Knowledge graph processing method and apparatus

Abstract

Embodiments of the present application provide a knowledge graph processing method and device, wherein the knowledge graph processing method comprises: obtaining one or more videos watched by a user and feature information of the user; determining a viewing time of each video in the one or more videos watched by the user, and taking one or more videos with a viewing time greater than or equal to a preset time threshold as initial videos; determining one or more target videos corresponding to the user according to attribute information of the one or more initial videos and the feature information of the user; extracting one or more video tags of the one or more target videos, and establishing or updating a personal knowledge graph for the user based on the one or more video tags.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a knowledge graph processing method. One or more embodiments of this application also relate to a knowledge graph processing apparatus, a computing device, and a computer-readable storage medium. Background Technology

[0002] With the rapid development of video platforms, the way of learning knowledge based on video platforms has gradually gained attention and use. In order to facilitate different types of users to learn based on videos, video platforms will collect a large number of different types of learning videos. As the types and number of learning videos increase, users need to make adaptive adjustments to their learning plans and learning time based on their historical learning records. However, at present, users need to spend a lot of time and energy to view and analyze their learning content in a huge amount of historical learning videos, resulting in low learning efficiency and a poor user experience when using video platforms for knowledge learning. Summary of the Invention

[0003] In view of this, embodiments of this application provide a knowledge graph processing method. One or more embodiments of this application also relate to a knowledge graph processing apparatus, a computing device, and a computer-readable storage medium, to address the technical deficiency in the prior art where users cannot quickly and accurately understand the content they are learning by watching historical videos.

[0004] According to a first aspect of the embodiments of this application, a knowledge graph processing method is provided, including:

[0005] Obtain one or more videos watched by the user and the user's characteristic information;

[0006] Determine the viewing time of each of the one or more videos watched by the user, and select one or more videos whose viewing time is greater than or equal to a preset time threshold as the initial videos;

[0007] Based on the attribute information of the one or more initial videos and the characteristic information of the user, determine one or more target videos corresponding to the user;

[0008] Extract one or more video tags from the one or more target videos, and build or update a personal knowledge graph for the user based on the one or more video tags.

[0009] According to a second aspect of the embodiments of this application, a knowledge graph processing apparatus is provided, comprising:

[0010] The acquisition module is configured to acquire one or more videos viewed by the user and the user's characteristic information;

[0011] The initial video determination module is configured to determine the viewing time of each of the one or more videos watched by the user, and to select one or more videos whose viewing time is greater than or equal to a preset time threshold as the initial videos;

[0012] The target video determination module is configured to determine one or more target videos corresponding to the user based on the attribute information of the one or more initial videos and the user's feature information;

[0013] The knowledge graph processing module is configured to extract one or more video tags from the one or more target videos, and to build or update a personal knowledge graph for the user based on the one or more video tags.

[0014] According to a third aspect of the embodiments of this application, a computing device is provided, comprising:

[0015] Memory and processor;

[0016] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions, which, when executed by the processor, implement the steps of the knowledge graph processing method.

[0017] According to a fourth aspect of the embodiments of this application, a computer-readable storage medium is provided that stores computer-executable instructions, which, when executed by a processor, implement the steps of the knowledge graph processing method.

[0018] One embodiment of this application implements a knowledge graph processing method and apparatus. The knowledge graph processing method includes: acquiring one or more videos watched by a user and the user's feature information; determining the viewing time of each video watched by the user, and selecting one or more videos with a viewing time greater than or equal to a preset time threshold as initial videos; determining one or more target videos corresponding to the user based on the attribute information of the one or more initial videos and the user's feature information; extracting one or more video tags from the one or more target videos, and establishing or updating a personal knowledge graph for the user based on the one or more video tags. Specifically, after acquiring the videos watched by the user, the knowledge graph processing method determines the final valid videos consumed by the user through the viewing time of the videos watched by the user, the attribute information of the videos, and the user's feature information, and then creates or updates the user's personal knowledge graph based on the video tags of the user's valid consumed videos. Subsequently, the user can directly and quickly understand their historical learning content based on this personal knowledge graph, so as to adjust their learning plan in a timely manner and enhance the user experience. Attached Figure Description

[0019] Figure 1 This is a flowchart of a knowledge graph processing method provided in one embodiment of this application;

[0020] Figure 2 This is a schematic diagram of the knowledge tag level in a knowledge graph processing method provided in one embodiment of this application;

[0021] Figure 3 This is a schematic diagram illustrating the display of recommended videos in the form of a Rubik's Cube in a knowledge graph processing method provided in one embodiment of this application;

[0022] Figure 4 This is a schematic diagram illustrating the recommendation of videos to users in the form of a Rubik's Cube in a knowledge graph processing method provided in one embodiment of this application;

[0023] Figure 5 This is a schematic diagram of a user knowledge graph in a knowledge graph processing method provided in one embodiment of this application;

[0024] Figure 6 This is a schematic diagram of a knowledge consumption plaza displayed to video uploaders in a knowledge graph processing method provided in one embodiment of this application;

[0025] Figure 7 This is a schematic diagram of a knowledge consumption plaza displayed to a user in a knowledge graph processing method provided in one embodiment of this application;

[0026] Figure 8 This is a schematic diagram of a recommended video work in a knowledge graph processing method provided in one embodiment of this application;

[0027] Figure 9 This is a flowchart illustrating video recommendation in a knowledge graph processing method according to one embodiment of this application;

[0028] Figure 10 This is a flowchart illustrating the updating of a knowledge video tag graph according to an embodiment of the knowledge graph processing method provided in this application;

[0029] Figure 11 This is a schematic diagram of the structure of a knowledge graph processing device provided in one embodiment of this application;

[0030] Figure 12 This is a structural block diagram of a computing device provided in one embodiment of this application. Detailed Implementation

[0031] Many specific details are set forth in the following description to provide a full understanding of this application. However, this application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of this application; therefore, this application is not limited to the specific embodiments disclosed below.

[0032] The terminology used in one or more embodiments of this application is for the purpose of describing particular embodiments only and is not intended to limit the scope of one or more embodiments of this application. The singular forms “a,” “the,” and “the” used in one or more embodiments of this application and in the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” used in one or more embodiments of this application refers to and includes any or all possible combinations of one or more associated listed items.

[0033] It should be understood that although the terms first, second, etc., may be used to describe various information in one or more embodiments of this application, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, first may also be referred to as second without departing from the scope of one or more embodiments of this application, and similarly, second may also be referred to as first. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to a determination."

[0034] First, the terms and concepts involved in one or more embodiments of this application will be explained.

[0035] Consumption: refers to video playback. The concept of consumption used in the embodiments of this application can be considered as a user's act of consuming content by watching videos.

[0036] Learning knowledge: Learning subjects such as physics, geography, mathematics, language, or animal encyclopedias can increase one's knowledge.

[0037] Knowledge tags: Tags added to videos that teach knowledge can be called knowledge tags, such as physics, chemistry, Chinese language, dance, etc.

[0038] Knowledge graph: A visualization or mapping map of a knowledge domain, it is a series of different graphics that show the development process and structural relationships of knowledge.

[0039] Python is a cross-platform computer programming language. It is a high-level scripting language that combines interpreted, compiled, interactive, and object-oriented features.

[0040] Currently, when users watch videos on other video platforms for knowledge learning, their viewing history is fragmented and weakly presented, primarily through lists. While users can understand the video content they've watched through the history and course outlines in the lists, the breadth and depth of their knowledge learning cannot be displayed. Therefore, the knowledge graph processing method described in this application builds a personal knowledge graph for each user based on their completed video consumption information. Users can then clearly understand the knowledge points and depth of their learning through their personal knowledge graph, improving the user experience. This application provides a knowledge graph processing method. One or more embodiments of this application also relate to a knowledge graph processing apparatus, a computing device, and a computer-readable storage medium, which will be described in detail in the following embodiments.

[0041] See Figure 1 , Figure 1 A flowchart of a knowledge graph processing method according to an embodiment of this application is shown, which specifically includes the following steps.

[0042] Step 102: Obtain one or more videos watched by the user and the user's characteristic information.

[0043] The video includes, but is not limited to, any type of video, such as entertainment videos, educational videos, and news videos. In this embodiment, the knowledge graph processing method is described using educational videos as an example. The knowledge graph processing is applied to the server side of the video platform.

[0044] User characteristics include, but are not limited to, the user's name, gender, age, interests, and historical video viewing history.

[0045] Specifically, obtaining one or more videos watched by a user and the user's characteristic information can be understood as obtaining one or more educational videos watched by a user and characteristic information such as the user's name, gender, age, interests, and historical video viewing records.

[0046] Step 104: Determine the viewing time of each of the one or more videos watched by the user, and select one or more videos whose viewing time is greater than or equal to a preset time threshold as the initial videos.

[0047] The preset time threshold can be set according to the actual application, such as 2 minutes, 5 minutes or 10 minutes, etc., and this application does not impose any restrictions on it.

[0048] In practical applications, although we may obtain one or more learning videos that a user has watched, it's possible that the user was attracted by the video title and clicked to watch. However, after opening the video, they may find that the content is not what they want to learn, and they may close the video after watching for about 30 seconds, indicating that they are not interested in the learning content. In this case, if we were to build or update the user's personal knowledge graph based on the video tags, it would greatly increase the space occupied by the user's personal knowledge graph and reduce its accuracy.

[0049] Therefore, in order to avoid the above problems, the viewing time of each video will be obtained, and one or more videos with a viewing time greater than or equal to a preset time threshold will be used as the initial videos.

[0050] Taking a preset time threshold of 5 minutes as an example, after obtaining one or more videos watched by a user, if it is determined that the user watched a certain video for 2 minutes, then it can be determined that the user's viewing time for that video is less than the preset time threshold. In this case, it can be determined that the video is not a valid consumption video, and there is no need to do any subsequent processing based on the video, thus avoiding waste of resources. However, if it is determined that the user watched the video for 8 minutes, then it can be determined that the user's viewing time for that video is greater than the preset time threshold. In this case, it can be determined that the video is the initial video, that is, a valid consumption video.

[0051] Step 106: Based on the attribute information of the one or more initial videos and the feature information of the user, determine one or more target videos corresponding to the user.

[0052] Specifically, determining one or more target videos corresponding to the user based on the attribute information of the one or more initial videos and the user's feature information includes three implementation methods:

[0053] The first method involves determining one or more target videos corresponding to the user based on the attribute information of the one or more initial videos and the user's feature information, including:

[0054] Match the attribute information of each of the one or more initial videos with the user's feature information;

[0055] One or more initial videos whose attribute information matches the user's feature information are used as one or more target videos corresponding to the user.

[0056] The initial video's attribute information includes its type, description, and duration.

[0057] In practice, the attribute information of each initial video is matched with the user's feature information. If a match is found, the initial video is used as the target video, and all target videos are determined in this way.

[0058] For example, if the initial video's attribute information includes that the video is a period drama, and the user's historical video viewing history indicates that the user prefers period dramas, then the initial video's attribute information matches the user's attribute information. Or, if the initial video's attribute information includes that the video is a romance drama, targeting women aged 20-30, and the user's age and gender information indicates that the user is a 25-year-old woman, then the initial video's attribute information matches the user's attribute information.

[0059] In this embodiment, when the attribute information of the initial video matches the user's feature information, the initial video is used as the target video to enable subsequent updates to the user's personal knowledge graph based on the target video. The number of video views and duration of the corresponding node of the target video are updated on the user's personal knowledge graph. Through real-time updates to the user's personal knowledge graph, the accuracy of the user's personal knowledge graph is ensured.

[0060] The second method involves determining one or more target videos corresponding to the user based on the attribute information of the one or more initial videos and the user's feature information, including:

[0061] Match the attribute information of each of the one or more initial videos with the user's feature information;

[0062] One or more initial videos whose attribute information does not match the user's feature information are selected as one or more target videos corresponding to the user.

[0063] In practice, the attribute information of each initial video is matched with the user's feature information. If they do not match, the initial video can be used as the target video. In this way, all target videos can be determined.

[0064] For example, if the initial video's attribute information includes that the video is a period drama, but the user's historical video viewing history indicates that the user prefers rural dramas, then the initial video's attribute information does not match the user's attribute information. Or, if the initial video's attribute information includes that the video is a romance drama and the target audience is women aged 20-30, but the user's age and gender information indicates that the user is a 55-year-old male, then the initial video's attribute information does not match the user's attribute information.

[0065] In practical applications, if a user effectively consumes a video, it can be determined that the user may be interested in this type of video and want to follow this type of video in the future. Therefore, even if this type of video does not match the user's feature information, a new node is created for the initial video on the user's personal knowledge graph based on the initial video's video tags, thus realizing the establishment of the personal knowledge graph. Alternatively, if the user is a new user with no prior video viewing history, then even if the attribute information of the initial video does not match the user's feature information, a new personal knowledge graph needs to be created for the user based on the initial video's video tags. This allows the user to quickly understand their learning progress or viewed content on the video learning platform based on this personal knowledge graph.

[0066] In this embodiment of the application, even if the attribute information of the initial video does not match the user's feature information, the initial video can still be used as the target video. This allows the user to create or update their personal knowledge graph based on the target video, enabling them to quickly understand their learning progress or viewing content on the video learning platform.

[0067] The third method involves determining one or more target videos corresponding to the user based on the attribute information of the one or more initial videos and the user's feature information, including:

[0068] Match the attribute information of each of the one or more initial videos with the user's feature information;

[0069] One or more initial videos whose attribute information does not match the user's feature information are used as one or more candidate videos corresponding to the user.

[0070] If the number of one or more candidate videos is greater than or equal to a preset threshold, the one or more candidate videos will be used as one or more target videos corresponding to the user.

[0071] In practice, the attribute information of each initial video is matched with the user's feature information. If they do not match, the initial video can be used as a candidate video. When the duration or number of times a user watches this type of candidate video reaches a preset threshold, the candidate video can be used as the target video. The preset threshold can be set according to the actual application and is not required here. For example, it can be 5 or 8, etc.

[0072] In practical applications, if the attribute information of the initial video does not match the user's feature information, the initial video can be placed in the candidate list as a candidate video. When the user watches the same type of initial video again, the candidate video is added back to the candidate list, and the number of candidate videos is accumulated. When the number accumulates to a certain level, it is determined that the user may indeed be interested in this type of video and want to learn about it in the future. In this case, these candidate videos can be used as target videos to expand the user's personal knowledge graph and improve the user experience.

[0073] Step 108: Extract one or more video tags from the one or more target videos, and build or update a personal knowledge graph for the user based on the one or more video tags.

[0074] Specifically, after identifying one or more target videos, video tags are extracted from each target video. Based on these video tags, the video category of each target video is determined. Then, a personal knowledge graph is built for the user based on all video categories. If the user already has a personal knowledge graph, it is updated based on all video categories to ensure the accuracy and effectiveness of the user's personal knowledge graph. In practical applications, the user knowledge graph can be in the form of a radar chart or a heat map.

[0075] The knowledge graph processing method provided in this application, after acquiring the videos watched by the user, determines the final effective videos consumed by the user based on the viewing time of the videos, the attribute information of the videos, and the user's characteristic information. Then, based on the video tags of the effective videos consumed by the user, the user's personal knowledge graph is newly created or updated. Subsequently, the user can directly and quickly understand their historical learning content based on this personal knowledge graph, so as to adjust their learning plan in a timely manner and enhance the user experience.

[0076] In another embodiment of this application, after determining the viewing time of each of the one or more videos watched by the user, and selecting one or more videos with a viewing time greater than or equal to a preset time threshold as the initial videos, the method further includes:

[0077] Extract the video tags from the initial video, and match the video tags of the initial video with the knowledge tags in the knowledge video tag graph to determine the knowledge tags that match the video tags of the initial video;

[0078] The recommended videos corresponding to the matched knowledge tags are obtained from the video library, and the recommendation format of the recommended videos is determined based on the matched knowledge tags;

[0079] The recommended video is recommended to the user according to the recommended format.

[0080] Specifically, after selecting one or more videos with a viewing time greater than or equal to a preset time threshold as initial videos, the system can also extract video tags from these initial videos and recommend videos to users based on these tags, thereby increasing users' reliance on the video learning platform.

[0081] If the video is an educational video, after extracting the video tags of one or more initial videos, it is determined whether the video tags of the initial videos are educational tags. That is, the initial video is determined to be a knowledge video based on its video tags. Since the embodiments of this application aim to solve the problem of recommending knowledge videos, effective knowledge video recommendations can only be made if the initial video watched by the user is a knowledge video. The specific implementation method is as follows:

[0082] The step of matching the video tags of the initial video with the knowledge tags in the knowledge video tag graph includes:

[0083] Determine whether the video tags of the initial video match tags in a preset learning tag library.

[0084] If so, the video tags of the initial video are matched with the knowledge tags in the knowledge video tag graph.

[0085] If not, then the preset recommended videos corresponding to the initial video will be recommended to the user in the form of a list.

[0086] The preset learning tag library contains multiple learning tags, such as physics, mathematics, Chinese, chemistry, and other learning tags.

[0087] Specifically, after extracting the video tags of the initial video, each video tag of the initial video is matched with tags in the preset learning tag library. If a match is found, the video tags of the initial video are matched with knowledge tags in the knowledge video tag graph. If no match is found, it means that the video tags of the initial video are not learning tags. In this case, it can be basically determined that the initial video is probably not a learning-related knowledge video, and there is no need to match it with knowledge tags in the knowledge video tag graph. At this time, the previously preset recommended videos corresponding to the initial video can be recommended to users who watch the initial video.

[0088] The preset recommended videos can be understood as recommended videos pre-configured for each video. For example, if the video is an entertainment video, then the pre-configured recommended videos may be entertainment news or celebrity gossip videos.

[0089] Specifically, when the initial video watched by a user is a video of effective consumption type, even if the initial video is not an educational video, in order to ensure the user experience, the user should still be recommended to the user based on the preset recommended videos. This allows the user to continue watching video content of interest based on the recommended videos, thereby enhancing the user's stickiness with the video platform.

[0090] In practical applications, a knowledge video tag graph can be understood as a graph constructed based on the tags of knowledge videos, where the knowledge video is a video for learning knowledge. Before extracting the video tags of the initial video and matching the video tags of the initial video with the knowledge tags in the knowledge video tag graph to determine the knowledge tags that match the video tags of the initial video, the process further includes:

[0091] Obtain knowledge videos from the video library and determine the knowledge tags of the knowledge videos;

[0092] The knowledge tags of the knowledge videos are used to build a knowledge video tag map based on preset knowledge tag levels, wherein the knowledge video tag map is a tree-shaped knowledge video tag map.

[0093] The video library stores various types of videos, while the knowledge videos are educational videos. The preset knowledge tag levels are designed to facilitate the categorization and retrieval of videos within the knowledge video tag graph. The knowledge video tag graph is set to different levels, allowing knowledge tags from different levels to be displayed in different locations. For example, if the knowledge video tag graph has four levels, when recommending videos to users after watching a video, the knowledge tags displayed will be the main category tags (second-level tags) and fourth-level tags. In the later-established learning plaza, the knowledge tags displayed will be from levels one to four.

[0094] Specifically, first, all knowledge videos in the video library are obtained, then the knowledge tags for each knowledge video are determined, and finally, a tree-like knowledge video tag map is built based on the preset knowledge tag levels for all the knowledge tags.

[0095] See Figure 2 , Figure 2 This diagram illustrates the knowledge tag levels in a knowledge graph processing method according to an embodiment of this application.

[0096] pass Figure 2As we can see, knowledge tag levels can be divided into four levels: Level 1, Level 2, Level 3, and Level 4 tags. For example, a Level 1 tag might be "learning," a Level 2 tag might be "the self-cultivation of a programmer," a Level 3 tag might be "Python," and a Level 4 tag might be "Python introduction" or "Python platform," etc. The Level 1 tag serves as the sole benchmark, representing the type of knowledge video tag graph. Level 2 tags represent the major categories of videos, Level 3 tags represent the sub-subjects within the videos, and Level 4 tags represent the specific knowledge points. A single video can have a maximum of two Level 2 tags, and each Level 2 tag can correspond to only one Level 3 tag and one Level 4 tag. A Level 4 tag can have multiple videos, generally those with the most likes.

[0097] In this embodiment, a knowledge video tag graph is pre-established based on the knowledge tags of knowledge videos in the video library. Subsequently, when recommending similar knowledge videos based on the initial videos watched by the user, the recommendation can be made quickly and accurately based on the knowledge video tag graph, thereby improving the efficiency of video recommendation.

[0098] In practice, the video tags of the initial video are matched with the knowledge tags in the knowledge video tag graph established in the above manner to determine the knowledge tags that match the video tags of the initial video. For example, if the video tags of the initial video are "Python Introduction" and "Python Platform", and the fourth-level knowledge tags in the knowledge video tag graph established in the above manner also include the knowledge tags "Python Introduction" and "Python Platform", then it means that the video tags of the initial video match the knowledge tags in the knowledge video tag graph established in the above manner. At this time, the knowledge tags "Python Introduction" and "Python Platform" are determined. Each knowledge tag in the knowledge video tag graph is associated with multiple videos corresponding to this knowledge tag.

[0099] After identifying the knowledge tag that matches the video tag, the recommended video corresponding to the matching knowledge tag is retrieved from the video library based on the association between the knowledge tag and the knowledge video. In practical applications, each knowledge tag in the knowledge video tag graph is associated with multiple corresponding videos. In order to reduce the memory usage of the knowledge video tag graph, the knowledge video tag graph stores the video identifiers of the multiple corresponding videos associated with each knowledge tag. Then, the recommended video corresponding to the matching knowledge tag is retrieved from the video library based on the video identifiers.

[0100] In addition, after obtaining the recommended videos, the number of major categories to which the video belongs, i.e., the number of secondary tags, needs to be determined based on the matching knowledge tags. Then, the recommendation format of these recommended videos is determined based on the number of major categories to which the video belongs. For example, if the video belongs to only one major category, then the recommended videos can be directly displayed to the user in the form of a single page. If the video belongs to two or more major categories, in order to facilitate the user's understanding of the recommended videos, i.e., to help the user understand why these recommended videos are recommended to them, these recommended videos can be displayed to the user through a Magic Cube recommendation format.

[0101] See Figure 3 , Figure 3 The illustration shows a schematic diagram of a knowledge graph processing method according to an embodiment of this application, in which recommended videos are displayed in the form of a Rubik's Cube.

[0102] Figure 3 If the initial video tags extracted are "cat encyclopedia" and "physics," with "cat encyclopedia" belonging to the animal knowledge category and "physics" belonging to the hard science category, and the recommended videos corresponding to the "cat encyclopedia" tag are things like "spinning and landing" and "spreading XX virus," while the recommended videos corresponding to the "physics" tag are things like "the law of conservation of energy," "the law of conservation of angular momentum," and "thermocouples," then displaying these recommended videos to users through a magic cube-based recommendation system can be done as follows: Figure 3 The recommended videos corresponding to the cat encyclopedia, such as "Rotation and Landing" and "Infecting XX Virus," are displayed on one face of the Rubik's Cube. The recommended videos corresponding to the physics tags, such as "Law of Conservation of Energy," "Law of Conservation of Angular Momentum," and "Thermocouple," are displayed on the other face of the Rubik's Cube.

[0103] Specifically, the recommendation format is a Rubik's Cube recommendation format;

[0104] Accordingly, recommending the recommended video to the user according to the recommendation format includes:

[0105] Determine the recommended video corresponding to each matched knowledge tag, and display the recommended video corresponding to each matched knowledge tag on one side of the magic cube to recommend it to the user.

[0106] Specifically, a Rubik's Cube image can be understood as a diagram based on the shape of a Rubik's Cube. If the recommendation format is a Rubik's Cube image recommendation format, after determining the recommended video corresponding to each matching knowledge tag, the recommended video corresponding to each matching knowledge tag will be displayed on one side of the Rubik's Cube image and recommended to the user.

[0107] See Figure 4 , Figure 4 The illustration shows a schematic diagram of recommending videos to users in the form of a Rubik's Cube in a knowledge graph processing method according to an embodiment of this application.

[0108] Taking the initial video tags extracted above as "cat encyclopedia" and "physics" as an example, the knowledge tags that match the initial video tags are also "cat encyclopedia" and "physics". Based on the above method, the recommended videos corresponding to the cat encyclopedia, such as "rotation and landing" and "spreading XX virus", are displayed on one face of the Rubik's Cube. The recommended videos corresponding to the physics tag, such as "law of conservation of energy", "law of conservation of angular momentum", and "thermocouple", are displayed on the other face of the Rubik's Cube.

[0109] In practice, if other matching knowledge tags exist, the recommended videos corresponding to those knowledge tags will be displayed on the other faces of the Rubik's Cube, and the corresponding knowledge tags will be displayed at the bottom of each face so that users can understand why these recommended videos are recommended to them, thus improving the user experience.

[0110] In this embodiment of the application, after acquiring the video watched by the user, the knowledge graph processing method will only extract video tags from the initial video if it is determined that the initial video watched by the user is a valid consumption video. This ensures that when the user is interested in the video watched, the method will automatically recommend similar videos to the user based on the video tags of the initial video, thereby improving the user's learning rate of the recommended videos. Furthermore, the method will also recommend videos to the user in a special recommendation format based on the video tags, thereby improving the user's personalized experience of video recommendations.

[0111] In another embodiment of this application, after recommending the recommended video to the user according to the recommendation format, the method further includes:

[0112] Based on the user's personal knowledge graph, it is determined whether the recommended video is a video that the user has already watched.

[0113] If so, if the recommended video is determined to be the initial video, the first annotation information of the recommended video in the user's personal knowledge graph is displayed; or if the recommended video is determined to be the target video, the second annotation information of the recommended video in the user's personal knowledge graph is displayed.

[0114] If not, after receiving the user's instruction to open the recommended video, the recommended video will be opened for knowledge learning.

[0115] Among the recommended videos, some may already be viewed by the user. To prevent users from repeatedly watching videos they are unaware of, thus creating a poor viewing experience, the system will determine the videos the user has already watched based on their historical video viewing records displayed in their personal knowledge graph after the recommended videos are determined. When a recommended video is determined to be an initial video that the user has already watched, the system will display the first annotation information of that recommended video in the user's personal knowledge graph; or when a recommended video is determined to be a target video that the user has already watched, the system will display the second annotation information of that recommended video in the user's personal knowledge graph, in order to increase the user's positive viewing experience. The first annotation information can be a highlighted "Learned" tag, and the second annotation information can be a highlighted "Completed" tag, etc.

[0116] In practical applications, if the recommended video is determined to be a video that the user has already watched based on the historical video viewing records displayed in the user's personal knowledge graph, the recommended video will be specially displayed. If the recommended video is determined not to be a video that the user has already watched based on the user's historical video viewing records, after receiving the user's instruction to open a certain recommended video, the user will be redirected to the playback interface of that recommended video for knowledge learning.

[0117] See Figure 4 Using the previous example, recommended videos are displayed on a face of the Rubik's Cube. Videos that the user has already watched can be displayed to the user with added first or second label information, while videos that the user has not watched are displayed with normal video titles or video descriptions. When a user wants to learn about a recommended video that they have not watched, they can directly click on a recommended video on a face of the Rubik's Cube, and will be directly redirected to the playback interface of that recommended video, allowing the user to learn knowledge based on that recommended video.

[0118] In this embodiment, recommended videos are personalized based on the user's historical video viewing records displayed in their personal knowledge graph. This allows users to select recommended videos that interest them for learning, and ensures that users do not repeatedly study videos they have already watched, thus enhancing the user experience. Furthermore, it prevents users from closing recommended videos they have already watched, which would increase their workload and the system's processing efficiency.

[0119] See Figure 5 , Figure 5 This illustration shows a user's personal knowledge graph in a knowledge graph processing method according to an embodiment of this application.

[0120] Figure 5The user's personal knowledge graph is built based on the above knowledge graph processing method. Figure 5 The user's personal knowledge graph can display the dynamic value of the user's learning time, and the user's personal knowledge graph is displayed in the form of a radar chart. Brand connotation, employment relationship, work environment, welfare benefits, etc., in the user's personal knowledge graph are the secondary tags (i.e., major categories) in the aforementioned knowledge video tag graph. Taking the creation of a user's personal knowledge graph with 10 major categories as an example, when creating the user's personal knowledge graph, the top 10 categories with the longest consumption time (secondary video tags) can be displayed in reverse order. The consumption time of the first major category is taken as 100% progress, and the consumption time of the other 9 major categories is calculated and constructed according to the proportion of the consumption time of the first major category, which is the user's personal knowledge graph. After the user's personal knowledge graph is created, each point in the radar chart displays the name of the secondary video tag. Users can enter the corresponding knowledge consumption square by clicking on the name area. The concept of the knowledge consumption square can be seen in the figure below. Figure 6 and Figure 7 When a user clicks on the name of a specific secondary video tag, the name (category) will display the total deduplicated consumption time of the corresponding video, for example, in the format hhhhh:mm:ss, with a maximum of 10,000 hours. If the total time reaches 10,000 hours, minutes and seconds will not be displayed. In order to ensure the simplicity of the user's personal knowledge graph, any longer durations will not be included in the statistics.

[0121] See Figure 6 , Figure 6 This illustration shows a knowledge consumption plaza displayed to a video uploader in a knowledge graph processing method according to an embodiment of this application.

[0122] Figure 6 In the knowledge consumption plaza, there is a foundation: introduction and environment setup, which can be understood as a broad category; the first layer: basic knowledge; the second layer: detailed explanation of basic knowledge: basic syntax, variable types, operators; the third layer: detailed explanation of basic knowledge: if statements, loop statements, while statements, for statements; after a video uploader uploads a video, the associated tags for that video are used to build the knowledge consumption plaza based on the above levels. At this time, the knowledge consumption plaza version for the corresponding video uploader has a selectable state: 1. The video uploader enters the tag selection page, Figure 6 All the left-side hierarchy tabs are to be selected. Video uploaders can select the left-side hierarchy tabs to establish a knowledge consumption plaza structure; 2. After selecting the hierarchy... Figure 6The knowledge point tags corresponding to the relevant level on the right are now selectable, and video uploaders can select multiple tags within the same level. Furthermore, the tagging strategies for video uploaders include: 1. Supporting video uploaders to tag a single video with multiple tags, but not supporting multiple tags across multiple levels for a single video; 2. If a video uploader tags a single video with more than three tags at once, the number of likes received in the last three months * 0.8 will be used for ranking.

[0123] See Figure 7 , Figure 7 The diagram illustrates a knowledge consumption plaza displayed to a user in a knowledge graph processing method according to an embodiment of this application.

[0124] Continuing with the previous example, in practical applications, Figure 7 Knowledge Consumption Plaza and Figure 6 The knowledge consumption plaza has the same style, but for users, there is no "to be checked" state, only "to be lit" state: 1. Users start from the bottom layer (i.e., the foundation). When they watch a video with a certain tag, they light up a square, and the lit square displays a "learned" stamp; 2. An animation needs to be displayed to show the lighting path. For example, if all levels from 0 to 1 are lit, the lighting path is 1.1 / 1.2 / 1.3, and there is a lighting animation between levels; and if the user skips to a level to watch a video, a skipped-level learning stamp will appear.

[0125] The specific rules for advancement or skipping levels include: 1. Users advance by completing learning tasks, meaning that users can unlock a higher level by completing more than 50% of the videos within their current learning level; 2. Users skip levels by completing learning tasks, meaning that users can also unlock a higher level by completing more than 50% of the videos within their skipped level, but skipping levels does not unlock lower levels.

[0126] Specifically, the backend determines user completion of learning in the following ways: 1. If a user selects a single video corresponding to a single tag, the user can light up the square and stamp the tag with "learned" once the user has effectively consumed 50% of the video's duration; 2. If a user selects a single video corresponding to multiple tags, the user can check in for the first tag corresponding to the video once the user's effective consumption time reaches 1 / x*50% (where x is the number of tags), and so on, until the user completes all tags corresponding to all single videos.

[0127] In practical applications, when a user clicks on any tag box, recommended video works under that tag are displayed as a bottom overlay. See details. Figure 8 , Figure 8 The illustration shows a schematic diagram of a recommended video work in a knowledge graph processing method according to an embodiment of this application.

[0128] Specifically, the floating layer pushes up the page content. Clicking any area other than the bottom floating layer or clicking the down button will collapse the floating layer. The recommended video works display video block information including: cover, title (if the title exceeds the limit, it will be truncated to only two lines), number of views, upload time, and video uploader; and videos that the user has already watched are marked with a "viewed" stamp in the upper right corner.

[0129] In addition, the floating video sorting rules include: 1. When a user clicks on any tag square, the top three videos with the most new likes in the past three months will appear, arranged from right to left; 2. Videos that the user has already watched are marked with a "watched" stamp in the upper right corner (but users can still jump to the video page). Watched videos are arranged from right to left in order of consumption time. If a user has consumed videos 1 and 2, but not video 3, the order will be 3, 1, 2.

[0130] In this embodiment of the application, a personalized user knowledge graph can be created or updated for the user based on the user's video viewing history on the video platform, which makes it easier to understand the user's knowledge learning status through the user knowledge graph and improve the user experience.

[0131] In specific implementation, the step of extracting one or more video tags from the one or more target videos and establishing or updating a personal knowledge graph for the user based on the one or more video tags includes:

[0132] Extract one or more video tags from the one or more target videos, and the viewing position of each video in the one or more target videos viewed by the user;

[0133] A personal knowledge graph is created or updated for the user based on the viewing location and one or more video tags.

[0134] The viewing position of each video in the one or more target videos can be understood as the IP (Internet Protocol) address of the terminal on which the user watches the video. Then, a user knowledge graph is built based on the IP address and video tags. This makes it easier for users to clearly identify their knowledge learning position and understand their learning progress at any point when reviewing knowledge videos based on their personal knowledge graph.

[0135] In another embodiment of this application, after establishing or updating the user's personal knowledge graph based on the viewing location and the one or more video tags, the method further includes:

[0136] Based on the viewing location, the user's learning zone in the learning square is determined, and the user's personal knowledge graph of all users in the learning zone is displayed to the user. The learning square is constructed by combining the personal knowledge graphs of all video viewers with an electronic map.

[0137] The Learning Plaza is constructed by combining the personal knowledge graphs of all video viewers with an electronic map. Through this Learning Plaza, users can see the knowledge learning status of users in every part of the country. For example, it can display the consumption tag graphics of users in City A's knowledge learning graph.

[0138] Specifically, learning zones can be divided by city. Based on the user's viewing location, it can be determined which learning zone the user is in within the learning square, i.e., which city the user is studying in. Then, the user is shown the personal knowledge graph of all users in their learning zone. This allows the user to see what knowledge videos other users in their learning zone are watching and what knowledge they are learning, so that the user can refer to the video content of others to learn new types of knowledge points and expand their knowledge.

[0139] In another embodiment of this application, the knowledge video tag map includes multiple categories, each category corresponds to multiple knowledge tags, and each knowledge tag corresponds to multiple knowledge points;

[0140] Accordingly, after establishing the knowledge video tag map based on the knowledge tags of the knowledge video according to the preset knowledge tag levels, the method further includes:

[0141] Receive uploaded videos, and if the uploaded videos meet preset recommendation conditions, obtain the video tags of the uploaded videos;

[0142] If the video tag of the uploaded video belongs to an existing category of knowledge tag in the knowledge video tag graph, determine whether the video tag of the uploaded video belongs to the knowledge point corresponding to the knowledge tag in the knowledge video tag graph.

[0143] If so, the uploaded video will be added to the video queue corresponding to the knowledge point.

[0144] If not, a knowledge point is created in the knowledge video tag graph based on the video tag of the uploaded video, and the uploaded video is added to the video queue corresponding to the knowledge point.

[0145] The preset recommendation criteria include, but are not limited to, whether the current video is a learning-related knowledge video.

[0146] Specifically, after receiving an uploaded video, the system checks if it is a new video (i.e., one that does not exist in the video library). If so, it checks if the uploaded video is a knowledge video related to learning. If not, the process ends. If so, the system extracts the tags for the uploaded video. These tags can be set by the uploader or automatically obtained by the backend through an algorithm, and there must be at least one tag. After extracting the tags, the system checks if the tags belong to an existing category in the knowledge video tag graph (i.e., the aforementioned knowledge video tag graph). If not, a corresponding category is created in the knowledge video tag graph based on the tag to update and enrich the graph. If so, the system checks if the tag is a knowledge point not found in the corresponding category. If not, the knowledge point is sorted in the knowledge point queue according to the algorithm. If so, a knowledge point is created based on the tag and added to the recommendation queue for recommendation.

[0147] In this embodiment of the application, a series of judgments and reviews are performed on each new video, and the original knowledge video tag graph is updated based on each new video to make the knowledge video tag graph richer, so as to improve the number of videos recommended to users of the same type and the accuracy of the recommendation.

[0148] The following is in conjunction with the appendix Figure 9 Taking the knowledge graph processing method provided in this application as an example in the application of learning-type knowledge videos, a flowchart of the video recommendation process within the knowledge graph processing method is provided. Specifically, it includes the following steps.

[0149] Step 902: Obtain videos that the user has effectively consumed.

[0150] Specifically, videos that a user effectively consumes can be understood as videos whose viewing time exceeds a preset time threshold, such as initial videos that are more than 3 minutes long.

[0151] Step 904: Determine the video tags for the video and check if there are any learning tags in the video tags. If yes, proceed to step 908; otherwise, proceed to step 906.

[0152] Specifically, a learning tag library can be pre-built, and the tags in this learning tag library can be considered as learning tags.

[0153] Step 906: Show the user recommended videos from the preset original recommendation list.

[0154] Step 908: Determine the learning tags for this video: Physics, Cat Encyclopedia.

[0155] Step 910: Based on the knowledge video tag graph, extract videos containing physics and cat encyclopedia from the video library as recommended videos.

[0156] Step 912: Determine a list of videos that include physics and cat encyclopedia.

[0157] Step 914: Display the videos from the list of videos containing physics on one face of the Rubik's Cube.

[0158] Step 916: Display the videos from the video list containing cat encyclopedia on another face of the Rubik's Cube.

[0159] Step 918: Determine if the user has already viewed any videos in the list of videos including physics and cat encyclopedia. If yes, proceed to step 920; otherwise, proceed to step 922.

[0160] Step 920: Display the videos that the user has watched as highlighted.

[0161] Step 922: Receive the user's click command to click on a face of the Rubik's Cube.

[0162] Step 924: Display all recommended videos on the side of the Rubik's Cube that the user clicks on, with a learning progress bar displayed below each video.

[0163] Step 926: When a user clicks on a video on that face of the Rubik's Cube, redirect the user to the learning interface for that video.

[0164] Furthermore, in practical implementation, if new videos are received, the knowledge video tag graph can be updated based on the new videos. See [link to relevant documentation]. Figure 10 , Figure 10 This paper illustrates a flowchart of the knowledge video tag graph update process according to an embodiment of the present application.

[0165] Step 1002: Obtain the new video uploaded by the user.

[0166] Specifically, new videos can be understood as videos that users have not watched before.

[0167] Step 1004: Determine whether the new video meets the criteria for recommendation. If not, proceed to step 1006; if yes, proceed to step 1008.

[0168] Step 1006: End.

[0169] Step 1008: Extract video tag 1, video tag 2 and video tag 3 from the new video.

[0170] Step 1010: Determine whether video tag 1 belongs to any of the existing major categories in the knowledge video tag graph. If yes, proceed to step 1014; otherwise, proceed to step 1012.

[0171] Specifically, the major categories of the knowledge video tag graph can be understood as the secondary tags of the knowledge video tag graph.

[0172] Step 1012: Create a new major category in the knowledge video tag graph based on video tag 1.

[0173] Step 1014: Determine whether video tag 1 is a knowledge point that is not present in the existing major categories of the knowledge video tag graph. If yes, proceed to step 1018; otherwise, proceed to step 1016.

[0174] Step 1016: Arrange video tag 1 in the corresponding knowledge point queue according to the preset sorting algorithm.

[0175] Step 1018: Based on video tag 1, create knowledge points under the major category of the knowledge video tag graph corresponding to video tag 1.

[0176] Step 1020: Add the video corresponding to video tag 1 to the created knowledge point queue and wait for recommendation.

[0177] Specifically, the subsequent processing of video tag 2 and video tag 3 can be found in steps 1010 to 1020 for the processing of video tag 1, and will not be repeated here.

[0178] The knowledge graph processing method described in this application categorizes videos by tagging them. The knowledge tree connects the tags involved in the videos by standardizing the tag system for knowledge-based videos. When it is found that a user has watched similar videos, videos with the same tags can be pushed to the user to encourage the user to consume video content in the knowledge graph. Furthermore, when a new video is uploaded by the user, the knowledge video tag graph can be updated based on the video tags of the new video, making the knowledge video tag graph richer and increasing the number and types of videos recommended to the user in the future, thereby enhancing the user experience.

[0179] Corresponding to the above method embodiments, this application also provides embodiments of a knowledge graph processing apparatus. Figure 11 A schematic diagram of the structure of a knowledge graph processing apparatus according to an embodiment of this application is shown. Figure 11 As shown, the device includes:

[0180] The acquisition module 1102 is configured to acquire one or more videos viewed by the user and the user's feature information;

[0181] The initial video determination module 1104 is configured to determine the viewing time of each of the one or more videos watched by the user, and to select one or more videos whose viewing time is greater than or equal to a preset time threshold as the initial videos;

[0182] The target video determination module 1106 is configured to determine one or more target videos corresponding to the user based on the attribute information of the one or more initial videos and the feature information of the user;

[0183] The knowledge graph processing module 1108 is configured to extract one or more video tags from the one or more target videos, and to build or update a personal knowledge graph for the user based on the one or more video tags.

[0184] Optionally, the target video determination module 1106 is further configured to:

[0185] Match the attribute information of each of the one or more initial videos with the user's feature information;

[0186] One or more initial videos whose attribute information matches the user's feature information are used as one or more target videos corresponding to the user.

[0187] Optionally, the target video determination module 1106 is further configured to:

[0188] Match the attribute information of each of the one or more initial videos with the user's feature information;

[0189] One or more initial videos whose attribute information does not match the user's feature information are selected as one or more target videos corresponding to the user.

[0190] Optionally, the target video determination module 1106 is further configured to:

[0191] Match the attribute information of each of the one or more initial videos with the user's feature information;

[0192] One or more initial videos whose attribute information does not match the user's feature information are used as one or more candidate videos corresponding to the user.

[0193] If the number of one or more candidate videos is greater than or equal to a preset threshold, the one or more candidate videos will be used as one or more target videos corresponding to the user.

[0194] Optionally, the device further includes:

[0195] The extraction module is configured to extract video tags from the initial video and match the video tags of the initial video with knowledge tags in the knowledge video tag graph to determine the knowledge tags that match the video tags of the initial video.

[0196] The video acquisition module is configured to retrieve recommended videos corresponding to the matched knowledge tags from the video library, and determine the recommendation format of the recommended videos based on the matched knowledge tags;

[0197] The recommendation module is configured to recommend the recommended videos to the user according to the recommendation format.

[0198] Optionally, the extraction module is further configured to:

[0199] Determine whether the video tags of the initial video match tags in a preset learning tag library.

[0200] If so, then match the video tags of the initial video with the knowledge tags in the knowledge video tag graph.

[0201] If not, then the preset recommended videos corresponding to the initial video will be recommended to the user in the form of a list.

[0202] Optionally, the recommendation format is a Rubik's Cube recommendation format;

[0203] Accordingly, the recommendation module is further configured as follows:

[0204] Determine the recommended video corresponding to each matched knowledge tag, and display the recommended video corresponding to each matched knowledge tag on one side of the magic cube to recommend it to the user.

[0205] Optionally, the device further includes:

[0206] The tag determination module is configured to acquire knowledge videos from the video library and determine the knowledge tags of the knowledge videos;

[0207] The knowledge video tag building module is configured to build a knowledge video tag map based on a preset knowledge tag level, wherein the knowledge video tag map is a tree-structured knowledge video tag map.

[0208] Optionally, the device further includes:

[0209] The judgment module is configured as follows:

[0210] Based on the user's personal knowledge graph, it is determined whether the recommended video is a video that the user has already watched.

[0211] If so, if the recommended video is determined to be the initial video, the first annotation information of the recommended video in the user's personal knowledge graph is displayed; or if the recommended video is determined to be the target video, the second annotation information of the recommended video in the user's personal knowledge graph is displayed.

[0212] If not, after receiving the user's instruction to open the recommended video, the recommended video will be opened for knowledge learning.

[0213] Optionally, the map processing module 1108 is further configured to:

[0214] Extract one or more video tags from the one or more target videos, and the viewing position of each video in the one or more target videos viewed by the user;

[0215] A personal knowledge graph is created or updated for the user based on the viewing location and one or more video tags.

[0216] Optionally, the knowledge video tag graph includes multiple categories, each category corresponds to multiple knowledge tags, and each knowledge tag corresponds to multiple knowledge points;

[0217] Accordingly, the device further includes:

[0218] The map update module is configured as follows:

[0219] Receive uploaded videos, and if the uploaded videos meet preset recommendation conditions, obtain the video tags of the uploaded videos;

[0220] If the video tag of the uploaded video belongs to an existing category of knowledge tag in the knowledge video tag graph, determine whether the video tag of the uploaded video belongs to the knowledge point corresponding to the knowledge tag in the knowledge video tag graph.

[0221] If so, the uploaded video will be added to the video queue corresponding to the knowledge point.

[0222] If not, a knowledge point is created in the knowledge video tag graph based on the video tag of the uploaded video, and the uploaded video is added to the video queue corresponding to the knowledge point.

[0223] The knowledge graph processing device provided in this application embodiment, after acquiring the videos watched by the user, determines the final valid videos consumed by the user based on the viewing time of the videos, the attribute information of the videos, and the user's characteristic information. Then, based on the video tags of the valid videos consumed by the user, the device creates or updates the user's personal knowledge graph. Subsequently, the user can directly and quickly understand their historical learning content based on this personal knowledge graph, so as to adjust their learning plan in a timely manner and enhance the user experience.

[0224] The above is a schematic scheme of a knowledge graph processing device according to this embodiment. It should be noted that the technical solution of this knowledge graph processing device and the technical solution of the knowledge graph processing method described above belong to the same concept. For details not described in detail in the technical solution of the knowledge graph processing device, please refer to the description of the technical solution of the knowledge graph processing method described above.

[0225] Figure 12 A structural block diagram of a computing device 1200 according to an embodiment of this application is shown. The components of the computing device 1200 include, but are not limited to, a memory 1210 and a processor 1220. The processor 1220 is connected to the memory 1210 via a bus 1230, and a database 1250 is used to store data.

[0226] The computing device 1200 also includes an access device 1240, which enables the computing device 1200 to communicate via one or more networks 1260. Examples of these networks include a Public Switched Telephone Network (PSTN), a Local Area Network (LAN), a Wide Area Network (WAN), a Personal Area Network (PAN), or a combination of communication networks such as the Internet. The access device 1240 may include one or more of any type of wired or wireless network interface (e.g., a Network Interface Card (NIC)), such as an IEEE 802.11 Wireless Local Area Network (WLAN) interface, a Wi-MAX interface, an Ethernet interface, a Universal Serial Bus (USB) interface, a cellular network interface, a Bluetooth interface, a Near Field Communication (NFC) interface, and so on.

[0227] In one embodiment of this application, the aforementioned components of the computing device 1200 and Figure 12 Other components, not shown, can also be connected to each other, for example, via a bus. It should be understood that... Figure 12 The block diagram of the computing device shown is for illustrative purposes only and is not intended to limit the scope of this application. Those skilled in the art can add or replace other components as needed.

[0228] The computing device 1200 can be any type of stationary or mobile computing device, including mobile computers or mobile computing devices (e.g., tablet computers, personal digital assistants, laptop computers, notebook computers, netbooks, etc.), mobile phones (e.g., smartphones), wearable computing devices (e.g., smartwatches, smart glasses, etc.) or other types of mobile devices, or stationary computing devices such as desktop computers or PCs. The computing device 1200 can also be a mobile or stationary server.

[0229] The processor 1220 is configured to execute the following computer-executable instructions, wherein the processor executes the computer-executable instructions to implement the steps of the knowledge graph processing method.

[0230] The above is an illustrative scheme of a computing device according to this embodiment. It should be noted that the technical solution of this computing device and the technical solution of the knowledge graph processing method described above belong to the same concept. For details not described in detail in the technical solution of the computing device, please refer to the description of the technical solution of the knowledge graph processing method described above.

[0231] An embodiment of this application also provides a computer-readable storage medium storing computer instructions that, when executed by a processor, implement the steps of the knowledge graph processing method.

[0232] The above is an illustrative scheme of a computer-readable storage medium according to this embodiment. It should be noted that the technical solution of this storage medium and the technical solution of the knowledge graph processing method described above belong to the same concept. For details not described in detail in the technical solution of the storage medium, please refer to the description of the technical solution of the knowledge graph processing method described above.

[0233] The foregoing has described specific embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0234] The computer instructions include computer program code, which may be in the form of source code, object code, executable file, or certain intermediate forms. The computer-readable medium may include any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium may be appropriately added to or subtracted according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media may not include electrical carrier signals and telecommunication signals.

[0235] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of this application are not limited to the described order of actions, because according to the embodiments of this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the embodiments of this application.

[0236] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0237] The preferred embodiments disclosed above are merely illustrative of this application. The optional embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the embodiments of this application. These embodiments are selected and specifically described in this application to better explain the principles and practical applications of the embodiments of this application, thereby enabling those skilled in the art to better understand and utilize this application. This application is limited only by the claims and their full scope and equivalents.

Claims

1. A knowledge graph processing method, characterized in that, include: Obtain one or more videos watched by the user and the user's characteristic information; Determine the viewing time of each of the one or more videos watched by the user, and select one or more videos whose viewing time is greater than or equal to a preset time threshold as the initial videos; Obtain knowledge videos from the video library and determine the knowledge tags of the knowledge videos; The knowledge tags of the knowledge videos are used to build a knowledge video tag map based on a preset knowledge tag level, wherein the knowledge video tag map is a tree-structured knowledge video tag map; Based on the attribute information of the one or more initial videos and the characteristic information of the user, determine one or more target videos corresponding to the user; Extract one or more video tags from the one or more target videos, and build or update a personal knowledge graph for the user based on the one or more video tags; the personal knowledge graph is displayed in the form of a radar chart, and each point of the radar chart displays the name of a secondary video tag in the knowledge video tag graph. Clicking on the name area can enter the corresponding knowledge consumption square, which has a hierarchical structure.

2. The knowledge graph processing method according to claim 1, characterized in that, The step of determining one or more target videos corresponding to the user based on the attribute information of the one or more initial videos and the user's feature information includes: Match the attribute information of each of the one or more initial videos with the user's feature information; One or more initial videos whose attribute information matches the user's feature information are used as one or more target videos corresponding to the user.

3. The knowledge graph processing method according to claim 1, characterized in that, The step of determining one or more target videos corresponding to the user based on the attribute information of the one or more initial videos and the user's feature information includes: Match the attribute information of each of the one or more initial videos with the user's feature information; One or more initial videos whose attribute information does not match the user's feature information are selected as one or more target videos corresponding to the user.

4. The knowledge graph processing method according to claim 1, characterized in that, The step of determining one or more target videos corresponding to the user based on the attribute information of the one or more initial videos and the user's feature information includes: Match the attribute information of each of the one or more initial videos with the user's feature information; One or more initial videos whose attribute information does not match the user's feature information are used as one or more candidate videos corresponding to the user. If the number of one or more candidate videos is greater than or equal to a preset threshold, the one or more candidate videos will be used as one or more target videos corresponding to the user.

5. The knowledge graph processing method according to claim 1, characterized in that, After determining the viewing time of each of the one or more videos watched by the user, and selecting one or more videos with a viewing time greater than or equal to a preset time threshold as the initial videos, the method further includes: Extract the video tags from the initial video, and match the video tags of the initial video with the knowledge tags in the knowledge video tag graph to determine the knowledge tags that match the video tags of the initial video; The recommended videos corresponding to the matched knowledge tags are obtained from the video library, and the recommendation format of the recommended videos is determined based on the matched knowledge tags; The recommended video is recommended to the user according to the recommended format.

6. The knowledge graph processing method according to claim 5, characterized in that, The step of matching the video tags of the initial video with the knowledge tags in the knowledge video tag graph includes: Determine whether the video tags of the initial video match the tags in the preset learning tag library. If so, then match the video tags of the initial video with the knowledge tags in the knowledge video tag graph. If not, then the preset recommended videos corresponding to the initial video will be recommended to the user in the form of a list.

7. The knowledge graph processing method according to any one of claims 5 or 6, characterized in that, The recommended format is a Rubik's Cube recommendation format; Accordingly, recommending the recommended video to the user according to the recommendation format includes: Determine the recommended video corresponding to each matched knowledge tag, and display the recommended video corresponding to each matched knowledge tag on one side of the magic cube to recommend it to the user.

8. The knowledge graph processing method according to claim 5, characterized in that, After recommending the recommended video to the user according to the recommendation format, the method further includes: Based on the user's personal knowledge graph, it is determined whether the recommended video is a video that the user has already watched. If so, if the recommended video is determined to be the initial video, the first annotation information of the recommended video in the user's personal knowledge graph is displayed; or if the recommended video is determined to be the target video, the second annotation information of the recommended video in the user's personal knowledge graph is displayed. If not, after receiving the user's instruction to open the recommended video, the recommended video will be opened for knowledge learning.

9. The knowledge graph processing method according to any one of claims 1-4, characterized in that, The step of extracting one or more video tags from the one or more target videos and building or updating a personal knowledge graph for the user based on the one or more video tags includes: Extract one or more video tags from the one or more target videos, and the viewing position of each video in the one or more target videos viewed by the user; A personal knowledge graph is created or updated for the user based on the viewing location and one or more video tags.

10. The knowledge graph processing method according to claim 1, characterized in that, The knowledge video tag map includes multiple categories, each category corresponds to multiple knowledge tags, and each knowledge tag corresponds to multiple knowledge points; Accordingly, after establishing the knowledge video tag map based on the knowledge tags of the knowledge video according to the preset knowledge tag levels, the method further includes: Receive uploaded videos, and if the uploaded videos meet preset recommendation conditions, obtain the video tags of the uploaded videos; If the video tag of the uploaded video belongs to an existing category of knowledge tag in the knowledge video tag graph, determine whether the video tag of the uploaded video belongs to the knowledge point corresponding to the knowledge tag in the knowledge video tag graph. If so, the uploaded video will be added to the video queue corresponding to the knowledge point. If not, a knowledge point is created in the knowledge video tag graph based on the video tag of the uploaded video, and the uploaded video is added to the video queue corresponding to the knowledge point.

11. A knowledge graph processing device, characterized in that, include: The acquisition module is configured to acquire one or more videos viewed by the user and the user's characteristic information; The initial video determination module is configured to determine the viewing time of each of the one or more videos watched by the user, and to select one or more videos whose viewing time is greater than or equal to a preset time threshold as the initial videos; The target video determination module is configured to determine one or more target videos corresponding to the user based on the attribute information of the one or more initial videos and the user's feature information; The knowledge graph processing module is configured to extract one or more video tags from the one or more target videos, and to build or update a personal knowledge graph for the user based on the one or more video tags. The personal knowledge graph is displayed in the form of a radar chart, and each point of the radar chart displays the name of a secondary video tag in the knowledge video tag graph. Clicking on the name area will take you to the corresponding knowledge consumption square, which has a hierarchical structure. The device further includes: The tag determination module is configured to acquire knowledge videos from the video library and determine the knowledge tags of the knowledge videos; The knowledge video tag building module is configured to build a knowledge video tag map based on a preset knowledge tag level, wherein the knowledge video tag map is a tree-structured knowledge video tag map.

12. A computing device, comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions, which, when executed by the processor, implement the steps of the knowledge graph processing method according to any one of claims 1-10.

13. A computer-readable storage medium storing computer instructions that, when executed by a processor, implement the steps of the knowledge graph processing method according to any one of claims 1-10.

14. A computer program product comprising computer instructions, characterized in that, When executed by a processor, the computer instructions implement the steps of the knowledge graph processing method according to any one of claims 1-10.

Images