Recommendation content generation method, object recommendation method, computing device, storage medium and program product
By determining the content preference characteristics of the target user and using the content generation model to generate personalized recommendation materials, the problem of low accuracy of recommended content is solved, efficient and accurate recommendation results are achieved, and user interaction and conversion rates are improved.
Patent Information
- Application Number
- CN202510007102.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-02
- Publication Date
- 2025-05-27
Smart Images
Figure CN120045777A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of electronic information processing technology, and in particular to a method for generating recommended content, a method for recommending objects, a computing device, a computer storage medium, and a computer program product. Background Art
[0002] In some online interactive systems that provide objects for users to perform interactive behaviors, the objects are usually provided by the object provider. In order to help improve the object conversion rate, etc., object recommendation becomes a solution.
[0003] In traditional implementations, in order to recommend an object, the recommended content is usually generated by the object provider. However, the generated recommended content is not accurate enough and cannot effectively attract users. Summary of the invention
[0004] Multiple aspects of the present application provide a recommended content generation method, an object recommendation method, a computing device, a storage medium, and a program product to solve the technical problem of low accuracy of recommended content in the prior art.
[0005] In a first aspect, an embodiment of the present application provides a method for generating recommended content, including:
[0006] In response to a recommendation request triggered by an object provider, determining a target object to be recommended;
[0007] Determining content preference characteristics of a target user to be recommended; wherein the content preference characteristics are predicted based on user feature data of the target user;
[0008] Based on the content preference feature and the object-related information of the target object, using a content generation model to generate a recommendation material matching the content preference feature;
[0009] Based on the recommendation material, recommendation content for the target object is generated; wherein the recommendation content is used to recommend the target object to the target user.
[0010] In a second aspect, an embodiment of the present application provides an object recommendation method, comprising:
[0011] In response to a recommendation request triggered by an object provider, determining a target object to be recommended;
[0012] Determining content preference characteristics of a target user to be recommended; wherein the content preference characteristics are predicted based on user feature data of the target user;
[0013] Based on the content preference feature and the object-related information of the target object, using a content generation model to generate a recommendation material matching the content preference feature;
[0014] Based on the recommendation material, generating recommendation content for the target object;
[0015] The recommended content is sent to the target user to recommend the target object to the target user.
[0016] In a third aspect, an embodiment of the present application provides a method for generating recommended content, including:
[0017] Displaying recommended prompt information in the display interface;
[0018] The detection object provider sends a recommendation request to the server in response to the recommendation operation triggered by the recommendation prompt information; the recommendation request is used to determine the target object to be recommended and the content preference characteristics of the target user to be recommended; the content preference characteristics and the object-related information of the target object are used to generate recommendation materials matching the content preference characteristics using a content generation model; the recommendation materials are used to generate recommended content for the target object; wherein the content preference characteristics are predicted based on the user characteristic data of the target user.
[0019] In a fourth aspect, an embodiment of the present application provides a computing device, including a processing component and a storage component;
[0020] The storage component stores a computer program; the processing component is coupled to the storage component and is used to execute the computer program stored in the storage to implement the recommended content generation method as described in the first aspect above, the object recommendation method as described in the second aspect above, or the recommended content generation method as described in the third aspect above.
[0021] In a fifth aspect, a computer-readable storage medium is provided in an embodiment of the present application, on which a computer program is stored. When the computer program is executed by a processing component, the recommended content generation method as described in the first aspect above, the object recommendation method as described in the second aspect above, or the recommended content generation method as described in the third aspect above is implemented.
[0022] In a sixth aspect, a computer program product is provided in an embodiment of the present application, comprising a computer program / instruction, which, when executed by a processing component, implements the recommended content generation method as described in the first aspect above, the object recommendation method as described in the second aspect above, or the recommended content generation method as described in the third aspect above.
[0023] In an embodiment of the present application, for a recommendation request triggered by an object provider, the target object to be recommended and the target user to be recommended are first determined, and the content preference characteristics of the target user are determined. The content preference characteristics of the target user can be predicted based on the user feature data of the target user, so that based on the content preference characteristics and the object-related information of the target object, a content generation model can be used to generate recommendation materials that match the content preference characteristics. Finally, recommended content can be generated based on the recommendation material, and the recommended content can be sent to the target user in a targeted manner to achieve personalized recommendations for the target user, effectively improving the relevance and attractiveness of the recommended content, making the recommended content generated by combining user characteristics and generative artificial intelligence technology more accurate, which helps to further improve the object conversion rate.
[0024] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0026] Figure 1 A schematic diagram showing the structure of an embodiment of a content generation system provided by the present application is shown;
[0027] Figure 2 A flowchart of an embodiment of a method for generating recommended content provided by the present application is shown;
[0028] Figure 3 A flowchart of an embodiment of an object recommendation method provided by the present application is shown;
[0029] Figure 4 A flowchart of another embodiment of a method for generating recommended content provided by the present application is shown;
[0030] Figure 5 A schematic diagram of a display interface provided in an actual application according to an embodiment of the present application is shown;
[0031] Figure 6 A schematic diagram showing a display of recommended content provided in an actual application according to an embodiment of the present application is shown;
[0032] Figure 7 A schematic diagram of scene interaction in an actual application of an embodiment of the present application is shown;
[0033] Figure 8 A schematic diagram showing the structure of an embodiment of a recommended content generation device provided by the present application is shown;
[0034] Fig. 9 A schematic diagram showing the structure of an embodiment of an object recommendation device provided by the present application is shown;
[0035] Fig.10 A schematic diagram showing the structure of an embodiment of a recommended content generation device provided by the present application is shown;
[0036] Fig.11 A schematic diagram of the structure of an embodiment of a computing device provided by the present application is shown. DETAILED DESCRIPTION
[0037] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only 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 without making creative work are within the scope of protection of the present application.
[0038] It should be noted that, in the case where the embodiments of the present application involve user information, the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with the relevant laws, regulations and standards of the relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse. In addition, the various models involved in this application (including but not limited to language models or large models) are in compliance with relevant laws and standards.
[0039] In addition, it should be noted that, in the case where the embodiments of the present application involve user interaction operations or trigger operations, the user interaction operations or trigger operations involved in the embodiments of the present application include but are not limited to: touch operation, gesture operation, voice operation, head movement operation, eye movement operation and other interactive operations in various ways; among which, touch operation includes but is not limited to: click operation, double-click operation, long press operation, sliding operation, pinch operation or mouse hover operation, etc. Sliding operation includes but is not limited to: straight line sliding, curve sliding, etc.
[0040] The technical solution of the embodiment of the present application can be applied to the object recommendation scenario. With the vigorous development of Internet technology and computer technology, the scale of online users and the scale of data have shown rapid development. Some online interactive systems can provide objects such as goods, content, or web pages for users to consume. There are usually a large number of objects in the online interactive system. In order to improve the conversion rate of certain objects, the object provider often needs to rely on recommendation operations to achieve this, such as paying for the recommended content of the object to be placed on a specific promotion channel or specific promotion platform such as a third-party media platform provided by the online interactive system, so as to achieve the advertising effect and realize the object recommendation. The recommended content becomes the key information that affects the user interaction rate and consumption conversion rate.
[0041] In the traditional implementation method, the recommended content is mainly controlled by the object provider and is generated by the object provider based on personal preferences. However, in the process of implementing this application, the inventors found that different users have different needs. The recommended content with personal color produced by the object provider cannot meet the needs of different users, and requires the object provider to have certain design capabilities, which is costly and ultimately cannot guarantee the recommendation effect.
[0042] Therefore, in order to improve the effectiveness and accuracy of content and improve the recommendation effect, personalized recommended content has become the core means to improve the accuracy of content. Based on this, the inventor has proposed the technical solution of this application after a series of studies. In the embodiment of this application, combined with user characteristics and generative artificial intelligence technology, unique recommended content is automatically generated for users for the target object to be recommended, thereby accurately meeting the needs of different users, making the recommended content more accurate, improving the relevance and attractiveness of the content, and being able to bring better interaction rates, which helps to further improve the object conversion rate, and does not require the cumbersome production process of the object provider, reducing the content generation cost and improving the content generation efficiency. And the generated recommended content can be automatically delivered to achieve recommendation to the target user, realizing one-stop operations such as generation and delivery, and also simplifying the recommendation process to ensure user experience.
[0043] The object in this article may refer to a virtual object provided online for users to perform interactive behaviors, and the user in this article may refer to a consumer. For example, the object may be a commodity provided by an online trading system, and the commodity may refer to a tangible item or an intangible service, etc. Users may perform interactive behaviors such as browsing, collecting, adding to cart, and purchasing, and the object provider refers to the merchant.
[0044] It should be noted that although the present invention describes users and object providers, those skilled in the art will appreciate that users and object providers actually correspond to different user accounts, which can be obtained by pre-registration by users; user interaction operations are implemented based on user accounts, and corresponding data received or sent to users are also implemented based on user accounts, and actually the client corresponding to the user account receives or sends corresponding data to the server, etc.
[0045] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only 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 those skilled in the art without creative work are within the scope of protection of this application.
[0046] The technical solution of the embodiment of the present application can be applied to Figure 1 In the content generation system, the content generation system may include a first client 101 and a server 102. The technical solution of the embodiment of the present application is based on the fact that the content generation system can generate effective and accurate recommended content in a portable and efficient manner.
[0047] In actual applications, the content generation system can be connected to an online interactive system that provides object interaction, so as to obtain and process relevant data of the online interactive system, such as user feature data, object-related information, etc., and to deliver recommended content to the online interactive system. Of course, in an actual application, the server of the content generation system can be the server of the online interactive system, etc.
[0048] The first client 101 is oriented to the object provider, and can be used by the object provider to trigger a recommendation operation, etc. The first client 101 can be an object publishing end used by the object provider, which is used to publish objects in the server of the online interactive system for users to perform interactive behaviors, etc. Of course, in the embodiment of the present application, the first client 101 can be a client independent of the object publishing end, etc.
[0049] The online interactive system may include a second client 103 for users, and the recommended content generated by the server 102 may be delivered to the online interactive system, and the target user may view the content through the second client 103 .
[0050] Of course, when the content generation system is an online interactive system, the second client 103 is also included in the content generation system.
[0051] The first client and the second client can interact with the server through the network to receive or send messages, etc. For example, the first client can perceive the recommended operation of the object provider, etc., and send a corresponding recommendation request to the server; the second client can perceive the interactive behavior performed by the user, and send a corresponding interactive request to the server, etc.
[0052] The first client or the second client may be a browser, an APP (Application), or a web application such as an H5 (HyperText Markup Language 5, version 5 of Hypertext Markup Language) application, or a light application (also known as a mini-program, a lightweight application) or a cloud application, etc. The first user terminal or the second user terminal may be deployed in an electronic device and may rely on the device to run or some APPs in the device to run, etc. The electronic device may, for example, have a display screen and support information browsing, etc., such as a personal mobile terminal such as a mobile phone, a tablet computer, a personal computer, a desktop computer, a smart speaker, a smart watch, etc.
[0053] The above-mentioned server may include a server that provides various services, such as a server for generating recommended content, a server for publishing objects or processing user interaction behaviors, etc.
[0054] It should be noted that the server can be implemented as a distributed server cluster consisting of multiple servers, or as a single server. The server can also be a server of a distributed system, or a server combined with a blockchain. The server can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms, or an intelligent cloud computing server or intelligent cloud host with artificial intelligence technology.
[0055] It should be understood that Figure 1 The number of clients and servers in the figure is only for illustration. Any number of clients and servers may be provided according to the implementation requirements.
[0056] The implementation details of the technical solution of the embodiment of the present application are elaborated in detail below.
[0057] Figure 2 This is a flowchart of an embodiment of a method for generating recommended content provided in an embodiment of the present application. The technical solution of this embodiment can be Figure 1 The method may be performed by the server in the system architecture shown in the figure. The method may include the following steps:
[0058] 201: In response to a recommendation request triggered by an object provider, determine a target object to be recommended.
[0059] The recommendation request may be generated by the first client in response to a recommendation operation of the object provider. The server may send recommendation prompt information to the first client, and the recommendation operation may be triggered by the recommendation prompt information.
[0060] As an optional method, the target object may be any object corresponding to the object provider.
[0061] As another optional method, the target object can be an object that meets the recommendation condition corresponding to the object provider. The recommendation condition can be, for example, the latest published object. The latest published object often has a lower conversion rate due to the lack of user interaction records, so the latest published object can be recommended. In addition, the recommendation condition can also be, for example, an object pre-set by the object provider. The object provider can pre-set a list of objects to be recommended, and objects in the recommendation category can be used as target objects, etc.
[0062] As another optional manner, the target object may be an object specified in the recommendation request. For example, the recommendation request may include object identification type information, and the corresponding target object may be determined according to the object identification type information.
[0063] The recommendation prompt information can be specifically used to prompt the object provider to provide the target object, etc. Optionally, the recommendation prompt information can include selection prompt information of multiple objects corresponding to the object provider, so that the object provider can select the target object and trigger the recommendation operation, etc. Of course, the recommendation prompt information can also include input controls, etc., so that the object provider can input object identification information such as object identification or object address; wherein the object address can be linked to the corresponding object details page, etc. The object details page is a page provided by the online interactive system for introducing the object function, style, material and other information in detail to facilitate users to understand the relevant information of the object, and provide corresponding interactive controls such as purchase, collection, and add to cart to facilitate users to perform corresponding interactive behaviors.
[0064] 202: Determine content preference characteristics of the target user to be recommended.
[0065] Among them, the content preference characteristics can be predicted based on the user characteristic data of the target user.
[0066] The target user may be any registered user in the online interactive system, or any user among the registered users of the online interactive system who has a corresponding relationship with the object provider, wherein the object provider may correspond to multiple users, and the multiple users may be, for example, users who have a history of interaction with the object provided by the object provider, or may be users pre-configured for the object provider according to the recommended targets of the object provider, or may be users who match the category to which the object provider belongs; or may be users in the target population corresponding to the object provider, etc.
[0067] In addition, the target user may also be provided by the object provider, and the user characteristic data may also be provided by the object provider, etc. For example, when the target user is a registered user of the promotion platform, the object provider may also provide user characteristic data, etc., so as to determine the target user and predict the content preference characteristics.
[0068] Optionally, the content preference feature may be predicted using a feature prediction model.
[0069] The content preference feature may be generated and saved in advance by prediction; optionally, the content preference feature may also be generated in real time, so step 202 may include: obtaining user feature data of the target user; and using a feature prediction model to predict the content preference feature of the target user from the user feature data.
[0070] Among them, the user feature data may include user behavior data and / or user attribute data of the target user, wherein the user behavior data may include historical behavior data. In addition, when the target user is online, the user behavior data may also include real-time behavior data; wherein the online status may be that the target user is logged in, etc.
[0071] Therefore, the feature prediction model may be used to predict the content preference feature based on at least one of the historical behavior data, user attribute data, and real-time behavior data of the target user.
[0072] The historical behavior data may include object-related information of the historical operation object; the historical operation object may refer to an object that has historically performed any type of interactive behavior, such as purchase, browsing, collection, or add-to-purchase, etc. The historical operation object may refer to the object of the most recent operation or the object with the most historical operations, etc., or may refer to any one or more objects of the historical operation, etc., which is not limited in this application.
[0073] The real-time behavior data may include, for example, current time information, device-related information of the device in use, and / or object-related information of the current operation object. The device-related information may refer to, for example, the device type, etc. The current operation object may refer to the object on which the target user currently performs any type of interactive behavior, etc.
[0074] The user attribute data may include, for example, age, occupation, gender, location, etc.
[0075] The feature prediction model may be obtained by pre-training. Optionally, the feature prediction model may be obtained by training, for example, in the following manner:
[0076] Obtain sample user data and sample preference features corresponding to the sample user data; and train a feature prediction model using the sample user data and the sample preference features.
[0077] The feature prediction model can be any machine learning model, such as a neural network model, a deep learning model, etc., and this application does not limit this.
[0078] Among them, the sample user data and the sample preference features corresponding to the sample user data can be used as training data. When using a supervised training method, the sample user data can be used as a model input, and the sample preference features can be used as training labels to train a feature prediction model.
[0079] In addition, the content preference feature may also be the content preference feature of the target group to which the target user belongs. Therefore, as another optional method, step 202 may include: determining the target group corresponding to the target object; and using the content preference feature corresponding to the target group as the content preference feature of the target user belonging to the target group. The recommended content may be specifically used to be sent to the target group to recommend the target object to the target user belonging to the target group.
[0080] The target population may be any group pre-set by the online interactive system, and the content preference characteristics of the target population may be obtained by statistically analyzing user characteristic data of different users in the target population.
[0081] In addition, when the recommended content of the target object is used to be delivered to the promotion platform, the object provider can specify the promotion platform. Therefore, the recommendation request can include the platform identifier of the promotion platform, etc. The target population can also be the user group corresponding to the promotion platform. For example, the promotion platform can be a third-party media platform. The preference characteristics of user groups corresponding to different third-party media platforms may be different. The content preference characteristics of the target population can be determined according to the platform type of the promotion platform. Of course, the relevant personnel can also pre-configure the content preference characteristics of the target populations corresponding to different promotion platforms and directly search for them.
[0082] Among them, the above-mentioned content preference features may include style preference features, type preference features, visual preference features, and / or emotional preference features, etc.; style preference features may include, for example, two-dimensional, humanities, retro, technology, fashion, or simplicity, etc.; type preference features may include, for example, text, images, videos, etc.; visual preference features may include, for example, light pink, red, white, etc.; emotional preference features may include, for example, family affection, love, friendship, humor, inspiration, etc. It should be noted that this is only an example, and the present application is not limited thereto.
[0083] 203: Based on the content preference features and the object-related information of the target object, a content generation model is used to generate recommendation materials matching the content preference features.
[0084] Among them, the content generation model can be a generative artificial intelligence (AI) model, such as a generative adversarial network, a variational autoencoder, a recurrent neural network, a diffusion model, etc. In addition, the content generation model can be a large model, which refers to a machine learning model with a large number of parameters and a complex structure, capable of processing massive data and completing various complex tasks, such as language generation, computer vision, speech recognition, etc. The large model can be implemented using a large language model (English: Large Language Model, abbreviated: LLM) or a large multimodal model (English: Large Multimodal Model, abbreviated: LMM), etc. For example, a generative pre-trained model, a bidirectional encoder model, etc. can be used. This application does not limit this.
[0085] The content generation model may be implemented using a pre-trained model, and may be fine-tuned to adapt to content generation. Therefore, in some embodiments, the content generation model may be pre-trained in the following manner:
[0086] Obtain sample preference features, object-related information of sample objects, and sample materials corresponding to the sample objects; and train a content generation model based on the sample preference features, object-related information of sample objects, and sample materials corresponding to the sample objects.
[0087] There are many ways to implement fine-tuning training of large models, such as full fine-tuning (Fine-tuning), LoRA (Low-Rank Adaptation, low-rank adaptation), etc., and this application does not limit this. Sample preference features, object-related information of sample objects, and sample materials corresponding to sample objects are used as training data. In a supervised training method, sample preference features and object-related information of sample objects can be used as input data, and sample materials can be used as training labels to train content generation models.
[0088] The content preference features and the object-related information of the target object may be input into the content generation model to instruct the content generation model to generate recommendation materials that match the content preference features.
[0089] The object-related information may include object detail information in an object detail page, which may be in the form of images, videos and / or texts, and may include object characteristic data such as functions, uses, and other information.
[0090] 204: Generate recommended content for the target object based on the recommended material.
[0091] The recommended material may include material data of at least one media type among images, videos and texts. These material data may be directly used as recommended content, or may be fused to generate final recommended content.
[0092] The recommended content can be used to recommend the target object to the target user. Since the recommended content matches the content preference characteristics of the target user, it is more attractive to the target user and can significantly stimulate the target user to interact with the target object, thereby helping to improve the user interaction rate and purchase conversion rate. Therefore, the technical solution of this embodiment provides personalized recommended content, ensures the validity and accuracy of the content, can improve the recommendation effect, and improves the user experience.
[0093] Since actual users may also be affected by the current scene, which may affect the interactive behavior towards the object, etc., therefore, in addition to combining content preference features to generate recommended materials, current scene data may also be considered, such as holiday information, the geographic location of the target user, weather data corresponding to the current time, etc., so that the recommended materials meet the current scene requirements and further enhance the content relevance and attractiveness to the target user. Therefore, in some embodiments, the above-mentioned recommendation materials matching the content preference features generated by the content generation model based on the content preference features and the object-related information of the target object may include:
[0094] Based on the content preference features, the object-related information of the target object and the current scene data, a content generation model is used to generate recommendation materials that match the content preference features and the current scene data.
[0095] In some embodiments, combined with the above description, it can be seen that the content generation model can be a generative artificial intelligence macro model;
[0096] The above-mentioned generating the recommendation material matching the content preference feature and the current scene data by using the content generation model based on the content preference feature, the object related information of the target object and the current scene data may include:
[0097] Based on the content preference features, the object-related information of the target object and the current scene data, a prompt instruction is generated; and the prompt instruction is input into the content generation model to generate a recommended material matching the content preference features.
[0098] Prompts can be information input to the big model, which can be a natural language input. They mainly prompt the big model with the context of the input information and the parameter information of the input model, so as to prompt or guide the big model to give expected output, etc., to help the model better understand the input intention and respond accordingly. Prompts can also improve the interpretability and accessibility of the model.
[0099] Among them, the prompt instruction may be generated according to the target structure, and the target structure may mainly include introduction instructions, clear requirements, information provision, and expected output. For ease of understanding, taking the target object as a commodity and the recommended content as an advertising content as an example, an example of a prompt instruction is given below. It should be noted that this is only an example, and the actual application can be set in combination with actual needs. The prompt instruction may be, for example:
[0100] "You are a professional advertising content writing assistant. Now you are given a product details content (as shown below). I will also provide the user's content preference characteristics and the current scene information. Please generate an attractive advertising content for the product based on these contents. The advertising content needs to highlight the characteristics of the product that meet the user's preferences and closely fit the current scene, so as to attract the user's attention and stimulate the desire to buy.
[0101] Product details:
[0102] [XXX (detailed description of the product here, including the product's functions, materials, appearance, specifications, advantages, etc.)]
[0103] User content preference characteristics:
[0104] [XXX (clearly list the user's content preference characteristics here, such as two-dimensional, realistic, cartoon, etc.)]
[0105] Current scene:
[0106] [XXX (describe the specific scenario here, such as holidays, geographic location, weather data, etc.)]
[0107] Please carefully generate the advertising content according to the requirements, and the language should be concise, vivid, and appealing.
[0108] As can be seen from the foregoing description, the recommended material may include first material data of at least one media type. In addition, the content generation system of the embodiment of the present application may also provide a setting function for the object provider to set the corresponding material data. Therefore, the recommended content may also integrate the material data preset by the object provider. If the target object corresponds to a brand, the brand logo data may also be integrated into the recommended content. Therefore, in some embodiments, the above-mentioned generation of the recommended content of the target object based on the recommended material may include:
[0109] Determine first material data of at least one media type in the recommended material; determine second material data preset by the object provider for the target object; determine brand identification data corresponding to the target object; and fuse the first material data with the second material data and / or the brand identification data to generate recommended content.
[0110] That is, the first material data may be fused with the second material data, or the first material data may be fused with the brand logo data, or the first material data may be fused with the second material data and the brand logo data, etc., to generate recommended content.
[0111] Among them, the first material data, the second material data and the brand logo data can usually be divided into image data and text data, or video data and text data, and the fusion processing can be to superimpose the text data into the image data, or to superimpose the text data into the image frame of the video data.
[0112] Optionally, the first material data and the second material data and / or the brand logo data may be fused using a feature fusion model to generate recommended content. Of course, the fusion process may also be performed using an image processing algorithm.
[0113] The feature fusion model can be a machine learning model, such as various neural network models or artificial intelligence large models, etc., which is not limited in this application.
[0114] In some embodiments, the feature fusion model can be trained, for example, in the following manner:
[0115] Obtain image sample data, text sample data and fusion sample data; use the image sample data and text sample data as input data, and the fusion sample data as training labels to train the feature fusion model.
[0116] In some embodiments, after the recommended content is sent to the target user, the target user's feedback data may be detected so as to optimize the recommended content, etc., so as to continuously improve the accuracy of the content and the recommendation effect, etc. Therefore, the method may further include:
[0117] Detect the interactive behavior data triggered by the target user for the recommended content; determine the feedback data corresponding to the recommended content based on the interactive behavior data; and perform one or more of the following adjustment operations based on the feedback data:
[0118] Adjusting the user feature data to re-determine the content preference features of the target user to be recommended;
[0119] Adjusting the content generation model to re-execute the generation of recommendation materials matching the content preference characteristics using the content generation model;
[0120] Adjusting the feature prediction model to re-execute the prediction of the target user's content preference features from the user feature data using the feature prediction model;
[0121] Retrain the content generation model;
[0122] as well as,
[0123] Retrain the feature prediction model.
[0124] The interactive behavior data may include, for example, the number of clicks or click-through rate on the recommended content, etc. If the click-through rate is low, the feedback data may be negative feedback, and the above one or more adjustment operations may be performed when the feedback data is negative feedback.
[0125] In addition, in order to better understand the target user's feedback information on the recommended content, in some embodiments, determining the feedback data corresponding to the recommended content based on the interactive behavior data may include: sending feedback prompt information to the target user when the click-through rate corresponding to the interactive behavior data is less than a preset value; and determining the feedback data corresponding to the recommended content based on the target user's feedback request.
[0126] The feedback prompt information may, for example, prompt the target user to provide the matching degree between the recommended content and the target user, and may also prompt the target user to input the recommended data, and different matching degrees may correspond to different types of adjustment operations. In addition, optimization prompt information may be sent to relevant personnel such as operation and maintenance personnel in combination with the recommended data to manually optimize the user feature data, feature prediction model and / or content generation model involved.
[0127] Optionally, the above adjustment of user feature data may be, for example, obtaining user feature data generated in the most recent period of time, or when the user feature data includes user data of different dimensions, it may be obtaining user data of one or more dimensions, etc. The above adjustment of the content generation model may be, for example, adjusting model parameters of the content generation model, etc. The above adjustment of the feature prediction model may be, for example, adjusting model parameters of the feature prediction model, etc. The above retraining of the content generation model may be, for example, reacquiring training data and training the content generation model, etc. The above retraining of the feature prediction model may be, for example, reacquiring training data and training the feature prediction model, etc.
[0128] In some embodiments, the method may further include: determining content generation requirements corresponding to the object provider;
[0129] Then, based on the content preference features and the object-related information of the target object, the recommended materials matching the content preference features generated by the content generation model may include:
[0130] Based on the content preference characteristics, content generation requirements and object-related information of the target object, a content generation model is used to generate recommendation materials that match the content preference characteristics and content generation requirements.
[0131] That is, in addition to content preference features, the generation of recommended content can also take into account the content generation requirements of the object provider, so that the recommended content finally obtained can meet the personalized needs of the target users while also taking into account the personalized requirements of the object provider.
[0132] The content generation requirement may include, for example, image size, etc. The content generation requirement may be a specific requirement including a feature dimension different from the feature type of the content preference feature, etc. For example, when the content preference feature includes a style preference feature, the content generation requirement may include content type, color type, emotion type, etc. This application does not limit this.
[0133] As can be seen from the previous description, the current scene data may also be considered. Therefore, in some embodiments, the content generation model may be used to generate recommended materials that match the content preference characteristics, content generation requirements, current scene data, and object-related information of the target object.
[0134] In some embodiments, in combination with the foregoing description, it can be known that the object provider can trigger a recommendation request through the first client, so the above-mentioned determination of the target object to be recommended in response to the recommendation request triggered by the object provider may include: sending recommendation prompt information to the first client corresponding to the object provider; obtaining the recommendation request sent by the first client, and determining the target object; the recommendation request is generated based on the recommendation operation triggered by the object provider;
[0135] In some embodiments, the method may further include:
[0136] The recommendation target set by the object provider is obtained; and according to the recommendation target, a target user to be recommended is determined from a plurality of users corresponding to the object provider.
[0137] Among them, the recommendation goals may include, for example, increasing click-through rate or increasing exposure, and different recommendation goals may correspond to different numbers of users. Therefore, according to the recommendation goal, multiple users corresponding to the object provider configurator can be selected, and any one of the multiple users can be used as the target user to be recommended, etc., in order to determine the recommendation scope of the recommended content in combination with the recommendation goal of the object provider to ensure the recommendation effect, etc.
[0138] Figure 3 This is a flowchart of an embodiment of an object recommendation method provided in an embodiment of the present application. The technical solution of this embodiment can be executed by the server. The method may include the following steps:
[0139] 301: In response to a recommendation request triggered by an object provider, determine a target object to be recommended.
[0140] 302: Determine content preference characteristics of the target user to be recommended; wherein the content preference characteristics are predicted based on user characteristic data of the target user.
[0141] 303: Based on the content preference features and the object-related information of the target object, a content generation model is used to generate recommendation materials matching the content preference features.
[0142] 304: Generate recommended content for the target object based on the recommended material.
[0143] The operations of step 301 to step 304 can be found in detail in the operations of step 201 to step 204 in the above embodiment, and will not be described in detail here.
[0144] 305: Send the recommended content to the target user to recommend the target object to the target user.
[0145] Among them, the recommended content can be sent to the target user to achieve the purpose of recommending the target object to the target user. The technical solution of this embodiment can ensure the accuracy of the recommended content, improve the relevance and attractiveness of the recommended content to the target user, thereby encouraging the target user to interact with the target object, which can significantly improve the user interaction rate and conversion rate, etc.
[0146] Among them, the recommended content can be indexed to the object details page of the target object, and the click operation on the recommended content can trigger the output of the object details page to further facilitate the target user to perform corresponding interactive operations, such as purchase, etc.
[0147] As an optional method, sending the recommended content to the target user may include:
[0148] In response to an object search request triggered by a target user based on search information, if the target object matches the search information, a search result page is sent to the target user, and recommended content of the target object is displayed in the search result page.
[0149] That is, when the target object is an object recalled based on search information, the recommended content of the target object can be directly displayed in the search results page, making the recommended content of the target object more attractive, so as to stimulate the target user to perform further operations on the recommended content. For example, a click operation on the recommended content can trigger entry into the object details page of the target object.
[0150] Optionally, in actual applications, the recommended content of the target object may be displayed in a fixed list position in the search results page. Of course, the target object may also be ranked with other objects and the corresponding recommended content may be displayed in the search results page according to the ranking position. This application does not limit this.
[0151] The object search request may be sent by the target user through the second client, and the search result page may be sent to the second client so that the search result page is displayed in the second client and the recommended content is displayed in the search result page.
[0152] As another optional method, the above-mentioned sending of the recommended content to the target user may include:
[0153] In response to a page acquisition request triggered by a target user for an object recommendation page, the object recommendation page is sent to the target user, and recommended content of the target object is displayed on the object recommendation page.
[0154] Among them, the object recommendation page can be, for example, an object aggregation page for object promotion, etc. When the target user requests to obtain the object recommendation page, the recommended content of the target object can be displayed in the object recommendation page to implement personalized promotion operations, etc.
[0155] As another optional method, the above-mentioned sending of the recommended content to the target user may include:
[0156] In response to a page acquisition request triggered by a target user for a target object, an object detail page of the target object is sent to the target user, and recommended content is displayed in the object detail page.
[0157] That is, the recommended content can be displayed in the object details page, so that different users see different displayed content when entering the object details page of the same object. Displaying recommended content in the object details page can make the object details page more attractive, and can further encourage target users to perform interactive operations, such as purchase operations, so as to improve the purchase conversion rate, etc.
[0158] As another optional manner, the above-mentioned determination of the content preference characteristics of the target user to be recommended may include:
[0159] Determine the target population corresponding to the promotion platform; use the content preference characteristics corresponding to the target population as the content preference characteristics of the target users belonging to the target population;
[0160] Then the above-mentioned sending of recommended content to target users may include:
[0161] Place the recommended content on the promotion platform.
[0162] That is, the content preference characteristics of the target user can be the content preference characteristics of the target population to which he belongs. When the target population is the target population corresponding to the promotion platform, the recommended content can be delivered to the third-party media platform.
[0163] Among them, in actual applications, the promotion platform may refer to, for example, a third-party media platform. The third-party media platform may provide information publishing and dissemination functions and is highly interactive and social. It is a media platform where users can share their lives, obtain information, participate in discussions, etc. It may be a social media platform, a short video media platform, etc.
[0164] The recommendation request triggered by the object provider may include, for example, a platform identifier to determine the promotion platform for the recommended content delivery.
[0165] Figure 4 This is a flowchart of another embodiment of a method for generating recommended content provided in an embodiment of the present application. The technical solution of this embodiment can be executed by a first client. The method may include the following steps:
[0166] 401: Displaying recommendation prompt information in the display interface.
[0167] 402: The detection object provider sends a recommendation request to the server in response to the recommendation operation triggered by the recommendation prompt information.
[0168] Among them, the recommendation request can be used to determine the target object to be recommended and the content preference characteristics of the target user to be recommended; the content preference characteristics and the object-related information of the target object are used to generate recommendation materials matching the content preference characteristics using the content generation model; the recommendation materials are used to generate recommended content for the target object; wherein, the content preference characteristics are predicted based on the user characteristic data of the target user.
[0169] The specific generation operation of the recommended content can be found in the above-mentioned related embodiments, and will not be repeated here.
[0170] Through the technical solution of this embodiment, the object provider can trigger the automatic generation of recommended content through the first client, and ensure the accuracy of the recommended content to meet the personalized needs of different users.
[0171] In some embodiments, the method may further include:
[0172] Displaying content requirement prompt information in a display interface; determining content generation requirements in response to a setting operation triggered by the object provider for the content requirement prompt information, and sending the content generation requirements to a server.
[0173] The server may generate recommendation materials matching the content preference features and content generation requirements using a content generation model based on the content preference features, content generation requirements and object-related information of the target object.
[0174] For ease of understanding, Figure 5 The following is a schematic diagram showing a display interface provided by the first client provided in an embodiment of the present application. It should be noted that: Figure 5 This is just an example, and the specific interface style can be set according to the actual situation, and this application does not limit this.
[0175] like Figure 5 As shown, in an online transaction scenario, the object may refer to a commodity, and the object provider may be a merchant. In the display interface 500, recommendation prompt information may be displayed, and the recommendation prompt information may include, for example, selection prompt information 501 for multiple commodities, and may also include an input control 502 for the merchant to input commodity identification type information such as a commodity address. The commodity address may refer to, for example, a link address of a commodity details page. In addition, the recommendation prompt information may also include confirmation prompt information 503, so as to trigger the generation of a recommendation request based on the target commodity provided by the merchant in response to the triggering operation of the confirmation prompt information, and the recommendation request is sent to the server, and the server can generate recommendation content for the target user and automatically send the recommendation content to the target user.
[0176] Through the technical solution of the embodiment of the present application, merchants only need to select target products to automatically generate personalized recommendation content and automatically deliver them to achieve recommendations to target users, realizing one-stop operations such as generation and delivery, simplifying the recommendation process, and ensuring the merchant experience.
[0177] In addition, for ease of understanding, Figure 6 A schematic diagram showing a display of recommended content is shown. It should be noted that: Figure 6 This is just an example, and the present application is not limited thereto. The object-related information is assumed to include an object image 601 of the target object, and may also include text information or video, etc., which will not be demonstrated. Assuming that the content preference feature of the target user is a two-dimensional style, the generated recommended material may include, for example, a material image 602, which is a situational image in a natural element style and generated based on the object-related information. The material image 602 may be fused with the material copy to obtain the corresponding recommended content 603, etc.
[0178] In a practical application, the technical solution of the embodiment of the present application can be applied to an online transaction scenario, where the object provider is a merchant, the object is the product sold by the merchant, and the user is a consumer. The merchant can generate personalized recommendation content for the product with the help of the content generation system provided by the present application. The technical solution of the embodiment of the present application is introduced below by taking an online transaction scenario as an example.
[0179] like Figure 7 As shown, the merchant can recommend a target through the first client 101 , and can specify a target product, and trigger a recommendation request to the server 102 .
[0180] The server 102 can determine the target user to be recommended and collect user feature data, such as historical behavior data of the target user such as browsing, purchasing, or clicking, and real-time behavior data of the target user such as browsing, purchasing, or clicking, as well as user attribute data such as age and gender. In addition, the dynamic environment data of the target user such as location, current time period, etc. can also be detected. Based on the user feature data, the content preference characteristics of the target user are predicted using a feature prediction model (step 701).
[0181] Afterwards, the server 102 can use the pre-trained content generation model to generate recommendation materials that are highly matched with the content preference features based on the content preference features, the product-related information of the target product, the current scene data, etc. (step 702). The recommendation materials can include, for example, product display scene images or situational advertising images, etc. The media type of the recommendation material can be determined in combination with the content type preferred by the target user, etc.
[0182] Afterwards, the server 102 may merge the image type material data with the text type material data to generate recommended content (step 703). The text type material data may include, for example, the reason for recommending the target product, product description, etc., which may be pre-set by the merchant or generated by the content generation model.
[0183] After the server 102 generates the recommended content, the recommended content can be sent to the target user (step 704). For example, when it is detected that the target user has issued a page acquisition request triggered by the second client 103, an object recommendation page including recommended content of the target product can be sent to the second client 103, so that the second client 103 can display the object recommendation page and display the recommended content in the object recommendation page, so as to achieve the purpose of recommending the target product to the target user.
[0184] The server 102 can also detect the target user's interactive behaviors such as clicking, browsing, etc. on the recommended content, so as to determine the feedback data for the recommended content based on the interactive behavior data, and thus optimize the recommended content in combination with the feedback data (step 705), such as adjusting model parameters, retraining the model, or adjusting related data.
[0185] Through the technical solution of the embodiment of the present application, combined with user characteristics and generative artificial intelligence technology, unique recommendation content is automatically generated for users for the target products to be recommended, thereby accurately meeting the needs of different users, making the recommended content more accurate, improving the relevance and attractiveness of the content, and bringing better interaction rates, which helps to further improve the conversion rate of goods, and does not require the cumbersome production process of merchants, reducing the cost of content generation and improving the efficiency of content generation. In addition, the generated recommended content can be automatically delivered to achieve recommendations to target users, realizing one-stop operations such as generation and delivery, simplifying the recommendation process, and ensuring the experience of users and merchants.
[0186] It should be noted that in some of the processes described in the above embodiments and the accompanying drawings, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel, and the sequence numbers of the operations, such as 201, 202, etc., are only used to distinguish between different operations, and the sequence numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.
[0187] Figure 8A schematic diagram of a structure of an embodiment of a recommended content generation device provided in an embodiment of the present application, the device may include:
[0188] The object determination module 801 is used to determine the target object to be recommended in response to the recommendation request triggered by the object provider;
[0189] A feature determination module 802 is used to determine the content preference features of the target user to be recommended; wherein the content preference features are predicted based on the user feature data of the target user;
[0190] A material generation module 803 is used to generate a recommended material matching the content preference characteristics using a content generation model based on the content preference characteristics and the object related information of the target object;
[0191] The content generation module 804 is used to generate recommended content of the target object based on the recommended material; wherein the recommended content is used to recommend the target object to the target user.
[0192] In some embodiments, the feature determination module may be specifically used to obtain user feature data of a target user; and use a feature prediction model to predict the content preference features of the target user from the user feature data.
[0193] In some embodiments, the content generation model is a generative artificial intelligence macro model;
[0194] The material generation module can be specifically used to generate prompt instructions based on content preference features, object-related information of the target object and current scene data; and input the prompt instructions into the content generation model to generate recommended materials that match the content preference features and the current scene data.
[0195] In some embodiments, the content generation module can be specifically used to: determine first material data of at least one media type in the recommended material; determine second material data preset by the object provider for the target object; determine brand identification data corresponding to the target object; and fuse the first material data with the second material data and / or the brand identification data to generate recommended content.
[0196] In some embodiments, the device may further include:
[0197] A first model training module is used to obtain sample preference features, object-related information of sample objects, and sample materials corresponding to the sample objects; based on the sample preference features, object-related information of sample objects, and sample materials corresponding to the sample objects, a training content generation model is generated;
[0198] The second model training module is used to obtain sample user data and sample preference features corresponding to the sample user data; and train a feature prediction model using the sample user data and the sample preference features.
[0199] In some embodiments, the device may further include:
[0200] An adjustment and optimization module is used to detect the interactive behavior data triggered by the target user for the recommended content; determine the feedback data corresponding to the recommended content based on the interactive behavior data; adjust the user feature data in combination with the feedback data to re-execute the determination of the content preference characteristics of the target user to be recommended; or adjust the content generation model to re-execute the use of the content generation model to generate recommended materials that match the content preference characteristics; or adjust the feature prediction model to re-execute the use of the feature prediction model to predict the content preference characteristics of the target user from the user feature data; or retrain the content generation model; or retrain the feature prediction model.
[0201] In some embodiments, the feature determination module can be specifically used to determine the target population corresponding to the target object; use the content preference features corresponding to the target population as the content preference features of the target users belonging to the target population; wherein the recommended content is used to be sent to the target population to recommend the target object to the target users belonging to the target population.
[0202] In some embodiments, the feature determination module obtaining user feature data of the target user may include: obtaining historical behavior data, user attribute data and real-time behavior data of the target user; wherein the historical behavior data includes object-related information of historical operation objects; the real-time behavior data includes current time information, device-related information of the device used, and / or object-related information of the current operation object.
[0203] In some embodiments, the device may further include:
[0204] A requirement determination module, used to determine the content generation requirements corresponding to the object provider;
[0205] The material generation module can be specifically used to generate recommended materials matching the content preference characteristics and content generation requirements using a content generation model based on content preference characteristics, content generation requirements and object-related information of a target object.
[0206] In some embodiments, the object determination module may be specifically used to send recommendation prompt information to the first client corresponding to the object provider; obtain a recommendation request sent by the first client to determine the target object; and generate the recommendation request based on the recommendation operation triggered by the object provider.
[0207] In some embodiments, the device may further include:
[0208] The user determination module is used to obtain the recommendation target set by the object provider; according to the recommendation target, determine the target user to be recommended from multiple users corresponding to the object provider.
[0209] Figure 8 The recommended content generating device can execute Figure 2 The implementation principle and technical effects of the recommended content generation method described in the illustrated embodiment will not be described in detail. The specific manner in which each module and unit performs operations in the recommended content generation device in the above embodiment has been described in detail in the embodiment of the method, and will not be described in detail here.
[0210] Fig. 9 A schematic diagram of the structure of an embodiment of an object recommendation device provided in an embodiment of the present application, the device may include:
[0211] The object determination module 901 is used to determine the target object to be recommended in response to the recommendation request triggered by the object provider;
[0212] A feature determination module 902 is used to determine the content preference features of the target user to be recommended; wherein the content preference features are predicted based on the user feature data of the target user;
[0213] A material generation module 903 is used to generate a recommended material matching the content preference characteristics by using a content generation model based on the content preference characteristics and the object related information of the target object;
[0214] The content generation module 904 is used to generate recommended content of the target object based on the recommended material; wherein the recommended content is used to recommend the target object to the target user.
[0215] The content recommendation module 905 is used to send the recommended content to the target user to recommend the target object to the target user.
[0216] In some embodiments, the content recommendation module may be specifically used to:
[0217] In response to an object search request triggered by a target user based on search information, if the target object matches the search information, a search result page is sent to the target user, and recommended content of the target object is displayed on the search result page;
[0218] or,
[0219] In response to a page acquisition request triggered by a target user for an object recommendation page, the object recommendation page is sent to the target user, and recommended content of the target object is displayed on the object recommendation page;
[0220] or,
[0221] In response to a page acquisition request triggered by a target user for a target object, an object detail page of the target object is sent to the target user, and recommended content is displayed in the object detail page.
[0222] In some embodiments, the feature determination module may be specifically used to determine a target population corresponding to the promotion platform; and use the content preference features corresponding to the target population as content preference features of target users belonging to the target population.
[0223] The content recommendation module can be specifically used to deliver recommended content to the promotion platform.
[0224] Fig. 9 The object recommendation device can execute Figure 3 The implementation principle and technical effect of the object recommendation method described in the illustrated embodiment will not be described in detail. The specific manner in which each module and unit performs operations in the object recommendation device in the above embodiment has been described in detail in the embodiment of the method, and will not be described in detail here.
[0225] Fig.10 A schematic diagram of a structure of another embodiment of a recommended content generation device provided in an embodiment of the present application, the device may include:
[0226] Display module 1001, used to display recommendation prompt information in the display interface;
[0227] The recommendation trigger module 1002 is used to detect the recommendation operation triggered by the object provider in response to the recommendation prompt information, and send a recommendation request to the server; the recommendation request is used to determine the target object to be recommended and the content preference characteristics of the target user to be recommended; the content preference characteristics and the object-related information of the target object are used to generate recommendation materials matching the content preference characteristics using the content generation model; the recommendation materials are used to generate recommended content for the target object; wherein the content preference characteristics are predicted based on the user characteristic data of the target user.
[0228] Fig.10 The recommended content generating device can execute Figure 4 The implementation principle and technical effects of the recommended content generation method described in the illustrated embodiment will not be described in detail. The specific manner in which each module and unit performs operations in the recommended content generation device in the above embodiment has been described in detail in the embodiment of the method, and will not be described in detail here.
[0229] Fig.11 A schematic diagram of a structure of an embodiment of a computing device provided in an embodiment of the present application, such as Fig.11 As described above, the computing device may include a storage component 1101 and a processing component 1102 .
[0230] Storage component 1101 is used to store computer programs and can be configured to store various other data to support operations on the computing platform. Examples of such data include instructions, data structures, contact data, phone book data, messages, pictures, videos, etc. for any application or method operating on the computing platform.
[0231] The processing component 1102 is coupled to the storage component 1101 and is used to execute the computer program in the storage component 1101 to implement the recommended content generation method in the above-mentioned related embodiments or the object recommendation method in the above-mentioned related embodiments.
[0232] Further, if Fig.11 As shown, the computing device may also include: a communication component 1103, a display component 1104, a power component 1105, an audio component 1106 and other components. Fig.11 Only some components are shown schematically, and it does not mean that the computing device only includes Fig.11 In addition, Fig.11 The components in the dashed box are optional components, not mandatory components, and the specific components depend on the product form of the working node. The working node of this embodiment can be implemented as a terminal device such as a desktop computer, a laptop computer, a smart phone, or an IOT device, or a server-side device such as a conventional server, a cloud server, or a server array. If the working node of this embodiment is implemented as a terminal device such as a desktop computer, a laptop computer, a smart phone, etc., it can include Fig.11 If the working node of this embodiment is implemented as a server device such as a conventional server, a cloud server or a server array, it may not include Fig.11 Components within the dashed box.
[0233] The above-mentioned storage components can be implemented by any type of volatile or non-volatile storage devices or their combinations, such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), erasable programmable read only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.
[0234] The communication component is configured to facilitate wired or wireless communication between the device where the communication component is located and other devices. The device where the communication component is located can access a wireless network based on a communication standard, such as a mobile communication network such as 2G, 3G, 4G / LTE, 5G, or a combination thereof. In an exemplary embodiment, the communication component receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel.
[0235] The above-mentioned display component includes a screen, and the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation.
[0236] The power supply assembly provides power to various components of the device where the power supply assembly is located. The power supply assembly may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the device where the power supply assembly is located.
[0237] The above-mentioned audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (Microphone, MIC), and when the device where the audio component is located is in an operating mode, such as a call mode, a recording mode, and a speech recognition mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in a memory or sent via a communication component. In some embodiments, the audio component also includes a speaker for outputting an audio signal.
[0238] Accordingly, the embodiment of the present application also provides a computer-readable storage medium storing a computer program, and when the computer program is executed by the processing component, the processing component is enabled to implement each step in the above method embodiment. Among them, the computer-readable storage medium includes volatile or non-volatile or a combination thereof, and can be removable or non-removable. Examples of computer-readable storage media include, but are not limited to, phase-change random access memory (Phase-change Random Access Memory, PRAM), static random access memory (SRAM), dynamic random access memory (Dynamic Random Access Memory, DRAM), other types of random access memory (Random-Access Memory, RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), flash memory or other memory technology, read-only compact disc read-only memory (CD-ROM), digital versatile disc (Digital Video Disc, DVD) or other optical storage, magnetic cassette, tape disk storage or other magnetic storage device or any other non-transmission medium
[0239] Accordingly, the present application embodiment also provides a computer program product, the computer program product includes a computer program or an instruction, when the computer program or the instruction is executed by the processing component, the processing component is enabled to implement each step in the above method embodiment. It should be understood that each process or a combination of multiple processes in the above method flow can be implemented by a computer program or an instruction. In addition, these computer programs or instructions can be applied to a processing component of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device, so that the processor of the general-purpose computer, the special-purpose computer, the embedded processor, or other programmable data processing device can be implemented as a device for implementing the corresponding functions in the above method embodiment.
[0240] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.
[0241] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.
[0242] Finally, it should be noted that the above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.
Claims
1. A method for generating recommended content, characterized in that: include: In response to a recommendation request triggered by an object provider, determining a target object to be recommended; Determining content preference characteristics of a target user to be recommended; wherein the content preference characteristics are predicted based on user feature data of the target user; Based on the content preference feature and the object-related information of the target object, using a content generation model to generate a recommendation material matching the content preference feature; Based on the recommendation material, recommendation content for the target object is generated; wherein the recommendation content is used to recommend the target object to the target user.
2. The method according to claim 1, characterized in that Determining the content preference characteristics of the target user to be recommended includes: Acquire user characteristic data of the target user; The feature prediction model is used to predict the content preference features of the target user from the user feature data.
3. The method according to claim 1, characterized in that The content generation model is a generative artificial intelligence large model; The generating of the recommendation material matching the content preference feature by using the content generation model based on the content preference feature and the object related information of the target object comprises: Generate a prompt instruction based on the content preference feature, the object-related information of the target object, and the current scene data; The prompt instruction is input into the content generation model to generate a recommendation material matching the content preference feature and the current scene data.
4. The method according to claim 1, characterized in that: The generating the recommended content of the target object based on the recommended material includes: Determine first material data of at least one media type in the recommended materials; Determine second material data preset by the object provider for the target object; Determining brand identification data corresponding to the target object; The first material data is fused with the second material data and / or brand logo data to generate recommended content.
5. The method according to claim 2, characterized in that: The content generation model is trained in the following manner: Acquire sample preference characteristics, object-related information of sample objects, and sample materials corresponding to the sample objects; Training the content generation model based on the sample preference features, object-related information of the sample object, and sample materials corresponding to the sample object; The feature prediction model is trained in the following way: Obtaining sample user data and sample preference features corresponding to the sample user data; The feature prediction model is trained using the sample user data and the sample preference features.
6. The method according to claim 5, characterized in that Also includes: Detecting interactive behavior data triggered by the target user in response to the recommended content; Determining feedback data corresponding to the recommended content according to the interactive behavior data; In combination with the feedback data, adjusting the user characteristic data to re-execute the step of determining the content preference characteristics of the target user to be recommended; Or adjust the content generation model to re-execute the step of using the content generation model to generate recommendation materials matching the content preference features; or adjust the feature prediction model to re-execute the step of using the feature prediction model to predict the content preference features of the target user from the user feature data; or retrain the content generation model; or retrain the feature prediction model.
7. The method according to claim 1, characterized in that Determining the content preference characteristics of the target user to be recommended includes: Determine a target population corresponding to the target object; Using the content preference features corresponding to the target population as content preference features of target users belonging to the target population; The recommended content is used to be sent to the target group, so as to recommend the target object to the target users belonging to the target group.
8. The method according to claim 2, characterized in that: The acquiring of user characteristic data of the target user comprises: Acquire the historical behavior data, user attribute data and real-time behavior data of the target user; wherein the historical behavior data includes object-related information of the historical operation object; the real-time behavior data includes current time information, device-related information of the device used, and / or object-related information of the current operation object.
9. The method according to claim 1, characterized in that: Also includes: Determining content generation requirements corresponding to the object provider; The generating of the recommendation material matching the content preference feature by using the content generation model based on the content preference feature and the object related information of the target object comprises: Based on the content preference feature, the content generation requirement and the object related information of the target object, a content generation model is used to generate recommendation materials matching the content preference feature and the content generation requirement.
10. The method according to claim 1, characterized in that The step of determining the target object to be recommended in response to the recommendation request triggered by the object provider includes: Sending recommendation prompt information to the first client corresponding to the object provider; Acquire the recommendation request sent by the first client, and determine the target object; the recommendation request is generated based on the recommendation operation triggered by the object provider; The method further comprises: Obtaining the recommended target set by the object provider; According to the recommendation target, a target user to be recommended is determined from a plurality of users corresponding to the object provider.
11. An object recommendation method, characterized in that: include: In response to a recommendation request triggered by an object provider, determining a target object to be recommended; Determining content preference characteristics of a target user to be recommended; wherein the content preference characteristics are predicted based on user feature data of the target user; Based on the content preference feature and the object-related information of the target object, using a content generation model to generate a recommendation material matching the content preference feature; Based on the recommendation material, generating recommendation content for the target object; The recommended content is sent to the target user to recommend the target object to the target user.
12. The method according to claim 11, characterized in that The sending the recommended content to the target user comprises: In response to an object search request triggered by the target user based on search information, if the target object matches the search information, a search result page is sent to the target user, and recommended content of the target object is displayed on the search result page; or, In response to a page acquisition request triggered by the target user for an object recommendation page, the object recommendation page is sent to the target user, and recommended content of the target object is displayed on the object recommendation page; or, In response to a page acquisition request triggered by the target user for the target object, an object details page of the target object is sent to the target user, and the recommended content is displayed in the object details page.
13. The method according to claim 11, characterized in that Determining the content preference characteristics of the target user to be recommended includes: Determine the target group for the promotion platform; Using the content preference features corresponding to the target population as content preference features of target users belonging to the target population; The sending the recommended content to the target user comprises: The recommended content is delivered to the promotion platform.
14. A method for generating recommended content, characterized in that: include: Displaying recommended prompt information in the display interface; The detection object provider sends a recommendation request to the server in response to the recommendation operation triggered by the recommendation prompt information; The recommendation request is used to determine the target object to be recommended and the content preference characteristics of the target user to be recommended; The content preference features and the object-related information of the target object are used to generate recommendation materials matching the content preference features using a content generation model; the recommendation materials are used to generate recommended content for the target object; wherein the content preference features are predicted based on user feature data of the target user.
15. A computing device, characterized in that: including a processing component and a storage component; The storage component stores a computer program; the processing component is coupled to the storage component and is used to execute the computer program in the storage to implement the recommended content generation method as described in any one of claims 1 to 10 or the object recommendation method as described in any one of claims 11 to 13 or the recommended content generation method as described in claim 14.
16. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by the processing component, the recommended content generating method according to any one of claims 1 to 10 or the object recommendation method according to any one of claims 11 to 13 or the recommended content generating method according to claim 14 is implemented.
17. A computer program product, characterized in that It includes a computer program / instruction, which, when executed by a processing component, implements the recommended content generation method according to any one of claims 1 to 10, the object recommendation method according to any one of claims 11 to 13, or the recommended content generation method according to claim 14.
Citation Information
Cited By
Recommended content generation method, recommended content generation model training method, recommended content generation model training device, computer equipment, medium and program product
CN121765143A