Conversion method and device for hotspot information, equipment and storage medium

Abstract

The invention provides a hotspot information conversion method and device, equipment and a storage medium, and the method comprises the steps: obtaining hotspot information and corresponding hotspot videos from a plurality of target platforms at preset time intervals, extracting hotspot copywriting in the hotspot videos, constructing a hotspot vector based on the hotspot copywriting and the hotspot information, and carrying out the conversion of the hotspot information according to the hotspot vector. The hotspot vector is updated to a Milvus vector database; then, obtaining a user portrait in a community and a corresponding user video, extracting a user document in the user video, and constructing a user feature vector based on the user document and the user portrait; then, a plurality of hot spot vectors similar to the user characteristic vector are selected from the Milvus vector database, and the hot spot vectors are sorted and then output; and finally, obtaining a selected hot spot vector, and generating a creation document associated with the user characteristics based on the hot spot vector. The problem that a self-media user is difficult to identify quickly and capture and convert hotspot information is solved.

Description

Technical Field

[0001] The present invention relates to the field of hot information conversion, and particularly relates to a method, device, equipment and storage medium for converting hot information. Background Art

[0002] Currently, the mining of network hotspots mainly relies on keyword extraction, clustering analysis and statistical model processing of massive social media, news and video data. Although these methods can capture hot information within a certain range, their main defect is that they cannot deeply understand the core semantics behind multimedia content such as videos, and only stay at the matching of surface data, resulting in inaccurate extraction of hot information. At the same time, they cannot reflect the dynamic changes of hot content and the personalized needs of users in a timely manner.

[0003] In practical applications, for some self-media users, they cannot quickly identify and capture the heat of hot information, which in turn affects their digestion of hot information.

[0004] In view of this, the present application is proposed. Summary of the Invention

[0005] The present invention discloses a method, device, equipment and storage medium for converting hot information, aiming to solve the problem that self-media users are difficult to quickly identify and capture and convert hot information.

[0006] The first embodiment of the present invention provides a method for converting hot information, including:

[0007] Obtain hot information and its corresponding hot videos from multiple target platforms at preset time intervals, extract the hot copywriting in the hot videos, construct hot vectors based on the hot copywriting and the hot information, and update the hot vectors to the Milvus vector database;

[0008] Obtain the user portraits and their corresponding user videos in the community, extract the user copywriting in the user videos, and construct user feature vectors based on the user copywriting and the user portraits;

[0009] Select multiple hot vectors similar to the user feature vectors from the Milvus vector database, and output the multiple hot vectors after sorting;

[0010] Obtain the selected hot vectors, and generate creative copywriting associated with the user features based on the hot vectors.

[0011] Preferably, the extracting the hot copywriting in the hot videos is specifically:

[0012] Separate the audio signal from the hot video, and use a convolutional neural network to extract acoustic features from the audio signal to generate a high-dimensional representation of the input audio;

[0013] Call the Transformer encoder to map the high-dimensional representation of the input audio into a high-dimensional semantic representation;

[0014] Decode the high-dimensional semantic representation, and during the decoding process, use the conditional independence assumption to generate target text symbols in parallel;

[0015] Call the length prediction module to predict the high-dimensional semantic representation to generate the target text length, integrate based on the target text symbols and the target text length, and generate a hot copywriting for reflecting the core content of the hot video based on the integration result.

[0016] Preferably, the sorting of the hot vectors is specifically as follows:

[0017] Obtain the number of times users in the community select the hot vectors in the Milvus vector database, and sort multiple hot vectors similar to the user feature vectors according to the number of selections.

[0018] Preferably, the user portrait includes personal tags and behavioral data.

[0019] Preferably, the hot information includes topic tags and interaction data.

[0020] Preferably, it further includes:

[0021] Construct a classification vector from the creator portrait of the hot video, the heat value of the hot information and its change trend;

[0022] Merge the hot classification vector with the hot vector and update it to the Milvus vector database.

[0023] Preferably, after calling the length prediction module to predict the high-dimensional semantic representation to generate the target text length, it further includes:

[0024] When it is determined that the target text length value is lower than the preset value, reselect other hot videos under the current hot topic.

[0025] The second embodiment of the present invention provides a conversion device for hot information, including:

[0026] A hot spot extraction unit, configured to obtain hot information and its corresponding hot videos from multiple target platforms at preset time intervals, extract the hot copywriting in the hot videos, construct hot vectors based on the hot copywriting and the hot information, and update the hot vectors to the Milvus vector database;

[0027] A user profile unit, configured to obtain user profiles within a community and their corresponding user videos, extract user copywriting in the user videos, and construct user feature vectors based on the user copywriting and the user profiles;

[0028] A vector sorting unit, configured to select multiple hot vectors similar to the user feature vectors from the Milvus vector database, and output the multiple hot vectors after sorting;

[0029] A copywriting generation unit, configured to obtain the selected hot vectors, and generate creative copywriting associated with the user features based on the hot vectors.

[0030] A third embodiment of the present invention provides a conversion device for hot information, including a memory and a processor. A computer program is stored in the memory and can be executed by the processor to implement a method for converting hot information as described in any one of the above.

[0031] A fourth embodiment of the present invention provides a computer-readable storage medium storing a computer program, which can be executed by a processor of a device where the computer-readable storage medium is located to implement a method for converting hot information as described in any one of the above.

[0032] Based on a method, device, equipment, and storage medium for converting hot information provided by the present invention, by obtaining hot information and its corresponding hot videos from multiple target platforms at preset time intervals, extracting hot copywriting in the hot videos, constructing hot vectors based on the hot copywriting and the hot information, and updating the hot vectors to the Milvus vector database; then, obtaining user profiles within a community and their corresponding user videos, extracting user copywriting in the user videos, and constructing user feature vectors based on the user copywriting and the user profiles; then, selecting multiple hot vectors similar to the user feature vectors from the Milvus vector database, and outputting the multiple hot vectors after sorting; finally, obtaining the selected hot vectors, and generating creative copywriting associated with the user features based on the hot vectors. The problem that it is difficult for self-media users to quickly identify and capture and convert hot information is solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 is a flowchart of a method for converting hot information provided by the first embodiment of the present invention;

[0034] Figure 2 is a module diagram of a device for converting hot information provided by the second embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0035] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0036] For a better understanding of the technical solutions of the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0037] It should be clear that the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0038] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms "a", "the" and "said" used in the embodiments of the present invention and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise.

[0039] It should be understood that the term "and / or" used herein is only a kind of association relationship describing associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " herein generally represents an "or" relationship between the front and rear associated objects.

[0040] Depending on the context, the word "if" as used herein can be interpreted as "when" or "while" or "in response to determining" or "in response to detecting". Similarly, depending on the context, the phrase "if determined" or "if detected (stated condition or event)" can be interpreted as "when determined" or "in response to determining" or "when detected (stated condition or event)" or "in response to detecting (stated condition or event)".

[0041] The "first / second" mentioned in the embodiments is only to distinguish similar objects, and does not represent a specific order for the objects. It can be understood that the "first / second" can be interchanged in a specific order or sequence when allowed. It should be understood that the objects distinguished by the "first / second" can be interchanged under appropriate circumstances so that the embodiments described herein can be implemented in an order other than those illustrated or described herein.

[0042] The following will give a detailed description of the specific embodiments of the present invention with reference to the accompanying drawings.

[0043] The present invention discloses a method, device, equipment and storage medium for converting hot information, aiming to solve the problem that it is difficult for self-media users to quickly identify and capture and convert hot information.

[0044] Please refer to Figure 1 , a first embodiment of the present invention provides a method for converting hot information, which can be executed by a device for converting hot information (hereinafter referred to as the conversion device), and particularly, by one or more processors in the conversion device, to at least implement the following steps:

[0045] S101, obtain hot information and its corresponding hot video from multiple target platforms at preset time intervals, extract the hot copywriting in the hot video, construct a hot vector based on the hot copywriting and the hot information, and update the hot vector to the Milvus vector database, where the hot information includes topic tags and interaction data;

[0046] In this embodiment, the conversion device may be a terminal with data processing capabilities such as a server, a workstation, a desktop computer, a laptop computer, etc. The conversion device may be installed with corresponding operating systems and application software, and the functions required in this embodiment are realized through the combination of the operating system and the application software.

[0047] In this embodiment, the conversion device automatically collects hot information and corresponding hot videos from multiple target platforms at preset time intervals. The collection process performs real-time data interaction through network interfaces with major social platforms and news portals to obtain current popular topic tags and interaction data, and at the same time identifies video resources associated with the hot information. The collected videos contain a large amount of voice information;

[0048] To efficiently extract the hot copywriting in the video, first, the audio signal is separated from the hot video, and the convolutional neural network is used to extract the acoustic features of the audio signal, thereby generating a high-dimensional input audio representation. Among them, the CNN can extract multi-scale and multi-level feature representations from complex audio signals through multiple convolutional and pooling operations, such as subtle features like phoneme boundaries, speech rhythms, and pitch changes, and then form a high-dimensional and information-rich audio representation vector. Next, this high-dimensional representation is fed into the Transformer encoder and mapped into a high-dimensional semantic representation through multiple self-attention mechanisms, adopting a parallel encoding mechanism so that the conversion device can simultaneously capture the global and local semantic associations in the audio signal. Through the self-attention mechanism, different semantic weights are assigned to each audio segment, making the subsequent text generation more accurate, and the context is correlated. Subsequently, when decoding the high-dimensional semantic representation, the conversion device adopts the conditional independence assumption to achieve parallel generation of target text symbols. This parallel decoding method significantly improves the processing speed and ensures that the generated symbols do not affect each other during the decoding process, improving the overall recognition accuracy. To ensure the correspondence between the generated text and the actual content of the hot video, in a possible implementation of this embodiment, a length prediction module is also called, which uses the high-dimensional semantic representation to predict the length of the target text and provides accurate length information for subsequent text integration. After obtaining the target text symbols and the predicted text length, the conversion device efficiently integrates the two to generate hot copywriting that can comprehensively reflect the core content of the hot video.

[0049] After obtaining the hot copywriting, the conversion device fuses the copywriting with the collected hot information (where the hot information mainly includes topic tags and interaction data), and converts the text and structured hot information into a hot vector that comprehensively expresses the characteristics of the hot event through a pre-designed vector construction model. During the construction process, the topic tags, as keywords, provide semantic guidance for the hot event, while the interaction data reflects the actual degree of user attention through popularity weights. The fused hot vector can not only accurately express the core semantics of the hot content but also reflect the dissemination heat and dynamic trend of the hot event on the network. The conversion device updates the constructed hot vector to the Milvus vector database in real time.

[0050] S102, obtain the user portrait and the corresponding user video in the community, extract the user copywriting in the user video, and construct a user feature vector based on the user copywriting and the user portrait, where the user portrait includes personal tags and behavior data;

[0051] In this embodiment, the conversion device first obtains user portraits from the community. These user portraits integrate user personal tags (such as age, gender, interest preferences, etc.) and user behavior data (such as the types of videos posted by the user on the platform, browsing history, interaction frequency, etc.), thus forming a multi-dimensional set of user basic information. Subsequently, the conversion device obtains user videos corresponding to the above user portraits. These videos contain the copywriting of these self-media users during the creation process. The conversion device uses technical means similar to those for extracting the copywriting of hot videos to obtain the user copywriting, which will not be elaborated here;

[0052] These copywritings not only accurately reflect the thoughts and emotions expressed by the user in the video but also capture the user's potential interest points. Next, the conversion device deeply integrates the extracted user copywriting with the original user portrait. Through natural language processing technology and vector encoding methods, the semantic information in the user copywriting is combined with the structured data (personal tags and behavior data) in the user portrait to construct a high-dimensional user feature vector. It can fully express the user's personalized characteristics and interest tendencies, providing data support for subsequent personalized recommendations and content matching.

[0053] S103, Select multiple hot vectors similar to the user feature vector from the Milvus vector database, and output the multiple hot vectors after sorting;

[0054] In this embodiment, the conversion device first extracts multiple hot vectors similar to it based on the user feature vector from the Milvus vector database. It can use vector retrieval technology to quickly select a hot candidate set that best matches the user's interests by calculating the similarity between the user feature vector and each hot vector in the database.

[0055] To further improve the accuracy of the recommendation results, the conversion device records the actual selection times of each hot vector by users in the community as a key indicator to measure the popularity of the hot topics and the real preferences of users. It should be noted that the core of the sorting mechanism is the introduction of the dynamic weight of "user selection times". It realizes intelligent recommendation based on the actual behavior of users. By continuously tracking and analyzing the selection frequency of users for specific hot vectors in the community, a continuously adaptive and real-time updated recommendation model is constructed. Each user interaction becomes a signal to optimize the recommendation strategy, enabling the recommendation results to follow the subtle changes in user interests. The sorting method based on user selection times is essentially a feedback closed-loop system. It not only passively matches user characteristics but also actively learns and predicts user preferences.

[0056] S104, Obtain the selected hot vector, and generate a creative copywriting associated with the user characteristics based on the hot vector.

[0057] In this embodiment, the conversion device adopts a multi-dimensional semantic matching strategy. Specifically, the essential features and core values of the hot topic are analyzed through the hot topic vector, and the key information context is extracted from it. Subsequently, a dynamic creation template is constructed by combining the subtle features contained in the user feature vector, such as personal style, creative preferences, language habits, etc. It should be understood that this template is not fixed, but an intelligent generation framework that can be adjusted in real time according to different users and different hot topics.

[0058] Specifically, the prompt engineering in the intelligent generation framework can guide the large language model to inject user personalization elements while maintaining the integrity of the hot topic. For example, for users who are good at humorous creation, the prompt will guide the model to generate more relaxed and witty copywriting; for rigorous professional creators, it will tend to generate more formal and in-depth content expressions.

[0059] In a possible implementation embodiment of the present invention, it further includes:

[0060] Construct a classification vector from the creator portrait of the hot topic video, the popularity value of the hot topic information, and its change trend;

[0061] Merge the hot topic classification vector with the hot topic vector and update it to the Milvus vector database.

[0062] It should be noted that the conversion device extracts relevant information of the creator from the hot topic video. This information covers data such as the basic portrait of the creator, content style, historical release records, and audience feedback. These information are encoded through multi-dimensional feature extraction technology to form a creator portrait vector. At the same time, the conversion device also records not only the popularity value but also its change trend (such as complex dynamic features such as growth rate, peak duration, decay curve, etc.) in the hot topic information. Through time series analysis and popularity weight calculation, the dynamic change characteristics of the hot topic information are quantified into numerical features and encoded into a popularity trend vector. Subsequently, the conversion device fuses the creator portrait vector and the popularity trend vector to construct a classification vector that comprehensively reflects the overall characteristics and dynamic changes of the hot topic video. This classification vector not only reflects the core semantics of the hot topic content, but also reflects the creator's style and the evolution trend of the hot topic during the dissemination process. To achieve efficient vector retrieval and intelligent matching, the conversion device merges the constructed hot topic classification vector with the original hot topic vector to form a composite vector with richer semantics and dynamic adaptability. Finally, the fused composite vector is updated to the Milvus vector database in real time.

[0063] In a possible implementation of the present invention, after calling the length prediction module to predict the high-dimensional semantic representation to generate the target text length, the following steps are further included:

[0064] When it is determined that the target text length value is lower than the preset value, other hot videos under the current hot topic are reselected.

[0065] It should be noted that mainly considering that some hot videos may have the problem of scarce voice or dialogue content. For example, videos like street piano performances lack sufficient text information during the process, making the generated hot copywriting may not fully reflect the core content of the video. Therefore, after calling the length prediction module to predict the high-dimensional semantic representation to generate the target text length, the conversion device will judge the prediction result. If it is detected that the generated target text length is lower than the preset threshold, it is considered that the information contained in the current video is not sufficient to form an effective hot copywriting. At this time, the re-selection process will be automatically triggered to select more appropriate materials from other videos under the current hot topic.

[0066] Please refer to Figure 2 , the second embodiment of the present invention provides a conversion device for hot information, including:

[0067] A hot topic extraction unit 201, configured to obtain hot topic information and its corresponding hot videos from multiple target platforms at preset time intervals, extract the hot copywriting in the hot videos, construct a hot topic vector based on the hot copywriting and the hot topic information, and update the hot topic vector to the Milvus vector database;

[0068] A user portrait unit 202, configured to obtain the user portrait and its corresponding user videos in the community, extract the user copywriting in the user videos, and construct a user feature vector based on the user copywriting and the user portrait;

[0069] A vector sorting unit 203, configured to select multiple hot topic vectors similar to the user feature vector from the Milvus vector database, and output the multiple hot topic vectors after sorting;

[0070] A copywriting generation unit 204, configured to obtain the selected hot topic vector and generate a creative copywriting associated with the user feature based on the hot topic vector.

[0071] The third embodiment of the present invention provides a conversion device for hot information, including a memory and a processor. The memory stores a computer program, and the computer program can be executed by the processor to implement a conversion method for hot information as described in any one of the above.

[0072] The fourth embodiment of the present invention provides a computer-readable storage medium storing a computer program, which can be executed by a processor of a device where the computer-readable storage medium is located to implement the method for converting hot-spot information as described in any one of the above.

[0073] Based on the method, device, equipment and storage medium for converting hot-spot information provided by the present invention, hot-spot information and its corresponding hot-spot videos are obtained from multiple target platforms at preset intervals, hot-spot copywriting in the hot-spot videos is extracted, hot-spot vectors are constructed based on the hot-spot copywriting and the hot-spot information, and the hot-spot vectors are updated to the Milvus vector database; then, user portraits and their corresponding user videos in the community are obtained, user copywriting in the user videos is extracted, and user feature vectors are constructed based on the user copywriting and the user portraits; then, multiple hot-spot vectors similar to the user feature vectors are selected from the Milvus vector database, and the multiple hot-spot vectors are sorted and output; finally, the selected hot-spot vectors are obtained, and creative copywriting associated with the user features is generated based on the hot-spot vectors. The problem that self-media users are difficult to quickly identify and capture and convert hot-spot information is solved.

[0074] Exemplarily, the computer program described in the third and fourth embodiments of the present invention can be divided into one or more modules, and the one or more modules are stored in the memory and executed by the processor to complete the present invention. The one or more modules can be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program in the device for implementing the method for converting hot-spot information. For example, the device described in the second embodiment of the present invention.

[0075] The so-called processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), off-the-shelf programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The processor is the control center of the method for converting hot-spot information, and connects all parts of the device for implementing the method for converting hot-spot information through various interfaces and lines.

[0076] The memory can be used to store the computer program and / or module. By running or executing the computer program and / or module stored in the memory, and invoking the data stored in the memory, the processor can implement various functions of a method for converting hotspot information. The memory may mainly include a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function (such as a sound playback function, a text conversion function, etc.); the data storage area can store data created according to the use of the mobile phone (such as audio data, text message data, etc.). In addition, the memory can include high-speed random access memory, and can also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0077] Among them, if the implemented module is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, to implement all or part of the processes in the above-mentioned embodiment methods of the present invention, it can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, an executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device that can carry the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disc, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0078] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, in the attached drawings of the device embodiments provided by the present invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0079] As mentioned above, the above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for converting hot information, characterized in that Including: Fetching hot information and its corresponding hot videos from multiple target platforms at preset time intervals, extracting hot copywriting from the hot videos, constructing hot vectors based on the hot copywriting and the hot information, and updating the hot vectors to the Milvus vector database; Obtaining user portraits and their corresponding user videos in the community, extracting user copywriting from the user videos, and constructing user feature vectors based on the user copywriting and the user portraits; Selecting multiple hot vectors similar to the user feature vectors from the Milvus vector database, and sorting and outputting the multiple hot vectors; Obtaining the selected hot vectors, and generating creative copywriting associated with the user features based on the hot vectors.

2. The conversion method for hot-spot information according to claim 1, wherein, The specific operation of extracting the hot copywriting from the hot videos is as follows: Separating the audio signal from the hot video, and using a convolutional neural network to extract acoustic features from the audio signal to generate a high-dimensional representation of the input audio; Invoking a Transformer encoder to map the high-dimensional representation of the input audio to a high-dimensional semantic representation; Decoding the high-dimensional semantic representation, where conditional independent assumptions are used during the decoding process to generate target text symbols in parallel; Invoking a length prediction module to predict the high-dimensional semantic representation to generate a target text length, integrating based on the target text symbols and the target text length, and generating hot copywriting reflecting the core content of the hot video based on the integration result.

3. The conversion method of hot-spot information according to claim 1, characterized in that The specific sorting of the hot vectors is as follows: Obtaining the number of times the users in the community select the hot vectors in the Milvus vector database, and sorting the multiple hot vectors similar to the user feature vectors according to the number of selections.

4. A method for converting hot information according to claim 1, characterized in that The user portrait includes personal tags and behavior data.

5. A method for converting hot-spot information according to claim 1, characterized in that, The hot information includes topic tags and interaction data.

6. The conversion method for hotspot information according to claim 1, wherein Also including: Constructing a classification vector from the creator portrait of the hot video, as well as the heat value of the hot information and its change trend; Merging the hot classification vector with the hot vector and updating it to the Milvus vector database.

7. The conversion method for hotspot information according to claim 2, wherein After invoking the length prediction module to predict the high-dimensional semantic representation to generate a target text length, it also includes: When it is determined that the target text length value is lower than the preset value, reselecting other hot videos under the current hot topic.

8. A conversion device for hot information, characterized in that, Including: A hot extraction unit for fetching hot information and its corresponding hot videos from multiple target platforms at preset time intervals, extracting hot copywriting from the hot videos, constructing hot vectors based on the hot copywriting and the hot information, and updating the hot vectors to the Milvus vector database; A user portrait unit for obtaining user portraits and their corresponding user videos in the community, extracting user copywriting from the user videos, and constructing user feature vectors based on the user copywriting and the user portraits; A vector sorting unit for selecting multiple hot vectors similar to the user feature vectors from the Milvus vector database, and sorting and outputting the multiple hot vectors; A copywriting generation unit for obtaining a selected hot vector and generating a creative copy associated with the user characteristics based on the hot vector.

9. A conversion device for hot information, characterized in that, It includes a memory and a processor. A computer program is stored in the memory and can be executed by the processor to implement a method for converting hot information as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, A computer program is stored and can be executed by the processor of the device where the computer-readable storage medium is located to implement a method for converting hot information as described in any one of claims 1 to 7.

Images