Pushing method, device, equipment and program product

Abstract

This disclosure provides a method, apparatus, computer device, and computer program product for pushing resources. The pushing method includes: acquiring a corpus set related to at least one target information; inputting the corpus set into a resource generation model to generate structured content to be pushed; constructing a resource index library based on the content to be pushed; determining the target content to be pushed associated with the triggering event from the resource index library in response to a triggering event; and outputting the target content to be pushed.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, specifically to a resource delivery method, a resource delivery device, a computer device, and a computer program product. Background Technology

[0002] In some related technical solutions, AI-enhanced search tools based on large language models have been introduced to more accurately meet user needs. These tools can perform online retrieval, semantic understanding, and content streaming generation in real time after a user initiates a query, providing a structured response. However, this "query-real-time generation" model has significant technical drawbacks: each user query requires triggering a complete, computationally intensive model inference and content generation process on the server side. This leads to frequent and high-load usage of server computing resources (such as GPUs / CPUs), high data processing pressure, and inevitably introduces significant response latency. This latency not only affects interaction efficiency but also leaves the user terminal in a resource-idling waiting state while awaiting a response, resulting in low overall system resource utilization. Summary of the Invention

[0003] One embodiment of this disclosure provides a method for pushing resources, a device for pushing resources, a computer device, and a computer program product. By constructing a resource index library of content to be pushed using a resource generation model that processes a corpus, the target content to be pushed can be quickly recalled and delivered in response to a triggering event, thereby reducing response and waiting time and improving the user experience.

[0004] On one hand, one embodiment of this disclosure provides a method for pushing resources. The method provides a graphical user interface through a terminal. The method includes: acquiring a corpus related to at least one target information; inputting the corpus into a resource generation model to generate structured content to be pushed; constructing a resource index library based on the content to be pushed; in response to a triggering event, determining the target content to be pushed associated with the triggering event from the resource index library; and outputting the target content to be pushed.

[0005] On the other hand, one embodiment of this disclosure provides a resource push device. The push device provides a graphical user interface via a terminal, and includes a display unit and a control unit. The control unit is connected to the display unit. The control unit is configured to: acquire a corpus related to at least one target information; input the corpus into a resource generation model to generate structured content to be pushed; construct a resource index library based on the content to be pushed; in response to a triggering event, determine the target content to be pushed associated with the triggering event from the resource index library; and output the target content to be pushed.

[0006] On the other hand, one embodiment of this disclosure provides a computer-readable storage medium storing a computer program adapted for loading by a processor to execute a resource push method. The resource push method provides a graphical user interface via a terminal. The push method includes: acquiring a corpus related to at least one target information; inputting the corpus into a resource generation model to generate structured content to be pushed; constructing a resource index library based on the content to be pushed; determining, in response to a triggering event, the target content to be pushed associated with the triggering event from the resource index library; and outputting the target content to be pushed.

[0007] On the other hand, one embodiment of this disclosure provides a computer device including a processor and a memory. The memory stores a computer program, and the processor executes a resource push method by calling the computer program stored in the memory. The resource push method provides a graphical user interface through a terminal. The push method includes: acquiring a corpus related to at least one target information; inputting the corpus into a resource generation model to generate structured content to be pushed; constructing a resource index library based on the content to be pushed; determining, in response to a triggering event, the target content to be pushed associated with the triggering event from the resource index library; and outputting the target content to be pushed.

[0008] On the other hand, one embodiment of this disclosure provides a computer program product, including computer instructions, which, when executed by a processor, implement a resource push method. The resource push method provides a graphical user interface through a terminal. The push method includes: acquiring a corpus set related to at least one target information; inputting the corpus set into a resource generation model to generate structured content to be pushed; constructing a resource index library based on the content to be pushed; in response to a triggering event, determining the target content to be pushed associated with the triggering event from the resource index library; and outputting the target content to be pushed.

[0009] In the resource push method, resource push device, computer-readable storage medium, computer device, and computer program product provided in this disclosure, a corpus is obtained by acquiring and integrating data related to target information. This corpus is then input into a resource generation model to generate content to be pushed, including questions and short texts, and a resource index library is constructed. In response to a trigger event, the target content to be pushed, which is associated with the trigger event and may correspond to the user's request, can be directly determined from the resource index library and output for the user to access. Therefore, by using pre-generated content to be pushed, the response and waiting time for users in obtaining content related to their requests is reduced, allowing users to quickly obtain relevant content and improving the smoothness and user experience of the resource interaction platform. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this disclosure, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a schematic diagram of a resource push system provided in one embodiment of the present disclosure.

[0012] Figure 2 This is a flowchart illustrating a resource push method provided in one embodiment of the present disclosure.

[0013] Figure 3 This is a schematic diagram of a first application scenario of the resource push method provided in one embodiment of the present disclosure.

[0014] Figure 4 This is a flowchart illustrating a resource push method provided in one embodiment of the present disclosure.

[0015] Figure 5 This is a flowchart illustrating a resource push method provided in one embodiment of the present disclosure.

[0016] Figure 6 This is a flowchart illustrating a resource push method provided in one embodiment of the present disclosure.

[0017] Figure 7 This is a flowchart illustrating a resource push method provided in one embodiment of the present disclosure.

[0018] Figure 8 This is a flowchart illustrating a resource push method provided in one embodiment of the present disclosure.

[0019] Figure 9This is a flowchart illustrating a resource push method provided in one embodiment of the present disclosure.

[0020] Figure 10 This is a flowchart illustrating a resource push method provided in one embodiment of the present disclosure.

[0021] Figure 11 This is a flowchart illustrating a resource push method provided in one embodiment of the present disclosure.

[0022] Figure 12 This is a schematic diagram of a second application scenario of the resource push method provided in one embodiment of this disclosure.

[0023] Figure 13 This is a flowchart illustrating a resource push method provided in one embodiment of the present disclosure.

[0024] Figure 14 This is a flowchart illustrating a resource push method provided in one embodiment of the present disclosure.

[0025] Figure 15 This is a schematic diagram of the structure of a resource delivery device provided in one embodiment of the present disclosure.

[0026] Figure 16 This is a schematic diagram of the structure of a computer device provided in one embodiment of the present disclosure. Detailed Implementation

[0027] The technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0028] It should be noted that the terms "first," "second," and "third," etc., used in various parts of the embodiments and drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. These terms can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in a sequence other than that illustrated or described herein. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0029] The AI-assisted interaction mode provided by resource interaction platforms typically relies on users actively initiating lengthy text queries and waiting for the system to complete the entire process of online retrieval, content reasoning, and streaming generation. This results in significant response delays, disrupting the smooth experience for users to quickly obtain the information they need. To address this issue, embodiments of this disclosure provide a resource push method, a resource push device, a computer-readable storage medium, a computer device, and a computer program product.

[0030] Specifically, the resource push method of this disclosure embodiment can be executed by a computer device, which can be a terminal or a server. The terminal can be a smartphone, tablet, laptop, smart TV, wearable smart device, smart vehicle terminal, etc., and may also include a client, such as a game client, browser client, instant messaging client, or mini-program. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0031] For example, when the resource push method runs on a terminal device, the terminal device may include a display screen and a processor. The display screen is used to present a graphical user interface (GUI) and receive user commands to the GUI. The GUI may include a trigger event input interface, a resource display interface, etc. The processor is used to store the resource push program, run the resource push method, generate the GUI, respond to trigger events, and control the display of the GUI on the display screen. When the user operates the GUI through the display screen, the GUI can control the local content of the terminal device in response to the received operation commands. The terminal device can provide the GUI to the user in various ways, such as rendering it on the terminal device's display screen or presenting the GUI through holographic projection.

[0032] For example, when the resource push method runs on a server, it can be implemented and executed based on a cloud-based generation system. A cloud-based generation system refers to a generation method based on cloud computing. It includes servers and client devices. The main body running the resource push program and the main body presenting the graphical user interface are separate. The storage and execution of the resource push method are completed on the server. The presentation of the graphical user interface is completed on the client. The client is mainly used for receiving and sending data and for the presentation of the graphical user interface. For example, the client can be a display device with data transmission capabilities located close to the user, such as a mobile terminal, television, computer, PDA, personal digital assistant, head-mounted display device, etc. However, the terminal device for data processing is the server in the cloud. When pushing resources, the user operates the client to send instructions to the server. The server controls the execution of the resource push method according to the instructions, encodes and compresses the data such as the graphical user interface, returns it to the client via the network, and finally, the client decodes and outputs the graphical user interface.

[0033] It should be noted that in this embodiment of the disclosure, the executing entity of the resource push method can be a terminal device or a server. The terminal device can be a local terminal device or a client device in the aforementioned cloud generation system. This embodiment of the disclosure does not limit the type of executing entity.

[0034] For example, in conjunction with the above description, Figure 1 This disclosure illustrates a resource push system 1000 for implementing a resource push method, according to one embodiment of the present disclosure. The resource push system 1000 may include at least one terminal 1001, at least one server 1002, at least one database 1003, and a network. The user-held terminal 1001 can connect to different servers via the network. The terminal is any device with computing hardware capable of supporting and executing software application tools corresponding to the resource push method.

[0035] In the aforementioned resource push system 1000, terminal 1001 is used to install and run the resource push program. In some cases, terminal 1001 may not need to have the resource push program or its corresponding client pre-installed; users can directly access and obtain the pushed resources through a browser or other client. During the process of users obtaining the pushed resources through the resource push program, terminal 1001 and server 1002 interact. Terminal 1001 sends various information to server 1002. Server 1002 determines the display data for terminal 1001 based on the storage mechanism and the received information, and sends the display data back to terminal 1001 so that terminal 1001 can display the data sent by server 1002 to the user. The network can be a wireless network or a wired network, such as a wireless local area network (WLAN), local area network (LAN), cellular network, 2G network, 3G network, 4G network, 5G network, etc. Additionally, the terminal can also use its own Bluetooth network or hotspot network to connect to other terminals or to the server. Furthermore, the resource push system 1000 can include multiple databases, which are coupled to different servers.

[0036] It should be noted that, Figure 1 The schematic diagram of the resource push system shown is merely an example. The resource push system 1000 described in this disclosure is intended to more clearly illustrate the technical solutions of this disclosure and does not constitute a limitation on the technical solutions provided in this disclosure. As those skilled in the art will know, with the evolution of resource push systems and the emergence of new business scenarios, the technical solutions provided in this disclosure are also applicable to similar technical problems.

[0037] It should be noted that the triggering operations mentioned in the subsequent detailed description of the resource push method provided in the embodiments of this disclosure can all be regarded as triggering operations performed by the user through a finger or by controlling a medium such as a mouse, keyboard, or stylus. The specific medium used can be determined according to the type of computer device. For example, when the computer device is a touchscreen device such as a mobile phone, tablet computer, or game console, the user can operate on the touchscreen using any suitable object or accessory such as a finger or stylus. When the terminal device is a non-touchscreen terminal device such as a desktop computer or laptop computer, the user can operate using an external device such as a mouse or keyboard.

[0038] The technical solutions of this disclosure will be described in detail below through specific embodiments. It should be noted that the following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0039] Please see Figure 2 and Figure 3 , Figure 2This is a flowchart illustrating a resource push method provided in one embodiment of the present disclosure. Figure 3 This diagram illustrates a resource push method according to one embodiment of the present disclosure. It should be noted that the steps shown may be executed in a logical order different from that shown in the flowchart of the resource push method. The resource push method may include the following steps: Step 01: Obtain a corpus set related to at least one target piece of information; Step 03: Input the corpus into the resource generation model to generate structured content to be pushed; Step 05: Build a resource index library based on the content to be pushed; and Step 07: In response to the triggering event, determine the target content to be pushed that is associated with the triggering event from the resource index; and Step 091: Output the target content to be pushed.

[0040] Specifically, resources refer to digital content or services that can be accessed, utilized, or interacted with by users. Users can interact with resource interaction platforms, such as software clients, browser clients, instant messaging clients, or mini-programs, whether online or offline. This includes entering keywords to search for resources or browsing or filtering the content provided by the platform to obtain resources. Obtaining resources here can include, but is not limited to, downloading resources, browsing resources, reposting resources, or extracting at least a portion of the content of resources.

[0041] In the current digital service ecosystem, resource recommendation plays a crucial role in enhancing user experience. Its core objective is to accurately match users' potential needs with a vast amount of resources, thereby improving both user experience and platform efficiency. Traditional recommendation methods often rely on simple keyword matching, making it difficult to capture the deeper, unspoken intentions of users. Furthermore, the types of resources recommended are limited, failing to meet users' underlying needs. In recent years, with the rapid development of natural language processing and large-scale language model technologies, some advanced platforms have begun to introduce AI-powered resource recommendation tools. These tools can perform deep semantic understanding and intent parsing of user-input query text, proactively initiating multiple rounds of queries or generating structured answers to directly address users' complex resource demands.

[0042] For ease of understanding and explanation, this disclosure uses a music platform as an example to illustrate the resource interaction platform, but it is not limited to resources only related to music. For example, if the resource interaction platform is a shopping platform, the music and related content can be replaced with goods and related content for interpretation; if the resource interaction platform is a game platform, the music and related content can be replaced with games and related content for interpretation; if the resource interaction platform is a reading platform, the music and related content can be replaced with books and related content for interpretation.

[0043] For example, on a music platform, a user might simply enter "songs suitable for a rainy night." Traditional recommendation methods might only return tracks whose titles or lyrics contain the words "rain" or "night," while AI resource recommendation tools can understand the underlying emotions, scenes, and even cultural contexts, thereby generating a list of song resources or descriptive text resources, including genres such as nostalgic jazz and soothing instrumental music.

[0044] However, AI resource recommendation tools have inherent limitations such as high interactivity and latency. Interacting with these tools to obtain resources typically requires users to actively initiate and construct relatively complete and detailed textual questions, such as explicitly requesting "Chinese songs from the 1990s with melancholic piano melodies." The AI ​​resource recommendation tool then must perform a series of computationally intensive steps online, including retrieval, contextual reasoning, and content streaming generation, resulting in a significant delay of several seconds or even longer. This waiting interrupts the smooth experience users desire for quick resource access, which is particularly prominent in efficiency-driven mobile interaction scenarios. To address this issue, this disclosure proposes a resource recommendation method that balances meeting user needs with rapid response.

[0045] To better provide users with rich, relevant, and relatable resources, the first step is to acquire a corpus. A corpus is a collection of data comprising various types of data, such as text, audio, and video. Target information serves as the basis for acquiring and collecting data from the corpus. For example, target information could include song titles, artist names, album names, excerpts from lyrics, or phrases related to the song and / or lyrics and / or artist.

[0046] At this point, data related to the target information is obtained from the client (i.e., software, or the corresponding webpage or mini-program client in a browser client, hereinafter referred to as the internal platform) or other clients (such as other music software, social clients, news clients, etc., hereinafter referred to as the external platform) used by the push method of the resources disclosed herein through information retrieval, data collection, etc. For example, the data that the internal platform can use to construct the corpus includes, but is not limited to, song text attribute information, song reviews, album introductions, playlist titles, song descriptions, related podcast explanations, and related note texts related to the target information. The data that the external platform can use to construct the corpus includes post content, song reviews, audiovisual resources, or book materials related to the target information. These data from the internal and external platforms may contain content that users are interested in. Furthermore, through preprocessing steps such as data integration, filtering, or selection, a corpus with high relevance to the target information and high-quality content is obtained, which is then used in the subsequent steps of generating the content to be pushed.

[0047] A resource generation model is a model used to generate content to be pushed based on a corpus. The resource generation model is a pre-defined model, which can be an immutable, fixed model stored in memory during the design and development of the resource push method, or a model that can be written to, adjusted, or replaced by operators in memory. The resource generation model performs semantic understanding and data integration on the input corpus to generate one or more pieces of content to be pushed based on the content of the corpus. For example, the resource generation model is an AI model, such as GPT-3.5, GPT-4, GPT-4o, Gemini, PaLM 2, Claude, Llama 2, and Llama 3.

[0048] It's understandable that the content to be pushed is generated by the resource generation model to cater to user needs and potentially pique their interest. Due to the limited display space of the graphical user interface, in order to better attract users' attention and quickly convey the core content of the resource, the content to be pushed is structured, meaning it has a pre-set format.

[0049] In some implementations, the content to be pushed includes a title and a short article corresponding to the title. The title is a concise and descriptive text within the content that summarizes and encapsulates the resource's content. The main characteristics of a title are brevity, a clear theme, and strong appeal. The short article is content that corresponds to the title and further elaborates on it to resonate with the user. The main characteristics of the short article may include, but are not limited to, rich content, eye-catching appeal, and high readability.

[0050] In other implementations, the content to be pushed includes a title, a short article corresponding to the title, and related images. The images are the content that corresponds to the title and is presented visually in relation to the title. The main characteristics of the images may include, but are not limited to, simplicity, intuitiveness, and good aesthetics.

[0051] For example, taking "Song A (a song by band A)" ​​as the target information, the corpus consists of various types of data related to band A, such as the lyrics of songs by band A, multiple song reviews by band A, albums mentioned in song A, podcasts, notes, playlists, etc. related to song A. In this case, the title of a generated piece of content to be pushed is "Band A: The Musical Story of Song A," and the short text is: "Band A's album 'Album A' records their real-life experiences on the road, with footprints covering Inner Mongolia, Xinjiang, Yunnan, Tibet, Nepal, and even traversing the dangerous Xinjiang-Tibet Highway. They also visited the beautiful West Lake in Suzhou and Hangzhou. Songs in the album, such as 'Song A,' 'Song B,' and 'Song C,' showcase various emotions such as dreams, friendship, love, and family affection." A resource index is a data structure or database specifically designed for efficient retrieval of content to be pushed. Instead of directly storing the complete original content to be pushed, a resource index extracts key features of the content (such as keywords, tags, metadata, and vector representations) and establishes a mapping relationship between these features and the data location or identifier of the original content to be pushed, thus forming a "quick query directory."

[0052] For example, the steps for constructing a searchable resource index library using the content to be pushed as the data source can be as follows: First, perform text analysis on the original content to be pushed, such as word segmentation and normalization; then extract terms and establish a mapping relationship from terms to the content to be pushed, usually using an inverted index structure; then calculate the term weights and may perform index compression and optimization. The resource index library is used to achieve accurate association and fast query, providing support for the essential information retrieval steps in the push method.

[0053] Trigger events can be triggered by users inputting keywords via typing or voice input, or by the terminal periodically issuing trigger signals through a preset program. The resource push system 1000 can respond to trigger events, matching the information reflected in the trigger event with content in a pre-built resource index library, such as keyword matching and semantic matching, and identifying one or more pieces of content associated with the trigger event as target content that reflects the user's needs. In other words, the target content to be pushed is the filtered content associated with the trigger event.

[0054] For example, the triggering event is when the user enters "song A" in the client, which will determine the content to be pushed in the example above as the target content to be pushed.

[0055] For example, if the client determines that the user is interested in "Song A" based on the user's historical usage data, the trigger event can be an automatic trigger signal initiated by the client's backend system. The trigger signal conveys the keyword "Song A". In this case, the content to be pushed in the example above can also be determined as the target content to be pushed.

[0056] Once it is determined that the target content to be pushed may arouse user interest and meet user needs, the target content to be pushed will be output, for example, by displaying it in a pop-up window, floating window, prompt box, or directly in the current or next screen of the graphical user interface, so that users can quickly know about it.

[0057] In some implementations, only the title of the target content to be pushed is displayed directly on the graphical user interface (GUI). The short article and other related content are then displayed after the user clicks on the title. In this case, the target content occupies a small portion of the GUI, does not interfere with or affect the display of other existing resources, and is less likely to cause user annoyance or resistance. In other implementations, the title or at least part of the short article of the target content to be pushed is displayed on the GUI. The complete short article and other related content are then displayed after the user clicks on the title. In this case, the target content displays more content on the GUI, increasing the probability of attracting the user's attention.

[0058] In some implementations, the target content to be pushed may also be displayed on the graphical user interface along with other related resources, such as related songs, thus providing users with a wider variety of resources.

[0059] For example, please refer to Figure 3 The graphical user interface (GUI) features an input box at the top where users can enter their target information, such as "Song A". The GUI then displays related content to be pushed to "Song A" along with other relevant resources, providing users with a diverse and comprehensive view of the information. As shown in the image, the target content is interspersed between song links and song comments. Furthermore, when a user shows interest in the target content and clicks on it, the GUI switches from displaying the left image to the right image, showing only the title or short text of the target content for direct and quick access to information of interest.

[0060] It is understandable that in the aforementioned push methods, the content in the corpus is thematically consistent and rich in content. Based on this, the content to be pushed generated by the powerful resource generation model is rich in content, in-depth, and easily appealing to users, thus better catering to their interests and needs. Since the resource index is pre-generated, the content to be pushed in the resource index can be quickly identified and output to the graphical user interface, greatly reducing the user's waiting time.

[0061] In the resource push method provided in this disclosure, a corpus is obtained by acquiring and integrating data related to the target information. This corpus is then input into a resource generation model to generate pushable content containing questions and short texts, and a resource index library is constructed. In response to a trigger event, the target pushable content, which is associated with the trigger event and may correspond to the user's request, can be directly determined from the resource index library and output for the user to access. Therefore, by using pre-generated pushable content, the response and waiting time for users in obtaining content related to their requests is reduced, enabling users to quickly access relevant content and improving the smoothness and user experience of the resource interaction platform.

[0062] Please see Figure 2 and Figure 4 In some implementations, step 01 can be achieved through step 011, specifically as follows: Step 011: Obtain corpus related to the target information from multiple data sources.

[0063] Specifically, to ensure a richer data source for the corpus, the corpus acquisition process disclosed herein utilizes multiple data sources. It can be understood that the corpus refers to the data within the corpus set. It should be noted that "multiple" here means "more than," and the data source is not singular; the number of data sources can be, but is not limited to, two, three, four, or five. For example, the data source can be the client used by the resource push method, i.e., the "internal platform" mentioned above, or other clients or data sources besides the client used by the resource push method, i.e., the "external platform" mentioned above.

[0064] At this point, the sources of the corpus are more diverse, and the forms and content of the corpus are correspondingly more varied. To ensure relevance to the target information, only information related to the target information obtained from multiple data sources is used as the corpus, ensuring that the generated corpus can match the user's actual needs. It can be understood that the more comprehensive and richer data set formed by integrating corpora from multiple data sources is the corpus.

[0065] Therefore, in the resource push method disclosed herein, the corpus is obtained from multiple data sources, and the corpus is highly relevant to the target information. In this case, the sources of the corpus are richer and the content is more comprehensive. The generation of the content to be pushed has rich information sources and can better match the actual needs of users.

[0066] Please see Figure 4 and Figure 5 In some implementations, step 011 can be achieved through steps 0111, 0113, and 0115, specifically as follows: Step 0111: Obtain the first type of corpus from the first data source to form a subset of the first corpus; Step 0113: Obtain a second type of corpus from the second data source to form a subset of the second corpus. The second data source is different from the first data source; and Step 0115: Integrate the first and second subsets of the corpus to obtain the corpus set.

[0067] Specifically, the first data source is the client used by the resource push method, i.e., the "internal platform" mentioned above. It can be understood that the target information corresponds to at least one song and / or more artists. In the first data source, the content associated with the target information is the content related to the corresponding song and / or artist, including but not limited to lyrics, song reviews, the album, associated podcasts, associated notes, and associated playlists. This part of the content constitutes the first type of corpus. The set of data from the first data source obtained after extracting, refining, filtering, and organizing the first type of corpus is the first subset of the corpus.

[0068] For example, the target information is "Song A". Song comments obtainable from the first data source include: "Fans of band A also like listening to band B, right?" and "This is band A's best song." Related notes obtainable from the first data source include: "Band A's 'Song A' expresses nostalgia for past experiences, reflections on the passage of time, and emotional reflections on life during a journey by depicting the natural and cultural landscape of Song A..." The text record formed after sorting and summarizing the above content constitutes the first corpus subset.

[0069] The second data source refers to clients or data sources other than those used in the resource push method, i.e., the "external platforms" mentioned above. It can be understood that in the second data source, the content associated with the target information is content related to the corresponding song and / or singer, including but not limited to comments on social media platforms, user posts, and song comments in other clients. This part of the content constitutes the second type of corpus. The set of data from the second data source obtained after extracting, refining, filtering, and organizing the second type of corpus is the subset of the second corpus.

[0070] For example, taking a social media platform as the second data source, the target information is "#SingerA SongD". Posts from netizens that can be obtained from the second data source include: "#SingerASongD# Singer A's latest cover single 'Song D' is released on all platforms! Looking forward to its midnight release," "#SingerASongD# He's worked hard all along, he deserves a bigger stage," and "Feels like he's going to release an album this year! He's almost done his concert tour, which usually requires 20-35 songs! Maybe he should release an album this year." A concert, perhaps? I'm amazed by how distinctive Singer A's cover is; I was pleasantly surprised by the arrangement style. Singer A's confidence always comes from their work! #SingerASongD...”, “#SingerASongD# Looking forward to it!” The text record formed after sorting and summarizing the above content constitutes the second subset of the corpus.

[0071] The corpus is the more comprehensive and richer data set formed by integrating the first and second subsets of the corpus. In some implementations, the corpus is the sum of the first and second subsets. In other implementations, the corpus is the content obtained after further processing the overall content formed by concatenating the first and second subsets, such as removing duplicate and contradictory content. In this case, the content of the corpus is less than or equal to the sum of the content of the first and second subsets.

[0072] Therefore, in the resource push method disclosed herein, the acquisition of the corpus integrates the content related to the target information from the first data source and the content related to the target information from the second data source. The data content is more comprehensive, providing more detailed and high-quality raw data for the generation of the content to be pushed. This helps the resource generation model to better analyze the content that users are currently interested in, so as to ensure that the generated content to be pushed can better meet the user's needs.

[0073] Please see Figure 2 and Figure 6 In some implementations, step 03 can be achieved through steps 031 and 033, specifically as follows: Step 031: Format the corpus to generate prompts that meet the input requirements of the resource generation model; and Step 033: Input the prompt information into the resource generation model to generate structured content to be pushed.

[0074] Specifically, generating content to be pushed is the core step in the resource push method. During this process, the corpus needs to be interpreted to extract and refine attractive content. Hints are words, phrases, or sentences that define the content and format of the generated content to be pushed, reflecting the specific content that aligns with the generation goals and improving the accuracy and efficiency of the corpus interpretation.

[0075] Formatting refers to generating corresponding output content from input content according to a preset format. When a corpus is obtained, the formatting process is performed on the corpus to ensure that the generated prompts have a specific format corresponding to the formatting process. It should be noted that the format here conforms to the input requirements of the resource generation model, meaning it ensures that the prompts generated from the formatted corpus can be directly used as input; the input refers to the resource generation model.

[0076] In some implementations, the formatting process includes key field extraction and format adjustment. When inputting a corpus, key fields with high relevance to the target information are first extracted from the corpus. Then, the extracted key fields are filled into the format template according to a preset format template to generate prompt information.

[0077] In other implementations, the formatting process includes embedding a corpus into a prompt word template to generate resource generation prompt words. The prompt word template includes at least one of the following: task description, formatting requirements, precautions, and reference examples. Here, the resource generation prompt words are the prompt information.

[0078] It is understandable that when the format of the prompt message meets the input requirements of the resource generation model, the resource generation model processes the prompt message efficiently, and the quality and format of the generated content to be pushed are superior. Furthermore, the resource generation model performs semantic parsing and content integration on the prompt message to generate structured content to be pushed that meets the requirements.

[0079] Therefore, in the resource push method disclosed herein, when the corpus is formatted, the generated prompt information can better meet the input requirements of the resource generation model. At this time, inputting the prompt information into the resource generation model can generate structured content to be pushed. Therefore, the generation process of the content to be pushed is highly standardized, and the quality of the content to be pushed can be well guaranteed.

[0080] Please see Figure 6 and Figure 7 In some implementations, step 031 can be achieved through step 0311, specifically as follows: Step 0311: Embed the corpus into the prompt word template to generate resource prompt words.

[0081] Specifically, resource generation prompts are the prompt information. Prompt templates are formatted templates used to improve the standardization and interpretability of resource generation prompts. By embedding a corpus into a pre-defined prompt template, the corpus is transformed into a more standardized form conducive to intent recognition and semantic interpretation. To standardize the content and form of resource generation prompts, the prompt template must include at least one of the following: a description of the purpose / task, formatting requirements, considerations for improving standardization, and a reference case for example.

[0082] The task description is a prompt word that instructs the resource generation steps, guiding the resource generation model through the specific steps to generate the content to be pushed. Format requirements specify the format of the generated content to be pushed. Notes highlight important points that may affect the output during the content generation process. Reference cases provide standardized transformation examples to guide the content generation process, ensuring that the generated resource generation prompt words can produce qualified content to be pushed. It should be noted that the prompt word template can also take other forms; this disclosure does not restrict this, as long as the corpus can be embedded into the prompt word template and resource generation prompt words can be generated.

[0083] For example, the prompt template includes a task description, format requirements, and notes. The task description is: "I want you to write a short article for the music and entertainment section. You need to write an article title of no more than 15 characters and a short article of no more than 100 characters. I will provide you with basic background information on the song, related trending information, podcast audio commentary on the song, and notes written by users when sharing the song (i.e., a corpus)." The format requirements are: "1. The content must include a title and a corresponding short article. 2. Please use a JSON string format for the article title and short article, such as {"title":"*","short article":"*"}. 3. For the short text content, use Markdown format for easier reading. Highlight key information related to the topic and book titles in bold and red. If line-by-line output is required, display it line-by-line, adjusting spacing for clarity. Notes: 1. The focus should be on creating an attractive title that will entice users to click. You can mimic the title style of podcast audio recordings. 2. The short text needs to extract the best content from the material, such as the viewpoints from the "audio" and "notes." However, do not explicitly state that it originates from the "audio" or "notes." 3. Some content may be unrelated to the songs. You need to filter out this content during the creation process. For example, the prompt template includes task description, format requirements, precautions, and reference examples. The task description is: "You are a blogger. Based on the given trending terms, descriptions, official media articles, detailed explanations, and user comments (i.e., a corpus), generate one candidate post. The text content of the candidate post must include: 1. Factual information section. Summarize the specific content of the trending topic and generate an objective factual description of the trending content. However, if the facts come from official media and ordinary users, please prioritize using the original official media articles and do not summarize or rewrite them to avoid risks. 2." Emotional Impression Section. After the factual information, summarize the facts and opinions in the blogger's voice, expressing emotions about the hot topic to attract user attention. Avoid using terms like "sisters" or similar terms of endearment. The format requirements are: "The content must include a title and a short text corresponding to the title. Return one post, and the result must be in strict JSON format." Important notes: "1. Posts should be under 400 characters. Use appropriate line breaks, and add emojis to enrich the text. However, reduce emojis for controversial topics. 2. To mitigate risk, only use detailed statistics from official media articles, not detailed statistics from ordinary user comments, such as "ranking on the list." If the article does not contain statistical data, then the summary should not include statistical data either. 3. Factual paragraphs and emotional paragraphs should be completely separated, and the factual paragraphs should occupy the majority of the text. Factual paragraphs should not contain sentences expressing emotions." A sample example is: "For example, taking the hot topic 'Singer B's new album 'Album B' released unexpectedly, the following is a sample JSON format:" { "Candidate Posts": [ [Title: Singer B's New Album "Album B" Released] "[Factual Information Section: According to widespread online discussion and studio news, singer B's new album "Album B" will be released globally today at 12 PM! The album includes 20 brand new singles, including the pre-release track "Song E" and several unreleased tracks. This is after 2019's "Song..." Qu F Four years after [previous album], he released a full-length album, which topped charts in multiple regions worldwide during its pre-sale period. "[Emotional Highlights: Four years of waiting have yielded 20 amazing tracks! From early fame to struggling through lows, from healing through marriage to musical rebirth, singer B powerfully responds to all expectations with this album. Those fans who flooded the comments section, those who waited through time zones for the album's release, will finally get their reward today—get your headphones ready and let singer B's voice once again dominate your playlist!]" ] }” Furthermore, the generated resource generation prompts are input into the resource generation model. The model then uses semantic parsing and content integration to generate content that meets the requirements for push notifications. Taking the prompt template (including task description, format requirements, precautions, and reference examples) and a corpus associated with the target information "#SingerASongD" as an example, the generated content to be pushed is as follows: { "Candidate Posts": [ [Title: Singer A Covers Old Love Song D, Musical Passion Blossoms] "[Factual Information Section: Singer A recently covered the song 'Song D' and publicly stated that it was a song he had always liked. Now, he has the opportunity to perform this song in public and share it with a wider audience. Fans have expressed their support and anticipation for his musical performance.]" "[Emotional Portrayal: From youthful days to the spotlight, singer A uses her bold voice to relive old dreams, every melody carrying perseverance and passion. Listen to this pure musical spirit, and let 'Song D' awaken the warm memories in your heart!]" ] }” Therefore, the resource push method disclosed herein inputs resource generation prompts containing a corpus and having a standardized format into the resource generation model to generate content to be pushed that has a high degree of standardized format and strong content appeal. The content to be pushed can better meet the user's needs and provide high-quality alternative content for the target content to be pushed in the future.

[0084] Please see Figure 2 and Figure 8 In some implementations, step 07 can be achieved through steps 071, 073, and 075, specifically as follows: Step 071: Determine the set of candidate content to be pushed that is associated with the triggering event from the resource index library; Step 073: Based on the sorting strategy, sort the candidate content to be pushed in the set of content to be pushed; and Step 075: Select at least one piece of content to be pushed according to the sorting result, as the target content to be pushed.

[0085] Specifically, since the resource index stores a large amount of content to be pushed, to ensure that the content pushed to users matches their needs well, it is necessary to filter the content determined from the resource index in conjunction with trigger events to determine the association between the selected content and the trigger events. In the resource index, each piece of content to be pushed contains multiple words, phrases, or statements, and the trigger events also contain words and / or phrases and / or statements.

[0086] By analyzing the trigger event and each piece of content to be pushed, quantifiable correlation data can be obtained. The correlation degree between the trigger event and each piece of content to be pushed is obtained separately, and the obtained correlation degree is compared with a preset threshold to filter out the content to be pushed that is highly correlated with the trigger event. In some implementations, the correlation degree is calculated by obtaining the matching degree between the trigger event and each piece of content to be pushed at the word / phrase level, such as determining whether each piece of content to be pushed contains the same words and / or phrases and / or sentences as the trigger event. In other implementations, the correlation degree is calculated by obtaining the matching degree between the trigger event and each piece of content to be pushed at the semantic level, such as performing semantic parsing on each piece of content to be pushed and calculating the correlation between the semantic parsing result and the trigger event.

[0087] A preset threshold represents the relevance between a triggered event and the content to be pushed. The preset threshold can be a fixed value or a variable value adjusted by developers based on factors such as user habits. When the relevance between a triggered event and a piece of content to be pushed exceeds the preset threshold, it can be determined that the content to be pushed contains words, phrases, or sentences that are the same as or similar to the triggered event. This allows for the precise location of content related to the triggered event from the resource retrieval database, and the set of this portion of content to be pushed is defined as the candidate set of content to be pushed.

[0088] For example, if the triggering event is the user's input of "Song A", the system calculates the relevance of this triggering event to each content to be pushed in the resource index. The preset threshold can be set according to different strategies, such as: (1) Fixed threshold strategy: the relevance must be greater than 0.7; (2) Dynamic threshold strategy: the threshold is set to the upper quartile of the relevance distribution based on the average matching success rate of recent queries; (3) Category threshold strategy: the threshold is set to 0.6 for the "music" category and 0.8 for the "movie" category. When the relevance of the content to be pushed exceeds the set preset threshold, it will be selected into the candidate content to be pushed set.

[0089] For example, if the triggering event is the user's input of "Song A", then the relevance of content A to be pushed, which contains "Song A" and "Band A" in both its title and text, is 0.8. Content B to be pushed, which only contains "Song A" in its text but actually discusses "Song B", has a relevance of 0.3. If the preset threshold is a fixed 0.7 (or 0.6 for "music" topics), then content A will be selected into the candidate set of content to be pushed. In this case, the content in the candidate set has a high relevance to the triggering event, reducing the likelihood of mistakenly selecting content that is irrelevant or has low relevance to the user's needs, thus improving the targeting and accuracy of the push method.

[0090] When a candidate set of content to be pushed is obtained, since the set may contain multiple content items, but the number of content items that can be displayed in the graphical user interface is limited, it is necessary to carefully select the content items from the candidate set. The sorting strategy is a pre-defined algorithm that ranks the "value" of each content item in the candidate set. This value may include factors such as "relevance to the triggering event" and "attractiveness to the user." The sorting strategy can be a fixed, unmodifiable strategy stored in memory after the push method of this disclosure is developed, or it can be a strategy that can be written, adjusted, or replaced by the operator in memory.

[0091] It should be noted that since "attractiveness" is a parameter with a strong subjective factor, when the ranking strategy involves "attractiveness", it is only a ranking of "attractiveness" in the sense of data processing and statistics. In actual application, it may not completely match the actual situation of users. The ranking strategy focuses on analyzing probability and allows for situations where it does not perfectly match the user's preferences.

[0092] After sorting the candidate content in the push notification set, the content is then sorted in ascending or descending order of "value." Based on the sorting result, at least one content that meets the push criteria is selected and designated as the target content. The push criteria prioritize content with higher "value" to better meet user needs. For example, if the content is sorted in ascending order of "value," the content ranked first is selected as the target content.

[0093] In some implementations, content with a comprehensive score greater than a filtering threshold is selected from the ranking results and identified as target content to be pushed. The filtering threshold is a threshold used to evaluate whether the content to be pushed has a significant "comprehensive advantage." The comprehensive score of each piece of content to be pushed is compared with the filtering threshold, and content with a comprehensive score greater than the filtering threshold is selected. It can be understood that this part of the content to be pushed is a resource with a significant "comprehensive advantage." Therefore, this part of the content to be pushed is identified as target content to be pushed.

[0094] It should be noted that the filtering threshold can be set according to different strategies, such as: (1) Fixed threshold strategy: the relevance must be greater than 0.8; (2) Dynamic threshold strategy: the filtering threshold is the product of the average matching success rate of the candidate content set and a certain coefficient; (3) Category threshold strategy: for the "music" category, the filtering threshold is set to 0.6, and for the "movie" category, the filtering threshold is set to 0.8.

[0095] It's understandable that the target content to be pushed can be zero, one, or more. If there are more than one target content, all of it can be displayed in the graphical user interface, or other algorithms can be used, such as re-sorting and filtering based on relevance or popularity metrics, or a certain number of target content can be randomly selected and displayed in the graphical user interface. In this case, the method of determining the target content is more flexible, and the selected content may be quite broad, potentially showing users some unexpected and novel content.

[0096] In other implementations, the content ranked first in the sorting results is identified as the target content to be pushed. This means directly filtering the content with the highest overall score and recognizing it as the resource with the greatest "overall advantage." This content is then designated as the target content and displayed in the graphical user interface. In this case, the method of identifying the target content is more accurate, and the selected content is more likely to better match the user's needs.

[0097] Therefore, the push method of the resources disclosed herein determines the set of candidate push content based on the correlation between the triggering event and the content to be pushed, and then selects the target push content based on the sorting strategy and push conditions, so as to ensure that the selected target push content can better meet the user's needs and fully improve the user experience.

[0098] Please see Figure 8 and Figure 9 In some implementations, step 073 can be achieved through steps 0731 and 0733, specifically as follows: Step 0731: Obtain evaluation metrics for each piece of content to be pushed from multiple dimensions; Step 0733: Determine the sorting strategy for each piece of content to be pushed based on evaluation indicators from multiple dimensions.

[0099] Specifically, to ensure that the ranking of content to be pushed accurately reflects how well the content meets user needs, a comprehensive evaluation of each piece of content across multiple dimensions is required during the ranking process. Evaluation metrics are quantitative indicators reflecting the degree to which each piece of content meets user needs, used to provide a basis and guidance for the ranking process.

[0100] In some implementations, the evaluation dimensions for push content may include the degree of relevance between the content and the triggering event. This ensures that the content to be pushed is highly relevant to the triggering event, avoiding the mistaken push of irrelevant content to the user. For example, if the triggering event is the user's input of "Song A," and the content to be pushed is unrelated to "Song A" but related to "Song D," then this content should not be selected as the target push content. Instead, it should be evaluated as content with low relevance and receive a lower (or higher) evaluation metric.

[0101] In other implementations, the evaluation dimensions for the content to be pushed can include the quality of the content, thereby ensuring that the content to be pushed is of high quality and avoiding the push of low-quality content to users. For example, if the content to be pushed has many typos or is grammatically incorrect, then the content to be pushed should not be selected as the target content to be pushed, but should be evaluated as low-quality content and receive a lower (or higher) evaluation score.

[0102] At this point, multiple evaluation metrics are integrated, such as through weighted averaging or table lookup, to obtain comprehensive evaluation indicators. Based on these multi-dimensional evaluation metrics, the suitability of each piece of content to be pushed for user needs is determined, thus providing a reliable and accurate basis for the ranking process of each piece of content to be pushed. Therefore, the resource push method disclosed in this paper comprehensively considers multiple dimensions of evaluation metrics, enabling a comprehensive evaluation and consideration of each piece of content to be pushed in the candidate content set, in order to achieve a more accurate and reliable ranking and ensure that the target content to be pushed meets user needs.

[0103] Please see Figure 9 and Figure 10 In some implementations, the evaluation metrics across multiple dimensions include relevance and popularity. Step 0731 can be achieved through steps 07311 and 07313, specifically as follows: Step 07311: Obtain the correlation metric between each piece of content to be pushed and the triggering event; and Step 07313: Obtain the popularity metrics for each piece of content to be pushed.

[0104] Specifically, this embodiment is used to illustrate the specific algorithm of the sorting strategy.

[0105] To ensure that the content to be pushed is related to the triggering event, a relevance metric is calculated for each triggering event and each piece of content to be pushed. A relevance metric is a specific score used to evaluate the degree of relevance between two compared texts. The higher the relevance metric, the more relevant the content to be pushed is to the query conditions, and the more likely it is to match the user's needs. The relevance metric can be used to quantify the degree of matching between the triggering event and each piece of content to be pushed in terms of text content. This score can be calculated by comprehensively evaluating multiple dimensions such as the similarity of character sequences between the two texts, the proportion of shared words, and the continuity and compactness of matching keywords appearing in the text. Specifically, one or more text similarity calculation methods known in the art can be used for comprehensive evaluation and weighted calculation. For example, in the process of analyzing the relevance metric, the edit distance score, the single-word intersection ratio score, the continuous minimum coverage score, and the compactness score can be comprehensively used to quantify the similarity between texts from different dimensions.

[0106] The edit distance score is calculated based on the Levenstein distance, by statistically determining the minimum number of single-character editing operations (including insertion, deletion, and replacement) required to transform one string into another. This score is typically normalized using Formula 1 - (edit distance / maximum string length), directly reflecting the similarity between the two texts at the character sequence level. The word intersection ratio score focuses on the overlap of vocabulary sets. First, the text is segmented into sets of independent characters or words, then the ratio of their intersection to their union size (i.e., the Jaccard similarity coefficient) is calculated. This score effectively measures the proportion of shared vocabulary between texts.

[0107] The continuous minimum coverage score focuses on the shortest continuous segment required to cover a query term in a longer text. Its calculation typically involves locating the smallest interval in the long text where the first occurrences of all words in the query term constitute a complete query. The ratio or reciprocal of this interval length to the query length serves as the score, assessing the compactness and completeness of the query's appearance in the text. The density score assesses the spatial concentration of matching words in the text. By calculating the variance or standard deviation of the sequence of matching words' positions in the text and performing appropriate normalization, a higher score indicates a denser distribution of matching words and stronger local relevance of the text. The relevance measure can be a weighted average of the edit distance score, the percentage of word intersection, the continuous minimum coverage score, and the density score. The weights can be set according to the specific application to achieve a robust assessment of text relevance.

[0108] To improve ranking accuracy, the ranking strategy also incorporates a popularity metric for the content to be pushed into the ranking process. The popularity metric is a quantitative measure of how much a piece of content to be pushed has been interacted with by a user before the current moment. Since any piece of content to be pushed may have been previously selected and delivered to a user's client, data regarding the interaction between that content and the user, such as whether the user clicked on the content and the time spent browsing it after clicking, can be used or included in the calculation of the popularity metric.

[0109] Furthermore, by combining relevance metrics and popularity indicators, such as by weighted averaging, a comprehensive score is obtained that reflects the "relevance to the triggering event" and "attractiveness" of the content to be pushed (hereinafter referred to as comprehensive advantage, which is to take into account both "relevance to the triggering event" and "attractiveness"). This allows the content to be pushed to be sorted according to its comprehensive score, so as to select target content to be pushed in the future.

[0110] Therefore, the push method for the resources disclosed herein uses a sorting strategy that comprehensively considers relevance metrics and popularity indicators. This allows for accurate sorting of each content to be pushed in the candidate content set, ensuring that the selection of target content is accurate and reliable. Consequently, the push method can better meet user needs and significantly improve the user experience.

[0111] Please see Figure 9 and Figure 10 In some implementations, popularity metrics include at least one of the following: click-through rate, effective click-through rate, content playback duration, number of shares, and number of favorites.

[0112] Specifically, click-through rate (CTR) is the percentage of clicks on content to be pushed out of the total number of times the content is displayed within a specific period. Effective click-through rate (CTR), on the other hand, is the percentage of clicks generated by genuine potential users out of the total number of impressions after filtering out invalid or non-targeted clicks (such as bot clicks, accidental clicks, and fraudulent clicks). Effective CTR more accurately measures the actual appeal of the content to be pushed to the target audience compared to CTR.

[0113] Content playback duration is the duration the user remains on the graphical user interface (GUI) after the content to be pushed is clicked and fully displayed. Share count is the total number of times a piece of content to be pushed has been shared by users to internal and / or external platforms. Favorite count is the total number of times a piece of content to be pushed has been favorited by users.

[0114] Click-through rate (CTR), effective CTR, content playback time, number of shares, and number of favorites are all key metrics for measuring the effectiveness of content in attracting user clicks. The popularity metric can be a weighted score of CTR, effective CTR, content playback time, number of shares, and number of favorites (if a certain item is not included, its weight can be considered as 0).

[0115] For example, if the relevance metric of content A to be pushed is 0.5 and the popularity metric is 0.5, and the relevance metric of content B to be pushed is 0.6 and the popularity metric is 0.4, with the relevance metric having a weight of 30% and the popularity metric having a weight of 70%, then the overall score of content A to be pushed is 0.5*30%+0.5*70%=0.5, and the overall score of content B to be pushed is 0.6*30%+0.4*70%=0.46. Content A to be pushed has a greater overall advantage than content B to be pushed, and should be ranked higher (or lower).

[0116] Therefore, in the resource push method disclosed herein, the sorting strategy calculates popularity indicators based on at least one of click-through rate, effective click-through rate, content playback duration, number of shares, and number of favorites, to ensure that the selected target content to be pushed can better meet the user's needs and fully improve the user experience.

[0117] Please see Figure 2 and Figure 11 In some implementations, the push method further includes: Step 0931: Obtain at least one associated resource item related to the triggering event; and Step 0933: Mix the target content to be pushed with at least one associated resource item and output it.

[0118] Specifically, in response to a triggering event, in addition to recalling the target content to be pushed, at least one associated resource item related to the triggering event is also recalled. The regular resource library is at least a portion of the databases of the resource push system 1000, excluding the resource index library used to execute the resource push method of this disclosure. Associated resource items are stored in the regular resource library; exemplarily, associated resource items may include, but are not limited to, at least one of song links, song comments, album titles, related blogs, and related notes.

[0119] The song link is a link that users can click to play music associated with the triggering event. Song comments are textual or other formatted comments from users or other users about the song associated with the triggering event. Album refers to the album to which the song associated with the triggering event belongs, and other content within that album. Related blogs are blogs related to the song associated with the triggering event. Related notes are notes related to the song associated with the triggering event.

[0120] When the target content to be pushed and associated resource items are obtained, the target content to be pushed and at least one associated resource item are mixed and sorted in a certain distribution. For example, the target content to be pushed is interspersed among multiple associated resource items, or the target content to be pushed is placed above all associated resource items. At this time, the graphical user interface displays the target content to be pushed and associated resource items in an ordered distribution based on the above mixed sorting result. For example, please refer to Figure 3 In the image on the left, the target content to be pushed is interspersed between song links and song comments. Users can click on the target content or related resources according to their interests and personal habits.

[0121] Therefore, in the resource push method disclosed herein, the target content to be pushed and related resource items are sorted together and displayed together in the graphical user interface. At this time, users can flexibly select the content they are interested in in the graphical user interface. Users have a large selection space and can enjoy multiple types of resources at the same time, resulting in a better user experience.

[0122] Please see Figures 11 to 13 In some implementations, step 0933 can be achieved through steps 09331 and 09333, specifically as follows: Step 09331: Obtain the popularity metrics of the target content to be pushed and the popularity metrics of related resource items; and Step 09333: When the popularity index of the target content to be pushed is higher than or equal to the popularity index of the associated resource item, the target content to be pushed will be displayed before the associated resource item.

[0123] Specifically, in the resource push method disclosed herein, since there is a certain probability that the target content to be pushed may not accurately match the user's needs, it is not attractive enough to the user. In order to improve the user experience, maintain the number of users, and improve user satisfaction, this implementation method proposes the following optimization strategy.

[0124] The definition of popularity metrics has been described above and will not be repeated here. Here, popularity metrics are used to represent the "attractiveness" of the target content to be pushed and related resource items to users. By obtaining the popularity metrics of the target content to be pushed and the related resource items, the two are compared to determine the degree of user interest in the target content to be pushed and the related resource items presented in the graphical user interface.

[0125] When the popularity metric of the target content to be pushed is higher than or equal to the popularity metric of the associated resource items, it can be determined that the user's interest in the target content exceeds that in the associated resource items. This may indicate that the user has a habit of accessing resources other than those associated, and that the target content to be pushed is more attractive to the user. To improve the user experience and avoid making users feel overly disturbed by uninteresting content, the target content to be pushed should be displayed before the associated resource items so that users can quickly access content that interests them. Figure 12 As shown, compared to Figure 3 At this point, the target content to be pushed is changed from being interspersed between song links and song comments to being placed at the top of the page.

[0126] Therefore, in the resource push method disclosed herein, the popularity index is used as the basis for optimizing the push process of the target content to be pushed. By changing the display position of the target content to be pushed, it can be ensured that the content pushed to the user can well match the user's needs and usage habits. Users can quickly and conveniently browse the content they are really interested in in the graphical user interface, and the user experience is further improved.

[0127] Please see Figure 2 and Figure 14 In some implementations, the push method further includes: Step 061: Obtain the popularity metrics for each piece of content to be pushed; and Step 063: When the popularity index of the target content to be pushed is lower than the preset index threshold, remove the target content to be pushed from the resource index library.

[0128] Specifically, after building the resource index repository, it needs to be managed. This includes operations such as data statistics, analysis, and removal of content to be pushed, to ensure that all content is of high quality. At this point, the resource index repository can be managed using popularity metrics for each piece of content to be pushed.

[0129] The popularity metric threshold is a threshold used to determine a user's interest in a particular resource. A value below the threshold indicates a lack of user interest. If the popularity metric of the target content to be pushed is below the preset threshold, it can be determined that the user lacks interest in that content. This could be due to the content's low relevance to the triggering event or its low appeal to the user. In this case, the target content should be removed from the resource index to prevent it from being subsequently pushed to other users who entered the same triggering event, thus reducing user satisfaction.

[0130] Optionally, the metric threshold is used to determine whether the appeal of the content to be pushed is too low. The setting methods include, but are not limited to: Method 1, setting it to the median historical click-through rate of all content to be pushed; Method 2, setting different absolute thresholds for different resource themes, such as setting the click-through rate threshold for music resources to 1% and for podcast resources to 0.5%; Method 3, setting it to 50% of the average popularity metric of the top 10 related resource items most relevant to this triggered event. When the popularity metric of the target content to be pushed is lower than the preset metric threshold, it is removed from the resource index.

[0131] Therefore, in the resource push method disclosed herein, using popularity metrics as the basis for managing the resource index library and the strategy of removing low-attractive target content to be pushed can ensure that the content pushed to users is of high quality and attractiveness, significantly improving the user experience.

[0132] All of the above technical solutions can be combined in any way to form optional embodiments of this disclosure, and will not be described in detail here.

[0133] In the resource push method provided in this disclosure, a corpus is obtained by acquiring and integrating data related to the target information. This corpus is then input into a resource generation model to generate pushable content containing questions and short texts, and a resource index library is constructed. In response to a trigger event, the target pushable content, which is associated with the trigger event and may correspond to the user's request, can be directly determined from the resource index library and output for the user to access. Therefore, by using pre-generated pushable content, the response and waiting time for users in obtaining content related to their requests is reduced, enabling users to quickly access relevant content and improving the smoothness and user experience of the resource interaction platform.

[0134] To facilitate better implementation of the resource push method of the embodiments of this disclosure, one embodiment of this disclosure also provides a resource push device 200. This resource push device can be applied to a server in software or hardware to perform resource push operations. Since the device embodiment is basically similar to the method embodiment, the description is relatively simple. For details of the relevant technical features, please refer to the corresponding descriptions of the method embodiments provided above. The following description of the device embodiment is merely illustrative.

[0135] Please see Figure 15 Understanding this embodiment, Figure 15 As shown in the figure, the resource push device 200 provided in this embodiment includes: Display unit 201 is used to display a graphical user interface.

[0136] The control unit 203 is connected to the display unit 201. The control unit 203 is used to acquire a corpus set related to at least one target information, input the corpus set into a resource generation model to generate structured content to be pushed, construct a resource index library based on the content to be pushed, determine the target content to be pushed associated with the trigger event from the resource index library in response to a trigger event, and output the target content to be pushed.

[0137] In some embodiments, the control unit 203 is used to acquire corpus related to target information from multiple data sources.

[0138] In some embodiments, the control unit 203 is configured to acquire a first type of corpus from a first data source to form a first corpus subset; acquire a second type of corpus from a second data source to form a second corpus subset, wherein the second data source is different from the first data source; and integrate the first corpus subset and the second corpus subset to obtain a corpus set.

[0139] In some embodiments, the control unit 203 formats the corpus to generate prompt information that meets the input requirements of the resource generation model; and inputs the prompt information into the resource generation model to generate structured content to be pushed.

[0140] In some embodiments, the control unit 203 is used to embed a corpus into a prompt word template to generate resource generation prompt words.

[0141] In some embodiments, the control unit 203 is configured to determine a set of candidate content to be pushed associated with the triggering event from a resource index library; sort each content to be pushed in the set of candidate content to be pushed based on a sorting strategy; and select at least one content to be pushed as the target content to be pushed based on the sorting result.

[0142] In some embodiments, the control unit 203 is used to obtain multiple dimensions of evaluation metrics for each content to be pushed; and, based on the multiple dimensions of evaluation metrics, to determine the sorting strategy for each content to be pushed.

[0143] In some embodiments, the control unit 203 is used to obtain a correlation metric between each content to be pushed and the triggering event; and to obtain a popularity metric for each content to be pushed.

[0144] In some embodiments, the control unit 203 is configured to acquire at least one associated resource item associated with the triggering event; and to output the target content to be pushed along with at least one associated resource item.

[0145] In some embodiments, the control unit 203 is used to obtain the popularity index of the target content to be pushed and the popularity index of the associated resource item; and, when the popularity index of the target content to be pushed is higher than or equal to the popularity index of the associated resource item, the target content to be pushed is displayed on the display unit 201 before the associated resource item.

[0146] In some embodiments, the control unit 203 is used to obtain the popularity index of each content to be pushed; and, when the popularity index of the content to be pushed is lower than a preset index threshold, the content to be pushed is removed from the resource index library.

[0147] Each unit in the aforementioned resource delivery device 200 can be implemented entirely or partially through software, hardware, or a combination thereof. Each unit can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can invoke and execute the operations corresponding to each unit.

[0148] The resource push device 200 can be integrated into a terminal or server that has storage and a processor and thus computing power, or the resource push device 200 can be the terminal or server.

[0149] Optionally, this disclosure also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0150] Figure 16 This is a schematic diagram of the structure of a computer device provided in one embodiment of the present disclosure. The computer device may be a terminal or a server. Figure 16 As shown, the computer device 300 includes a processor 301 with one or more processing cores, a memory 302 with one or more computer-readable storage media, and a program for a method of pushing resources stored in the memory 302 and executable on the processor. The processor 301 and the memory 302 are electrically connected. Those skilled in the art will understand that the computer device structure shown in the figures does not constitute a limitation on the computer device, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0151] The processor 301 is the control center of the computer device 300. It connects various parts of the computer device 300 through various interfaces and lines. By running or loading software programs and / or modules stored in the memory 302, and calling data stored in the memory 302, it performs various functions of the computer device 300 and processes data, thereby performing overall processing of the computer device 300.

[0152] In one embodiment of this disclosure, the processor 301 in the computer device 300 loads the instructions corresponding to the processes of one or more resource push method programs into the memory 302 according to the following steps, and the processor 301 runs the resource push method program stored in the memory 302 to realize various functions: Acquire a corpus related to at least one target piece of information; The corpus is input into a preset resource generation model to generate content to be pushed, the content to be pushed including at least a title and a short text corresponding to the title; A searchable resource index library is constructed based on the content to be pushed; In response to a triggering event, retrieve the target content to be pushed associated with the triggering event from the resource index; and The target content to be pushed is pushed and displayed on the graphical user interface.

[0153] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0154] Optional, such as Figure 16 As shown, the computer device 300 also includes: a display screen 303, a radio frequency circuit 304, an audio circuit 305, an input unit 306, and a power supply 307. The processor 301 is electrically connected to the display screen 303, the radio frequency circuit 304, the audio circuit 305, the input unit 306, and the power supply 307. Those skilled in the art will understand that... Figure 16 The computer device structure shown does not constitute a limitation on the computer device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0155] The display screen 303 can be used for a graphical user interface (GUI) and to receive operation commands generated by the user interacting with the GUI. The display screen 303 may include a display panel and a touch panel. The display panel can be used to display information input by the user or information provided to the user, as well as various graphical user interfaces of the computer device. These graphical user interfaces can be composed of graphics, text, icons, video, and any combination thereof. The touch panel can be used to collect touch operations performed by the user on or near it (such as operations performed by the user using a finger, stylus, or any suitable object or accessory on or near the touch panel), and generate corresponding operation commands, which then execute the corresponding program. Optionally, the touch panel may include a touch detection device and a touch controller. The touch detection device detects the user's touch location and the signal generated by the touch operation, transmitting the signal to the touch controller. The touch controller receives touch information from the touch detection device, converts it into touch point coordinates, sends it to the processor 301, and can also receive and execute commands from the processor 301. The touch panel can cover the display panel. When the touch panel detects a touch operation on or near it, it transmits the information to the processor 301 to determine the type of touch event. Subsequently, the processor 301 provides corresponding visual output on the display panel based on the type of touch event. In this embodiment, the touch panel and the display panel can be integrated into the display screen 303 to achieve input and output functions. However, in some embodiments, the touch panel and the display panel can be implemented as two independent components to achieve input and output functions. That is, the display screen 303 can also be used as part of the input unit 306 to achieve input functions.

[0156] The radio frequency circuit 304 can be used to transmit and receive radio frequency signals to establish wireless communication with network devices or other computer devices, and to transmit and receive signals with network devices or other computer devices.

[0157] Audio circuitry 305 can be used to provide an audio interface between a user and a computer device via a speaker and a microphone. Audio circuitry 305 converts received audio data into electrical signals, transmits them to the speaker, and the speaker converts them into sound signals for output. Conversely, the microphone converts collected sound signals into electrical signals, which are then received by audio circuitry 305, converted back into audio data, and output to processor 301 for processing. The audio data is then transmitted via radio frequency circuitry 304 to, for example, another computer device, or output to memory 302 for further processing. Audio circuitry 305 may also include an earphone jack to facilitate communication between peripheral headphones and the computer device.

[0158] The input unit 306 can be used to receive input numbers, characters, or object feature information (such as fingerprints, irises, facial information, etc.), and to generate keyboard, mouse, joystick, optical, or trackball signal inputs related to user settings and function control.

[0159] Power supply 307 is used to supply power to various components of computer device 300. Optionally, power supply 307 can be logically connected to processor 301 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. Power supply 307 may also include one or more DC or AC power supplies, recharging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.

[0160] although Figure 16 As not shown in the diagram, computer equipment 300 may also include a camera, sensor, wireless fidelity module, Bluetooth module, etc., which will not be described in detail here.

[0161] This disclosure also provides a computer-readable storage medium for storing a computer program. This computer-readable storage medium can be applied to a computer device, and the computer program causes the computer device to execute corresponding processes in the display control method of the embodiments of this disclosure; for brevity, further details are omitted here.

[0162] This disclosure also provides a computer program product including 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 corresponding flow in the resource pushing method of the embodiments of this disclosure, such as executing the methods in steps 01, 011, 0111, 0113, 0115, 03, 031, 0311, 033, 05, 061, 063, 07, 071, 073, 0731, 07311, 07313, 0733, 075, 091, 093, 0931, 0933, 09331, and 09333. For simplicity, these details are not elaborated here.

[0163] It should be understood that the processor disclosed herein may be an integrated circuit chip with signal processing capabilities. In implementation, the steps of the above method embodiments can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor described above may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this disclosure. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this disclosure can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules may reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0164] It is understood that the memory in the embodiments of this disclosure can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DR RAM). It should be noted that the memory used in the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0165] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this disclosure.

[0166] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0167] In this disclosure, the terms "module" or "unit" refer to a computer program or part of a computer program that has a predetermined function and works with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0168] In the several embodiments provided in this disclosure, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0169] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0170] In addition, the functional units in this disclosure can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0171] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer or a server) to execute all or part of the steps of the methods described in the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0172] The above description is merely a specific embodiment of this disclosure, but the scope of protection of this disclosure is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this disclosure should be included within the scope of protection of this disclosure. Therefore, the scope of protection of this disclosure should be determined by the scope of the claims.

Claims

1. A resource push method, characterized in that, The methods include: Acquire a corpus related to at least one target piece of information; The corpus is input into the resource generation model to generate structured content to be pushed. Build a resource index library based on the content to be pushed; In response to a triggering event, the target content to be pushed associated with the triggering event is determined from the resource index library; and Output the target content to be pushed.

2. The method according to claim 1, characterized in that, The acquisition of the corpus related to at least one target information includes: Corpus related to the target information is obtained from multiple data sources.

3. The method according to claim 2, characterized in that, The step of obtaining corpus related to the target information from multiple data sources includes: The first type of corpus is obtained from the first data source to form a subset of the first corpus; A second type of corpus is obtained from a second data source to form a second subset of corpus, wherein the second data source is different from the first data source; and, The first subset of corpus and the second subset of corpus are integrated to obtain the corpus set.

4. The method according to claim 1, characterized in that, The step of inputting the corpus into the resource generation model to generate structured content to be promoted includes: The corpus is formatted to generate prompt information that meets the input requirements of the resource generation model; and, The prompt information is input into the resource generation model to generate the structured content to be pushed.

5. The push method according to claim 1, characterized in that, The step of determining the target content to be pushed associated with the triggering event from the resource index includes: Determine a set of candidate content to be pushed that is associated with the triggering event from the resource index library; Based on the sorting strategy, the content to be pushed in the candidate content set is sorted; and, At least one piece of content to be pushed is selected based on the sorting results, and is designated as the target content to be pushed.

6. The push method according to claim 1, characterized in that, Also includes: Retrieve at least one associated resource item related to the triggering event; and The target content to be pushed is mixed with the at least one associated resource item and output.

7. The push method according to claim 1, characterized in that, Also includes: Obtain the popularity metrics for each of the aforementioned content items to be pushed; and When the popularity index of the content to be pushed is lower than a preset index threshold, the content to be pushed is removed from the resource index library.

8. A resource delivery device, characterized in that, The push device provides a graphical user interface via a terminal and includes: Display unit; and The control unit, connected to the display unit, is used for: Acquire a corpus related to at least one target piece of information; The corpus is input into the resource generation model to generate structured content to be pushed. Build a resource index library based on the content to be pushed; In response to a triggering event, determine the target content to be pushed associated with the triggering event from the resource index; and Output the content to be pushed.

9. A computer device, characterized in that, The computer device includes a processor and a memory, the memory storing a computer program, and the processor executing the resource push method according to any one of claims 1-7 by calling the computer program stored in the memory.

10. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the resource push method according to any one of claims 1-7.

Images