Content generation method and device and computer readable storage medium

Abstract

The invention relates to a content generation method and device and a computer readable storage medium, and relates to the technical field of computers. The content generation method comprises the following steps: determining key information corresponding to a target event; selecting a candidate multimedia resource associated with the target event as a target multimedia resource from a resource library according to the correlation between at least one-dimensional feature of the candidate multimedia resource in the resource library and the key information; and generating recommended content of the target multimedia resource according to the key information and the information of the target multimedia resource.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to a content generation method and device, and a computer-readable storage medium. Background Art

[0002] With the rapid development of Internet technology, the consumption of multimedia content has been deeply integrated into all aspects of modern life and has become an important carrier for people to obtain information, leisure and entertainment, and social interaction.

[0003] With the popularization of 5G networks and the rapid iteration and upgrading of smart terminal devices, the consumption scenarios of multimedia content are constantly expanding. Users can play videos, listen to audiobooks, and read novels anytime and anywhere. Summary of the Invention

[0004] According to some embodiments of the present disclosure, a content generation method is provided, including: determining key information corresponding to a target event; selecting candidate multimedia resources associated with the target event from the resource library as target multimedia resources based on the correlation between the characteristics of at least one dimension of the candidate multimedia resources in the resource library and the key information; and generating recommended content for the target multimedia resources based on the key information and information of the target multimedia resources.

[0005] According to other embodiments of the present disclosure, a content generation device is provided, including: a determination module, configured to determine key information corresponding to a target event; a selection module, configured to select, from a resource library, candidate multimedia resources associated with the target event as target multimedia resources based on the correlation between the characteristics of at least one dimension of the candidate multimedia resources in the resource library and the key information; and a generation module, configured to generate recommended content for the target multimedia resource based on the key information and information of the target multimedia resource.

[0006] According to some other embodiments of the present disclosure, a content generation device is provided, including: a memory; and a processor coupled to the memory, wherein the processor is configured to execute the content generation method according to some embodiments of the present disclosure based on instructions stored in the memory.

[0007] According to some other embodiments of the present disclosure, a computer-readable storage medium is provided, on which computer program instructions are stored. When the instructions are executed by a processor, the content generation method according to some embodiments of the present disclosure is implemented.

[0008] According to some other embodiments of the present disclosure, a computer program product is provided, including computer program instructions, which, when executed by a processor, implement the content generation method according to some embodiments of the present disclosure.

[0009] Other features, aspects and advantages of the present disclosure will become apparent from the following detailed description of exemplary embodiments of the present disclosure with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The following describes embodiments of the present disclosure with reference to the accompanying drawings. It should be understood that the drawings described below only relate to some embodiments of the present disclosure and do not constitute a limitation to the present disclosure. In the accompanying drawings:

[0011] Figure 1 A flowchart showing a content generation method according to some embodiments of the present disclosure is shown;

[0012] Figure 2 A schematic diagram of a process for determining a target event according to some embodiments of the present disclosure is shown;

[0013] Figure 3 A schematic diagram illustrating recommending resources associated with hot events according to some embodiments of the present disclosure is shown;

[0014] Figure 4 A block diagram illustrating a content generation device according to some embodiments of the present disclosure is shown;

[0015] Figure 5 A block diagram illustrating a content generation device according to some embodiments of the present disclosure is shown;

[0016] Figure 6 A block diagram of an electronic device according to some other embodiments of the present disclosure is shown.

[0017] It should be understood that, for ease of description, the dimensions of the various parts shown in the drawings are not necessarily drawn to scale. The same or similar reference numerals are used throughout the drawings to indicate the same or similar parts. Therefore, once an item is defined in one drawing, it may not be discussed further in subsequent drawings. DETAILED DESCRIPTION

[0018] The following will be combined with the accompanying drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. It should be understood that the present disclosure can be implemented in various forms and should not be interpreted as being limited to the embodiments described herein.

[0019] It should be understood that the various steps described in the method embodiments of the present disclosure can be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect. Unless otherwise specifically stated, the relative arrangement of the steps set forth in these embodiments should be interpreted as being merely exemplary and not limiting the scope of the present disclosure.

[0020] The term “including” and its variations used in the present disclosure are open terms that include at least the following elements / features but do not exclude other elements / features, that is, “including but not limited to.” The term “based on” means “at least in part based on.”

[0021] It should be noted that the modifications of "one" and "multiple" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, they should be understood as "one or more".

[0022] The following detailed description of the embodiments of the present disclosure is provided in conjunction with the accompanying drawings, but the present disclosure is not limited to these specific embodiments. The following specific embodiments may be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments. In addition, in one or more embodiments, specific features, structures, or characteristics may be combined in any suitable manner that will be apparent to those skilled in the art from this disclosure.

[0023] Figure 1 A flowchart of a content generation method according to some embodiments of the present disclosure is shown.

[0024] like Figure 1 As shown, the content generation method includes: step S1, determining key information corresponding to the target event; step S2, selecting candidate multimedia resources associated with the target event from the resource library as target multimedia resources based on the correlation between the characteristics of at least one dimension of the candidate multimedia resources in the resource library and the key information; step S3, generating recommended content for the target multimedia resource based on the key information and information of the target multimedia resource.

[0025] Target events are, for example, current hot topics, highly-watched events, and events that reflect the user's short-term interests. Key information about a target event includes, for example, the time, location, subject (person or organization), cause, process, and outcome of the event. Key information can be a sentence describing the target event or a title.

[0026] Multimedia resources (i.e., candidate multimedia resources and target multimedia resources) include, for example, novels, short plays, TV series, animation, and audiobooks. The characteristics of candidate multimedia resources can be determined based on various dimensions, such as theme, subject matter, plot, characters, narrative techniques, mood, and time. Target multimedia resources are the multimedia resources to be recommended to users. Recommended content includes, for example, keywords, titles, text, dialogue, cover art, and videos used to recommend multimedia resources.

[0027] A target event can match one or more target multimedia resources at the same time, and one or more recommendation copy can be generated for each target multimedia resource.

[0028] For example, using technologies like natural language processing, we analyze hot events and extract key information, such as time (e.g., specific date, time period), location (city, venue), and theme. Using this extracted key information as a benchmark, we screen multimedia resources within the resource library. For example, we annotate novels and short plays within the resource library with multidimensional features. Through manual semantic analysis or structured modeling, we transform the specific content of these novels and short plays into abstract features at the dimensional level. During matching, we compare the semantic relevance and structural similarity of these dimensional features with the key information, selecting novels and short plays that align with the core of the event in a specific dimension.

[0029] After matching the target multimedia resources, the inherent connection between them and the hot event is explored. By combining information from the resources themselves, such as novel titles, video highlights, and image elements, recommendations are generated, including text descriptions, images, and videos. These recommendations can be displayed to users to recommend novels, short plays, and other content. During the development of a hot event, recommendations for novels and short plays that are more relevant to the hot event can be prioritized. For example, consider the hot event "XX team releases the world's first XXXX artificial intelligence model, which is used in XX scenarios and has XXX performance..." The keyword "artificial intelligence" can be extracted from the text of the hot event. The resource library is then searched for novels and short plays related to "artificial intelligence" in terms of theme, subject matter, plot, characters, narrative techniques, emotion, and time. For example, a novel (or short play) titled "Digital Times" is found, and combined with the hot event keyword "artificial intelligence," a recommendation for "Digital Times" can be generated, such as "Where is future technology heading?"

[0030] The content generation method of this embodiment can be executed on the client side, or partially executed on the server side.

[0031] The content generation method disclosed herein matches the characteristics of a target event and a certain dimension of multimedia resources to select target multimedia resources closely related to the target event and generate recommended content for the target multimedia resources. This method can quickly recommend multimedia resources associated with the target event, especially for target events with short popularity cycles. This can more accurately meet users' current browsing preferences and increase the click-through rate of the target multimedia resources.

[0032] Figure 2 A schematic diagram of a process for determining a target event according to some embodiments of the present disclosure is shown.

[0033] In some embodiments, the content generation method further includes step S0: selecting a target event from the candidate events based on user feedback information on the candidate events. Figure 2 , introduces a method for determining a target event according to some embodiments of the present disclosure.

[0034] like Figure 2 As shown, step S0 includes: step S01, obtaining user feedback information on candidate events; step S02, calculating the scores of the candidate events according to the feedback information; step S03, selecting a target event from the candidate events according to the scores of the candidate events.

[0035] For example, event information posted on social media is collected and cleaned to remove noise to obtain candidate events. User feedback on candidate events includes metrics such as page views, comments, favorites, and reposts. Weights are assigned based on the importance of each metric, and a weighted calculation is performed to determine the candidate event score. The top N candidate events with the highest scores are then identified as target events, where N is, for example, a positive integer. Furthermore, candidate events can be reviewed to exclude those that do not meet the requirements.

[0036] Taking hot tech events as an example, we collect real-time information on cutting-edge fields like machine learning. We clean this information and filter out low-quality content. We can also merge similar events. For example, if "Newly released language model" and "XX team releases XXX model" actually refer to the release of the same model, we can merge them. For candidate events, we can build a multi-dimensional quantitative evaluation system that comprehensively considers the value of candidate events in multiple aspects and then assigns a score to each.

[0037] In some embodiments, determining key information corresponding to a target event includes: identifying key entities from information about the target event, wherein the key entities include at least one of a person, time, place, and the course of an event; and generating text with a length not exceeding a threshold as key information based on the key entities.

[0038] For example, a machine learning model first extracts structured key entities from the collected raw information. It then analyzes the semantic relationships between these entities, clarifies the logic of the event's development, and generates a coherent and accurate summary of the event as key information. This key information can be a keyword, a short sentence, a title, or a summary paragraph.

[0039] For example, consider the hot topic "XX Team releases the world's first XXXX artificial intelligence model. The model is used in XX scenarios and has XXX performance..." Key structured entity information is extracted from related news reports and user discussions, such as the author or organization that released the model (e.g., "XX Team"), the model's functionality (e.g., "the world's first XXXX"), and the application scenario (e.g., "XX scenario"). After obtaining this entity information, semantic analysis is used to identify the relationships between these entities. Based on these entities and relationships, a deep understanding of the text semantics is achieved, generating a summary as key information. Because "artificial intelligence" is a core concept throughout the event description, entities such as the model name, functionality description, and application scenarios all have close semantic connections to "artificial intelligence." Through in-depth text semantic mining and analysis, "artificial intelligence" is ultimately extracted as a keyword, providing more precise targeting for subsequent matching.

[0040] The following describes a method for determining a target multimedia resource associated with a target event.

[0041] In some embodiments, based on the correlation between the characteristics of at least one dimension of the candidate multimedia resources in the resource library and the key information, a candidate multimedia resource associated with the target event is selected from the resource library as the target multimedia resource, including: in response to the key information including the event process of the target event, the plot dimension characteristics of the candidate multimedia resource are obtained as the characteristics of at least one dimension of the candidate multimedia resource; step S22, based on the correlation between the event process of the target event and the plot dimension characteristics of the candidate multimedia resource, the target multimedia resource is selected from the resource library.

[0042] If the extracted key information includes the cause, outcome, turning point, or conflict of the target event, similar plots in candidate multimedia resources are searched for as target multimedia resources associated with the target event. For example, if the plot dimension of the target event is characterized by "bizarre," candidate multimedia resources with suspenseful plots are searched for as target multimedia resources to be recommended that match the target event. If the user has previously followed the target event, they may have recently become interested in the conflicts within it, and recommending target multimedia resources with similar plots may better align with their interests.

[0043] In the context of novel and short play recommendations, after extracting the event-process dimensional features from hot events, a multi-level analysis of the novel and short play narratives can be used to establish a deep connection between hot events and novels and short plays. This connection not only considers the matching of plot elements but also explores the similarities in the story development logic. A mapping model is constructed from "event" to "novel plot" or "short play plot," mapping people's real-life attention to the development process of hot topics with the story context in the fictional narrative. For example, for the "emotional relationship" hot event, not only novels and short plays with similar character relationships are matched, but also novels and short plays that resonate with the hot event at their core can be connected by analyzing the matching of deeper narratives such as the emotional changes of the characters.

[0044] In some embodiments, based on the correlation between the characteristics of at least one dimension of the candidate multimedia resources in the resource library and the key information, a candidate multimedia resource associated with the target event is selected from the resource library as the target multimedia resource, including: in response to the key information including the theme of the target event, obtaining background dimension characteristics of the candidate multimedia resource as the characteristics of at least one dimension of the candidate multimedia resource; based on the correlation between the theme of the target event and the background dimension characteristics of the candidate multimedia resource, the target multimedia resource is selected from the resource library.

[0045] If the extracted key information is the theme of the target event, such as a phenomenon or viewpoint, then candidate multimedia resources with similar story settings can be searched for as target multimedia resources associated with the hot event. For example, if the target event's theme dimension is "technology," novels and short plays with a futuristic technology background can be matched as target multimedia resources to be recommended that match the target event's theme. If the user has followed the target event, it indicates that they may have recently shown interest in its theme, and recommending target multimedia resources with similar background settings may better suit their interests.

[0046] When recommending novels and short plays, after extracting thematic features from hot events, the system then establishes deep connections with the novels and short plays through multi-level semantic analysis. This connection not only considers keyword matching but also explores similarities in thematic content, constructing a mapping model from "event" to "novel setting" or "short play story setting," mapping people's real-life concerns about hot topics with the worldviews in fictional narratives. For example, for the "technology" hot event, not only are novels and short plays set in the future matched, but novels and short plays with ideological connotations that resonate with the hot event can also be linked through matching deeper themes such as the relationship between man and machine.

[0047] In some embodiments, based on the correlation between the characteristics of at least one dimension of the candidate multimedia resources in the resource library and the key information, a candidate multimedia resource associated with the target event is selected from the resource library as the target multimedia resource, including: in response to the key information including the character information of the target event, obtaining the character dimension characteristics of the candidate multimedia resource; based on the correlation between the character information of the target event and the character dimension characteristics of the candidate multimedia resource, selecting the target multimedia resource from the resource library.

[0048] If the extracted key information includes information about the characters in the target event (such as their upbringing, personality, and relationship networks), we can search for candidate multimedia resources with similar character traits as target multimedia resources associated with the hot event. If the core character trait of the target event is "hero," we can match novels and short plays with similar protagonists as target multimedia resources to be recommended. If the user has followed the target event, it indicates that they may have recently followed similar characters, and recommending target multimedia resources with similar character settings may better match their interests.

[0049] When recommending novels and short plays, after extracting character-level features from hot events, the system then establishes deep connections with the novels and short plays through multi-layered personality analysis. This connection not only considers matching character labels but also explores similarities in the characters' underlying traits, constructing a mapping model from "event" to "novel character" or "short play protagonist," thereby mapping people's attention to hot figures in real life with the characterizations in fictional narratives. For example, for the hot figure "Technology Leader," not only are novels and short plays featuring scientists matched, but also novels and short plays whose values ​​resonate with the hot figure through matching deeper traits like ideals and spirituality.

[0050] In some embodiments, based on the correlation between the features of at least one dimension of the candidate multimedia resources in the resource library and the key information, a candidate multimedia resource associated with the target event is selected from the resource library as the target multimedia resource, including: extracting features of multiple dimensions of the multimedia resource; splicing the features of multiple dimensions; and determining the correlation based on the semantic similarity between the features of the spliced ​​multiple dimensions and the key information.

[0051] For example, consider a book as a candidate multimedia resource. First, determine the various text features to be extracted, such as the title, author, description, and category. For target events, you can set a maximum length for each extracted text feature; any excess length will be truncated. For example, the maximum length of a book title is 50 characters, and the maximum length of a description is 250 characters. The maximum length can be determined based on coverage. For example, if a length of 50 characters covers 95% of actual book title data, then the maximum length of the title is 50 characters.

[0052] Key information is also text, such as a summary of the target event, a title, or a short sentence describing the target event. After extracting the text features of the candidate multimedia resources, they are concatenated. The semantic similarity between the key information and the concatenated text is calculated, and the candidate multimedia resource corresponding to the concatenated text with the highest semantic similarity is selected as the target multimedia resource.

[0053] For example, the key information of a hot event, "artificial intelligence," is matched with the text features of a novel resource. First, the text information of a novel, such as the title "AAAAA," the story synopsis "BBBBBB," and the character settings "CCCCCCC," is extracted and concatenated into the text "AAAAABBBBBBCCCCCCC." Then, the semantic similarity of the feature vector of "AAAAABBBBBBCCCCCCC" is calculated with the feature vector of "artificial intelligence." If the similarity exceeds the threshold, the novel is selected as data to be recommended. This matching method based on semantic similarity can break through the limitations of simple keyword matching, covering more comprehensive and critical novel information, thereby improving the accuracy of matching. If the multimedia resource is a short play, the content of the short play can be analyzed through image semantic analysis, an introduction text for the short play can be generated, and text features can be extracted from the introduction text.

[0054] In some embodiments, the information of the target multimedia resource includes the name and introduction of the target multimedia resource, and recommended content of the target multimedia resource is generated based on the key information and the information of the target multimedia resource, including: generating prompt information based on the key information, the name and introduction of the target multimedia resource; and generating recommended content based on the prompt information using a machine learning model.

[0055] For example, the target multimedia resource is analyzed to extract key features, and a name and description are generated based on these features. Then, a prompt message is constructed that includes key information, the name of the target multimedia resource, and a description. This prompt message serves as input to a machine learning model to generate recommended content for the target multimedia resource.

[0056] For example, in the novel and short play recommendation scenario, the key information is "artificial intelligence", the title of the novel or short play is "Digital Age", and the story synopsis is "XXXXXX", then the following prompt information is generated: "Now we need to generate content copy to send to users, referring to hot events and related book information. The hot event is "artificial intelligence", the title of the book (or play) is "Digital Age", and the story synopsis is "XXXXXX". Restrictions: Focus on hot events, analyze the relationship between hot events and novels, and generate copy, but do not include information that explicitly points to hot events. The copy should not exceed 20 words, and at least 3 different copies need to be generated for each book." The prompt information is input into the model to generate recommendation copy for novels and short plays around hot events.

[0057] The machine learning model deeply analyzes the connection between the hot topic "artificial intelligence" and "Digital Times". Based on the knowledge of the cutting-edge scientific fields described by "artificial intelligence", as well as the thinking about digital technology in "Digital Times" and the exciting plot in the novel, it generates copy.

[0058] For the combination of the same target multimedia resource and an associated target event, the machine learning model can generate multiple candidate solutions in parallel, then score them based on multiple dimensions such as language fluency, expected page views, and relevance to the target event, ultimately selecting content to recommend to the user. In some embodiments, the content generation method further includes: categorizing the recommended content; configuring the effective duration of each type of recommended content; calculating the score of the recommended content based on user feedback on the target event corresponding to the recommended content; and determining whether to push the recommended content and the push time of the recommended content based on the effective duration and score of the recommended content.

[0059] For example, the generated recommended content is stored in the content pool, and the validity period of each type of recommended content in the content pool is set. The validity period refers to the period within which the content is pushed to the user. For example, the recommended content needs to be pushed to the user within three days from the time it enters the content pool. The type of recommended content can be determined based on the type of the target event associated with it. For example, if the target event is news, then the corresponding recommended content type is news, and the validity period of news can be set to be shorter, such as 3 days. For recommended content related to interpersonal relationships, the validity period can be set to be longer, such as half a month.

[0060] When recommending novels and short plays, you can also set different validity periods for different types of novels and short plays: for example, new books have a shorter validity period, while classic works have a longer validity period. The validity period of the corresponding copy can be calculated by combining the validity period of the target event and the validity period of the novel (or short play).

[0061] User feedback on the target event corresponding to the recommended content, for example, represents user attention to the target event, such as the number of user comments. The higher the user attention, the higher the score of the corresponding recommended content. Based on the recommended content's validity period and score, a decision is made as to whether to push the recommended content. If so, the push time is determined. For example, recommended content with a high score and a short validity period is given a higher priority and can be pushed to the user more quickly. Recommended content with a low priority may not be recommended.

[0062] For example, a hot topic event's key information is "artificial intelligence," a cutting-edge field of science and technology. The related novel (or short play) "Digital Age," as well as the recommended copy for "Digital Age," "Where is Future Technology Heading?", also revolve around themes of science and technology. Therefore, the recommended content type can be determined to be science and technology. Based on the preset validity period, recommended content related to science and technology hot topics has a shorter validity period, for example, three days. This means that the recommended content must be delivered within three days of entering the content pool.

[0063] Meanwhile, Digital Times recommends new books or dramas related to hot events, with a validity period of five days. Taking into account the validity period of the hot event (three days) and the validity period of the novel (five days), the shorter of the two is used as the validity period for the recommended content, which is three days.

[0064] Furthermore, user feedback on hot topics related to "AI" can be collected, such as the number of comments, likes, and shares. If user interest is high and there are a lot of comments and likes, the corresponding recommended content, "Where is Future Technology Heading?", will receive a higher score than other content.

[0065] In summary, "Where is Future Technology Heading" is a recommended content with a high score and a short validity period. It has a higher priority and will be pushed to users within 3 days.

[0066] Figure 3 A schematic diagram illustrating recommendation of resources associated with hot events according to some embodiments of the present disclosure is shown.

[0067] like Figure 3 As shown, based on the definition of hot events, hot events are categorized into categories such as festivals, sports, and entertainment. One or more categories are selected, and hot information for the selected categories is collected to mine hot events. For hot events, a machine learning model is used to extract common key phrases that are unrelated to the characters in the event and serve as key information for the hot events. A candidate multimedia resource pool stores multimedia resources such as novels, short plays, audiobooks, and short stories. Recommended multimedia resources are selected from these resources based on specified criteria and serve as candidate multimedia resources. A matching model is used to match the key information of the hot event with the candidate multimedia resources to select the target multimedia resource. Recommended content is generated for the target multimedia resource. Multiple content can be generated for the same target multimedia resource. These content is then reviewed by the review system based on relevance to the hot event and popularity. Approved content is visualized and entered into a distribution candidate pool for placement and push based on the characteristics of different communication channels. Content scores are calculated, and the content to be pushed and the push date are determined based on the score.

[0068] Figure 4 A block diagram of a content generating device according to some embodiments of the present disclosure is shown.

[0069] like Figure 4 As shown, the content generating device 4 includes: a determining module 41, configured to determine key information corresponding to the target event; a selecting module 42, configured to select candidate multimedia resources associated with the target event from the resource library as target multimedia resources based on the correlation between the characteristics of at least one dimension of the candidate multimedia resources in the resource library and the key information; a generating module 43, configured to generate recommended content of the target multimedia resource based on the key information and information of the target multimedia resource.

[0070] The determination module 41 of the content generation device 4 can be used to perform Figure 1 In step S1, the selection module 42 may be used to perform Figure 1 Step S2, the generating module 43 can be used to perform Figure 1 Step S3.

[0071] The content generation device disclosed herein selects target multimedia resources closely related to the target event by matching characteristics of a certain dimension of the multimedia resource with the target event, and generates recommended content for the target multimedia resource. This method allows for rapid recommendation of multimedia resources associated with the target event, particularly for target events with short popularity cycles. This allows for more accurate matching of users' current browsing preferences and improves the click-through rate of the target multimedia resource.

[0072] In some embodiments, the selection module 42 is further configured to: in response to the key information including the event process of the target event, obtain the plot dimension characteristics of the candidate multimedia resource as the characteristics of at least one dimension of the candidate multimedia resource; and select the target multimedia resource from the resource library based on the correlation between the event process of the target event and the plot dimension characteristics of the candidate multimedia resource.

[0073] In some embodiments, the selection module 42 is further configured to: in response to the key information including the theme of the target event, obtain background dimension features of the candidate multimedia resource as features of at least one dimension of the candidate multimedia resource; and select the target multimedia resource from the resource library based on the correlation between the theme of the target event and the background dimension features of the candidate multimedia resource.

[0074] In some embodiments, the selection module 42 is further configured to: obtain character dimension characteristics of the candidate multimedia resources in response to the key information including the character information of the target event; and select the target multimedia resource from the resource library based on the correlation between the character information of the target event and the character dimension characteristics of the candidate multimedia resources.

[0075] In some embodiments, the selection module 42 is further configured to: extract features of multiple dimensions of the multimedia resource; combine the features of multiple dimensions; and determine the relevance based on the semantic similarity between the combined features of multiple dimensions and the key information.

[0076] In some embodiments, the generation module 43 is further configured to: generate prompt information based on the key information, the name and introduction of the target multimedia resource; and generate recommended content based on the prompt information using a machine learning model.

[0077] In some embodiments, the content generation device 4 further includes an event selection module configured to: obtain user feedback information on candidate events; calculate scores of candidate events based on the feedback information; and select a target event from the candidate events based on the scores of the candidate events.

[0078] In some embodiments, the determination module 41 is further configured to: identify key entities from the information of the target event, wherein the key entities include at least one of a person, a time, a place, and the course of the event;

[0079] Based on the key entity, a text with a length not exceeding a threshold is generated as the key information.

[0080] In some embodiments, the content generation device 4 also includes a calculation module, which is configured to: classify recommended content; configure the effective duration of each type of recommended content; calculate the score of the recommended content based on the user's feedback information on the target event corresponding to the recommended content; determine whether to push the recommended content and the push time of the recommended content based on the effective duration and score of the recommended content.

[0081] Figure 5 A block diagram of a content generating device according to some embodiments of the present disclosure is shown.

[0082] like Figure 5 As shown, the content generating device 5 includes: at least one memory 51; and at least one processor 52 coupled to the at least one memory 51, and the at least one processor 52 is configured to execute the content generating method described in any of the foregoing embodiments based on instructions stored in the at least one memory 51.

[0083] The content generation device disclosed herein selects target multimedia resources closely related to the target event by matching characteristics of a certain dimension of the multimedia resource with the target event, and generates recommended content for the target multimedia resource. This method allows for rapid recommendation of multimedia resources associated with the target event, particularly for target events with short popularity cycles. This allows for more accurate matching of users' current browsing preferences and improves the click-through rate of the target multimedia resource.

[0084] Memory 51 is used to store one or more computer-readable instructions. Memory 51 may include any combination of various forms of computer-readable storage media, such as volatile and / or non-volatile memory, including but not limited to random access memory (RAM), dynamic random access memory (DRAM), static random access memory (SRAM), read-only memory (ROM), and flash memory. Memory 51 may store, for example, an operating system, applications, a boot loader, databases, and other programs, as well as various applications and data.

[0085] The processor 52 is configured to execute computer-readable instructions to implement the method described in any of the aforementioned embodiments. Detailed implementations of the various steps of the method can be found in the aforementioned embodiments, and repetitive details are omitted here.

[0086] The processor 52 may be configured to execute Figure 1 The processor 52 may be embodied as various processing devices, such as a central processing unit (CPU), a network processor (NP), etc.; it may also be 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. The central processing unit (CPU) may be of X86 or ARM architecture, etc.

[0087] The processor 52 and the memory 51 can communicate with each other directly or indirectly. For example, the processor 52 and the memory 51 can communicate via a network. The network can include a wireless network, a wired network, and / or any combination of wireless and wired networks. The processor 52 and the memory 51 can also communicate with each other via a system bus, which is not limited in this disclosure.

[0088] It should be noted that Figure 5 The components of the content generating device 5 shown are merely exemplary and non-limiting. The content generating device 5 may also have other components according to actual application requirements. The processor 52 may control the other components in the content generating device 5 to perform desired functions.

[0089] The content generating device 5 may be implemented by software, firmware and / or hardware, and may be integrated into a device installed with relevant application programs.

[0090] Figure 6 A block diagram of an electronic device according to some other embodiments of the present disclosure is shown.

[0091] Figure 6The electronic device 6 shown may be a computer system with a dedicated hardware structure, which can execute corresponding functions when a relevant application program is installed.

[0092] Electronic devices include, but are not limited to, mobile terminals such as smartphones, laptops, personal digital assistants (PDAs), tablet personal computers (Tablet PCs), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), wearable devices, etc., as well as fixed terminals such as digital televisions and desktop computers, etc.

[0093] like Figure 6 As shown, the central processing unit (CPU) 61 executes various processes according to the program stored in the read-only memory (ROM) 62 or the program loaded from the storage unit 68 to the random access memory (RAM) 63. In the RAM 63, data required when the CPU 61 executes various processes is stored as needed. The central processing unit is merely an example, and it may also be other types of processors, such as the various processors described above. The ROM 62, RAM 63 and storage unit 68 may be various forms of computer-readable storage media. It should be noted that although Figure 6 ROM 62, RAM 63 and storage portion 68 are shown separately in FIG, but one or more of them may be combined or located in the same or different memory or storage modules.

[0094] The CPU 61, the ROM 62, and the RAM 63 are connected to one another via a bus 64. To the bus 64, an input / output interface 65 is also connected.

[0095] The following components are connected to the input / output interface 65: an input section 66 such as a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output section 67 including a display such as a cathode ray tube (CRT), a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage section 68 including a hard disk, a magnetic tape, etc.; and a communication section 69 including a network interface card such as a LAN card, a modem, etc. The communication section 69 allows communication processing to be performed via a network such as the Internet. It is easy to understand that although Figure 6 Some of the electronic devices 6 are shown to communicate via a bus 64, but they may also communicate via a network or other means, where the network may include a wireless network, a wired network, and / or any combination of wireless networks and wired networks.

[0096] A drive 610 is also connected to the input / output interface 65 as needed. A removable medium 611 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc. is mounted on the drive 610 as needed so that a computer program read therefrom is installed in the storage section 68 as needed.

[0097] When the series of processing described above is implemented by software, the program constituting the software can be installed from a network such as the Internet or a storage medium such as the removable medium 611 .

[0098] According to embodiments of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, some embodiments of the present disclosure include a computer program product that, when executed on a computer, causes the computer to implement the method described in any of the aforementioned embodiments.

[0099] The computer program product disclosed herein selects target multimedia resources closely related to the target event by matching characteristics of a certain dimension of the target event and multimedia resources, and generates recommendations for the target multimedia resources. This method allows for rapid recommendation of multimedia resources associated with the target event, particularly for target events with short popularity cycles. This allows for more accurate matching of users' current browsing preferences and improves the click-through rate of the target multimedia resources.

[0100] The computer program product includes computer instructions carried on a computer-readable medium, including program code for executing the method shown in the flowchart. In such an embodiment, the computer instructions can be downloaded and installed from a network via the communication unit 69, or installed from the storage unit 68, or installed from the ROM 62. When the computer program is executed by the CPU 61, the method of the embodiment of the present disclosure is performed.

[0101] It should be noted that, in the context of the present disclosure, a computer-readable medium may be a tangible medium that may contain or store a program for use by an instruction execution system, apparatus, or device or for use in conjunction with an instruction execution system, apparatus, or device.

[0102] The computer readable medium may be a computer readable storage medium, or a computer readable signal medium, or any combination of the two.

[0103] Computer-readable storage media include, but are not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or components, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device or device. Computer instructions are stored on the computer-readable storage medium, and when the instructions are executed by the processor, the method described in any of the aforementioned embodiments is implemented.

[0104] The computer-readable storage medium disclosed herein matches characteristics of a target event and multimedia resources along a certain dimension, selects target multimedia resources closely related to the target event, and generates recommendations for the target multimedia resources. This approach allows for rapid recommendation of multimedia resources associated with the target event, particularly for target events with short popularity cycles. This more accurately meets users' current browsing preferences and increases the click-through rate of the target multimedia resources.

[0105] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries a computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination thereof.

[0106] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0107] In some embodiments, a computer program is further provided, comprising: instructions, which, when executed by a processor, cause the processor to perform the method described in any of the aforementioned embodiments. For example, the instructions may be embodied as computer program codes.

[0108] In embodiments of the present disclosure, computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or combinations thereof, including but not limited to object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0109] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the module, program segment, or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of the boxes in the block diagram and / or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0110] The functions described above may be performed at least in part by one or more hardware logic components. For example, and without limitation, exemplary hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0111] Although some specific embodiments of the present disclosure have been described in detail by way of examples, those skilled in the art will appreciate that the above examples are for illustrative purposes only and are not intended to limit the scope of the present disclosure. Those skilled in the art will appreciate that modifications may be made to the above embodiments without departing from the scope and spirit of the present disclosure. The scope of the present disclosure is defined by the appended claims.

Claims

1. A content generation method, comprising: Identify key information corresponding to the target event; Selecting, from the resource library, a candidate multimedia resource associated with the target event as the target multimedia resource based on a correlation between a feature of at least one dimension of the candidate multimedia resource in the resource library and the key information; Generate recommended content of the target multimedia resource according to the key information and information of the target multimedia resource.

2. The content generation method according to claim 1, wherein: The method further comprises: selecting, from the resource library, a candidate multimedia resource associated with the target event as the target multimedia resource based on a correlation between a feature of at least one dimension of the candidate multimedia resource and the key information; In response to the key information including the event process of the target event, acquiring plot dimension features of the candidate multimedia resource as features of the at least one dimension of the candidate multimedia resource; The target multimedia resource is selected from the resource library according to the correlation between the event process of the target event and the plot dimension features of the candidate multimedia resources.

3. The content generation method according to claim 1, wherein: The method further comprises: selecting, from the resource library, a candidate multimedia resource associated with the target event as the target multimedia resource based on a correlation between a feature of at least one dimension of the candidate multimedia resource and the key information; In response to the key information including the theme of the target event, obtaining background dimension features of the candidate multimedia resource as features of the at least one dimension of the candidate multimedia resource; The target multimedia resource is selected from the resource library according to the correlation between the theme of the target event and the background dimension features of the candidate multimedia resources.

4. The content generation method according to claim 1, wherein: The method further comprises: selecting, from the resource library, a candidate multimedia resource associated with the target event as the target multimedia resource based on a correlation between a feature of at least one dimension of the candidate multimedia resource and the key information; In response to the key information including the character information of the target event, obtaining a character dimension feature of the candidate multimedia resource as a feature of the at least one dimension of the candidate multimedia resource; The target multimedia resource is selected from the resource library according to the correlation between the character information of the target event and the role dimension characteristics of the candidate multimedia resources.

5. The content generation method according to claim 1, wherein: The method further comprises: selecting, from the resource library, a candidate multimedia resource associated with the target event as the target multimedia resource based on a correlation between a feature of at least one dimension of the candidate multimedia resource and the key information; Extracting features of multiple dimensions of the multimedia resource; splicing the features of the multiple dimensions; The correlation is determined based on the semantic similarity between the spliced ​​features of the multiple dimensions and the key information.

6. The content generation method according to claim 1, wherein: The information of the target multimedia resource includes the name and introduction of the target multimedia resource. Generating recommended content of the target multimedia resource according to the key information and the information of the target multimedia resource includes: Generate prompt information according to the key information and the name and introduction of the target multimedia resource; The recommended content is generated based on the prompt information using a machine learning model.

7. The content generation method according to claim 1, further comprising: Obtain user feedback on candidate events; Calculating the score of the candidate event according to the feedback information; The target event is selected from the candidate events according to the scores of the candidate events.

8. The content generation method according to claim 1, wherein: Identify key information corresponding to the target event, including: Identifying key entities from the target event information, wherein the key entities include at least one of person, time, place, and event process; According to the key entity, a text with a length not exceeding a threshold is generated as the key information.

9. The content generation method according to claim 1, further comprising: categorizing the recommended content; Configure the validity period of each type of recommended content; Calculating a score for the recommended content based on user feedback information on a target event corresponding to the recommended content; Whether to push the recommended content and the time to push the recommended content are determined according to the effective duration and score of the recommended content.

10. A content generation device, comprising: a determination module configured to determine key information corresponding to a target event; A selection module is configured to select, from the resource library, a candidate multimedia resource associated with the target event as a target multimedia resource based on a correlation between a feature of at least one dimension of the candidate multimedia resource and the key information; The generating module is configured to generate recommended content of the target multimedia resource according to the key information and the information of the target multimedia resource.

11. A content generation device, comprising: Memory; as well as A processor coupled to the memory, wherein the processor is configured to execute the content generating method according to any one of claims 1 to 9 based on instructions stored in the memory.

12. A computer-readable storage medium having computer program instructions stored thereon, wherein when the instructions are executed by a processor, the content generation method according to any one of claims 1 to 9 is implemented.

13. A computer program product comprising computer program instructions, wherein when the computer program instructions are executed by a processor, the content generation method according to any one of claims 1 to 9 is implemented.

Images