating system may be an Android operating system, an iOS operating system, or other possible operating systems, which are not specifically limited in the embodiment of the present application.

[0211] The device coordination apparatus provided in the embodiment of the present application can achieve Figure 1 The various processes implemented in the embodiment of the recommended method shown in FIG6 achieve the same technical effect, and will not be described again here to avoid repetition.

[0212] Based on this, the recommendation device provided in the embodiment of the present application can, when displaying the first page, determine the type of problem encountered by the user in operating the electronic device based on the user's operation information of the electronic device and the device information of the electronic device, and determine N function setting items related to the first page based on the problem type, operation information and device information through a semantic model, and display prompt information of the function setting items on the first page. In this way, when the user operates the electronic device, it can be determined that the user has encountered an operation problem based on the user's operation information of the electronic device and the device information of the electronic device, thereby triggering the determination of N function setting items related to the first page, and displaying prompt information of the N function setting items related to the first page on the first page viewed by the user, so as to realize the function of triggering the function setting items through the prompt information, solve the problem encountered by the user in operating the electronic device, eliminate the need for the user to manually search for the setting items, reduce user operations, and avoid the user being unable to use the function of the setting items to solve the problem encountered by the user in operating the electronic device when the user is unclear or does not understand the name and function of the setting items, thereby improving the convenience of the user in using the setting items in the electronic device, and thereby improving the efficiency of the user in using the system setting items in the electronic device.

[0213] Optional, such as Figure 8 As shown, an embodiment of the present application also provides an electronic device 80, including a processor 801 and a memory 802, wherein the memory 802 stores a program or instruction that can be run on the processor 801, and when the program or instruction is executed by the processor 801, the various steps of the above-mentioned recommended method embodiment are implemented and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0214] It should be noted that the electronic devices in the embodiments of the present application include the above-mentioned mobile electronic devices and non-mobile electronic devices.

[0215] Figure 9 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present application.

[0216] The electronic device 900 includes but is not limited to: a radio frequency unit 901, a network module 902, an audio output unit 903, an input unit 904, a sensor 905, a display unit 906, a user input unit 907, an interface unit 908, a memory 909, a processor 910 and other components.

[0217] Those skilled in the art will understand that the electronic device 900 may also include a power source (such as a battery) to power each component, and the power source may be logically connected to the processor 910 through a power management system, thereby implementing functions such as charging, discharging, and power consumption management through the power management system. Figure 9The electronic device structure shown in the figure does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently, which will not be repeated here.

[0218] In this embodiment of the present application, the processor 910, while displaying the first page, determines the type of problem encountered by the user operating the electronic device based on the user's operation information and device information of the electronic device. The processor 910 may also be configured to determine N function setting items related to the first page based on the problem type, operation information, and device information using a semantic model trained using sample data from a large language model, where N is a positive integer. The display unit 906 is configured to display prompt information for the function setting items on the first page.

[0219] The electronic device 900 is described in detail below, as shown below.

[0220] In some embodiments of the present application, the processor 910 may be specifically configured to determine abnormal operation information of the user operating the electronic device through a decision tree model based on the operation information and the device information;

[0221] The type of problem encountered by the user in operating the electronic device is determined based on the type of abnormal operation information.

[0222] In some embodiments of the present application, the processor 910 may be specifically configured to, when the operation information includes information of a user clicking an input button of the electronic device or information of a user sliding a display screen of the electronic device, and the device information includes at least one of the following: device configuration information when the electronic device displays a first page, and page information of the first page, determine, based on the operation information and the device information, a rationality index value of the user operating the electronic device through a decision tree model;

[0223] When the rationality index value is less than or equal to a preset threshold, the operation information is determined to be abnormal operation information of the user operating the electronic device.

[0224] In some embodiments of the present application, the processor 910 is configured to encode the question type, operation information, and device information using a semantic model to obtain encoded information;

[0225] determining, based on the coding information and the reference coding information of each reference function setting item in the at least two reference function setting items, correlation information between the coding information and each reference coding information;

[0226] According to the correlation information between the coding information and each reference coding information, N function setting items are determined, and the value of the correlation information between the reference coding information and the coding information of the function setting item is greater than or equal to a preset threshold.

[0227] In some embodiments of the present application, the processor 910 may be further configured to determine, using a multimodal large language model, recommendation item characteristic information of multiple reference function setting items in the electronic device, wherein the recommendation item characteristic information includes logical dependency information between the multiple reference function setting items, function information of the reference function setting items, and reference adjustment information of the reference function setting items.

[0228] Constructing sample data of a large language model based on the recommended item characteristic information and cold-start inference data corresponding to the recommended item characteristic information, wherein the cold-start inference data includes at least one of the following: reference operation information of a user operating an electronic device, a reference question type of the user operating the electronic device, and reference device information of the electronic device;

[0229] The reference semantic model is trained using sample data from the large language model until the reference training conditions are met to obtain a semantic model.

[0230] In some embodiments of the present application, the processor 910 is configured to traverse a plurality of reference function setting items in the electronic device using a multimodal large language model to construct a tree data structure of the plurality of reference function setting items;

[0231] Determining logical dependency relationship information between the multiple reference function setting items according to the tree data structure of the multiple reference function setting items;

[0232] Performing semantic analysis on page content related to the reference function setting item of each branch in the tree data structure to obtain first function information of the reference function setting item;

[0233] Based on the first function information, function information of the reference function setting item is determined.

[0234] In some embodiments of the present application, the processor 910 may further be configured to, when the reference function setting item is an adjustable function setting item, determine the second function information of the reference function setting item according to the reference adjustment information of the adjustable function setting item;

[0235] Function information of a reference function setting item is determined according to the first function information and the second function information.

[0236] In some embodiments of the present application, the display unit 906 is configured to, when the prompt information includes at least one of the following: a touch point of a function setting item, touch prompt text of the function setting item, the first page includes at least one of the following display areas: a first display area, a second display area, and the prompt information includes the touch point of the function setting item, display the touch point in the first display area;

[0237] In a case where the prompt information includes a touch prompt text of the function setting item, the touch prompt text is displayed in the second display.

[0238] In some embodiments of the present application, the user input unit 907 is configured to receive a first input of the user for prompt information;

[0239] A display unit 906 is configured to display a function control corresponding to a function setting item in response to the first input;

[0240] The processor 910 is configured to adjust the function corresponding to the function setting item according to the user input to the function control.

[0241] It should be understood that the input unit 904 may include a graphics processing unit (GPU) 9041 and a microphone 9042, and the graphics processor 9041 processes the image data of the static image or video obtained by the image capture device (such as a camera) in the video capture mode or the image capture mode. The display unit 906 may include a display panel, and the display panel may be configured in the form of a liquid crystal display, an organic light emitting diode, etc. The user input unit 907 includes a touch panel 9071 and at least one of the other input devices 9072. The touch panel 9071 is also called a touch screen. The touch panel 9071 may include two parts: a touch detection device and a touch display. Other input devices 9072 may include but are not limited to a physical keyboard, function keys (such as a volume display button, a switch button, etc.), a trackball, a mouse, and a joystick, which will not be repeated here.

[0242] The memory 909 can be used to store software programs and various data. The memory 909 may mainly include a first storage area for storing programs or instructions and a second storage area for storing data. The first storage area may store an operating system, applications or instructions required for at least one function (such as a sound playback function, an image playback function, etc.). In addition, the memory 909 may include a volatile memory or a non-volatile memory, or the memory 909 may include both volatile and non-volatile memory. The non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. Volatile memory can be random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct RAM bus random access memory (DRRAM). The memory 909 in the embodiment of the present application includes but is not limited to these and any other suitable types of memory.

[0243] Processor 910 may include one or more processing units. Optionally, processor 910 integrates an application processor and a modem processor. The application processor primarily handles operations related to the operating system, user pages, and applications, while the modem processor primarily processes wireless display signals, such as a baseband processor. It is understood that the modem processor may not be integrated into processor 910.

[0244] An embodiment of the present application also provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the various processes of the above-mentioned recommended method embodiment are implemented and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0245] The processor is the processor in the electronic device in the above embodiment. The readable storage medium includes a computer-readable storage medium, such as a computer read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0246] In addition, an embodiment of the present application further provides a chip, which includes a processor and a display interface. The display interface and the processor are coupled. The processor is used to run programs or instructions to implement the various processes of the above-mentioned recommended method embodiment and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0247] It should be understood that the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

[0248] An embodiment of the present application provides a computer program product, which is stored in a storage medium. The program product is executed by at least one processor to implement the various processes of the above-mentioned recommended method embodiment and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0249] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0250] Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of the present application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in reverse order depending on the functions involved. For example, the methods described may be performed in an order different from that described, and various steps may be added, omitted, or combined. Furthermore, features described with reference to certain examples may be combined in other examples.

[0251] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for enabling a terminal (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of each embodiment of the present application.

[0252] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of this application, ordinary technicians in this field can also make many forms without departing from the purpose of this application and the scope of protection of the claims, all of which are within the protection of this application.

Claims

1. A recommendation method, characterized in that: include: In the case of displaying the first page, determining the type of problem encountered by the user in operating the electronic device according to the operation information of the user operating the electronic device and the device information of the electronic device; Determining, using a semantic model, N function setting items related to the first page according to the question type, the operation information, and the device information, wherein the semantic model is trained using sample data of a large language model, and N is a positive integer; Display prompt information of the function setting item on the first page.

2. The method according to claim 1, characterized in that The determining, based on the operation information of the user operating the electronic device and the device information, the type of problem encountered by the user in operating the electronic device, includes: Determining abnormal operation information of the user operating the electronic device through a decision tree model based on the operation information and the device information; The type of problem encountered by the user in operating the electronic device is determined according to the type of the abnormal operation information.

3. The method according to claim 2, characterized in that The operation information includes click information of a user on an input button of the electronic device or sliding information of a user on a display screen of the electronic device, and the device information includes at least one of the following: device configuration information when the electronic device displays the first page, and page information of the first page; The determining, based on the operation information and the device information, abnormal operation information of the user operating the electronic device through a decision tree model includes: Determining a rationality index value of the user's operation of the electronic device through the decision tree model according to the operation information and the device information; When the rationality index value is less than or equal to a preset threshold, the operation information is determined to be abnormal operation information of the user operating the electronic device.

4. The method according to any one of claims 1 to 3, characterized in that The determining, using the semantic model and according to the question type, the operation information, and the device information, N function setting items related to the first page includes: Encoding the question type, the operation information, and the device information through a semantic model to obtain encoded information; determining, based on the coding information and reference coding information of each reference function setting item in at least two reference function setting items, correlation information between the coding information and each reference coding information; The N function setting items are determined according to correlation information between the coding information and each reference coding information, wherein a value of correlation information between the reference coding information of the function setting item and the coding information is greater than or equal to a preset threshold.

5. The method according to claim 1, characterized in that The method further comprises: determining, using a multimodal large language model, recommended item characteristic information for a plurality of reference function setting items in the electronic device, wherein the recommended item characteristic information includes logical dependency information between the plurality of reference function setting items, function information of the reference function setting items, and reference adjustment information of the reference function setting items; Constructing sample data of the large language model based on the recommendation item characteristic information and cold-start inference data corresponding to the recommendation item characteristic information, wherein the cold-start inference data includes at least one of the following: reference operation information of the user operating the electronic device, reference question type of the user operating the electronic device, and reference device information of the electronic device; The reference semantic model is trained using the sample data of the large language model until a reference training condition is met, thereby obtaining the semantic model.

6. The method according to claim 5, characterized in that The determining, by using a multimodal large language model, recommended item characteristic information of a plurality of reference function setting items in the electronic device includes: Traversing a plurality of reference function setting items in the electronic device by using a multimodal large language model to construct a tree data structure of the plurality of reference function setting items; determining logical dependency relationship information between the plurality of reference function setting items according to the tree data structure of the plurality of reference function setting items; Performing semantic analysis on page content related to the reference function setting item of each branch in the tree data structure to obtain first function information of the reference function setting item; Based on the first function information, function information of the reference function setting item is determined.

7. The method according to claim 6, characterized in that Before determining the function information of the reference function setting item based on the first function information, the method further includes: In a case where the reference function setting item is an adjustable function setting item, determining second function information of the reference function setting item according to reference adjustment information of the adjustable function setting item; The determining, based on the first function information, the function information of the reference function setting item includes: Function information of the reference function setting item is determined according to the first function information and the second function information.

8. The method according to claim 1, characterized in that The prompt information includes at least one of the following: a touch point of the function setting item, a touch prompt text of the function setting item, and the first page includes at least one of the following display areas: a first display area, a second display area; The displaying of the prompt information of the function setting item on the first page includes: In a case where the prompt information includes a touch point of the function setting item, displaying the touch point in the first display area; In a case where the prompt information includes a touch prompt text of the function setting item, the touch prompt text is displayed in the second display.

9. The method according to claim 1, characterized in that The method further comprises: receiving a first input of the prompt information from the user; In response to the first input, displaying a function control corresponding to the function setting item; According to the user's input to the function control, the function corresponding to the function setting item is adjusted.

10. A recommendation device, characterized in that: include: a determination module, configured to determine, when the first page is displayed, a type of problem encountered by the user in operating the electronic device based on operation information of the user operating the electronic device and device information of the electronic device; The determining module is further configured to determine, using a semantic model, N function setting items related to the first page according to the question type, the operation information, and the device information, wherein the semantic model is trained using sample data of a large language model, and N is a positive integer; A display module is used to display prompt information of the function setting item on the first page.

Images