Recommendation system, recommendation content display and generation method, device, equipment and product

Abstract

The invention discloses a recommendation system, a recommendation content display and generation method, device, equipment and product, and relates to the technical field of machine learning. The system comprises a server, a first client and a second client, a first account is logged in the first client, and a second account is logged in the second client; the server is used for sending first recommendation information to the first client based on the first account; and sending second recommendation information to the second client based on the second account, the first client is used for displaying a first content recommendation interface comprising a plurality of first recommendation contents and first interface elements which are displayed in an aggregated manner as a splash screen interface based on the first recommendation information; the second client is used for displaying a second content recommendation interface comprising a plurality of second recommendation contents and second interface elements which are displayed in an aggregated manner as a splash screen interface based on the second recommendation information; wherein the first recommendation information is different from the second recommendation information, and personalized aggregation recommendation of thousands of people can be provided.

Description

Technical Field

[0001] The present application relates to the field of machine learning technology, and in particular to a recommendation system, a method, apparatus, device, and product for displaying and generating recommended content. Background Art

[0002] With the development of computer technology and the Internet, more and more advertisers choose to place advertisements on online platforms (such as video platforms, novel reading platforms, etc.).

[0003] In related technologies, the content of an advertisement is provided by the advertiser. After the advertiser designs the content of the advertisement, the designed advertisement content is sent to the platform for delivery. Users of the platform can receive and view the advertisement content delivered by the advertiser during the use of the platform, and different users receive the same advertisement content delivered by the same advertiser.

[0004] However, in the above scheme, the advertising content is designed and delivered by the advertiser, and the advertising content and advertising format are relatively simple, which is not attractive enough to users. Users are likely to ignore the advertising content, resulting in a waste of advertising resources. The independent delivery of single advertising content also increases the advertising cost, and the content recommendation efficiency is low. Summary of the Invention

[0005] The embodiments of the present application provide a recommendation system, a method, apparatus, device, and product for displaying and generating recommended content, which can provide personalized recommendations to users. The technical solution is as follows.

[0006] In one aspect, a recommendation system is provided, comprising a server, a first client and a second client, wherein the first client is logged in with a first account, and the second client is logged in with a second account;

[0007] The server is configured to send first recommendation information to the first client based on the first account; and send second recommendation information to the second client based on the second account;

[0008] The first client is configured to display a first content recommendation interface as the splash screen interface based on the first recommendation information when the first client is started to display the splash screen interface, wherein the first content recommendation interface includes a plurality of first recommended contents displayed in an aggregated manner and a first interface element other than the first recommended contents;

[0009] the second client being configured to display a second content recommendation interface as the splash screen interface based on the second recommendation information when the second client is started to display the splash screen interface, the second content recommendation interface including a plurality of second recommended contents displayed in an aggregated manner and second interface elements other than the second recommended contents;

[0010] The first recommendation information is different from the second recommendation information.

[0011] In another aspect, a method for displaying recommended content is provided. The method is executed by a first client logged in to a first account, and the method includes:

[0012] Obtain first recommendation information from the server;

[0013] When the first client is started to display a splash screen interface, a first content recommendation interface is displayed as the splash screen interface based on the first recommendation information, wherein the first content recommendation interface includes a plurality of first recommended contents displayed in an aggregated manner and a first interface element other than the first recommended contents;

[0014] The first recommendation information is personalized content for the first account.

[0015] In another aspect, a method for generating recommended content is provided, the method being executed by a server and comprising:

[0016] In response to receiving a recommendation request from a first account, aggregating a plurality of first recommended contents based on personalization of the first account, wherein the recommendation request is a request sent when a first client logged in with the first account starts and displays a splash screen interface;

[0017] generating a first interface element based on the first account and at least one of the plurality of first recommended contents;

[0018] Send first recommendation information to the first client corresponding to the first account, where the first recommendation information is used to provide relevant information of a first content recommendation interface displayed by the first client, where the first content recommendation interface includes the multiple first recommended contents displayed in an aggregated manner and the first interface elements other than the first recommended contents.

[0019] In another aspect, a device for displaying recommended content is provided, the device comprising:

[0020] A receiving module, configured to obtain first recommendation information from a server;

[0021] a display module configured to display a first content recommendation interface as the splash screen interface based on the first recommendation information when the first client is started to display the splash screen interface, wherein the first content recommendation interface includes a plurality of first recommended contents displayed in an aggregated manner and first interface elements other than the first recommended contents;

[0022] The multiple first recommendation information are personalized content for the first account.

[0023] In another aspect, a device for generating recommended content is provided, the device comprising:

[0024] a processing module configured to, in response to receiving a recommendation request from a first account, aggregate a plurality of first recommended contents based on the personalized first account, wherein the recommendation request is a request sent when a first client logged in with the first account is started and displays a splash screen interface;

[0025] The processing module is further configured to generate a first interface element based on the first account and at least one of the plurality of first recommended contents;

[0026] A sending module is used to send first recommendation information to the first client corresponding to the first account, where the first recommendation information is used to provide relevant information of a first content recommendation interface displayed by the first client, where the first content recommendation interface includes the multiple first recommended contents displayed in an aggregated manner and the first interface elements other than the first recommended contents.

[0027] On the other hand, a computer device is provided, comprising a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set, or an instruction set, and the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by the processor to implement the method for displaying recommended content and / or the method for generating recommended content as described in any of the above-mentioned embodiments of the present application.

[0028] On the other hand, a computer-readable storage medium is provided, wherein the storage medium stores at least one instruction, at least one program, a code set, or an instruction set, and the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by a processor to implement the method for displaying recommended content and / or the method for generating recommended content as described in any of the above-mentioned embodiments of the present application.

[0029] In another aspect, a computer program product or computer program is provided. The computer program product or computer program includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the method for displaying recommended content and / or the method for generating recommended content described in any of the above embodiments.

[0030] The beneficial effects of the technical solutions provided in the embodiments of the present application include at least:

[0031] In the process of recommending content, the server performs personalized individual analysis on different accounts to determine multiple recommended content and interface elements corresponding to different accounts, such as: after analyzing the first account, multiple first recommended content and first interface elements are obtained, and after analyzing the second account, multiple second recommended content and second interface elements are obtained. Among them, the recommendation information used to indicate the above-mentioned multiple recommended content and interface elements is determined based on the personalized analysis of the account, so the aggregated content of the recommended content is different, and a personalized display scheme based on multiple aggregated content and / or interface elements in the content recommendation interface for different accounts is implemented, providing users with personalized recommendations for thousands of people, which can improve the fit between the content recommendation interface and the user's interest tendencies and increase its attractiveness to users. At the same time, the richness of the recommended content in the content recommendation interface is increased by the aggregated display of multiple recommended contents, thereby improving the recommendation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0033] Figure 1 is a schematic diagram of a recommendation system provided by an exemplary embodiment of the present application;

[0034] Figure 2 is an interactive flow chart of a recommendation system provided by an exemplary embodiment of the present application;

[0035] Figure 3 This is a comparative diagram of a content recommendation interface provided by an exemplary embodiment of the present application;

[0036] Figure 4 is a flowchart of a method for displaying recommended content provided by an exemplary embodiment of the present application;

[0037] Figure 5 is a comparative schematic diagram of a content recommendation interface provided by another exemplary embodiment of the present application;

[0038] Figure 6 This is a flowchart of a method for displaying interactive controls based on a content recommendation interface provided by an exemplary embodiment of the present application;

[0039] Figure 7 This is a schematic diagram of element animation effects provided by an exemplary embodiment of the present application;

[0040] Figure 8is a flowchart of an aggregate recommendation method provided by an exemplary embodiment of the present application;

[0041] Figure 9 This is a schematic diagram of a method for displaying sub-aggregate recommendations provided by an exemplary embodiment of the present application;

[0042] Figure 10 This is a schematic diagram of a method for displaying content decoration elements provided by an exemplary embodiment of the present application;

[0043] Figure 11 This is a schematic diagram of an awarding animation display provided by an exemplary embodiment of the present application;

[0044] Figure 12 is a flow chart of a content occlusion interaction method provided by an exemplary embodiment of the present application;

[0045] Figure 13 is a schematic diagram of an interactive method for occluding display provided by an exemplary embodiment of the present application;

[0046] Figure 14 is a flowchart of a content recommendation method based on topic switching provided by an exemplary embodiment of the present application;

[0047] Figure 15 This is a schematic diagram of a theme switching method provided by an exemplary embodiment of the present application;

[0048] Figure 16 is a flowchart of a method for generating recommended content provided by an exemplary embodiment of the present application;

[0049] Figure 17 This is a flow chart of a personalized content aggregation method provided by an exemplary embodiment of the present application;

[0050] Figure 18 is an interactive flow chart of a content recommendation method provided by an exemplary embodiment of the present application;

[0051] Figure 19 is a structural block diagram of a device for displaying recommended content provided by an exemplary embodiment of the present application;

[0052] Figure 20 This is a structural block diagram of a device for generating recommended content provided by an exemplary embodiment of the present application;

[0053] Figure 21 It is a structural block diagram of a terminal provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0054] In order to make the objectives, technical solutions and advantages of this application clearer, the implementation methods of this application will be further described in detail below with reference to the accompanying drawings.

[0055] It should be understood that although the terms first, second, etc. may be used in this disclosure to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, a first parameter may also be referred to as a second parameter, and similarly, a second parameter may also be referred to as a first parameter without departing from the scope of this disclosure. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0056] A brief introduction is given to the nouns involved in the embodiments of this application.

[0057] Artificial Intelligence (AI) is the theory, methods, techniques, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, to perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that seeks to understand the essence of intelligence and produce new intelligent machines that can respond in a manner similar to human intelligence. AI also studies the design principles and implementation methods of various intelligent machines, enabling them to possess the capabilities of perception, reasoning, and decision-making.

[0058] Artificial intelligence (AI) technology is a comprehensive discipline encompassing a wide range of fields, encompassing both hardware and software technologies. Foundational AI technologies generally include sensors, specialized AI chips, cloud computing, distributed storage, big data processing, pre-trained models, operating / interaction systems, and mechatronics. Pre-trained models, also known as large models or basic models, can be fine-tuned and widely applied to downstream tasks across various AI disciplines. AI software technologies primarily encompass computer vision, speech processing, natural language processing, and machine learning / deep learning.

[0059] Machine Learning (ML) is a multidisciplinary field that encompasses probability theory, statistics, approximation theory, convex analysis, and algorithmic complexity theory. It specifically studies how computers can simulate or implement human learning behaviors to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is at the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications span all areas of AI. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and learning through demonstration.

[0060] In related technologies, advertisements usually adopt a unified display mode, which makes it difficult to meet the personalized needs of different user groups; the relatively independent advertising delivery method makes the advertising cost high, making it difficult for small businesses and individuals to afford it; related background images and layouts require professional design, with high production costs and complex processes; the content and form of advertisements are single, and have low appeal to users, resulting in low user click-through rates on advertisements and low content recommendation efficiency.

[0061] The recommendation system provided in the embodiment of the present application recommends content by adopting a recommendation system including a server, a first client logged in with a first account, and a second client logged in with a second account. The server sends first recommendation information and second recommendation information to the first client and the second client respectively based on the first account and the second account. The first client displays a first content recommendation interface including multiple first recommended contents displayed in an aggregated manner and first interface elements other than the first recommended contents based on the first recommendation information. The second client displays a second content recommendation interface including multiple second recommended contents displayed in an aggregated manner and other than the second recommended contents based on the second recommendation information. The first recommendation information is different from the second recommendation information, thereby realizing a personalized aggregation display scheme based on the aggregated recommended contents and / or interface elements in the content recommendation interface for different accounts, which can improve the fit between the content recommendation interface and the user's interests and increase its attractiveness to users. At the same time, the richness of the recommended contents in the content recommendation interface is increased by the aggregated display of multiple recommended contents, thereby improving the recommendation efficiency.

[0062] The method for displaying recommended content provided in an embodiment of the present application obtains first recommendation information from a server by a first client logged in to a first account, and displays a first content recommendation interface based on the first recommendation information. The first content recommendation interface includes multiple first recommended contents displayed in an aggregated manner and first interface elements other than the first recommended contents. The first recommendation information is personalized content for the first account, thereby implementing a personalized aggregate display solution based on the recommendation information and achieving a content recommendation display effect that is tailored to each individual user, thereby increasing the appeal to users and improving the efficiency of content recommendation.

[0063] The method for generating recommended content provided in an embodiment of the present application is that, in response to a recommendation request received from a first account, a server aggregates multiple first recommended contents based on the personalized first account, generates a first interface element based on the first account and at least one of the first recommended contents, and sends first recommendation information to a first client corresponding to the first account. The first recommendation information is used to provide relevant information of a first content recommendation interface displayed by the first client. The first content recommendation interface includes multiple first recommended contents displayed in an aggregated manner and the first interface element other than the first recommended contents, thereby implementing a personalized aggregate recommendation scheme based on recommendation information for user accounts, making the generated recommendation information more consistent with the interest tendencies of the account, improving its attractiveness to users, improving the generation quality and efficiency of the recommendation information, and at the same time reducing interface design costs and content recommendation costs. The multiple recommended contents displayed in an aggregated manner improve the richness of recommended content in the content recommendation interface and improve the content recommendation efficiency.

[0064] First, let's introduce the implementation environment of this application. Figure 1 , which shows a schematic diagram of a recommendation system provided by an exemplary embodiment of the present application. The implementation environment includes: terminal 110, server 120, and terminal 130. Terminals 110 and 130 are connected to server 120 via a communication network. The communication network can be implemented as a wired network or a wireless network, which is not limited in the present embodiment.

[0065] In some embodiments, terminal 110 is a first client logged in with a first account, terminal 130 is a second client logged in with a second account, and server 120 is implemented as a background server for the first client and the second client. Server 120 is used to provide personalized recommendation information to the client based on different accounts, so that the client can display a personalized content recommendation interface based on the recommendation information, thereby realizing a personalized content recommendation solution.

[0066] Illustratively, the server 120 sends first recommendation information to the terminal 110 based on the first account, and sends second recommendation information to the terminal 130 based on the second account. The terminal 110 displays a first content recommendation interface 101 based on the first recommendation information. The first content recommendation interface 101 includes multiple first recommended contents 102 displayed in an aggregated manner and a first interface element other than the first recommended contents 102 (for example, a character sticker 103). The terminal 130 displays a second content recommendation interface 104 based on the second recommendation information. The second content recommendation interface 104 includes multiple second recommended contents 105 displayed in an aggregated manner and a second interface element other than the second recommended contents 105 (for example, a character sticker 106). The first recommendation information is different from the second recommendation information.

[0067] In some embodiments, the above-mentioned client can be a client terminal corresponding to the first application, and the first application can be implemented as an instant messaging application, a video application, a news information application, a comprehensive search engine application, a social application, a game application, a shopping application, a map navigation application, etc. The embodiments of the present application are not limited to this.

[0068] The above-mentioned content recommendation interface can be a recommendation interface displayed during the operation of the first application. Optionally, the content recommendation interface can be a welcome interface that serves as the startup interface when entering the first application; or, the content recommendation interface can be the homepage interface that is first displayed when logging into the first application; or, the content recommendation interface can be any interface displayed during the use of the first application; or, the content recommendation interface can be an interface displayed after a certain control on the web page is triggered, etc.

[0069] The embodiment of the present application is described by taking the above-mentioned content recommendation interface implemented as a splash screen interface displayed when the client is started as an example.

[0070] The recommended content can be any type of content resource, such as images, videos, audio, and text. For example, for a video playback platform, which plays TV series, movies, cartoons, live streams, and edited videos, the recommended content can be video resources on the platform, resources such as videos or images used to recommend a first product, or promotional resources for user accounts other than the terminal account on the platform.

[0071] The above-mentioned interface elements are interface display elements in the content recommendation interface other than the recommended content. Optionally, the interface elements include but are not limited to at least one of the layout, background image, and control elements of the content recommendation interface.

[0072] The above-mentioned terminal is optional. The terminal can be a desktop computer, a laptop computer, a mobile phone, a tablet computer, an e-book reader, a Moving Picture Experts Group Audio Layer III (MP3) player, a Moving Picture Experts Group Audio Layer IV (MP4) player, a smart TV, a smart car, and other terminal devices in various forms. The embodiments of the present application are not limited to this.

[0073] It is worth noting that the above-mentioned servers can be independent physical servers, or they can be server clusters or distributed systems composed of multiple physical servers. They can also be cloud servers that provide basic cloud computing services such as cloud services, cloud security, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), as well as big data and artificial intelligence platforms.

[0074] Among them, cloud technology refers to a hosting technology that unifies hardware, software, network and other resources within a wide area network or local area network to achieve data calculation, storage, processing and sharing.

[0075] In some embodiments, the above-mentioned server can also be implemented as a node in a blockchain system.

[0076] It should be noted that before collecting the user's relevant data and during the process of collecting the user's relevant data, this application can display a prompt interface, pop-up window or output voice prompt information. The prompt interface, pop-up window or voice prompt information is used to remind the user that its relevant data is currently being collected, so that this application only starts to execute the relevant steps of obtaining the user's relevant data after obtaining the user's confirmation operation on the prompt interface or pop-up window. Otherwise (that is, when the user's confirmation operation on the prompt interface or pop-up window is not obtained), the relevant steps of obtaining the user's relevant data are terminated, that is, the user's relevant data is not obtained. In other words, all user data collected by this application are collected with the user's consent and authorization, and the collection, use and processing of relevant user data need to comply with relevant laws, regulations and standards.

[0077] For illustration, please refer to Figure 2 , which shows an interactive flow chart of a recommendation system provided by an exemplary embodiment of the present application, such as Figure 2 As shown, the system includes a server, a first client and a second client, wherein the first client is logged in with a first account and the second client is logged in with a second account. The system interaction includes the following steps.

[0078] Step 211: The server sends recommendation information to the client based on the account corresponding to the client.

[0079] Wherein, step 211 includes step 211a, the server sends a first recommendation message to the first client based on the first account; and step 211b, the server sends a second recommendation message to the second client based on the second account.

[0080] In some embodiments, the recommendation information is used to indicate a plurality of recommended contents and interface elements that are aggregated and displayed by the client in the content recommendation interface.

[0081] Among them, the recommended information is generated for account personalization.

[0082] Optionally, at least one of the multiple recommended content and interface elements displayed in aggregate is generated for account personalization.

[0083] In some embodiments, the multiple recommended contents are recommended contents aggregated by the server based on the account through an artificial intelligence model, and / or the first interface element is an element generated by the server based on the account and at least one of the multiple recommended contents through an artificial intelligence model.

[0084] In some embodiments, the client may be a client terminal corresponding to the first application, and the server may be a background server of the first application.

[0085] Optionally, the first application can be implemented as an instant messaging application, a video application, a news information application, a comprehensive search engine application, a social application, a game application, a shopping application, a map navigation application, etc., which is not limited in the embodiments of the present application.

[0086] The content recommendation interface can be a recommendation interface displayed during the operation of the first application. Optionally, the content recommendation interface can be a welcome interface that serves as the startup interface when entering the first application; or, the content recommendation interface can be a homepage interface that is first displayed when logging into the first application; or, the content recommendation interface can be any interface displayed during the use of the first application; or, the content recommendation interface can be an interface that is displayed after a control on a web page is triggered, etc. In some embodiments, the content recommendation interface can be implemented as a web page interface, that is, a web page interface displayed in a browser application, such as an H5 web page.

[0087] The embodiment of the present application is described by taking the content recommendation interface implemented as a splash screen interface displayed when the client is started as an example.

[0088] The recommended content can be any type of content resource, such as images, videos, audio, and text. For example, for a video playback platform, which plays TV series, movies, cartoons, live streams, and edited videos, the recommended content can be video resources on the platform, resources such as videos or images used to recommend a first product, or promotional resources for user accounts other than the terminal account on the platform.

[0089] The above interface elements may be any display elements in the content recommendation interface except the recommended content, such as the layout, background image, control elements, etc. of the content recommendation interface.

[0090] Step 221 : When the first client is started to display a splash screen interface, the first client displays a first content recommendation interface as a splash screen interface based on the first recommendation information.

[0091] The first content recommendation interface includes a plurality of first recommended contents displayed in an aggregated manner and first interface elements other than the first recommended contents.

[0092] Optionally, the plurality of first recommended contents are recommended contents aggregated based on personalization of the first account.

[0093] Illustratively, a server stores a candidate recommendation set, which includes multiple candidate recommendation contents. The server identifies the recommendation features of the multiple candidate recommendation contents based on an artificial intelligence model, and obtains recommendation tags corresponding to the multiple candidate recommendation contents. The recommendation tags are used to indicate the recommendation topics corresponding to the candidate recommendation contents. For a first account, the server analyzes the first account's interest in the recommended contents based on the artificial intelligence model, and obtains a first content tag. Based on the recommendation tags, multiple first recommended contents corresponding to the first content tag are determined from the multiple candidate recommendation contents. The first content tag is used to indicate recommended content that matches the first account's interest.

[0094] Optionally, the first interface element is an element generated based on personalization of the first account.

[0095] In some embodiments, the plurality of first recommended contents correspond to a first recommended theme, and the first interface element is adapted to the first recommended theme.

[0096] Optionally, the first recommended topic is a personalized topic determined by the server based on the first account's interest tendency in the recommended topic.

[0097] Optionally, the first interface element includes at least one of the layout of the first content recommendation interface, the first background image, and the first control element.

[0098] The layout of the first content recommendation interface is used to indicate the display position, display size, layout, etc. of the first recommended content and other display elements in the interface (such as recommended topics, first control elements, etc.) in the first content recommendation interface.

[0099] Optionally, the first control element includes but is not limited to any control element such as an information display control, a container control, a navigation control, a multimedia control, a notification control, a progress control, and an interactive control.

[0100] Among them, information display controls are used to display static information or data, such as text labels, images, charts, etc.; container controls are used to organize and layout other control elements, such as panels, group boxes, etc.; navigation controls are used to assist users in navigating and operating in the interface, such as menu bars, toolbars, navigation bars, etc.; multimedia controls are used to display and play multimedia content, such as video players, audio players, and image viewers; notification controls are used to send prompts, warnings, or notification information to users, such as dialog boxes, message boxes, notification bars, etc.; progress controls are used to display the progress or status of tasks, such as progress bars, loading indicators, etc.; interactive controls are used to receive interactive operations for the content recommendation interface, such as interface jump controls, interface close controls, etc.

[0101] Optionally, the first interface element is artificial intelligence-generated content generated by the server based on at least one of the multiple first recommended content and the first account.

[0102] In some embodiments, the server analyzes the first account's interest in the recommended style based on the first account and at least one of the multiple first recommended contents based on an artificial intelligence model, obtains a first style label, and generates a first interface element based on the first style label based on the artificial intelligence model.

[0103] The above-mentioned first style tag is used to indicate the element display style of the first interface element, including but not limited to display color tone, contrast, recommended theme, interface layout, etc.

[0104] Step 231 : When the second client is started to display a splash screen interface, the second client displays a second content recommendation interface as a splash screen interface based on the second recommendation information.

[0105] The second content recommendation interface includes a plurality of second recommended contents displayed in an aggregated manner and second interface elements other than the second recommended contents.

[0106] The first recommendation information and the second recommendation information are different.

[0107] Optionally, the plurality of first recommended contents are different from the plurality of second recommended contents, and / or the first interface element is different from the second interface element.

[0108] In some embodiments, the plurality of second recommended contents correspond to a second recommended theme, and the second interface element is adapted to the second recommended theme.

[0109] Optionally, the second recommended topic is a personalized topic determined by the server based on the interest tendency of the second account in the recommended topic.

[0110] Taking the example that the plurality of first recommended contents are different from the plurality of second recommended contents, the plurality of second recommended contents may be recommended contents aggregated based on personalization of the second account.

[0111] Schematically, a candidate recommendation set is stored in the server, which includes multiple candidate recommendation contents. The server identifies the recommendation features of the multiple candidate recommendation contents based on an artificial intelligence model, and obtains recommendation tags corresponding to the multiple candidate recommendation contents. The recommendation tags are used to indicate the recommendation topics corresponding to the candidate recommendation contents. For the second account, the server analyzes the interest tendency of the second account in the recommended content based on the artificial intelligence model to obtain a second content tag. According to the recommendation tags, multiple second recommended contents corresponding to the second content tag are determined from the multiple candidate recommendation contents, wherein the second content tag is used to indicate recommended content that is in line with the interest tendency of the second account.

[0112] Taking the example where the first interface element is different from the second interface element, the second interface element may be artificial intelligence generated content generated by the server based on at least one of the multiple first recommended content and the first account.

[0113] Illustratively, the server analyzes the second account's interest in the recommended style based on the artificial intelligence model according to at least one of the second account and the second recommended content, obtains a second style label, and generates a second interface element based on the second style label based on the artificial intelligence model.

[0114] The above-mentioned second style tag is used to indicate the element display style of the second interface element, including but not limited to display color tone, contrast, recommended theme, interface layout, etc.

[0115] Optionally, the multiple first recommended contents are recommended contents aggregated by the server through an artificial intelligence model based on the first account, and the multiple second recommended contents are recommended contents aggregated by the server through an artificial intelligence model based on the second account; the first interface element is an element generated by the server through an artificial intelligence model based on the first account and at least one of the multiple first recommended contents, and the second interface element is an element generated by the server through an artificial intelligence model based on the second account and at least one of the multiple second recommended contents.

[0116] Optionally, the second interface element includes at least one of the layout of the second content recommendation interface, a second background image, and a second control element.

[0117] The layout of the second content recommendation interface is used to indicate the display position, display size, layout, etc. of the second recommended content and other display elements in the interface (such as recommended topics, second control elements, etc.) in the second content recommendation interface.

[0118] Optionally, the second control element includes but is not limited to any control element such as an information display control, a container control, a navigation control, a multimedia control, a notification control, a progress control, and an interactive control.

[0119] Optionally, the first account and the second account have different interests in the recommended content, and therefore, the multiple first recommended contents are different from the multiple second recommended contents. On this basis, the first account and the second account may have the same or different interests in the recommended style, that is, the first interface element and the second interface element may be the same or different. Alternatively, the first account and the second account have different interests in the recommended style, and therefore, the first interface element and the second interface element are different. On this basis, the first account and the second account may have the same or different interests in the recommended content, that is, the first recommended content and the second recommended content may be the same or different.

[0120] For example, take the movie recommendation scenario as an example, please refer to Figure 3 , Figure 3 This is a comparative diagram of a content recommendation interface provided by an exemplary embodiment of the present application. Figure 3 As shown in , the first account and the second account have different interests in recommended content. The first account is interested in comedy movies, and the second account is interested in suspense movies. The first account and the second account have different interests in recommendation styles. The first account tends to prefer a simple and light-hearted recommendation style, and the second account tends to prefer a complex and dark recommendation style. The first account corresponds to a first content recommendation interface 330, which includes a plurality of first recommended contents 331 displayed in an aggregated manner and a first interface element other than the first recommended contents 331. The second account corresponds to a second content recommendation interface 340, which includes a plurality of second recommended contents 341 displayed in an aggregated manner and a second interface element other than the second recommended contents 341. The plurality of first recommended contents 331 and the plurality of second recommended contents 341 are different. The plurality of first recommended contents 331 include comedy movie A and comedy movie B, and the plurality of second recommended contents 341 include suspense movie C, suspense movie D, and suspense movie E. The first interface element is different from the second interface element. The first interface element is adapted to a simple and light style. The interface layout of the first interface element is used to indicate that the plurality of first content recommendation interfaces 330 recommend comedy movie A and comedy movie B side by side in the focus display area. The second interface element is adapted to a complex and dark style. The interface layout of the second interface element is used to indicate that the second content recommendation interface 340 recommends suspense movie C, suspense movie D, and suspense movie E in an aggregated manner in the form of a list.

[0121] To summarize, the system provided by the embodiment of the present application, during the process of recommending content, performs personalized individual analysis on different accounts through the server, thereby determining multiple recommended content and interface elements corresponding to different accounts, such as: after analyzing the first account, multiple first recommended content and first interface elements are obtained, and after analyzing the second account, multiple second recommended content and second interface elements are obtained. Among them, the recommendation information used to indicate the above-mentioned multiple recommended content and interface elements is determined based on the personalized analysis of the account, so the aggregated content of the recommended content is different, and a personalized display scheme based on multiple aggregated content and / or interface elements in the content recommendation interface for different accounts is implemented, providing users with personalized recommendations for thousands of people, which can improve the fit between the content recommendation interface and the user's interest tendencies and increase its attractiveness to users. At the same time, the richness of the recommended content in the content recommendation interface is increased by the aggregated display of multiple recommended contents, thereby improving the recommendation efficiency.

[0122] This application also provides a method for displaying recommended content. Figure 4 , which shows a flowchart of a method for displaying recommended content provided by an exemplary embodiment of the present application. The method can be executed by a terminal, or by a server, or by both the terminal and the server. The embodiment of the present application takes the method executed by a terminal as an example for explanation. Figure 4 As shown, the method is executed by a first client logged in with a first account, and the method includes the following steps:

[0123] Step 410: Obtain first recommendation information from the server.

[0124] The server is a server in the recommendation system, and is configured to provide first recommendation information to a first client corresponding to a first account. The first recommendation information is sent by the server to the first client based on the first account.

[0125] In some embodiments, the first recommendation information is used to indicate the first recommended content and the first interface element, wherein the first recommended content is a content resource recommended to the first account, and the first interface element is a display element in the interface other than the first recommended content when the first recommended content is recommended to the first account through the first content interface.

[0126] The first recommendation information can be preloaded or loaded in real time. For example, in response to the network connection status being connected to a wireless fidelity network or a 5G network, the first recommendation information is obtained from the server in real time; or, in response to the network connection status being connected to a cellular network (for example, 2G, 3G, 4G, etc.), the locally preloaded first recommendation information is read.

[0127] Optionally, when the first client needs to display the first content recommendation interface, it sends a first recommendation request to the server, and the server sends first recommendation information to the first client based on the first recommendation request.

[0128] Step 420 : When the first client is started to display a splash screen interface, a first content recommendation interface is displayed as a splash screen interface based on the first recommendation information.

[0129] The first content recommendation interface includes a plurality of first recommended contents displayed in an aggregated manner and a first interface element other than the first recommended contents.

[0130] The first recommendation information is personalized content for the first account.

[0131] Optionally, the personalized content may be at least one of a plurality of first recommended contents aggregated based on personalization of the first account, and a first interface element generated based on personalization of the first account.

[0132] The first recommendation information is used to provide relevant information of the first content recommendation interface displayed by the first client.

[0133] The first recommendation information is different from the recommendation information sent by the server to other clients based on other accounts.

[0134] The personalized content includes a plurality of first recommended contents aggregated based on the first account, and a first interface element generated based on at least one of the first account and the plurality of first recommended contents. The plurality of first recommended contents are obtained based on the first account, and therefore the first interface element is generated specifically for the first account. For example, the first interface element is generated based on the first account's interest in the recommended style. If the first account and other accounts (e.g., the second account) have different interests in the recommended style, the first interface element generated for the first account is different from the interface elements generated for the other accounts.

[0135] The above-mentioned personalized content includes but is not limited to any interface elements other than recommended content in the first content recommendation interface, such as interface layout, background images (including content decoration elements and background rendering elements, such as mascot stickers in the background images), control elements (such as interactive controls, navigation controls), etc., which are in line with the interests of the first account.

[0136] Taking the first client as the client terminal corresponding to the first application as an example, the first application can be a social application, an audio player, a video player software, an electronic reading software, an online teaching platform, an accounting software, an online shopping software, or any other application that provides an interface display function. This application does not limit this.

[0137] In some embodiments, the first application has a content recommendation function and can recommend content to the first account.

[0138] Optionally, the first recommended content may be any type of content resource provided by the first application, such as pictures, videos, audio, text, etc.

[0139] Taking the first application as an example, which is a video playback platform, the video playback platform is used to play video resources provided by the platform, such as TV series, movies, cartoons, live videos, and edited videos. The recommended content can be video resources in the video playback platform, or it can be resources such as videos or pictures used to recommend the first product, or it can be promotional resources for user accounts other than the first account in the recommendation platform.

[0140] Optionally, the first recommendation information personalized for the first account includes the following three situations:

[0141] First, the multiple first recommended contents are recommended contents aggregated by the server based on the first account through an artificial intelligence model.

[0142] In some embodiments, an artificial intelligence model is used to analyze the interest tendency of the first account in recommended content to obtain a first content tag, and a plurality of first recommended contents that match the interest tendency of the first account are aggregated based on the first content tag.

[0143] Schematically, a candidate recommendation set is stored in the server, which includes multiple candidate recommendation contents. The server identifies the recommendation features of the multiple candidate recommendation contents based on an artificial intelligence model, and obtains recommendation tags corresponding to the multiple candidate recommendation contents. The recommendation tags are used to indicate the recommendation topics corresponding to the candidate recommendation contents. For the first account, the server analyzes the first account's interest tendency in the recommended content based on the artificial intelligence model to obtain a first content tag. According to the recommendation tag, the first recommended content corresponding to the first content tag is determined from the multiple candidate recommendation contents, wherein the first content tag is used to indicate the recommended content that is in line with the interest tendency of the first account.

[0144] In some embodiments, the interest tendency of the first account in the recommended content can be analyzed based on historical data of the first account, where the historical data is operation data of the first account in the first client within a historical time period.

[0145] The above-mentioned historical data includes but is not limited to the information data, access data, storage data, etc. input by the first account into the first client. It is worth noting that, as emphasized above, this application displays a prompt interface, pop-up window or outputs voice prompt information before collecting historical data and in the process of collecting the user's historical data. The prompt interface, pop-up window or voice prompt information is used to remind the user that its historical data is currently being collected, so that this application only starts to execute the relevant steps of obtaining the user's historical data after obtaining the user's confirmation operation on the prompt interface or pop-up window. Otherwise (that is, when the user's confirmation operation on the prompt interface or pop-up window is not obtained), the relevant steps of obtaining the user's historical data are terminated, that is, the user's historical data is not obtained.

[0146] Illustratively, historical data includes but is not limited to at least one of the browsing time, number of views, interaction frequency, and attention paid to the recommending entity (such as a social account, brand, etc.) by the first account in the first client during a historical time period.

[0147] Second, the first interface element is an element generated by the server based on the first account through an artificial intelligence model.

[0148] In some embodiments, the server analyzes the first account's interest in the recommended style based on at least one of the first account and the first recommended content based on an artificial intelligence model, obtains a first style label, and generates a first interface element based on the first style label based on the artificial intelligence model.

[0149] The above-mentioned first style tag is used to indicate the element display style of the first interface element, including but not limited to display color tone, contrast, recommended theme, interface layout, etc.

[0150] Optionally, the first interface element includes at least one of the layout of the first content recommendation interface, the first background image, and the first control element.

[0151] The layout of the first content recommendation interface is used to indicate at least one of the display position, display size, layout, text font, and display angle of the first recommended content and other display elements in the interface (such as recommended topics, first control elements, etc.) in the first content recommendation interface.

[0152] Optionally, the first control element includes but is not limited to any control element such as an information display control, a container control, a navigation control, a multimedia control, a notification control, a progress control, and an interactive control.

[0153] In some embodiments, the first account's interest tendency in the recommended style may be analyzed based on historical data of the first account.

[0154] Optionally, the historical data includes but is not limited to the browsing time, number of views, interaction frequency of the historical recommended content of the same recommendation style by the first account in the first client during the historical time period, as well as the avatar used by the first account during the historical time period, the layout of the account homepage, etc.

[0155] The third type is that the multiple first recommended contents and first interface elements are generated by the server through an artificial intelligence model for personalization of the first account.

[0156] In the case where the plurality of first recommended contents and the first interface element are both generated for personalization of the first account, the first interface element may be generated based on the plurality of first recommended contents.

[0157] Illustratively, a first recommendation theme is determined based on the recommendation features of multiple first recommended contents, and a first interface element adapted to the first recommendation theme is generated based on the first recommendation theme; or, a first interface element matching the historical recommendation style is generated based on the historical recommendation style of multiple first recommended contents; or, a first interface element is generated based on multiple first recommended contents and the first account's historical data on the multiple first recommended contents. For example, the multiple first recommended contents belong to film and television works, and the generated first interface element includes an interface layout in the form of a list and a film and television recommendation theme. Based on the first account's historical data on the multiple first recommended contents, a character map corresponding to the first account's favorite character A in multiple film and television works is generated as a background rendering element in the first background image in the first interface element.

[0158] For example, take the movie recommendation scenario as an example, please refer to Figure 5 , Figure 5 is a comparative diagram of a content recommendation interface provided by another exemplary embodiment of the present application. Figure 5 As shown, the first account and the second account have the same interest in recommended content. The multiple first recommended contents based on the personalized aggregation of the first account are the same as the multiple second recommended contents based on the personalized aggregation of the second account. The first account and the second account have different interest in the recommended style. For film and television work A among the multiple recommended contents, the first account likes the first character, while the second account likes the second character. The first account corresponds to a first content recommendation interface 510, which includes multiple first recommended contents 511 displayed in an aggregated manner and a first interface element other than the first recommended contents. The first interface element includes a character sticker 512 corresponding to the first character. The second account corresponds to a second content recommendation interface 520, which includes multiple second recommended contents 521 displayed in an aggregated manner and a second interface element other than the second recommended contents. The second interface element includes a character sticker 522 corresponding to the second character.

[0159] In some embodiments, the first interface element includes a first theme element.

[0160] The above step 420 may be implemented by displaying the first theme element and the plurality of first recommended contents in association with each other in the first content recommendation interface, where the first theme element is the first recommended theme corresponding to the plurality of first recommended contents.

[0161] Optionally, the first theme element may be implemented as a theme element in any form, such as theme text content, theme animation content, etc.

[0162] In some embodiments, the first recommendation topic is used to indicate the recommendation subject type and / or recommended content type corresponding to the first recommended content.

[0163] For example, the first recommendation theme is implemented as "Comedy Movie of the Year", and the recommended content type corresponding to multiple first recommended contents belongs to comedy movies, and the recommended subject type belongs to the movie producer or playback platform, etc.; the first recommendation theme is implemented as "Most Popular Beauty Blogger", and the recommended subject type corresponding to multiple first recommended contents is beauty blogger, and the recommended content type can be the blogger's account, avatar, etc.

[0164] To summarize, the method provided in the embodiment of the present application obtains first recommendation information from a server by a first client logged in to a first account, and displays a first content recommendation interface based on the first recommendation information. The first content recommendation interface includes first recommended content and a first interface element other than the first recommended content. The first interface element is personalized content for the first account, thereby realizing a personalized display solution based on interface elements and achieving a content recommendation display effect that is tailored to each individual, thereby increasing the attractiveness to users and improving the efficiency of content recommendation.

[0165] Taking the above method for displaying recommended content executed by the first client logged in with the first account as an example, in some embodiments, the first interface element includes a first control element, and the first control element includes an interactive control. Figure 6 , which shows a flowchart of an interactive control display method based on a content recommendation interface provided by an exemplary embodiment of the present application. The method can be executed by a terminal, a server, or both. The embodiment of the present application takes the method executed by the server as an example for explanation. Figure 6 As shown, the above step 420 includes the following steps:

[0166] Step 610 : In response to the first content recommendation interface meeting the interaction requirement, displaying an interactive control in the first content recommendation interface.

[0167] The interactive control is used to receive interactive operations on the content recommendation interface.

[0168] Optionally, the interactive operation can be an overall operation of the first content recommendation interface, such as an interface jump operation, an interface screenshot operation, etc., or an operation of the first recommended content, such as a recommended content viewing operation, a recommended content like operation, etc., or an operation of the first interface element, such as an animation trigger operation, an element cropping operation, etc. This application does not limit this.

[0169] The interactive requirements can be overall display requirements for the content recommendation interface, for example, the display time of the first content recommendation interface reaches a preset time, or requirements for multiple first recommended contents, for example, multiple first recommended contents include at least one preset content, or requirements for the first interface elements, for example, the first background image corresponds to a preset interactive animation, etc.

[0170] In some embodiments, multiple interactive controls are included. When different parts of the first content recommendation interface meet the interactive requirements, corresponding interactive controls are displayed. Optionally, the above step 610 includes at least the following three situations:

[0171] First, when the display duration of the first content recommendation interface meets the first interaction requirement, the first interaction control is displayed in the first content recommendation interface.

[0172] The first interactive control is used to trigger the display of a topic interactive interface corresponding to the recommended topic.

[0173] Optionally, the theme interactive interface may be an interactive interface provided by the first client according to the recommended theme, which may be used to view detailed content of the first recommended content, or may be used to purchase a recommended product corresponding to the first recommended content.

[0174] In some embodiments, the first interaction requirement means that the display time of the first content recommendation interface reaches a preset time, that is, in response to the display time of the first content recommendation interface reaching the preset time, the first interaction control is displayed in the first content recommendation interface.

[0175] The interaction tendency of the first account with respect to the first content recommendation interface is determined based on the display time of the first content recommendation interface. When the display time reaches a preset time, the first interactive control is displayed, thereby realizing a dynamic display scheme of the first interactive control. This scheme can intuitively remind the first account to trigger the display of the theme interactive interface, thereby improving interaction efficiency.

[0176] The second type is that when a plurality of first recommended contents meet the second interaction requirement, the second interactive control is displayed in association with the first recommended contents.

[0177] The second interactive control is used to receive an interactive operation on the first recommended content.

[0178] Optionally, the second interaction requirement includes at least one of the following situations:

[0179] 1. The second interaction requirement includes that at least one of the recommended subjects of the first recommended content belongs to the preset recommended subject list;

[0180] That is, when the recommendation subject corresponding to the at least one first recommended content belongs to the preset recommendation subject list, the second interactive control is displayed in association with the at least one first recommended content.

[0181] The preset recommendation subject list includes multiple preset interactive recommendation subjects.

[0182] Schematically, the first application is pre-set with a preset recommendation subject list, and the first application recommends multiple interactive recommendation subjects in the list with higher strength than other recommendation subjects. When the recommendation subject corresponding to at least one first recommended content belongs to the preset recommendation subject list, the second interactive control is displayed in association with at least one first recommended content, so as to achieve a prominent recommendation effect for at least one first recommended content among multiple first recommended contents and improve the recommendation strength. The first account can perform interactive operations such as likes and views on at least one first recommended content through the second interactive control, thereby increasing the attractiveness to users through the interactive scheme with the first recommended content, and also improving the interactive efficiency of specific recommended content in the aggregated recommendation scenario.

[0183] 2. The second interaction requirement includes that the recommendation score of at least one first recommended content reaches a preset score threshold;

[0184] Optionally, when the recommendation score corresponding to the at least one first recommended content reaches a preset score threshold, the second interactive control is displayed in association with the at least one first recommended content.

[0185] The recommendation score is used to indicate the relevance between the first recommended content and the recommended topic.

[0186] Illustratively, assuming that the first recommended topic corresponding to the first content recommendation interface is used to indicate the most popular beauty blogger in the historical recommendation period, the recommendation score is determined based on the popularity data of multiple beauty blogger accounts in the historical period. The higher the popularity, the higher the recommendation score. For the beauty blogger with the highest recommendation score, the second interactive control is displayed in association with the blogger's account avatar.

[0187] The third type is to display the third interactive control when the first background image meets the third interactive requirement.

[0188] The third interactive control is used to receive interactive operations on the first background image.

[0189] In some embodiments, the first background image includes content decoration elements and background rendering elements. The content decoration elements are used to decorate the first recommended content, and the background rendering elements are used to decorate the background screen of the first background image.

[0190] Optionally, the third interaction requirement includes that the first recommended content corresponding to the content decoration element meets the interaction content requirement, or the third interaction requirement includes that the background rendering element belongs to a preset element list.

[0191] Specifically, the interactive content requirement includes that the recommendation subject corresponding to the first recommended content belongs to a preset recommendation subject list, or that the recommendation score of the first recommended content reaches at least one of a preset score threshold.

[0192] In some embodiments, when the first recommended content corresponding to the content decoration element meets the interactive content requirements, the main decoration element and the third interactive control are displayed in association.

[0193] The third interactive control is used to trigger the display of a highlight animation corresponding to the first recommended content.

[0194] The interactive content requirements include that the recommendation subject corresponding to the first recommended content belongs to a preset recommendation subject list, or that the recommendation score corresponding to the first recommended content reaches a preset score threshold.

[0195] For example, taking the first recommended content decorated with a crown element as an example, if the recommended content A has the highest recommendation score among multiple first recommended contents, the crown element corresponding to the recommended content A is associated and displayed with the third interactive control. By triggering the third interactive control, the highlight animation corresponding to the recommended content A can be triggered, for example, the recommended content A is displayed in an enlarged manner, and the flashing animation corresponding to the crown element is played.

[0196] In some embodiments, when the background rendering element belongs to the preset element list, the background rendering element is displayed in association with the third interactive control.

[0197] The preset element list includes a correspondence between background rendering elements and preset element animations, and the third interactive control is used to trigger the display of the preset element animation corresponding to the background rendering element.

[0198] For illustration, please refer to Figure 7 , Figure 7 This is a schematic diagram of an element animation effect provided by an exemplary embodiment of the present application. Figure 7As shown, the first content recommendation interface 700 includes a plurality of first recommended contents 710 (account avatars corresponding to multiple official game accounts) displayed in an aggregated manner and a first interface element other than the first recommended content 710. The first interface element includes a first theme element 701, a mascot 730 (background rendering element) and a first interactive control 720. The first theme element 701 is used to indicate that the recommended theme is "hot games". The first interactive control 720 is used to jump to an interactive interface for interacting with multiple official game accounts. When the mascot 730 belongs to the preset element list, the third interactive control 740 is displayed. The third interactive control 740 is used to trigger the rotation animation effect of the mascot 730.

[0199] To sum up, the method provided in the embodiment of the present application provides an interactive solution based on the content recommendation interface by displaying interactive controls in the first content recommendation interface in response to the first content recommendation interface meeting the interactive requirements. It prompts users to participate in the interaction through explicit interactive controls, thereby improving the user's interaction rate based on the content recommendation interface.

[0200] The method provided in the embodiment of the present application provides a method for displaying different interactive controls based on whether different content in the content recommendation interface meets the interactive requirements, and provides multiple interactive solutions. It can increase the attractiveness of the content recommendation interface to users, enrich the interface elements of the content recommendation interface, and increase the user's click-through rate based on the content recommendation interface, thereby improving the efficiency of content recommendation.

[0201] In some embodiments, the first content recommendation interface includes multiple first recommended contents, the multiple first recommended contents belong to the same recommendation theme, the first interface element is adapted to the recommendation theme, and the first content recommendation interface implements an aggregate recommendation method based on the recommendation theme.

[0202] In some embodiments, for the aggregation recommendation scenario, a multiple aggregation recommendation scheme based on multiple aggregation dimensions can also be implemented. The first content recommendation interface includes aggregation recommendation areas corresponding to multiple aggregation dimensions, and the first interface elements include content decoration elements, which are used to decorate the recommended content. For an example, please refer to Figure 8 , Figure 8 This is a flowchart of an aggregate recommendation method provided by an exemplary embodiment of the present application. The method can be executed by a terminal, a server, or both. The embodiment of the present application takes the method executed by a first client logged in with a first account as an example. Figure 8 As shown, the above step 420 includes the following steps:

[0203] Step 810 : Displaying a plurality of first recommended contents belonging to an i-th aggregation dimension in an i-th aggregation recommendation area among a plurality of aggregation recommendation areas, where i is a positive integer.

[0204] Optionally, the aggregation dimensions include but are not limited to the popularity of the recommended content, the release time of the recommended content, the quality of the recommended content, etc.

[0205] In some embodiments, the aggregation dimensions are pre-defined by the first application.

[0206] The number of the multiple first recommended contents in each aggregate recommendation area may be the same or different, and this application does not impose any limitation on this.

[0207] For illustration, please refer to Figure 9 , Figure 9 This is a schematic diagram of a method for displaying sub-aggregate recommendations provided by an exemplary embodiment of the present application. Figure 9 As shown, taking the recommendation subject indicated by the recommendation theme as a beauty blogger as an example, the first content recommendation interface 900 is divided into three aggregate recommendation areas of "Hot Blogger of the Year", "New Blogger of the Year", and "Powerful Blogger of the Year" according to the three aggregation dimensions of recommendation subject popularity, recommended content release time, and recommended content quality. For multiple first recommended contents corresponding to the recommendation theme, they are aggregated and classified according to the above three aggregation dimensions. The account avatars of hot bloggers A, hot bloggers B, and hot bloggers C are displayed in the aggregate recommendation area 901 corresponding to "Hot Blogger of the Year", the account avatars of new bloggers D, new bloggers E, and new bloggers F are displayed in the aggregate recommendation area 902 corresponding to "New Blogger of the Year", and the account avatars of powerful bloggers M, powerful bloggers N, and powerful bloggers K are displayed in the aggregate recommendation area 903 corresponding to "Powerful Blogger of the Year".

[0208] Step 820 : For the first recommended content in the aggregated recommendation area that reaches a preset ranking threshold according to the aggregated score from high to low, display content decoration elements according to the aggregated score.

[0209] Among them, the aggregation score is used to represent the matching degree between the first recommended content and the aggregation dimension corresponding to the aggregation recommendation area, and the visual prominence of the content decoration elements is positively correlated with the recommendation score.

[0210] For illustration, please refer to Figure 10 , Figure 10 This is a schematic diagram of a method for displaying content decoration elements provided by an exemplary embodiment of the present application. Figure 10As shown, for the first recommended content 1011 and the first recommended content 1012 in the aggregated recommendation area 1010, which are ranked from high to low according to the aggregated scores and reach the preset ranking threshold, the content decoration element 1001 and the content decoration element 1002 are displayed according to the aggregated scores corresponding to the first recommended content 1011 and the first recommended content 1012 respectively, wherein, when the aggregated score of the first recommended content 1011 is higher than that of the first recommended content 1012, the visual prominence of the content decoration element 1001 is higher than that of the content decoration element 1002.

[0211] In some embodiments, after displaying the content decoration element according to the aggregated score corresponding to the first recommended content, an award animation display process is also included.

[0212] Optionally, multiple aggregate recommendation areas correspond to display priorities, and the award animation display process includes playing the award animation effect corresponding to the first recommended content in the aggregate recommendation area in sequence according to the display priority.

[0213] For example, the first recommended content with an aggregate score reaching a preset aggregate score threshold in multiple aggregate recommendation areas is played in sequence according to the display priority. Figure 11 , Figure 11 This is a schematic diagram of an award animation display provided by an exemplary embodiment of the present application. Figure 11 As shown, for the aggregate recommendation area 1110, the aggregate recommendation area 1120 and the aggregate recommendation area 1130, according to the display priority corresponding to the multiple sub-aggregation areas, fireworks effects are played in turn for the first recommended content A, the first recommended content B and the first recommended content C. The first recommended content A, the first recommended content B and the first recommended content C are the recommended contents with the highest aggregation scores among the first recommended contents in the aggregate recommendation area 1110, the aggregate recommendation area 1120 and the aggregate recommendation area 1130, respectively.

[0214] To sum up, the method provided in the embodiment of the present application realizes further classification and aggregation scheme in the aggregated recommendation scenario by displaying the first recommended content in multiple aggregated recommendation areas respectively, and for the first recommended content that ranks from high to low according to the aggregated score and reaches a preset ranking threshold in the aggregated recommendation area, displays content decoration elements according to the aggregated score corresponding to the first recommended content, thereby improving the diversity of aggregated recommendations.

[0215] In some embodiments, based on the display method of recommended content provided in the embodiments of the present application, an interactive method based on content blocking elements can also be implemented. The embodiments of the present application illustrate the content blocking interactive method in the aggregated recommendation scenario as an example.

[0216] In some embodiments, the first interface element includes multiple content blocking elements, and the multiple content blocking elements are used to block multiple first recommended contents. The multiple content blocking elements have corresponding blocking priorities. Figure 12 , which shows a flow chart of a content blocking interactive method provided by an exemplary embodiment of the present application. The method can be executed by a terminal, or by a server, or by both the terminal and the server. The embodiment of the present application takes the method executed by a first client logged in with a first account as an example for explanation. Figure 12 As shown, the above step 420 includes the following steps:

[0217] Step 1210 : Displaying a plurality of content blocking elements in association with a plurality of first recommended contents on a first content recommendation interface.

[0218] Among them, when the content blocking element is in a visible state, the first recommended content corresponding to the content blocking element is in an invisible state; when the content blocking element is in an invisible state, the first recommended content corresponding to the content blocking element is in a visible state.

[0219] In some embodiments, the first recommended content being in an invisible state is used to indicate that the first recommended content is blocked by a content blocking element, and the content blocking element being in an invisible state is used to indicate that the content blocking element is cancelled from being displayed.

[0220] Optionally, when the content blocking element blocks the first recommended content, it may be a partial block or a complete block, which is not limited in this application.

[0221] Optionally, when rendering the content recommendation interface, the content blocking elements can be directly overlaid and rendered on the basis of the complete rendering of multiple first recommended contents to simplify the occlusion rendering process; or the rendering process of the blocked part can be omitted based on the occlusion relationship between the content blocking elements and the first recommended contents to save interface rendering resources. This application does not limit this.

[0222] Step 1220: Receive a cancel display operation for multiple content-blocking elements.

[0223] The unshow operation switches a content-obstructing element from a visible state to an invisible state.

[0224] Optionally, a global display cancellation operation for multiple content obstructing elements may be received, or an independent cancellation operation for each content obstructing element may be received.

[0225] The overall display cancellation operation is used to automatically cancel the display of multiple content blocking elements in descending order of blocking priorities based on a single operation.

[0226] Step 1230 : cancel the display of multiple content blocking elements in descending order of blocking priority.

[0227] The blocking priority corresponding to the content blocking element is negatively correlated with the recommendation score of the first recommended content corresponding to the content blocking element. The recommendation score is determined in advance based on the correlation between the first recommended content and the recommended topic.

[0228] For illustration, please refer to Figure 13 , Figure 13 FIG. 1 is a schematic diagram of an interactive method for occluding display provided by an exemplary embodiment of the present application. Figure 13 As shown, multiple content blocking elements are associated with multiple first recommended contents in the first content recommendation interface 1300, wherein the content blocking element 1301 is used to block the first recommended content 1310, the content blocking element 1302 is used to block the first recommended content 1320, and the content blocking element 1303 is used to block the first recommended content 1330. The cancel display operation for the multiple content elements is received, and the content blocking element 1301, the content blocking element 1302 and the content blocking element 1303 are canceled in descending order according to the blocking priority.

[0229] To sum up, the method provided in the embodiment of the present application increases the mystery of the recommended content and stimulates the user's curiosity by associating the display content blocking elements with multiple recommended contents in the content interface, thereby increasing the attractiveness to users; by receiving the cancel display operation for multiple content blocking elements, the display content blocking elements are canceled in sequence from high to low according to the blocking priority, so that the corresponding recommended contents are revealed in sequence, providing users with an interface interaction channel, improving the interactive interest of human-computer interaction, and improving the efficiency of human-computer interaction.

[0230] In some embodiments, based on the method for displaying recommended content provided in the embodiments of the present application, a content recommendation method based on topic switching can also be implemented.

[0231] In some embodiments, the first content recommendation interface includes multiple first recommended contents, the multiple first recommended contents correspond to the first recommended topic, the first interface element is adapted to the first recommended topic, the first recommended topic corresponds to an interest tendency score, and the interest tendency score is used to represent the first account's interest tendency degree in the first recommended topic. Figure 14 , Figure 14 This is a flowchart of a content recommendation method based on topic switching provided by an exemplary embodiment of the present application. The method can be executed by a terminal, a server, or both. The embodiment of the present application takes the method executed by a first client logged in with a first account as an example. Figure 14 As shown, the above step 420 includes the following steps:

[0232] Step 1410: Display multiple first recommended contents corresponding to a first recommended theme and a first interface element adapted to the first recommended theme in a first content recommendation interface.

[0233] In some embodiments, the first interface element corresponds to a first element display style adapted to the first recommended theme.

[0234] Optionally, the first element display style can be used to indicate display parameters of the first interface element, such as element type, color scheme, display size, and color contrast.

[0235] Schematically, the first content recommendation interface corresponding to the first account A includes multiple first recommended contents and first interface elements displayed in an aggregated manner. The multiple first recommended contents correspond to the first recommendation theme of "recommended book list". The multiple first recommended contents include recommended books 01, recommended books 02 and recommended books 03, corresponding to the first recommendation theme of "recommended book list". The first interface element includes a book sticker element, which is adapted to the first recommendation theme of "recommended book list".

[0236] Step 1420 : In response to receiving a theme switching operation for the first content recommendation interface, multiple candidate recommended contents corresponding to the candidate recommended themes and candidate interface elements adapted to the candidate recommended themes are displayed in the first content recommendation interface.

[0237] Among them, the interest tendency score corresponding to the first recommended topic is higher than the interest tendency score corresponding to the candidate recommended topic, and the interest tendency score corresponding to the candidate recommended topic reaches the preset interest tendency score threshold.

[0238] In some embodiments, the candidate interface elements correspond to candidate element display styles that are adapted to the candidate recommendation themes.

[0239] Optionally, the theme switching operation may be a sliding operation, a triggering operation on a theme switching control, a shaking operation, etc., which is not limited in this application.

[0240] For illustration, please refer to Figure 15 , Figure 15 This is a schematic diagram of a theme switching method provided by an exemplary embodiment of the present application. Figure 15As shown, multiple first recommended contents 1510 corresponding to the first recommended theme and a first interface element 1520 adapted to the first recommended theme are aggregated and displayed in the first content recommendation interface 1500; in response to receiving a theme switching operation for the first content recommendation interface 1500, such as swiping left to switch the theme type, the first recommended theme corresponding to the first content recommendation interface 1500 is switched to a candidate recommended theme, and multiple candidate recommended contents 1530 corresponding to the candidate recommended theme and a candidate interface element 1540 adapted to the candidate recommended theme are displayed in the first content recommendation interface 1500.

[0241] In some embodiments, the above interest tendency score is obtained by analyzing the interest tendency of historical data of the first account through a machine learning model.

[0242] Historical data is the operation data of the first account in the first client during the historical time period, including but not limited to the information data, access data, storage data, etc. input by the first account into the first client. The first interest tendency label is used to indicate the first recommended topic. It is worth noting that, as emphasized above, this application displays a prompt interface, pop-up window or outputs voice prompt information before collecting historical data and in the process of collecting the user's historical data. The prompt interface, pop-up window or voice prompt information is used to prompt the user that his historical data is currently being collected, so that this application only starts to execute the relevant steps of obtaining the user's historical data after obtaining the user's confirmation operation on the prompt interface or pop-up window. Otherwise (that is, when the user's confirmation operation on the prompt interface or pop-up window is not obtained), the relevant steps of obtaining the user's historical data are terminated, that is, the user's historical data is not obtained.

[0243] In some embodiments, different content recommendation interfaces are provided for different aggregation dimensions, including but not limited to the popularity, release time, content quality, and release channel of the recommended content. For each aggregation dimension, multiple different types of recommended content are generated based on the interest propensity score of the first account.

[0244] Optionally, a priority relationship exists between the multiple aggregation dimensions, and the priority relationship may be preset by the first application.

[0245] Schematically, according to the priority relationship preset by the first application, the priority of the aggregation dimensions of preferential discounts, release time, and popularity is arranged from high to low. The interest tendency corresponding to the first account indicates multiple recommendation types, which are games, videos, and books in descending order according to the interest tendency score. Then, when displaying the content recommendation interface, the content recommendation interface with the aggregation dimension of preferential discounts and the recommendation type of games is displayed first. The content recommendation interface is used to aggregate and recommend multiple virtual games with preferential discounts to the first account; in response to receiving a horizontal sliding operation, the aggregation content type for the preferential discount dimension is switched, and the corresponding recommendation type is switched to videos or books in turn. According to the switched recommendation type, the content recommendation interface for recommending videos or the content recommendation interface for recommending books is displayed. A content recommendation interface for recommending books, wherein the content recommendation interface for recommending videos aggregates multiple videos with viewing rights and discounts, and the content recommendation interface for recommending books aggregates multiple books with purchase discounts; in the case of displaying a content recommendation interface with the aggregation dimension as discount and the recommendation type as game, in response to receiving a vertical sliding operation, the aggregation dimension for the game type is switched, and the aggregation dimension is switched to release time or popularity in turn. For example, upon receiving a first sliding operation, the aggregation dimension is switched to release time, and a content recommendation interface for recommending the latest released games is displayed. Upon receiving a second sliding operation, the aggregation dimension is switched to popularity, and a content recommendation interface for recommending games with the highest downloads in a historical time period is displayed.

[0246] To sum up, the method provided in the embodiment of the present application realizes a content recommendation scheme based on topic switching by displaying the first recommended content corresponding to the first recommended topic and the first interface element adapted to the first recommended topic in the first content recommendation interface, and in response to receiving a topic switching operation for the first content recommendation interface, displaying the candidate recommended content corresponding to the candidate recommended topic and the candidate interface element adapted to the candidate recommended topic in the first content recommendation interface. In the aggregated recommendation scenario, it can recommend aggregated recommended content under multiple recommended topics according to user interest tendencies, thereby improving recommendation efficiency.

[0247] This application also provides a method for generating recommended content. Figure 16 , which shows a flowchart of a method for generating recommended content provided by an exemplary embodiment of the present application. The method can be executed by a terminal, a server, or both. The embodiment of the present application takes the method executed by the server as an example for explanation. Figure 16 As shown, the method includes the following steps:

[0248] Step 1610 : In response to receiving a recommendation request from a first account, a plurality of first recommended contents are aggregated based on personalization of the first account.

[0249] The recommendation request is a request sent when the first client logged in with the first account is started and a splash screen interface is displayed.

[0250] In some embodiments, a candidate recommendation set is stored in the server, and the candidate recommendation set includes a plurality of candidate recommendation contents. The server reads the first recommendation content from the candidate recommendation set based on the first account.

[0251] Optionally, the first recommended content may also be generated by the server in real time based on the first account.

[0252] In some embodiments, an artificial intelligence model is used to analyze the first account's interest in recommended content based on historical data of the first account to obtain a first content tag, and multiple first recommended contents are aggregated according to the first content tag.

[0253] The first content tag is used to indicate recommended content that matches the interests of the first account.

[0254] In some embodiments, the server aggregates multiple candidate recommended contents matching the first content tag from the candidate recommendation set as multiple first recommended contents through an artificial intelligence model.

[0255] Step 1620: Generate a first interface element based on at least one of the first account and the first recommended content.

[0256] The first interface element is personalized content for the first account.

[0257] Optionally, the first interface element includes at least one of the layout of the first content recommendation interface, the first background image, and the first control element.

[0258] The layout of the first content recommendation interface is used to indicate at least one of the display position, display size, layout, text font, and display angle of the first recommended content and other display elements in the interface (such as recommended topics, first control elements, etc.) in the first content recommendation interface.

[0259] Optionally, the first control element includes but is not limited to any control element such as an information display control, a container control, a navigation control, a multimedia control, a notification control, a progress control, and an interactive control.

[0260] In some embodiments, the first interface element is generated by an artificial intelligence model based on historical data of the first account and at least one of the recommendation styles corresponding to multiple first recommended contents.

[0261] Optionally, the first interface element is generated based on the historical data of the first account, or based on the recommendation style corresponding to the first recommended content, or based on the historical data of the first account and the recommendation style corresponding to the first recommended content.

[0262] In some embodiments, step 1620 includes the following three steps:

[0263] The first step is to use an artificial intelligence model to analyze the first account's interest in the recommended style based on the first account's historical data to obtain a first style label.

[0264] Optionally, a pre-trained first style analysis model is used to analyze the first account's interest tendency in the recommended style based on historical data of the first account to obtain a first style label.

[0265] Optionally, the historical data includes but is not limited to the browsing time, number of views, interaction frequency of the historical recommended content of the same recommendation style by the first account in the first client during the historical time period, as well as the avatar used by the first account during the historical time period, the layout of the account homepage, etc.

[0266] In some embodiments, a training process of the first style analysis model is also included. Taking the example of the first style analysis model generating a first style label based on the historical data of the first account, in this training process, account sample data marked with a reference style label is obtained, and a sample style label is generated based on the account sample data by a first candidate style analysis model. The style prediction loss is determined based on the style difference between the reference style label and the sample style label, and the first candidate style analysis model is trained based on the style prediction loss until it meets the training requirements, thereby obtaining the above-mentioned first style analysis model.

[0267] Optionally, the training requirement includes but is not limited to at least one of the following: the number of training times reaches a preset number threshold, or the style prediction loss reaches a preset loss threshold, or the style prediction loss converges.

[0268] Illustratively, the first style tag includes style categories such as simple, gorgeous, and colorful that are consistent with the interests of the first account.

[0269] In the second step, the style features corresponding to the multiple first recommended contents are analyzed through an artificial intelligence model to obtain a second style label.

[0270] Optionally, the style features corresponding to the plurality of first recommended contents are analyzed by a pre-trained second style analysis model to obtain the first recommended theme as the second style label.

[0271] Illustratively, if multiple first recommended contents all belong to film and television works, the analysis results in the first recommendation theme being "film and television recommendation" as the second style label; further, if multiple first recommended contents all belong to horror movies, the analysis results in the first recommendation theme being "horror" + "movie recommendation" as the second style label.

[0272] In some embodiments, a training process of a second style analysis model is also included. Taking the second style analysis model analyzing the style features corresponding to multiple first recommended contents and obtaining the first recommended topic as the second style label as an example, in this training process, a sample recommended content group marked with a reference style label is obtained. The sample recommended content group includes multiple sample recommended contents corresponding to the same reference style label. A sample recommended topic is generated as a sample style label based on the sample recommended content group through a second candidate style analysis model. The style prediction loss is determined based on the style difference between the reference style label and the sample style label. The second candidate style analysis model is trained based on the style prediction loss until it meets the training requirements to obtain the above-mentioned second style analysis model.

[0273] Optionally, the training requirement includes but is not limited to at least one of the following: the number of training times reaches a preset number threshold, or the style prediction loss reaches a preset loss threshold, or the style prediction loss converges.

[0274] Optionally, the first and second steps may be performed sequentially, in reverse order, in parallel, or one of them may be performed at a time.

[0275] In the third step, a first interface element is generated according to at least one of the first style label and the second style label through an artificial intelligence model.

[0276] Optionally, the first interface element is generated according to the first style label through a pre-trained interface element generation model; or, the first interface element is generated according to the second style label through a pre-trained interface element generation model; or, the first interface element is generated according to the first style label and the second style label through a pre-trained interface element generation model.

[0277] Schematically, according to the first style tag, it is determined that the element display style of the first interface element belongs to the simple type, and according to the second style tag, it is determined that the first interface element is adapted to the recommendation theme of "funny" + "book recommendation", and a simple layout of the first content recommendation interface is generated, as well as a first background image including rendering elements of funny elements and book stickers.

[0278] In some embodiments, the interface element generation model generates at least one first interface element such as the layout of the first content recommendation interface, the first background image, the first control element, and the first recommended theme according to the first style tag.

[0279] In some embodiments, the server may also directly generate a first content recommendation interface, and send the first content recommendation interface to a first client corresponding to the first account for display.

[0280] Taking the aggregated recommendation scenario as an example, the recommendation interface generation model aggregates the recommended content corresponding to multiple recommendation subjects to generate a content recommendation interface.

[0281] The content recommendation interface includes an aggregate recommendation area and recommendation interface elements. The aggregate recommendation area includes recommended content corresponding to multiple recommendation subjects. The recommendation interface elements are used to decorate the content recommendation interface. The element display style of the recommendation interface elements corresponds to the first subject type.

[0282] Schematically, the recommendation interface generation model generates a display template for the aggregate recommendation area according to the recommendation requirements set by the first application and the first subject type. The recommendation requirements are used to indicate the number of recommended subjects in the aggregate recommendation area, the display position of the recommended content, the display size and other display parameters. The recommendation interface generation model determines the recommendation theme according to the first subject type, determines the element display style according to the recommendation theme, and then generates recommendation interface elements according to the element display style, such as theme elements, background rendering elements, etc., and fills the recommended contents corresponding to multiple recommendation subjects into the aggregate recommendation area according to the display template, and combines and renders the aggregate recommendation area and the recommendation interface elements to obtain a content recommendation interface.

[0283] Optionally, the recommendation interface generation model can be implemented as an image generation model, which generates a display template and recommendation interface elements according to the recommendation requirements, and renders the aggregated recommendation area and recommendation interface elements according to the display template combination to generate a content recommendation interface.

[0284] In some embodiments, a training process for a recommendation interface generation model is also included. During this training process, labeled data-image pairs are obtained, where the data-image pairs include sample data and reference recommended images, where the sample data includes user operation data and aggregated data sampled from a backend database of the first application, and the reference recommended images include reference aggregated regions and reference associated elements. Sample recommended images are generated based on the data-image pairs by a candidate recommendation interface generation model, an image generation loss is determined based on the image difference between the reference recommended image and the sample recommended image, and the candidate recommendation interface generation model is trained based on the image generation loss until the candidate recommendation interface generation model meets the training requirements, thereby obtaining the aforementioned recommendation interface generation model.

[0285] Optionally, the training requirements include but are not limited to at least one of the number of training times reaching a preset number threshold, or the image generation loss reaching a preset loss threshold, or the image generation loss converging.

[0286] The content recommendation interface is automatically generated through the above recommendation interface generation model, which reduces the complexity, workload and cost of manually constructing the content recommendation interface and improves the generation efficiency and quality of the content recommendation interface.

[0287] Step 1630: Send first recommendation information to the first client corresponding to the first account.

[0288] The first recommendation information is used to provide relevant information of a first content recommendation interface displayed by the first client. The first content recommendation interface includes first recommended content and a first interface element.

[0289] Optionally, the server may send the first recommendation information to the first client in real time based on the recommendation request sent by the first account, or may send the first recommendation information to the first client at a predetermined period.

[0290] By sending the first recommendation information to the first client in real time, the real-time performance of the multiple first recommended contents aggregated and displayed in the first content recommendation interface can be ensured when the first client starts and displays the splash screen interface; by sending the first recommendation information to the first client at regular intervals, the first client can still display the first content recommendation interface based on the first recommendation information of the previous cycle when the first client starts and displays the splash screen interface even when the network status is poor, thereby improving the display efficiency of the content recommendation interface.

[0291] To sum up, the method provided in the embodiment of the present application is that, by the server responding to a recommendation request received from the first account, the server obtains first recommended content based on the first account, generates a first interface element based on the first account and at least one of the first recommended content, and sends first recommendation information to the first client corresponding to the first account. The first recommendation information is used to indicate the first recommended content and the first interface elements other than the first recommended content in the multiple first content recommendation interfaces aggregated and displayed by the first client, thereby realizing a personalized aggregate recommendation scheme for the user account, making the generated recommendation information more in line with the interest tendencies of the account, improving the attractiveness to users, improving the generation quality and efficiency of the recommendation information, and at the same time reducing the interface design cost, reducing the content recommendation cost, and improving the content recommendation efficiency.

[0292] Please refer to Figure 17 , Figure 17 This is a flowchart of a personalized content aggregation method provided by an exemplary embodiment of the present application. The method can be executed by a terminal, a server, or both. The embodiment of the present application takes the method executed by the server as an example for explanation. Figure 17 As shown, the above step 1610 includes the following steps:

[0293] Step 1611: Obtain historical data of the first account.

[0294] The historical data is the operation data of the first account in the first client during a historical period.

[0295] Optionally, the historical data includes but is not limited to information data, access data, storage data, etc. input into the first application by the first account.

[0296] It is worth noting that, as emphasized above, this application displays a prompt interface, pop-up window or outputs voice prompt information before collecting historical data and during the process of collecting user historical data. The prompt interface, pop-up window or voice prompt information is used to remind the user that his or her historical data is currently being collected, so that this application only starts to execute the relevant steps of obtaining user historical data after obtaining the user's confirmation operation on the prompt interface or pop-up window. Otherwise (that is, when the user's confirmation operation on the prompt interface or pop-up window is not obtained), the relevant steps of obtaining user historical data are terminated, that is, the user's historical data is not obtained.

[0297] Illustratively, the historical data includes but is not limited to at least one of the types of recommended subjects that the first account follows during a historical period, the frequency of interaction with the recommended subjects, and the frequency of access to the same type of recommended content.

[0298] Step 1612: Analyze the first account's interest in the recommended content based on the first account's historical data through an artificial intelligence model to obtain a first content tag.

[0299] The first content tag is used to indicate recommended content that matches the interests of the first account.

[0300] Optionally, the first interest analysis model is used to analyze the interest tendency of the first account based on historical data such as the type of recommended subjects followed by the first account in a historical time period, the frequency of interaction with the recommended subjects, and the frequency of visits to the same type of recommended content to obtain the first content label.

[0301] For example, based on the recommended content access data of the first account within a historical time period, the recommended content category corresponding to the recommended content clicked by the first account within the historical time period is identified, and multiple content categories are obtained. The content category whose click-through rate reaches a preset threshold is determined as the tendency content category and used as the first content label.

[0302] Optionally, the first content tag may be used to indicate one or more recommended topics corresponding to the recommended content that the first account tends to access.

[0303] After obtaining the first content tag, the server associates and stores the first account and the first content tag.

[0304] In some embodiments, a training process of the first interest analysis model is also included. During this training process, sample account data marked with reference interest tags are obtained, and the interest tendencies corresponding to the sample account data are analyzed by the first candidate interest analysis model to obtain sample interest tags. Based on the difference between the reference interest tags and the sample interest tags, the interest analysis loss is determined, and the first candidate interest analysis model is trained based on the interest analysis loss until it meets the training requirements, thereby obtaining the above-mentioned first interest analysis model.

[0305] Optionally, the training requirements include but are not limited to at least one of the number of training times reaching a preset number threshold, or the interest analysis loss reaching a preset loss threshold, or the interest analysis loss converging.

[0306] The above-mentioned first interest analysis model realizes automatic interest tendency analysis, so that the first recommended content in the content recommendation interface can be more in line with the interest tendency of the first account based on the first interest tag, thereby realizing personalized content recommendation, improving the accuracy of content recommendation, and thus improving content recommendation efficiency.

[0307] Step 1613: Aggregate multiple first recommended contents according to the first content tag.

[0308] In some embodiments, a candidate recommendation set is stored in the server, and the candidate recommendation set includes multiple candidate recommendation contents. The candidate recommendation contents correspond to preset recommendation tags, which are used to indicate the recommendation characteristics of the candidate recommendation content, such as the recommendation subject type, recommended content type, recommended content style, and recommended theme.

[0309] Optionally, the above-mentioned recommendation tags are automatically generated by the server after identifying the recommendation features of the candidate recommended content through an artificial intelligence model.

[0310] Illustratively, candidate recommended content having the same recommendation tag as the first content tag or a similarity reaching a preset similarity threshold is retrieved from the candidate recommendation set as the first recommended content.

[0311] In some embodiments, step 1613 includes the following steps:

[0312] The first step is to obtain the candidate recommendation set.

[0313] The candidate recommendation set includes multiple candidate recommendation contents.

[0314] Optionally, the candidate recommended content corresponds to a preset recommendation tag, which is used to indicate the recommendation characteristics of the candidate recommended content, such as the recommendation subject type, recommended content type, recommended content style, and recommended theme.

[0315] Indicatively, recommendation tags include recommended topics, subject types, recommended content types, recommended content styles, download volumes and click-through rates of recommended content within a historical period, release time, and release channels.

[0316] The embodiment of the present application is described by taking the generation of the above-mentioned recommendation tags based on an artificial intelligence model as an example.

[0317] The second step is to use the artificial intelligence model to identify the recommendation features of multiple candidate recommendation contents and obtain multiple recommendation labels.

[0318] The recommendation tag may be used to indicate a recommendation topic corresponding to the candidate recommendation content.

[0319] For example, if the recommended subject is a film or TV series, the recommended tags can be implemented as genre tags, such as thriller, suspense, mystery, romance, etc. The recommended tags can also be implemented as the starring actors of the film or TV series, or the poster style of the film or TV series, such as warm, scary, etc. For example, if the recommended subject is a book, the recommended tags can be implemented as book subject category tags, such as novels, essays, poetry, etc. The civil engineering tag can also be implemented as the author of the book, or the cover style of the book, such as simple, gorgeous, etc.

[0320] Optionally, the features of multiple candidate recommendation contents are identified by a pre-trained recommendation recognition model to obtain the above-mentioned recommendation labels.

[0321] In some embodiments, the method further includes a training process for the above-mentioned recommendation recognition model. During the training process, sample recommendation content marked with a reference recommendation label is obtained, and the features of the sample recommendation content are identified by a candidate recommendation recognition model to obtain a sample recommendation label. Based on the difference between the reference recommendation label and the sample recommendation label, the feature recognition loss is determined, and the candidate recommendation recognition model is trained based on the feature recognition loss until it meets the training requirements to obtain the above-mentioned recommendation recognition model.

[0322] Optionally, the training requirements include but are not limited to at least one of the number of training times reaching a preset number threshold, or the feature recognition loss reaching a preset loss threshold, or the feature recognition loss converging.

[0323] The above-mentioned recommendation recognition model automatically identifies the recommended content features and obtains recommendation tags for participating in the generation of the content recommendation interface. This avoids the workload of manual labeling, improves the efficiency of obtaining recommendation tags, facilitates subsequent aggregation and classification, provides a basis for the implementation of aggregated recommendations, and improves the efficiency of content recommendation.

[0324] In the third step, an artificial intelligence model is used to aggregate multiple first recommended contents from the candidate recommendation set based on the matching degree between the recommended tags and the first content tags.

[0325] Optionally, a pre-trained aggregation model is used to automatically identify the text similarity between the first content tag and multiple recommended tags as the matching degree, and the candidate recommended contents corresponding to the multiple recommended tags whose matching degree ranking reaches a preset ranking threshold are aggregated into multiple first recommended contents.

[0326] The aggregation model described above automatically aggregates multiple first recommended contents, thereby improving content aggregation efficiency.

[0327] In summary, the method provided in the embodiment of the present application realizes an automated content aggregation solution through an artificial intelligence model, improves the efficiency of content aggregation recommendation, and while increasing the richness of recommended content in the content recommendation interface, ensures personalized aggregation for accounts, and can recommend content to increase its attractiveness to users, thereby improving recommendation efficiency.

[0328] In this way, we will illustrate the above recommendation system, the display and generation method of recommended content, and take the aggregated advertising scenario as an example. Figure 18 , Figure 18 This is an interactive flow chart of a content recommendation method provided by an exemplary embodiment of the present application. Figure 18 As shown, the interaction process is performed by the advertising server, the creative server, and the terminal. The advertising server is used to obtain push advertisements, the creative server is used to generate a content recommendation interface based on the advertisements provided by the advertising server, and the terminal is used to display the content recommendation interface. The above interaction process includes the following steps:

[0329] Step 1811: The advertising server labels and categorizes the advertisements.

[0330] The ad server uses a pre-trained machine learning model to identify ad content and assign a label to the ad. The label can be used to indicate the ad type, content, style, etc.

[0331] Indicatively, the recommended tags include advertising brand category, advertising product category, advertising format category (such as video, picture, text, etc.), advertising content style (such as simple or gorgeous display style of advertising pictures), publishing channel, etc.

[0332] Optionally, the brand category, product category, form category, style category, etc. of the advertisement can be identified separately through multiple pre-trained classification models.

[0333] Step 1812: The advertising server accesses the dimension data of the advertisement.

[0334] The advertising server obtains data such as the click-through rate of the advertisement, the purchase volume of the advertising product, the exposure volume of the advertisement, the release time, the release channel, etc. within the historical advertising period for later analysis and creative generation.

[0335] When classifying advertisement aggregation, advertisements that match the aggregation requirements can be aggregated based on these dimensional data.

[0336] Schematically, the aggregation requirement indicates that the advertisements whose click volume reaches a preset threshold are aggregated. Then, the advertising server reads the advertisements whose click volume reaches the preset threshold based on the click rate data in the dimensional data to implement aggregation classification.

[0337] Step 1813: The advertising server defines clustering information.

[0338] The ad server defines clustering information based on the ad's tags and dimension data, such as rankings, popular, and latest games, and clusters the ads based on the clustering information.

[0339] Schematically, the advertising server can determine the target clustering information based on the delivery needs of the advertiser or application. The target clustering information includes the target recommendation tag and the target dimension. For example, based on the advertising delivery needs of the game platform, the target recommendation tag is determined to be the game, and the target dimension is the release time, aggregating the latest released games.

[0340] Step 1814: The advertising server sends the advertising data to the creative server.

[0341] The advertisement server sends advertisement data corresponding to the clustered advertisements to the creative server.

[0342] Advertising data includes but is not limited to advertising categories, industries, aggregation dimensions, title information, etc.

[0343] Advertising data is used to participate in generating the content recommendation interface.

[0344] Step 1821: The creative server generates a recommendation template based on the data.

[0345] The creative server automatically generates a recommendation template based on the received advertising data through a machine learning model, and a content recommendation interface can be generated by filling the recommendation template with materials.

[0346] For example, the creative server determines, based on the received advertising data, that the content recommendation interface needs to recommend 10 popular games in the form of a list. The machine learning model then automatically generates a recommendation list template for filling in the 10 game advertisements. The content recommendation interface is generated by filling the received advertising data with the 10 popular game advertisement contents according to the recommendation list template.

[0347] Step 1822: The creative server generates materials.

[0348] The creative server generates materials with consistent style based on the clustering dimensions to fill the content recommendation interface, including the aggregated recommendation area and recommendation interface elements.

[0349] Optionally, the recommendation interface elements include but are not limited to theme elements, advertising decoration elements, background rendering elements, etc. The theme elements are the advertising themes corresponding to the content recommendation interface, the advertising decoration elements are used to decorate the advertising content, and the background rendering elements are used to decorate the background screen of the content recommendation interface.

[0350] Step 1823: Fill in the advertiser's avatar name.

[0351] The creative server fills the creative materials and the clustered advertising avatar names into the recommendation template to generate a content recommendation interface.

[0352] For example, taking the above-mentioned recommendation list template for recommending 10 popular games as an example, the game names and game posters of the 10 popular games are filled in the recommendation list template in order of the popularity of the games.

[0353] Step 1824: The creative server sends the creative image to the advertising server.

[0354] The creative server sends the generated content recommendation interface as a creative graph to the advertising server.

[0355] Optionally, the creative server may send the creative image to the advertising server immediately after generating the content recommendation interface, or may send the creative image to the advertising server in batches according to a preset period, which is not limited in this application.

[0356] Step 1815: The advertising server pushes the advertisement to the terminal.

[0357] The advertising server sends a content recommendation interface to the terminal for display based on the advertising delivery needs, thus realizing advertising push.

[0358] Optionally, the advertising server can respond to the delivery request sent by the terminal and send a content recommendation interface to the terminal in real time to ensure the real-time nature of the content recommendation. It can also send the content recommendation interface to the terminal according to a preset period, which is preloaded locally after being received by the terminal. When the content recommendation interface needs to be displayed, the terminal directly reads the preloaded content recommendation interface from the local computer.

[0359] In summary, the method provided in the embodiment of the present application analyzes the interest tendency of the first account based on historical data through a first machine learning model to obtain a first interest tendency label for indicating the first subject type, clusters multiple candidate recommendation subjects in the candidate recommendation set according to the first subject type, and obtains multiple recommendation subjects corresponding to the first subject type, thereby realizing the aggregation of recommendation subjects according to the interest tendency of the first account, aggregating the recommended content corresponding to the multiple recommendation subjects through a second machine learning model, and generating a content recommendation interface including an aggregated recommendation area and recommendation interface elements, realizing the generation scheme of the aggregated recommendation interface, improving the efficiency of recommendation content generation, and the display style of the recommendation interface elements corresponds to the first subject type, thereby reducing the interface design cost while ensuring the rationality of the interface display style, and realizing personalized recommendations for the first account as a whole, and at the same time aggregating the recommended content of multiple recommendation subjects, enriching the content recommendation interface, increasing the attractiveness to users, reducing the content recommendation cost, and improving the content recommendation efficiency.

[0360] In combination with the above embodiments, the application scenarios of the recommended content display method and the recommended content generation method provided in the embodiments of the present application are explained. The recommended content display method and the recommended content generation method provided in the embodiments of the present application can be applied to various content recommendation fields such as advertising recommendation, video recommendation, audio recommendation, book recommendation, and game recommendation, and the embodiments of the present application are not limited to this.

[0361] 1. Take the application of content recommendation method in the field of advertising recommendation as an example to illustrate.

[0362] Illustratively, the content recommendation interface provided in the embodiment of the present application is implemented as an advertisement recommendation interface. A personalized advertisement recommendation function is implemented through an advertisement recommendation system, which includes a server and multiple clients logged in with different accounts.

[0363] The server in the advertising recommendation system is used to send personalized advertising recommendation information to multiple corresponding clients for different accounts. Taking a first client logged in with a first account and a second client logged in with a second account as an example, the server is used to send first advertising recommendation information to the first client based on the first account, and to send second advertising recommendation information to the second client based on the second account; the first client is used to display the first advertising recommendation interface as a splash screen interface based on the first advertising recommendation information when the first client is started to display the splash screen interface, and the first advertising recommendation interface includes multiple first advertising contents displayed in an aggregate and first interface elements other than the first advertising content; the second client is used to display the second advertising recommendation interface as a splash screen interface based on the second advertising recommendation information when the second client is started to display the splash screen interface, and the second advertising recommendation interface includes multiple second advertising contents displayed in an aggregate and second interface elements other than the second advertising content; wherein the first recommendation information is different from the second recommendation information.

[0364] In some embodiments, multiple first advertising contents correspond to a first advertising theme, and the first interface element is adapted to the first advertising theme; multiple second advertising contents correspond to a second advertising theme, and the second interface element is adapted to the second advertising theme.

[0365] The multiple first advertising contents are recommended contents aggregated by the server based on the first account through the artificial intelligence model, and the multiple second advertising contents are recommended contents aggregated by the server based on the second account through the artificial intelligence model.

[0366] The first interface element is artificial intelligence-generated content generated by the server based on at least one of the first advertising content and the first account, and the second interface element is artificial intelligence-generated content generated by the server based on at least one of the second advertising content and the second account.

[0367] In some embodiments, the first interface element includes at least one of the layout of the first advertising recommendation interface, the first background image, and the first control element, and the second interface element includes at least one of the layout of the second advertising recommendation interface, the second background image, and the second control element.

[0368] A method for displaying advertising content based on an advertising recommendation system is described below, taking as an example a first client logged into the advertising recommendation system as the execution subject. The first client obtains first advertising recommendation information from a server. When the first client is started to display a splash screen interface, a first advertising recommendation interface is displayed as a splash screen interface based on the first advertising recommendation information. The first advertising recommendation interface includes multiple first advertising contents displayed in an aggregated manner and first interface elements other than the first advertising contents. The first recommendation information is personalized content for the first account.

[0369] Optionally, the multiple first advertising contents are recommended contents aggregated by the server based on the first account through an artificial intelligence model; or, the first interface element is an element generated by the server based on the first account through an artificial intelligence model; the multiple first advertising contents and first interface elements are personalized generated by the server for the first account through an artificial intelligence model.

[0370] Taking the first background image in the first interface element including the first theme element as an example, the first client associates and displays the first theme element and the first advertisement content in the first advertisement recommendation interface, and the first theme element is the first recommended theme corresponding to the first advertisement content.

[0371] Taking the example that the first control element in the first interface element includes an interactive control, the first client displays the interactive control in the first advertising recommendation interface in response to the first advertising recommendation interface meeting the interactive requirements. The interactive control is used to receive interactive operations for the first content recommendation interface.

[0372] Optionally, the first client displays a first interactive control in the first advertisement recommendation interface when the display duration of the first advertisement recommendation interface meets the interaction requirement. The first interactive control is used to trigger the display of a topic interactive interface corresponding to the recommended topic.

[0373] Optionally, when the first advertisement content meets the interaction requirement, the first client displays a second interactive control in association with the first advertisement content. The second interactive control is used to receive an interactive operation on the first advertisement content.

[0374] Optionally, when the first background image meets the interaction requirement, the first client displays a third interaction control for receiving an interaction operation on the first background image.

[0375] In some embodiments, the plurality of first recommended advertisements belong to the same recommended theme, and the first interface element is adapted to the recommended theme.

[0376] In the aggregated advertising recommendation scenario, taking the case where the first advertising recommendation interface includes aggregated recommendation areas corresponding to multiple aggregated dimensions, the first interface elements include advertising decoration elements, and the advertising decoration elements are used to decorate advertising content, as an example, the first client displays multiple first advertising contents belonging to the i-th aggregation dimension in the i-th aggregation recommendation area among the multiple aggregated recommendation areas, where i is a positive integer; for the first advertising content in the aggregated recommendation area that is ranked from high to low according to the aggregation score and reaches a preset ranking threshold, the advertising decoration element is displayed according to the aggregation score corresponding to the first advertising content, wherein the aggregation score is used to characterize the matching degree between the first advertising content and the aggregation dimension corresponding to the aggregated recommendation area, and the visual prominence of the advertising decoration element is positively correlated with the recommendation score.

[0377] In the aggregated advertising recommendation scenario, taking the case where a first interface element includes multiple advertising blocking elements, the multiple advertising blocking elements are used to block multiple first advertising contents, and the multiple advertising blocking elements have corresponding blocking priorities as an example, the first client associates and displays multiple advertising blocking elements with multiple first advertising contents in the first content recommendation interface, wherein when the advertising blocking element is in a visible state, the first advertising content corresponding to the advertising blocking element is in an invisible state, and when the advertising blocking element is in an invisible state, the first advertising content corresponding to the advertising blocking element is in a visible state; receiving a cancel display operation for the multiple advertising blocking elements, the cancel display operation is used to switch the advertising blocking elements from a visible state to an invisible state; canceling the display of the multiple advertising blocking elements in order from high to low according to the blocking priority, the blocking priority corresponding to the advertising blocking element is negatively correlated with the recommendation score of the first advertising content corresponding to the advertising blocking element, and the recommendation score is determined in advance based on the correlation between the first advertising content and the recommended topic.

[0378] In some embodiments, the first advertising content corresponds to the first advertising theme, the first interface element is adapted to the first advertising theme, the first advertising theme corresponds to an interest tendency score, and the interest tendency score is used to represent the first account's interest tendency degree in the first advertising theme. The first client displays multiple first advertising contents corresponding to the first advertising theme and the first interface element adapted to the first advertising theme in the first content recommendation interface; in response to receiving a theme switching operation for the first content recommendation interface, the first client displays multiple candidate advertising contents corresponding to the candidate advertising theme and the candidate interface elements adapted to the candidate advertising theme in the first content recommendation interface; wherein, the interest tendency score corresponding to the first advertising theme is higher than the interest tendency score corresponding to the candidate advertising theme, and the interest tendency score corresponding to the candidate advertising theme reaches a preset interest tendency score threshold.

[0379] A method for generating advertising content based on an advertising recommendation system is described using an example in which the execution entity is a server of the advertising recommendation system. In response to receiving a recommendation request from a first account, the server aggregates multiple first advertising contents based on the personalized first account, where the recommendation request is sent when a first client logged in with the first account starts and displays a splash screen interface; generates a first interface element based on the first account and at least one of the multiple first advertising contents; and sends first advertising recommendation information to a first client corresponding to the first account, where the first advertising recommendation information indicates the multiple first advertising contents aggregated and displayed in a first advertising recommendation interface displayed by the first client.

[0380] In some embodiments, for the method of aggregating multiple first advertising contents, the server obtains historical data of the first account, which is the operation data of the first account in the first client during a historical time period; analyzes the first account's interest in the advertising content based on the historical data of the first account through an artificial intelligence model to obtain a first content label; and aggregates multiple first advertising contents according to the first content label.

[0381] Schematically, a candidate advertising set is obtained, which includes multiple candidate advertising contents; the recommendation features of the multiple candidate advertising contents are identified through an artificial intelligence model to obtain multiple recommendation tags; and the artificial intelligence model is used to aggregate multiple first advertising contents from the candidate advertising set based on the matching degree between the recommendation tags and the first content tags.

[0382] In some embodiments, the server generates the first interface element through an artificial intelligence model based on historical data of the first account and at least one of the recommended styles corresponding to the plurality of first recommended contents.

[0383] Illustratively, an artificial intelligence model is used to analyze the first account's interest in the recommended style based on the historical data of the first account to obtain a first style label; an artificial intelligence model is used to analyze style features corresponding to multiple first advertising contents to obtain a second style label; and an artificial intelligence model is used to generate a first interface element based on at least one of the first style label and the second style label.

[0384] 2. Take the application of interface interaction method in the field of social recommendation as an example to illustrate.

[0385] Illustratively, the content recommendation interface provided in the embodiment of the present application is implemented as a social recommendation interface. A personalized social recommendation function is implemented through a social recommendation system, which includes a server and multiple clients logged in with different accounts.

[0386] The server in the social recommendation system is used to send personalized social recommendation information to multiple corresponding clients for different accounts. Taking a first client logged in with a first account and a second client logged in with a second account as an example, the server is used to send first social recommendation information to the first client based on the first account, and to send second social recommendation information to the second client based on the second account; the first client is used to display a first social recommendation interface as a splash screen interface based on the first social recommendation information when the first client is started to display a splash screen interface, and the first social recommendation interface includes multiple first social contents displayed in an aggregated manner and first interface elements other than the first social contents; the second client is used to display a second social recommendation interface as a splash screen interface based on the second social recommendation information when the second client is started to display a splash screen interface, and the second social recommendation interface includes multiple second social contents displayed in an aggregated manner and second interface elements other than the second social contents; wherein the first social recommendation information is different from the second social recommendation information.

[0387] Illustratively, the first social content may be content such as account avatars corresponding to social accounts other than the first account, and the second social content may be content such as account avatars corresponding to social accounts other than the second account.

[0388] In some embodiments, the first interface element includes at least one of the layout of the first social recommendation interface, the first background image, and the first control element, and the second interface element includes at least one of the layout of the second social recommendation interface, the second background image, and the second control element.

[0389] The first interface element is artificial intelligence-generated content generated by the server based on multiple first social content and at least one of the first account, and the second interface element is artificial intelligence-generated content generated by the server based on multiple second social content and at least one of the second account.

[0390] Specific implementation details can be found in the corresponding embodiments of the above-mentioned recommendation system, method for displaying recommended content, and method for generating recommended content, which will not be repeated here.

[0391] Illustratively, a social application includes multiple social accounts. With user permission, a content recommendation interface is displayed during the use of the application. The content recommendation interface is used to recommend other social accounts in the application to a first account. The content recommendation interface displays an aggregated recommendation area and recommendation-related elements. The aggregated recommendation area includes account avatars and account names corresponding to the multiple recommended accounts, each of which corresponds to a first account type. The recommendation-related elements are used to decorate the content recommendation interface, and the display style of the recommendation-related elements corresponds to the first account type. Specific implementation details can be found in the description of the above-mentioned embodiments of the recommended content display method and the recommended content generation method, and are not further described here. By displaying the content recommendation interface during the operation of the first application, social accounts are recommended to the first account. The aggregated recommendation area and recommendation-related elements are displayed in the content recommendation interface. The account avatars and account names corresponding to the multiple recommended accounts, each of which corresponds to the first account type, are displayed in the aggregated recommendation area. This implements a clustered display scheme for multiple recommended accounts of the same account type, increasing the richness of the recommended content in the content recommendation interface, thereby increasing user appeal and improving content recommendation efficiency.

[0392] 3. Take the application of interface interaction method in the field of game recommendation as an example to illustrate.

[0393] Illustratively, the content recommendation interface provided in the embodiment of the present application is implemented as a game recommendation interface. The personalized game recommendation function is implemented through a game recommendation system, which includes a server and multiple clients with different logged-in accounts.

[0394] The server in the game recommendation system is used to send personalized game recommendation information to multiple corresponding clients for different accounts. Taking a first client logged in with a first account and a second client logged in with a second account as an example, the server is used to send first game recommendation information to the first client based on the first account, and to send second game recommendation information to the second client based on the second account; the first client is used to display a first game recommendation interface as a splash screen interface based on the first game recommendation information when the first client is started to display a splash screen interface, and the first game recommendation interface includes multiple first game contents displayed in an aggregated manner and first interface elements other than the first game contents; the second client is used to display a second game recommendation interface as a splash screen interface based on the second game recommendation information when the second client is started to display a splash screen interface, and the second game recommendation interface includes second game contents and second interface elements other than the second game contents; wherein, the first game recommendation information and the second game recommendation information.

[0395] Illustratively, the first game content may be a game poster that matches the interests of the first account, and the second game content may be a game poster that matches the interests of the second account.

[0396] In some embodiments, the first interface element includes at least one of the layout of the first game recommendation interface, the first background image, and the first control element, and the second interface element includes at least one of the layout of the second game recommendation interface, the second background image, and the second control element.

[0397] The first interface element is artificial intelligence-generated content generated by the server based on multiple first game contents and at least one of the first account, and the second interface element is artificial intelligence-generated content generated by the server based on multiple second game contents and at least one of the second account.

[0398] Specific implementation details can be found in the corresponding embodiments of the above-mentioned recommendation system, method for displaying recommended content, and method for generating recommended content, which will not be repeated here.

[0399] Schematically, a game application is implemented as a game integration platform, including multiple virtual games. With the user's permission, a content recommendation interface is displayed during the use of the application. The content recommendation interface is used to recommend virtual games to the first account that logs into the application. An aggregate recommendation area and recommendation-related elements are displayed in the content recommendation interface. The aggregate recommendation area includes game posters corresponding to multiple recommended games, wherein the multiple recommended games correspond to a first game type. The recommendation-related elements are used to decorate the content recommendation interface, and the display style of the recommendation-related elements corresponds to the first game type. For example, the opening screen interface of the application is used as the content recommendation interface. In response to starting the application, the content recommendation interface is displayed, and game posters of multiple latest shooting games are displayed in the aggregate recommendation area of ​​the content recommendation interface. The content recommendation interface is decorated with recommendation-related elements with a display style corresponding to the latest shooting games. For specific implementation details, please refer to the description of the embodiments corresponding to the above-mentioned recommended content display method and recommended content generation method, which will not be repeated here. By displaying a content recommendation interface during the operation of the first application, a virtual game is recommended to the first account, an aggregated recommendation area and recommendation-related elements are displayed in the content recommendation interface, and game posters corresponding to multiple recommended games are displayed in the aggregated recommendation area, wherein the multiple recommended games correspond to the first game type, thereby realizing a clustered display scheme for multiple recommended games belonging to the same game type, thereby improving the richness of recommended content in the content recommendation interface, thereby increasing the attractiveness to users and improving the efficiency of content recommendation.

[0400] It is worth noting that the above application scenarios are only illustrative examples and are not limited here.

[0401] Figure 19 is a structural block diagram of a display device for recommended content provided by an exemplary embodiment of the present application, such as Figure 19 As shown, the device includes the following parts:

[0402] Receiving module 1910, configured to obtain first recommendation information from a server;

[0403] Display module 1920, configured to display a first content recommendation interface as the splash screen interface based on the first recommendation information when the first client is started to display the splash screen interface, wherein the first content recommendation interface includes a plurality of first recommended contents displayed in an aggregated manner and first interface elements other than the first recommended contents;

[0404] The multiple first recommendation information are personalized content for the first account.

[0405] In some embodiments, the multiple first recommended contents are recommended contents aggregated by the server based on the first account through an artificial intelligence model; or, the first interface element is an element generated by the server based on the first account through an artificial intelligence model; or, the multiple first recommended contents and the first interface element are personalized generated by the server for the first account through an artificial intelligence model.

[0406] In some embodiments, the first interface element includes a first theme element;

[0407] The display module 1920 is further configured to display the first theme element and the plurality of first recommended contents in association with each other in the first content recommendation interface, where the first theme element is a first recommendation theme corresponding to the plurality of first recommended contents.

[0408] In some embodiments, the first content recommendation interface includes aggregate recommendation areas corresponding to a plurality of aggregate dimensions, and the first interface elements include content decoration elements, which are used to decorate recommended content;

[0409] The display module 1920 is further used to display multiple first recommended contents belonging to the i-th aggregation dimension in the i-th aggregation recommendation area among the multiple aggregation recommendation areas, where i is a positive integer; for the first recommended contents in the aggregation recommendation area that are ranked from high to low according to the aggregation score and reach a preset ranking threshold, the content decoration element is displayed according to the aggregation score, wherein the aggregation score is used to represent the matching degree between the multiple first recommended contents and the aggregation dimension corresponding to the aggregation recommendation area, and the visual prominence of the content decoration element is positively correlated with the recommendation score.

[0410] In some embodiments, the first interface element includes a plurality of content blocking elements, the plurality of content blocking elements are used to block the plurality of first recommended content, and the plurality of content blocking elements have corresponding blocking priorities;

[0411] The display module 1920 is further configured to:

[0412] Associating and displaying the plurality of content-blocking elements with the plurality of first recommended contents in the first content recommendation interface, wherein when the content-blocking element is in a visible state, the first recommended contents corresponding to the content-blocking element are in an invisible state, and when the content-blocking element is in the invisible state, the first recommended contents corresponding to the content-blocking element are in the visible state;

[0413] receiving a cancel display operation for the plurality of content blocking elements, wherein the cancel display operation is used to switch the content blocking elements from the visible state to the invisible state;

[0414] The multiple content blocking elements are displayed in sequence from high to low according to the blocking priority. The blocking priority corresponding to the content blocking element is negatively correlated with the recommendation score of the first recommended content corresponding to the content blocking element. The recommendation score is determined in advance based on the correlation between the first recommended content and the recommended topic.

[0415] In some embodiments, the plurality of first recommended contents correspond to a first recommended topic, the first interface element is adapted to the first recommended topic, the first recommended topic corresponds to an interest tendency score, and the interest tendency score is used to represent the first account's interest tendency in the first recommended topic;

[0416] The display module 1920 is further configured to:

[0417] Displaying the plurality of first recommended contents corresponding to the first recommendation theme and the first interface elements adapted to the first recommendation theme in the first content recommendation interface;

[0418] In response to receiving a theme switching operation for the first content recommendation interface, displaying a plurality of candidate recommended contents corresponding to the candidate recommended themes and candidate interface elements adapted to the candidate recommended themes in the first content recommendation interface;

[0419] Among them, the interest tendency score corresponding to the first recommended topic is higher than the interest tendency score corresponding to the candidate recommended topic, and the interest tendency score corresponding to the candidate recommended topic reaches a preset interest tendency score threshold.

[0420] To sum up, the device provided in the embodiment of the present application, during the process of recommending content, performs personalized individual analysis on different accounts through the server, thereby determining the recommended content and interface elements corresponding to different accounts, such as: after analyzing the first account, the first recommended content and the first interface element are obtained, and after analyzing the second account, the second recommended content and the second interface element are obtained. Among them, the first interface element and the second interface element are determined based on the personalized analysis of the account, so the interface elements are performed differently, realizing a personalized display scheme for interface elements in the content recommendation interface for different accounts, providing users with personalized recommendations for each person, which can improve the fit between the content recommendation interface and the user's interest tendencies, increase the attractiveness to users, and thus improve the recommendation efficiency.

[0421] Figure 20 This is a structural block diagram of a device for generating recommended content provided by an exemplary embodiment of the present application. Figure 20 As shown, the device includes the following parts:

[0422] Processing module 2010, configured to, in response to receiving a recommendation request from a first account, aggregate a plurality of first recommended contents based on the personalization of the first account, wherein the recommendation request is a request sent when a first client logged in with the first account starts and displays a splash screen interface;

[0423] The processing module 2010 is further configured to generate a first interface element based on the first account and at least one of the plurality of first recommended contents;

[0424] Sending module 2020 is used to send first recommendation information to the first client corresponding to the first account, and the first recommendation information is used to provide relevant information of the first content recommendation interface displayed by the first client, and the first content recommendation interface includes the multiple first recommended contents displayed in an aggregated manner and the first interface elements other than the first recommended contents.

[0425] In some embodiments, the processing module 2010 is further configured to:

[0426] Acquire historical data of the first account, where the historical data is operation data of the first account in the first client within a historical time period;

[0427] analyzing the first account's interest in the recommended content based on the first account's historical data using an artificial intelligence model to obtain a first content label;

[0428] The plurality of first recommended contents are aggregated according to the first content tag.

[0429] In some embodiments, the processing module 2010 is further configured to:

[0430] Obtaining a candidate recommendation set, wherein the candidate recommendation set includes a plurality of candidate recommendation contents;

[0431] Identify the recommendation features of the plurality of candidate recommendation contents through an artificial intelligence model to obtain a plurality of recommendation tags;

[0432] Aggregate the plurality of first recommended contents from the candidate recommendation set based on a matching degree between the recommendation tag and the first content tag through an artificial intelligence model.

[0433] In some embodiments, the processing module 2010 is further configured to:

[0434] The first interface element is generated by the artificial intelligence model based on the historical data of the first account and at least one of the recommendation styles corresponding to the multiple first recommended contents.

[0435] In some embodiments, the processing module 2010 is further configured to:

[0436] analyzing, by the artificial intelligence model, the interest tendency of the first account in the recommended style based on the historical data of the first account, to obtain a first style label;

[0437] Analyzing the style features corresponding to the plurality of first recommended contents by the artificial intelligence model to obtain a second style label;

[0438] The first interface element is generated by the artificial intelligence model according to at least one of the first style label and the second style label.

[0439] In some embodiments, the first interface element includes at least one of a layout of the first content recommendation interface, a first background image, and a first control element.

[0440] To sum up, the device provided by the embodiment of the present application obtains the first recommended content based on the first account by the server, generates a first interface element based on the first account and at least one of the first recommended content, and sends the first recommendation information to the first client corresponding to the first account. The first recommendation information is used to indicate the first recommended content and the first interface element other than the first recommended content in the first content recommendation interface displayed by the first client, thereby realizing a personalized recommendation scheme based on interface elements for the user account, making the generated recommendation information more in line with the interest tendencies of the account, improving the attractiveness to users, improving the generation quality and generation efficiency of the recommendation information, and at the same time reducing the interface design cost, reducing the content recommendation cost, and improving the content recommendation efficiency.

[0441] It should be noted that the display device for recommended content and the generation device for recommended content provided in the above embodiments are only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0442] Figure 21 The following is a block diagram of a terminal 2100 according to an exemplary embodiment of the present application. Terminal 2100 may be a smartphone, tablet computer, MP3 player, MP4 player, laptop computer, or desktop computer. Terminal 2100 may also be referred to as user equipment, portable terminal, laptop terminal, desktop terminal, or other similar names.

[0443] Typically, the terminal 2100 includes a processor 2101 and a memory 2102 .

[0444] The processor 2101 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 2101 may be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), and programmable logic array (PLA). The processor 2101 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a central processing unit (CPU); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 2101 may be integrated with a graphics processing unit (GPU), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 2101 may also include an artificial intelligence (AI) processor, which is used to process computing operations related to machine learning.

[0445] The memory 2102 may include one or more computer-readable storage media, which may be non-transitory. The memory 2102 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices and flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in the memory 2102 is used to store at least one instruction, which is executed by the processor 2101 to implement the method for displaying recommended content and / or the method for generating recommended content provided in the method embodiment of the present application.

[0446] In some embodiments, the terminal 2100 further includes other components, which can be understood by those skilled in the art. Figure 21 The structure shown in the figure does not constitute a limitation on the terminal 2100, and the terminal 2100 may include more or fewer components than shown in the figure, or combine certain components, or adopt a different component arrangement.

[0447] The embodiment of the present application further provides a computer device, which can be implemented as follows: Figure 1 The terminal or server shown. The computer device includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, code set, or instruction set, and the at least one instruction, at least one program, code set, or instruction set is loaded and executed by the processor to implement the method for displaying recommended content and / or the method for generating recommended content provided by the above-mentioned method embodiments.

[0448] An embodiment of the present application also provides a computer-readable storage medium, on which is stored at least one instruction, at least one program, code set, or instruction set, and the at least one instruction, at least one program, code set, or instruction set is loaded and executed by a processor to implement the recommended content display method and / or recommended content generation method provided in the above-mentioned method embodiments.

[0449] Embodiments of the present application also provide a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the method for displaying recommended content and / or the method for generating recommended content provided in the above-described method embodiments.

[0450] Optionally, the computer-readable storage medium may include: Read Only Memory (ROM), Random Access Memory (RAM), Solid State Drives (SSD), or an optical disk. Among them, the random access memory may include resistance random access memory (ReRAM) and dynamic random access memory (DRAM). The serial numbers of the above embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0451] Those skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware, or by a program to instruct the relevant hardware, and the program may be stored in a computer-readable storage medium, which may be a read-only memory, a disk, or an optical disk, etc.

[0452] The above description is merely an optional embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A recommendation system, characterized in that: The system includes a server, a first client and a second client, wherein the first client is logged in with a first account and the second client is logged in with a second account; The server is configured to send first recommendation information to the first client based on the first account; and sending second recommendation information to the second client based on the second account; The first client is configured to display a first content recommendation interface as the splash screen interface based on the first recommendation information when the first client is started to display the splash screen interface, wherein the first content recommendation interface includes a plurality of first recommended contents displayed in an aggregated manner and a first interface element other than the first recommended contents; the second client being configured to display a second content recommendation interface as the splash screen interface based on the second recommendation information when the second client is started to display the splash screen interface, the second content recommendation interface including a plurality of second recommended contents displayed in an aggregated manner and second interface elements other than the second recommended contents; The first recommendation information is different from the second recommendation information.

2. The system according to claim 1, wherein: The multiple first recommended contents correspond to a first recommended theme, and the first interface element is adapted to the first recommended theme; the multiple second recommended contents correspond to a second recommended theme, and the second interface element is adapted to the second recommended theme.

3. The system according to claim 1 or 2, characterized in that The plurality of first recommended contents are recommended contents aggregated by the server based on the first account through an artificial intelligence model, and the plurality of second recommended contents are recommended contents aggregated by the server based on the second account through the artificial intelligence model; The first interface element is an element generated by the server through an artificial intelligence model based on the first account and at least one of the multiple first recommended contents, and the second interface element is an element generated by the server through an artificial intelligence model based on the second account and at least one of the multiple second recommended contents.

4. A method for displaying recommended content, characterized in that: The method is executed by a first client logged in with a first account, and includes: Obtain first recommendation information from the server; When the first client is started to display a splash screen interface, a first content recommendation interface is displayed as the splash screen interface based on the first recommendation information, wherein the first content recommendation interface includes a plurality of first recommended contents displayed in an aggregated manner and a first interface element other than the first recommended contents; The first recommendation information is personalized content for the first account.

5. The method according to claim 4, characterized in that The plurality of first recommended contents are recommended contents aggregated by the server based on the first account through an artificial intelligence model; or The first interface element is an element generated by the server through an artificial intelligence model based on the first account; Alternatively, the multiple first recommended contents and the first interface element are generated by the server through an artificial intelligence model for personalization of the first account.

6. The method according to claim 4 or 5, characterized in that The first interface element includes a first theme element; The displaying of the first content recommendation interface as the splash screen interface based on the first recommendation information includes: The first theme element and the plurality of first recommended contents are displayed in association with each other in the first content recommendation interface, where the first theme element is a first recommendation theme corresponding to the plurality of first recommended contents.

7. The method according to claim 4 or 5, characterized in that The first content recommendation interface includes a plurality of aggregation recommendation areas corresponding to the aggregation dimensions respectively, and the first interface elements include content decoration elements, which are used to decorate the recommended content; The displaying of the first content recommendation interface as the splash screen interface based on the first recommendation information includes: The i-th aggregate recommendation area among the multiple aggregate recommendation areas displays multiple first recommended contents belonging to the i-th aggregate dimension, where i is a positive integer; For the first recommended content in the aggregated recommendation area that is ranked from high to low according to the aggregated score and reaches a preset ranking threshold, the content decoration element is displayed according to the aggregated score, wherein the aggregated score is used to characterize the matching degree between the multiple first recommended contents and the aggregated dimension corresponding to the aggregated recommendation area, and the visual prominence of the content decoration element is positively correlated with the recommendation score.

8. The method according to claim 4 or 5, characterized in that The first interface element includes a plurality of content blocking elements, the plurality of content blocking elements are used to block the plurality of first recommended contents, and the plurality of content blocking elements have corresponding blocking priorities; The displaying of the first content recommendation interface as the splash screen interface based on the first recommendation information includes: Associating and displaying the plurality of content-blocking elements with the plurality of first recommended contents in the first content recommendation interface, wherein when the content-blocking element is in a visible state, the first recommended contents corresponding to the content-blocking element are in an invisible state, and when the content-blocking element is in the invisible state, the first recommended contents corresponding to the content-blocking element are in the visible state; receiving a cancel display operation for the plurality of content blocking elements, wherein the cancel display operation is used to switch the content blocking elements from the visible state to the invisible state; The multiple content blocking elements are displayed in sequence from high to low according to the blocking priority. The blocking priority corresponding to the content blocking element is negatively correlated with the recommendation score of the first recommended content corresponding to the content blocking element. The recommendation score is determined in advance based on the correlation between the first recommended content and the recommended topic.

9. The method according to claim 4 or 5, characterized in that The plurality of first recommended contents correspond to a first recommended topic, the first interface element is adapted to the first recommended topic, the first recommended topic corresponds to an interest tendency score, and the interest tendency score is used to represent the first account's interest tendency in the first recommended topic; The displaying of the first content recommendation interface as the splash screen interface based on the first recommendation information includes: Displaying the plurality of first recommended contents corresponding to the first recommendation theme and the first interface elements adapted to the first recommendation theme in the first content recommendation interface; In response to receiving a theme switching operation for the first content recommendation interface, displaying a plurality of candidate recommended contents corresponding to the candidate recommended themes and candidate interface elements adapted to the candidate recommended themes in the first content recommendation interface; Among them, the interest tendency score corresponding to the first recommended topic is higher than the interest tendency score corresponding to the candidate recommended topic, and the interest tendency score corresponding to the candidate recommended topic reaches a preset interest tendency score threshold.

10. A method for generating recommended content, characterized in that: The method is executed by a server, and includes: In response to receiving a recommendation request from a first account, aggregating a plurality of first recommended contents based on the personalized first account, wherein the recommendation request is a request sent when a first client logged in with the first account starts and displays a splash screen interface; generating a first interface element based on the first account and at least one of the plurality of first recommended contents; Send first recommendation information to the first client corresponding to the first account, where the first recommendation information is used to provide relevant information of a first content recommendation interface displayed by the first client, where the first content recommendation interface includes the multiple first recommended contents displayed in an aggregated manner and the first interface elements other than the first recommended contents.

11. The method according to claim 10, characterized in that The personalized aggregation of the plurality of first recommended contents based on the first account includes: Acquire historical data of the first account, where the historical data is operation data of the first account in the first client within a historical time period; analyzing the first account's interest in the recommended content based on the first account's historical data using an artificial intelligence model to obtain a first content label; The plurality of first recommended contents are aggregated according to the first content tag.

12. The method according to claim 11, characterized in that Aggregating the plurality of first recommended contents according to the first content tags includes: Obtaining a candidate recommendation set, wherein the candidate recommendation set includes a plurality of candidate recommendation contents; Identify the recommendation features of the plurality of candidate recommendation contents through an artificial intelligence model to obtain a plurality of recommendation tags; Aggregate the plurality of first recommended contents from the candidate recommendation set based on a matching degree between the recommendation tag and the first content tag through an artificial intelligence model.

13. The method according to any one of claims 10 to 12, characterized in that: Generating a first interface element based on the first account and at least one of the plurality of first recommended contents includes: The first interface element is generated by the artificial intelligence model based on the historical data of the first account and at least one of the recommendation styles corresponding to the multiple first recommended contents.

14. The method according to claim 13, characterized in that Generating the first interface element by the artificial intelligence model based on historical data of the first account and at least one of the recommendation styles corresponding to the plurality of first recommended contents includes: analyzing, by the artificial intelligence model, the interest tendency of the first account in the recommended style based on the historical data of the first account, to obtain a first style label; Analyzing the style features corresponding to the plurality of first recommended contents by the artificial intelligence model to obtain a second style label; The first interface element is generated by the artificial intelligence model according to at least one of the first style label and the second style label.

15. The method according to any one of claims 10 to 12, characterized in that: The first interface element includes at least one of the layout of the first content recommendation interface, a first background image, and a first control element.

16. A device for displaying recommended content, characterized in that: The device comprises: A receiving module, configured to obtain first recommendation information from a server; a display module configured to display a first content recommendation interface as the splash screen interface based on the first recommendation information when the first client is started to display the splash screen interface, wherein the first content recommendation interface includes a plurality of first recommended contents displayed in an aggregated manner and first interface elements other than the first recommended contents; The multiple first recommendation information are personalized content for the first account.

17. A device for generating recommended content, characterized in that: The device comprises: a processing module configured to, in response to receiving a recommendation request from a first account, aggregate a plurality of first recommended contents based on the personalized first account, wherein the recommendation request is a request sent when a first client logged in with the first account is started and displays a splash screen interface; The processing module is further configured to generate a first interface element based on the first account and at least one of the plurality of first recommended contents; A sending module is used to send first recommendation information to the first client corresponding to the first account, where the first recommendation information is used to provide relevant information of a first content recommendation interface displayed by the first client, where the first content recommendation interface includes the multiple first recommended contents displayed in an aggregated manner and the first interface elements other than the first recommended contents.

18. A computer device, characterized in that: The computer device includes a processor and a memory, wherein the memory stores at least one program, and the at least one program is loaded and executed by the processor to implement the method for displaying recommended content as described in any one of claims 4 to 9, or the method for generating recommended content as described in any one of claims 10 to 15.

19. A computer-readable storage medium, characterized in that The storage medium stores at least one program, which is loaded and executed by the processor to implement the method for displaying recommended content according to any one of claims 4 to 9, or the method for generating recommended content according to any one of claims 10 to 15.

20. A computer program product, characterized in that The invention comprises a computer program, which, when executed by a processor, implements the method for displaying recommended content according to any one of claims 4 to 9, or the method for generating recommended content according to any one of claims 10 to 15.

Images