Method for annotating contents of a label

Abstract

The application discloses a label content marking method. The method comprises the following steps: displaying target content to be marked on an operation interface; if the operation interface detects at least one target label selected from a label library, a determination instruction is generated, wherein the determination instruction carries the at least one target label, and the target label belongs to a target label system; in response to the determination instruction, an original label matched with the target label is determined based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system; and it is determined that material marked by the original label in the target content is material that needs to be marked by the target label. The application solves the technical problem that the label generation efficiency is low due to the fact that the label systems of different platforms are different and customized development needs to be carried out for different label systems.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of label annotation, in particular to a label content annotation method. BACKGROUND

[0002] In the current network, there are various types of functional platforms, which are a set of label systems that meet their own attributes built in their respective implementation scenarios for a long time. Therefore, the entity label systems of different platforms are often different. For example, for an e-commerce platform, the entity label is the hierarchical division of various types of goods, and for a short video platform, the entity label is usually based on the hierarchical division of video themes. In the face of different platform label annotation needs, the provider of entity labels needs to face the corresponding customized development of different platform label needs, resulting in high cost and low efficiency of label output.

[0003] In the related art, different platforms have different label systems, and customized development needs to be carried out for different label systems, resulting in low label generation efficiency. Currently, there is no effective solution to this problem. SUMMARY

[0004] The embodiments of the present application provide a label content annotation method to at least solve the technical problem of different label systems of different platforms in the related art, which requires customized development for different label systems, resulting in low label generation efficiency.

[0005] According to an aspect of an embodiment of the present application, a label content annotation method is provided, comprising: displaying target content to be annotated on an operation interface; if the operation interface detects that at least one target label in the label library is selected, a determination instruction is generated, wherein the determination instruction carries at least one target label, and the target label belongs to a target label system; in response to the determination instruction, an original label matching the target label is determined based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine the mapping relationship between a first entity in the original label system and a second entity in the target label system; and determining that the material annotated by the original label in the target content is the material that needs to be annotated by the target label.

[0006] According to an aspect of the embodiments of the present application, a method for labeling label content is provided, comprising: obtaining target content to be labeled; calling a selected target label from a label library, wherein the target label belongs to a target label system; determining a primary label matched with the target label based on a label knowledge graph, wherein the primary label belongs to an entity of a primary label system, and the label knowledge graph is used to record a mapping relationship between a first entity in the primary label system and a second entity in the target label system; labeling a material labeled with the primary label from the target content based on the primary label matched with the target label; and performing content labeling on the material labeled with the primary label by using the target label.

[0007] According to another aspect of the embodiments of the present application, a method for labeling label content is also provided, comprising: receiving at least one target label and target content input by a user through a client by a cloud server, wherein the target label belongs to a target label system; determining a primary label matched with the target label based on a label knowledge graph by the cloud server, wherein the primary label belongs to a primary label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the primary label system and a second entity in the target label system; determining, by the cloud server, that a material labeled with the primary label in the target content is a material needing to be labeled by the target label; and returning, by the cloud server, the material needing to be labeled by the target label to the client.

[0008] According to another aspect of the embodiments of the present application, a method for labeling label content is also provided, comprising: receiving at least one target label and target content input by a user through a client by a cloud server, wherein the target label belongs to a target label system; determining a primary label matched with the target label based on a label knowledge graph by the cloud server, wherein the primary label belongs to a primary label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the primary label system and a second entity in the target label system; determining, by the cloud server, that a material labeled with the primary label in the target content is a material needing to be labeled by the target label; and returning, by the cloud server, the material needing to be labeled by the target label to the client.

[0009] According to another aspect of the embodiments of the present application, a method for labeling label content is also provided, comprising: receiving at least one target label and target content input by a user through a client by a cloud server, wherein the target label belongs to a target label system; determining a primary label matched with the target label based on a label knowledge graph by the cloud server, wherein the primary label belongs to a primary label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the primary label system and a second entity in the target label system; determining, by the cloud server, that a material labeled with the primary label in the target content is a material needing to be labeled by the target label; and returning, by the cloud server, the material needing to be labeled by the target label to the client.

[0010] According to another aspect of the embodiments of the present application, a method for labeling label content is also provided, comprising: receiving at least one course theme label and teaching material input by a user through a client, wherein the course theme label belongs to a target label system of an online education platform; determining a primary label matched with the course theme label based on a label knowledge graph, wherein the primary label belongs to a primary label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the primary label system and a second entity in the target label system; determining teaching material in the teaching material that is labeled by the primary label as material that needs to be labeled by the course theme label; and returning the material that needs to be labeled by the course theme label to the client.

[0011] According to another aspect of the embodiments of the present application, a method for labeling label content is also provided, comprising: receiving at least one course theme label and teaching material input by a user through a client, wherein the course theme label belongs to a target label system of an online education platform; determining a primary label matched with the course theme label based on a label knowledge graph, wherein the primary label belongs to a primary label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the primary label system and a second entity in the target label system; determining teaching material in the teaching material that is labeled by the primary label as material that needs to be labeled by the course theme label; and returning the material that needs to be labeled by the course theme label to the client.

[0012] According to another aspect of the embodiments of the present application, a method for labeling label content is also provided, comprising: receiving at least one course theme label and teaching material input by a user through a client, wherein the course theme label belongs to a target label system of an online education platform; determining a primary label matched with the course theme label based on a label knowledge graph, wherein the primary label belongs to a primary label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the primary label system and a second entity in the target label system; determining teaching material in the teaching material that is labeled by the primary label as material that needs to be labeled by the course theme label; and returning the material that needs to be labeled by the course theme label to the client.

[0013] In the embodiment of the present application, the target content to be labeled is displayed on an operation interface; if the operation interface detects that at least one target label in a label library is selected, a determination instruction is generated, wherein the determination instruction carries the at least one target label, and the target label belongs to a target label system; in response to the determination instruction, a primary label matched with the target label is determined based on a label knowledge graph, wherein the primary label belongs to a primary label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the primary label system and a second entity in the target label system; and it is determined that the material labeled by the primary label in the target content is the material that needs to be labeled by the target label. The above scheme aims at the customized demand under different label systems, maintains its own primary label system, and then uses the knowledge graph to perform relationship reasoning through its own primary label capability, obtains the mapping relationship between the customized label and the self-owned label, and performs label labeling of different label systems based on the mapping relationship, so that it is not necessary to customize the development of each different label system, and the technical problem of low label generation efficiency caused by the different label systems of different platforms and the need for customized development of different label systems is solved. BRIEF DESCRIPTION OF DRAWINGS

[0014] The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the present application and serve to explain the present application. In the drawings:

[0015] Figure 1 Fig. 1 shows a hardware structure block diagram of a computing device (or mobile device) for implementing a labeling method of label content;

[0016] Figure 2 Fig. 2 is a flowchart of a labeling method of label content according to Embodiment 1 of the present application;

[0017] Figure 3 Fig. 3 is a schematic diagram of an operation interface of a label output platform according to Embodiment 1 of the present application;

[0018] Figure 4 Fig. 4 is a flowchart of a labeling method of label content according to Embodiment 2 of the present application;

[0019] Figure 5 Fig. 5 is a schematic diagram of a first entity and a second entity with a mapping relationship according to Embodiment 2 of the present application;

[0020] Figure 6 Fig. 6 is a schematic diagram of a plurality of first entities and second entities with a mapping relationship according to Embodiment 2 of the present application;

[0021] Figure 7 Fig. 7 is a schematic diagram of a combined mapping according to Embodiment 2 of the present application;

[0022] Figure 8 is a schematic diagram of a label marking according to an embodiment of the present application;

[0023] Figure 9 is a schematic diagram of a label marking device according to an embodiment of the present application;

[0024] Figure 10 is a schematic diagram of a label marking device according to an embodiment of the present application;

[0025] Figure 11 is a flowchart of a marking according to an embodiment of the present application;

[0026] Figure 12 is a schematic diagram of a label marking device according to an embodiment of the present application;

[0027] Figure 13 is a flowchart of a label content marking method according to an embodiment of the present application;

[0028] Figure 14 is a schematic diagram of a label marking device according to an embodiment of the present application;

[0029] Figure 15 is a flowchart of a label content marking method according to an embodiment of the present application;

[0030] Figure 16 is a schematic diagram of a label marking device according to an embodiment of the present application;

[0031] Figure 17 is a flowchart of a label content marking method according to an embodiment of the present application;

[0032] Figure 18 is a schematic diagram of a label marking device according to an embodiment of the present application;

[0033] Figure 19 is a flowchart of a label content marking method according to an embodiment of the present application;

[0034] Figure 20 is a schematic diagram of a label marking device according to an embodiment of the present application;

[0035] Figure 21 is a flowchart of a label content marking method according to an embodiment of the present application;

[0036] Figure 22 is a schematic diagram of a label marking device according to an embodiment of the present application;

[0037] Figure 23is a structural block diagram of a computing device according to an embodiment 15 of the present application. DETAILED DESCRIPTION

[0038] In order to make the personnel in the technical field better understand the present application scheme, the technical scheme in the embodiment of the present application will be clearly and completely described below in combination with the drawings in the embodiment of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor should belong to the scope of protection of the present application.

[0039] It should be noted that the terms "first", "second" and the like in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily have to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0040] First, some of the nouns or terms that appear in the description of the embodiments of the present application are applicable to the following explanations:

[0041] Label: an entity with specific meaning, mainly including characters, scenes, objects, events, signs, proper nouns, etc.

[0042] Knowledge graph: a network-like semantic network constructed by "entities" and "relationships" between entities, which structures data by embedding data into a graph to represent the association between entities.

[0043] Knowledge inference: obtaining new knowledge with the help of knowledge graph information.

[0044] BERT (Bidirectional Encoder Representation from Transformers): a deep learning method for extracting text feature vectors.

[0045] Graph Convolution Networks: A deep learning method mainly used to build the association and feature representation between "edges" and "nodes" on unstructured data.

[0046] Embodiment 1

[0047] According to the embodiments of the present application, an embodiment of a labeling method of label content is also provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0048] The method embodiment provided by the embodiment one of the present application can be executed in a mobile terminal, a computing device or similar computing device. Figure 1 A hardware structure block diagram of a computing device (or mobile device) for implementing the labeling method of label content is shown. As shown in the figure, Figure 1 The computing device 10 (or mobile device 10) can include one or more processors 102 (the processor 102 can include but not limited to a microprocessor MCU or a programmable logic device FPGA processing device), a memory 104 for storing data, and a transmission module 106 for communication function. In addition, it can also include a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which can be included as one of the ports of the I / O interface), a network interface, a power supply and / or a camera. Those skilled in the art can understand, Figure 1 The structure shown is only schematic, which does not limit the structure of the above-mentioned electronic device. For example, the computing device 10 can include more or less components than those shown in Figure 1 or have a different configuration than that shown in Figure 1 .

[0049] It should be noted that the one or more processors 102 and / or other data processing circuits described above can be referred to as "data processing circuits" herein. The data processing circuit can be embodied in whole or in part as software, hardware, firmware or any combination thereof. In addition, the data processing circuit can be a single independent processing module, or all or part of any one of the other elements combined into the computing device 10 (or mobile device). As referred to in the embodiments of the present application, the data processing circuit serves as a processor control (for example, selection of variable resistance terminal path connected with the interface).

[0050] The memory 104 can be used to store software programs of application software and modules, such as program instructions / data storage means corresponding to the labeling method of label content in the embodiments of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, implements the vulnerability detection method of the application program as described above. The memory 104 can include a high-speed random access memory, and can also include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some examples, the memory 104 can further include a memory remotely arranged with respect to the processor 102, which can be connected to the computing device 10 through a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0051] The transmission module 106 is used to receive or send data via a network. Specific examples of the above-mentioned network can include a wireless network provided by a communication provider of the computing device 10. In one example, the transmission module 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to be able to communicate with the Internet. In one example, the transmission module 106 can be a radio frequency (Radio Frequency, RF) module, which is used to communicate with the Internet in a wireless manner.

[0052] The display can be, for example, a touch screen type liquid crystal display (LCD), which can enable a user to interact with the user interface of the computing device 10 (or mobile device).

[0053] It should be noted that, in some optional embodiments, the above-mentioned Figure 1 The computer device (or mobile device) shown can include hardware elements (including circuitry), software elements (including computer code stored on a computer readable medium), or a combination of both hardware and software elements. It should be noted that, Figure 1 is only one example of a particular implementation and is intended to illustrate the types of components that can be present in the above-mentioned computer device (or mobile device).

[0054] In the above-mentioned operating environment, the present application provides a labeling method of label content as shown in Figure 2 . Figure 2 is a flowchart of a labeling method of label content according to Embodiment 1 of the present application.

[0055] Step S21, displaying the target content of the labeling method of label content and the label library on the operation interface.

[0056] The steps in this embodiment can be performed by a label output platform. The label output platform is used for labeling a plurality of types of multimedia content. For example, for video content, a labeling method of label content is needed to mark a specified part as a promotional use, i.e., the video content can be labeled according to the semantics of the specified part as a target label, and the corresponding output content can be obtained.

[0057] Specifically, the target content can be content uploaded by a user to the label output platform. It can be any one of video content, text content, audio content, picture content, etc.

[0058] The operation interface can be a human-computer interaction interface provided by the label output platform for the user. The operation interface is used to display the target content to be labeled and a label library. The user selects at least one label from the label library to label the target content. The label library is a label library of a specified function platform, which is determined according to the labeling requirements. For example, the user needs to apply the labeling result to an e-commerce platform, and the label library is the label library of the e-commerce platform. For another example, the user needs to apply the labeling result to a news platform, and the label library is the label library of the news platform. For another example, the user needs to apply the labeling result to a short video platform, and the label library is the label library of the short video platform.

[0059] Figure 3 is a schematic diagram of an operation interface of a label output platform according to Embodiment 1 of the present application. As shown in Figure 3 The operation interface 30 includes a content display area 31 and a label input area 32. The user can directly input the target label to be labeled from the target content in the label input area 32, or click the downward triangle symbol 321 provided in the label input area 32 to call up the drop-down menu 322 and select the target content from the drop-down menu 322.

[0060] In step S23, if the operation interface detects that at least one target label in the label library is selected, a determination instruction is generated. The determination instruction carries at least one target label, and the target label belongs to a target label system.

[0061] Specifically, the label system is composed of entities with a preset hierarchical relationship, and each entity represents a label. The target label system is the label system composed of the label library.

[0062] The user can select at least one target label in the label library in a specified area of the operation interface. The target label is the label to be labeled from the target content.

[0063] In an optional embodiment, in combination with Figure 3As shown, when the user inputs the label "fire" in the label input area 32, if it is determined that the label library contains this label, it is determined that the target label selected by the user is "fire". Then a determination instruction can be generated based on this label, and the label output platform receives the determination instruction, and performs label labeling according to the target label carried in the determination instruction.

[0064] In step S25, in response to the determination instruction, the original label matching the target label is determined based on the label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine the mapping relationship between the first entity in the original label system and the second entity in the target label system.

[0065] Specifically, the original label system described above is the label system of the label output platform itself, which can be constructed according to the preset multimedia resources, and does not depend on any other functional platform.

[0066] The original label system described above includes a plurality of first entities having a hierarchical relationship, and the target label system includes a plurality of second entities having a hierarchical relationship, wherein the first entity and the second entity have a mapping relationship based on the label knowledge graph, so that after the target label is determined, the original label matching the target label can be determined based on the preset relationship.

[0067] In an optional embodiment, the label output platform maintains its own original label system and a plurality of other label systems corresponding to a plurality of functional platforms, and a mapping table between the original label system and each other label system is constructed in advance to obtain a mapping table corresponding to each other label system. Each mapping table corresponding to each other label system includes a corresponding relationship between each first entity in the original label system and the second entity of the other label system. When the determination instruction is received, the target label system to which the target label belongs is determined, so that the mapping table corresponding to the target label system is obtained, and the target label is searched from the mapping table, so that the original label corresponding to the target label is obtained.

[0068] The label knowledge graph described above can be a knowledge graph corresponding to the original label system. Based on the label knowledge graph of the original label, knowledge reasoning is performed to obtain the mapping relationship between the first entity in the original label system and the second entity in the target label system.

[0069] It should be noted that the mapping relationship between the first entity and the second entity described above can be one-to-one, one-to-many or many-to-one, or combined mapping, that is, a combination of a plurality of first entities and a combination of a plurality of second entities have a mapping relationship.

[0070] In the above step, the label output platform starts to perform the labeling of the label in response to the determination instruction. It should be noted that in the scheme of the present application, the label output platform does not directly perform label determination instruction on the target content when performing the labeling of the label, but finds the original label matched with the target label, and then performs label labeling based on the found original label. Such a way can avoid customizing and developing the label output platform for each functional platform.

[0071] In step S27, it is determined that the material labeled by the original label in the target content is the material labeled by the target label.

[0072] Specifically, the material labeled by the original label can be obtained by the label output platform by labeling the target content through the label in the original label system.

[0073] For different types of target content, the display content labeled by the original label can be different. For text type target content, the display content labeled by the original label can be the text content corresponding to the target label; for video type target content, the display content labeled by the original label can be the start and end time of the video content corresponding to the target label; for target content containing multiple images, the display content labeled by the original label can be the identification of one or more images corresponding to the target label.

[0074] In an optional embodiment, taking the target label system as the label system of an e-commerce platform as an example, the target content is a large number of commodity images. The target label

women's clothing

dresses

skirts

dresses

skirts

[0075] In an optional embodiment, taking the target label system as the label system of a news platform as an example, the target content is a large number of news videos. The target label "fire" is obtained, and based on the target label, a plurality of original labels

fire situation

fire

fire situation

fire

[0076] Still in combination with Figure 3 It is shown that the display content corresponding to the target label can be displayed in 33 in the operation interface.

[0077] Therefore, the above embodiment of the present application displays the target content to be labeled on the operation interface. If the operation interface detects that at least one target label in the label library is selected, a determination instruction is generated, wherein the determination instruction carries the at least one target label, and the target label belongs to a target label system. In response to the determination instruction, a original label matching the target label is determined based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine the mapping relationship between a first entity in the original label system and a second entity in the target label system. It is determined that the material labeled by the original label in the target content is the material that needs to be labeled by the target label. The above scheme maintains its own original label system according to the customization demand of different label systems, and then uses the knowledge graph to perform relationship reasoning through its own original label capability to obtain the mapping relationship between the customized label and the self-owned label, and performs label labeling of different label systems based on the mapping relationship, thereby eliminating the need for customized development of each different label system, and solving the technical problem of low label generation efficiency caused by the difference between the label systems of different platforms and the need for customized development of different label systems.

[0078] As an optional embodiment, after determining that the material labeled by the original label in the target content is the material that needs to be labeled by the target label, the above method further includes: displaying the labeling result of the target content on the operation interface, wherein the labeling result is the content labeling of the material labeled by the original label using the target label; if the labeling result is incorrect labeling, calling an editing interface and displaying the editing interface on the operation interface; in response to an editing instruction, modifying the labeling result; and updating the mapping relationship between the first entity in the original label system and the second entity in the target label system based on the modified labeling result.

[0079] In the above scheme, the labeling result of the target label on the target content is displayed, and in the case that the labeling result is inaccurate, the editing interface can be used to update the labeling result. After adjusting the labeling result, the mapping relationship between the first entity in the original label system and the second entity in the target label system can also be adjusted according to the adjusted labeling result.

[0080] In an optional embodiment, taking the target label system as the label system of a news platform as an example, the target content is a large number of news videos. The target label is obtained, and based on the target label, a plurality of original labels corresponding to the target label are found. Video clips labeled by the plurality of original labels in the news videos are obtained, and these video clips are labeled as target labels. After that, the output video clips can be played, and the user checks whether the played video clips are accurate. In the case that the played video clips are inaccurate, the labeling result is updated through the editing interface.

[0081] In an alternative way, updating the labeling result can be adjusting the material labeled by the target label. Still in the above example, the user can enter the editing interface and in the editing interface, adjust the start and end time of the video segment on the news video timeline, thereby adjusting the video segment. After adjusting the video segment in the editing interface, the mapping relationship between the original label and the target label can be updated. For example, through the mapping relationship between the target label A and the original label B, a video segment is labeled as the target label A, but the user plays the video segment and finds that the video segment does not match the target label A, so it can be determined that the labeling result is wrong, and the mapping relationship between the target label A and the original label B can be removed.

[0082] In another alternative way, updating the labeling result can be adjusting the target label. Still in the above example, the user can enter the editing interface and in the editing interface, update the target label, and contact the mapping relationship between the original label and the original target label, and create the mapping relationship between the original label and the updated target label. For example, through the mapping relationship between the target label A and the original label B, a video segment is labeled as the target label A, but the user plays the video segment and finds that the video segment does not match the target label A, so it can be determined that the labeling result is wrong, and the target label A is changed to the target label C, and the mapping relationship between the target label A and the original label B is removed, and the mapping relationship between the target label C and the original label B is created.

[0083] As an optional embodiment, the labeling result is updated by inputting description information on the editing interface, and the updated target label is saved to the target label system.

[0084] Specifically, the above description information can be text information, voice information, or mouse operation information. The description information input to the editing interface updates the target label in the labeling result, and the updated target label is saved to the target label system.

[0085] In an alternative embodiment, still taking the example that the target label A and the original label B have a mapping relationship, the target content labeled with the original label B is labeled with the target label A, and in the case where it is determined that the labeling result is wrong, the target label A can be changed to a label C that matches the target content labeled with the original label B, and the label C is saved to the target label system, and the mapping relationship between the label C and the original label B is created.

[0086] As an optional embodiment, the original label system entity is combined based on the label knowledge graph to generate new label information, and the new label information is synchronized to the original label system.

[0087] The above scheme is used to perform knowledge reasoning within a preset domain in order to obtain new labels within that domain.

[0088] In one optional embodiment, candidate entities in a preset domain can be selected from the first entities within the original tagging system, and the first features of multiple candidate entities can be combined to form a joint representation. This can be a combination of candidate entities with connectivity relationships in a tagging knowledge graph. For example, taking the candidate entities "Obama" and "speech" as an example, there is a relationship "behavior" between "speech" and "Obama". Therefore, entities "Obama" and "speech" are taken as candidate entities, and their feature information is connected to obtain the aforementioned joint representation. Obtaining the corresponding new tag based on the joint representation can be achieved by predicting other entities corresponding to each candidate entity in the joint representation and combining these other entities to obtain a new tag. Again, taking the candidate entities "Obama" and "speech" as an example, based on the first entity "Obama", other entities "leaders" are predicted, and based on the candidate entity "speech", other entities "speaking" are predicted. Therefore, the new tag "leader speaking" can be obtained.

[0089] As an optional embodiment, if the similarity between the new tag information and the tag information already stored in the original tag system exceeds a threshold, the synchronization of the new tag information to the original tag system is terminated.

[0090] Adding all new tags to the original tag system would make it overly complex, and many first entity words in the original tag system might be too similar. Therefore, the above solution first checks the similarity between the new tag information and the existing tag information in the original tag system before adding new tags. If the similarity exceeds a threshold, the new tag is not added to the original tag system.

[0091] In one optional embodiment, a new tag "[Leader's Speech]" is generated, and the similarity between the new tag and other tags in the original tag system is obtained. If there are other tags whose similarity with the new tag "[Leader's Speech]" is greater than a threshold, then the addition of the new tag "[Leader's Speech]" to the original tag system is prohibited. If there are no other tags whose similarity with the new tag "[Leader's Speech]" is greater than the threshold, then the new tag "[Leader's Speech]" is added to the original tag system.

[0092] Example 2

[0093] This application provides, as follows: Figure 4 The labeling method shown. Figure 4 This is a flowchart of a labeling method according to Embodiment 2 of this application, the method including:

[0094] Step S41, obtaining target content to be labeled.

[0095] The steps in this embodiment can be performed by a label output platform. The label output platform is used for label annotation for various types of multimedia content. For example, for film and television content, the specified part needs to be labeled for promotion, i.e., the film and television content can be labeled according to the semantics of the specified part as a target label, i.e., the corresponding output content can be obtained.

[0096] Specifically, the above-mentioned target content to be labeled can be content uploaded by a user to the label output platform. It can be any one of video content, text content, audio content, picture content, etc.

[0097] Step S43, calling the selected target label from the label library, wherein the target label belongs to a target label system.

[0098] The label output platform can provide an operation interface for the user to select the target label. The above-mentioned label library is a label library of a specified function platform, which is determined according to the needs of label annotation. For example, the user needs to apply the annotation result to an e-commerce platform, and the label library is the label library of the above-mentioned e-commerce platform; for example, the user needs to apply the annotation result to a news platform, and the label library is the label library of the above-mentioned news platform; for example, the user needs to apply the annotation result to a short video platform, and the label library is the label library of the above-mentioned short video platform. The above-mentioned target label system is the label system constituted by the label library.

[0099] Step S45, determining a primary label matched with the target label based on a label knowledge graph, wherein the primary label belongs to an entity of a primary label system, and the label knowledge graph is used to record a mapping relationship between a first entity in the primary label system and a second entity in the target label system.

[0100] Specifically, the above-mentioned primary label system is the label system of the label output platform itself, which can be constructed according to the preset multimedia resources, and does not depend on any other function platform.

[0101] The above-mentioned primary label system includes a plurality of first entities having a hierarchical relationship, and the target label system includes a plurality of second entities having a hierarchical relationship, wherein the first entity and the second entity have a mapping relationship constructed based on the label knowledge graph, so that after the target label is determined, the primary label matched with the target label can be determined based on the preset relationship.

[0102] In an optional embodiment, the label output platform maintains its own original label system and a plurality of other label systems corresponding to a plurality of functional platforms, a mapping table between the original label system and each other label system is constructed in advance to obtain a mapping table corresponding to each other label system, and the mapping table corresponding to each other label system includes a corresponding relationship between each first entity in the original label system and a second entity of the other label system. When receiving the determination instruction, the target label is determined to belong to a target label system, so as to obtain the mapping table corresponding to the target label system, and then the target label is searched from the mapping table to obtain the original label corresponding to the target label.

[0103] The above label knowledge graph can be a knowledge graph corresponding to the original label system. Based on the label knowledge graph of the original label, knowledge reasoning is performed to obtain the mapping relationship between the first entity in the original label system and the second entity in the target label system.

[0104] It should be noted that the mapping relationship between the first entity and the second entity can be one-to-one, one-to-many or many-to-one, or combined mapping, that is, a combination of a plurality of first entities and a combination of a plurality of second entities have a mapping relationship.

[0105] Step S47, based on the original label matched with the target label, labeling the material labeled by the original label from the target content.

[0106] Step S49, using the target label to perform content labeling on the material labeled by the original label.

[0107] In the above steps, the target content is labeled in the target label system based on the mapping relationship between the target label and the original label.

[0108] It can be known that the above-mentioned embodiments of the present application obtain target content to be labeled; selected target labels are called from a label library, wherein the target labels belong to a target label system; original labels matched with the target labels are determined based on a label knowledge graph, wherein the original labels belong to entities of an original label system, and the label knowledge graph is used to record mapping relationships between first entities in the original label system and second entities in the target label system; materials labeled with the original labels are labeled from the target content based on the original labels matched with the target labels; and the materials labeled with the original labels are content-labeled by using the target labels. The above-mentioned scheme aims at customized needs under different label systems, maintains its own original label system, and then uses a knowledge graph to perform relationship reasoning through its own original label capability, obtains a mapping relationship between customized labels and self-owned labels, and performs label labeling of different label systems based on the mapping relationship, so that it is not necessary to customize development for each different label system, and the technical problem of low label generation efficiency caused by the need to customize development for different label systems in different platforms of related technologies is solved.

[0109] As an optional embodiment, the target content to be labeled is displayed on an operation interface, and if it is detected that a target label is selected on the operation interface, determination of original labels matched with the target labels based on a label knowledge graph is triggered, and materials labeled with the original labels in the target content are content-labeled by using the target labels.

[0110] The above-mentioned operation interface can be a man-machine interaction interface provided by a label output platform for a user, and the operation interface is used to display target content to be labeled and a label library, and the user selects at least one label from the label library to label the target content. The above-mentioned label library is a label library of a designated function platform, and is specifically determined according to the needs of label labeling. For example, the user needs to apply the labeling result to an e-commerce platform, and the label library is the label library of the above-mentioned e-commerce platform; for another example, the user needs to apply the labeling result to a news platform, and the label library is the label library of the above-mentioned news platform; for still another example, the user needs to apply the labeling result to a short video platform, and the label library is the label library of the above-mentioned short video platform.

[0111] In an optional embodiment, in combination with Figure 3 As shown in the figure, when the user inputs the label "fire" in the label input area 32, if it is determined that the label library contains this label, it can be determined that the target label selected by the user is "fire". Then a determination instruction can be generated based on this label, and the label output platform can receive the determination instruction to perform label labeling according to the target label carried in the determination instruction.

[0112] As an optional embodiment, the method further comprises: determining a mapping relationship between a first entity in the original label system and a second entity in the target label system, which comprises: obtaining the original label system, and obtaining a corresponding label knowledge graph based on the original label system; determining the mapping relationship based on the label knowledge graph and the plurality of second entities in the target label system through a preset neural network model.

[0113] Specifically, the neural network model can be a graph convolutional neural network model. Determining the mapping relationship based on the label knowledge graph and the plurality of second entities in the target label system through a preset neural network model can be feature extraction on the label knowledge graph and the plurality of second entities in the target label system to obtain a feature extraction result, wherein the feature extraction result comprises a first feature corresponding to each first entity in the label knowledge graph and a second feature corresponding to each second entity in the target label system; inputting the feature extraction result into a graph convolutional neural network, and determining the mapping relationship based on an output parameter of the graph convolutional neural network, wherein the output parameter of the graph convolutional neural network is used to represent the similarity between the first entity in the label knowledge graph and the second entity in the target label system.

[0114] Specifically, the first entity in the label knowledge graph and the second entity in the target label system can be extracted by a BERT model.

[0115] The input parameter of the graph convolutional neural network is a feature matrix of nodes in a graph and an adjacency matrix representing the relationship between the nodes. In an optional embodiment, the first entity in the label knowledge graph and the second entity in the target label system are extracted by a BERT to obtain a first feature corresponding to the label knowledge graph and a second feature corresponding to the target label system, wherein the first feature comprises a feature matrix of each node representing a first entity in the label knowledge graph and an adjacency matrix representing the relationship between the nodes; and the second feature comprises a feature corresponding to each second entity in the target label system. The first feature is inputted into the graph convolutional neural network as a feature matrix and an adjacency matrix of the graph convolutional neural network, and the second feature is inputted into the graph convolutional neural network as a propagation parameter of the graph convolutional neural network for operation, so as to obtain an output parameter of the graph convolutional neural network. The output parameter can be a visualized image. In the output image, similar first entities and similar second entities are automatically clustered. The distance between the nodes representing the first entities and the nodes representing the second entities in the output image is the similarity between the first entities and the second entities. According to the similarity and a preset similarity threshold, the first entity and the second entity having a preset relationship can be determined.

[0116] The knowledge reasoning based on the label knowledge graph specifically refers to finding original label feature information or combined feature information of original labels in an original label system that matches a given new label system (i.e., the target label system described above). The above scheme performs feature extraction on the second entity and the label knowledge graph, and embeds them into a GCN to perform correlation calculation on related entity nodes.

[0117] As an optional embodiment, the extraction result is input into a graph convolutional neural network, and a mapping relationship is determined based on an output parameter of the graph convolutional neural network, including at least one of the following:

[0118] According to the output parameter, a first feature and a second feature with a similarity higher than a first preset value are determined, and it is determined that an original label represented by the first feature and a target label represented by the second feature have a mapping relationship.

[0119] According to the output parameter, a joint representation with a similarity higher than a second preset value and the second feature are obtained, wherein the joint representation is obtained by connecting a plurality of first features; edge link prediction is performed based on a plurality of first feature information and second feature information constituting the joint representation, to constitute a pointing relationship between a plurality of first entities represented by a plurality of first feature information and a second entity represented by the second feature information; and it is determined that the first entity and the pointed second entity have a mapping relationship.

[0120] The above scheme is used to obtain a mapping relationship between a first entity and a second entity by performing knowledge reasoning based on a label knowledge graph according to a graph convolutional neural network. In this scheme, two ways of determining the mapping relationship are provided, which are described below.

[0121] In the first way, a first entity and a second entity with a similarity higher than a first preset value are determined according to an output parameter, and it is determined that an original label represented by the first entity and a target label represented by the second entity have a mapping relationship. In an optional embodiment, the output image of the graph convolutional neural network described above is still taken as an example, and each node representing the first entity is taken as the center, and a circle with a radius representing the first preset value is drawn, and the second entity within the circle and the first entity as the center have a mapping relationship. In addition, each node representing the second entity can be taken as the center, and a circle with a radius representing the first preset value is drawn, and the first entity within the circle and the second entity as the center have a mapping relationship.

[0122] The above scheme can obtain the relevance of one first entity and one second entity. Figure 5 is a schematic diagram of a first entity and a second entity with a mapping relationship according to Embodiment 2 of the present application, combined with Figure 5As shown, the label

The Palace Museum

Forbidden City

The Palace Museum

Forbidden City

[0123] In the second mode, the joint representation can be obtained by connecting the matrices representing the feature information of the plurality of first entities. Based on the similarity between the joint representation and the feature information of the second entity, edge link prediction can be performed on the plurality of first entities and the second entity. The edge link prediction is used to predict the missing edges in the output image of the graph convolutional neural network. Through the edge link prediction, the pointing of the plurality of first entities is obtained. The target second entity pointed by the plurality of first entities is the second entity having a mapping relationship with the plurality of first entities, thereby completing the combined mapping.

[0124] The above scheme can obtain the relevance of the plurality of first entities and the second entity. Figure 6 is a schematic diagram of a plurality of first entities and a second entity having a mapping relationship according to Embodiment 2 of the present application, combined with Figure 6 As shown, the edges of the nodes of the first entities

Entertainment Information

News Information

Film and Television Information

Information

Information

[0125] As an optional embodiment, the step of determining the mapping relationship between the first entity in the original label system and the second entity in the target label system further comprises: selecting candidate entities of a preset field from the first entities in the original label system, and constructing a joint representation from the first features of the plurality of candidate entities; obtaining a new label corresponding to the joint representation based on the joint representation; selecting a new label with a matching degree greater than a third preset value from the joint representation, and adding the selected new label to the original label system; determining that any one candidate entity constituting the joint representation has a mapping relationship with the new label

[0126] Specifically, the above scheme is used for knowledge reasoning in a preset field to obtain a new label in the preset field.

[0127] In the above step, the candidate entities of a preset field are selected from the first entities in the original label system, and the joint representation is constructed from the first features of the plurality of candidate entities. The candidate entities having a connection relationship in the label knowledge graph can be combined. For example, taking the candidate entities

Obama

Speech

Speech

Obama

Behavior

Obama

Speech

Obama

Speech

[0128] Still in the above step, based on the joint representation, a corresponding new label is obtained, which can be other entities corresponding to each candidate entity in the predicted joint representation, and the other entities corresponding to each candidate entity are combined to obtain the new label. Still taking the candidate entities

Obama

speech

Obama

leader

speech

speech

leader speech

[0129] Based on the fact that multiple other entities can be predicted for each candidate entity, multiple new labels can be predicted according to one joint representation, but not all new labels obtained have a mapping relationship with the candidate entities constituting the joint representation, so the obtained new labels need to be screened.

[0130] Specifically, the feature information of the new label can be extracted, and the feature information of the new label and the joint label are jointly input into a graph convolutional neural network. The similarity between the new label and the multiple candidate labels constituting the joint representation is determined according to the output parameters of the graph convolutional neural network. Then, the similarity is compared with a third preset value. When the similarity is greater than the third preset value, it is determined that the new label and the candidate label constituting the joint representation have a mapping relationship.

[0131] The above scheme is used for knowledge reasoning in a specific field, which refers to the process of obtaining a new label using the mutual relationship between entities in a given field. First, the related candidate labels in the field are obtained, the joint representation between the first entities is matched to obtain similar labels, and then the effective combination mapping is obtained through screening. Figure 7 is a schematic diagram of a combination mapping according to Embodiment 2 of the present application, as shown in Figure 7 The entities

Obama

speech

leader speech

[0132] The above scheme adds the new label to the original label system, and the new label is backflowed, so that the original label system becomes more and more rich, achieving the purpose of continuously expanding the original label system.

[0133] As an optional embodiment, the original label system is obtained by: mining entities from a preset corpus; generating first entities in the original label system according to the mined entities; performing hierarchical division on the first entities to obtain the original label system.

[0134] In an optional embodiment, the entity can be mined from the preset corpus, and the standard word under the preset field is obtained; the mined entity is clustered according to semantics to obtain a clustering result; the semantics with an entity quantity greater than a preset value are screened from the clustering result, and the screened semantics are converted into standard words; the standard words and the standard words constitute first entities in the original label system, and the standard words and the standard words are hierarchically divided to obtain the original label system.

[0135] Specifically, the above-mentioned preset corpus is a preset text corpus, which can be obtained from a large text pool. The above-mentioned preset field is a specified field, and the preset neighborhood standard word is used to represent the standard word known in this field. For example, in the e-commerce field, women's clothing, men's clothing, and electrical appliances are all standard words.

[0136] Some of the entities mined from the corpus have similar semantics, so the mined entities are clustered based on semantics to obtain a clustering result. Then the entities with a small quantity in the clustering result can be discarded, and the entities with a large quantity in the clustering result can be retained. For the retained entities with a large quantity, the semantics thereof are converted into standard words.

[0137] Through the above steps, the standard words of the conventional field and the standard words of the preset field can be obtained, and the standard words and the standard words are hierarchically divided to construct the above-mentioned original label system.

[0138] As an optional embodiment, a label knowledge graph corresponding to the original label system is obtained, including: obtaining text data corresponding to the first entity in the original label system in the corpus; information extraction is performed on the text data to obtain triple information of the first entity, wherein the triple information includes: entity name, attribute between different first entities, and relationship between different first entities; and generating a label knowledge graph based on the triple information.

[0139] The above-mentioned step constructs the label knowledge graph of the original label system by obtaining the triple information between the plurality of first entities. Specifically, the first entity is taken as a node in the label knowledge graph, and according to the attribute between different first entities and the relationship between different first entities, each first entity can be taken as a node, and the attribute and the relationship can be taken as different edges, so as to construct the label knowledge graph of the original label system.

[0140] Figure 8 is a schematic diagram for label annotation according to an embodiment of the present application. In combination with Figure 8As shown, first, a original label system of a label output platform itself is established, and a label knowledge graph is constructed based on the original label system. Through the original label system and the label knowledge graph, knowledge reasoning can be performed to obtain a mapping relationship between the original label system and a target label system, and further, a new label with specific meaning can be obtained. The generated new label can also be returned to the source label system to expand the source label system. When performing label annotation, relationship reasoning mapping is performed based on the mapping relationship between the original label system and the target label system, so that customized development is not required.

[0141] It should be noted that, for the foregoing method embodiments, in order to simply describe, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited to the action sequence described, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification all belong to preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0142] From the above description of the embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be realized by means of software and a general hardware platform as required, and of course, it can also be realized by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a storage medium (such as a ROM / RAM, a magnetic disk, or an optical disk), and includes a plurality of instructions for causing a terminal device (which can be a mobile phone, a computer, a server, or a network device) to execute the method of each embodiment of the present application.

[0143] Embodiment 3

[0144] According to the embodiments of the present application, a label annotation device for executing the label annotation method in embodiment 1 is also provided, Figure 9 is a schematic diagram of a label annotation device according to embodiment 3 of the present application, as Figure 9 shown, the device 900 includes:

[0145] The display module 902 is configured to display the target content to be annotated on an operation interface.

[0146] The generation module 904 is configured to generate a determination instruction if the operation interface detects that at least one target label in the label library is selected, wherein the determination instruction carries the at least one target label, and the target label belongs to a target label system.

[0147] The response module 906 is configured to determine, in response to the determination instruction, a original label matching the target label based on the label knowledge graph, where the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system.

[0148] The determination module 908 is configured to determine that the material labeled by the original label in the target content is a material that needs to be labeled by the target label.

[0149] It should be noted that the display module 902, the generation module 904, the response module 906, and the determination module 908 correspond to steps S21 to S27 in Embodiment 1, and the four modules have the same instances and application scenarios as the corresponding steps, but are not limited to the content disclosed in Embodiment 1. It should be noted that the above modules can run in the computing device 10 provided in Embodiment 1 as part of the device.

[0150] As an optional embodiment, the apparatus further includes a display module configured to display a labeling result of the target content on the operation interface after determining that the material labeled by the original label in the target content is a material that needs to be labeled by the target label, where the labeling result is content labeling of the material labeled by the original label using the target label; a calling module configured to call an editing interface and display the editing interface on the operation interface if the labeling result is incorrect labeling; a response module configured to modify the labeling result in response to an editing instruction; and an updating module configured to update the mapping relationship between the first entity in the original label system and the second entity in the target label system based on the modified labeling result.

[0151] As an optional embodiment, the labeling result is updated by inputting description information on the editing interface, and the updated target label is saved to the target label system.

[0152] As an optional embodiment, the original label system entity is combined based on the label knowledge graph to generate new label information, and the new label information is synchronized to the original label system.

[0153] As an optional embodiment, if it is detected that the similarity between the new label information and the existing label information in the original label system exceeds a threshold, the new label information is terminated from being synchronized to the original label system.

[0154] Embodiment 4

[0155] According to the embodiment of the present application, a label marking device for marking the label in the marking method of the label in embodiment 2 is further provided, Figure 10 is a schematic diagram of a label marking device according to embodiment 4 of the present application, as shown in Figure 9 The device 1000 comprises:

[0156] The acquisition module 1002 is configured to acquire target content to be marked.

[0157] The calling module 1004 is configured to call a selected target label from a label library, wherein the target label belongs to a target label system.

[0158] The determination module 1006 is configured to determine a source label matched with the target label based on a label knowledge graph, wherein the source label belongs to an entity of a source label system, and the label knowledge graph is used to record a mapping relationship between a first entity in the source label system and a second entity in the target label system.

[0159] The first marking module 1008 is configured to mark out material marked with the source label from the target content based on the source label matched with the target label.

[0160] The second marking module 11010 is configured to mark the material marked with the source label in content with the target label.

[0161] It should be noted that the acquisition module 1002, the calling module 1004, the determination module 1006, the first marking module 1008 and the second marking module 11010 correspond to steps S41 to S49 in embodiment 2, and the five modules have the same instances and application scenarios as the corresponding steps, but are not limited to the content disclosed in the above embodiment. It should be noted that the above modules as part of the device can run in the computing device 10 provided in embodiment 1.

[0162] As an optional embodiment, the target content to be marked is displayed on an operation interface, and if it is detected that a target label is selected on the operation interface, the determination of the source label matched with the target label based on the label knowledge graph is triggered, and the material marked with the source label in the target content is marked in content with the target label.

[0163] As an optional embodiment, the apparatus further comprises a second determining module configured to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system, wherein the second determining module comprises: a first obtaining sub-module configured to obtain the original label system and obtain a corresponding label knowledge graph based on the original label system; and a first determining sub-module configured to determine the mapping relationship based on the label knowledge graph and the plurality of second entities in the target label system by using a preset neural network model.

[0164] As an optional embodiment, the second determining module further comprises: a first screening sub-module configured to screen out candidate entities of a preset field from the first entities in the original label system, and form a joint representation of first features of the plurality of candidate entities; a second obtaining sub-module configured to obtain corresponding new labels based on the joint representation; a second screening sub-module configured to screen out new labels having a matching degree greater than a third preset value with the joint representation, and add the screened new labels to the original label system; and a second determining sub-module configured to determine that any one of the candidate entities constituting the joint representation has a mapping relationship with the new labels.

[0165] As an optional embodiment, the first obtaining sub-module comprises: a mining unit configured to mine entities from a preset corpus; a generating unit configured to generate the first entities in the original label system according to the mined entities; and a dividing unit configured to perform hierarchical division on the first entities to obtain the original label system.

[0166] Embodiment 5

[0167] The present application provides a label content annotation method as shown in Figure 11 . Figure 11 is a flowchart of a label content annotation method according to Embodiment 5 of the present application, in combination with Figure 11 , the method comprises:

[0168] In step S111, the cloud server receives at least one target label and target content input by the user through the client, wherein the target label belongs to the target label system.

[0169] The steps in this embodiment can be executed by the cloud server of the label output platform. The label output platform is used for label annotation of various types of multimedia content. For example, for film and television content, the specified part needs to be annotated as promotional use, that is, the specified part can be annotated as target label according to the semantic content, and the corresponding output content can be obtained.

[0170] The label output platform can provide an operation interface for a user to input at least one target label and target content. The operation interface can be a man-machine interface provided by the label output platform for the user, which is used to display target content to be labeled and a label library from which the user selects at least one label to label the target content. The label library is a label library of a specified function platform, which is determined according to the labeling requirements. For example, if the user needs to apply the labeling result to an e-commerce platform, the label library is the label library of the e-commerce platform; for another example, if the user needs to apply the labeling result to a news platform, the label library is the label library of the news platform; for another example, if the user needs to apply the labeling result to a short video platform, the label library is the label library of the short video platform.

[0171] Specifically, the label system is composed of entities with a preset hierarchical relationship, each of which represents a label. The target label system is a label system composed of the label library.

[0172] In step S112, the cloud server determines a source label matched with the target label based on the label knowledge graph, wherein the source label belongs to a source label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the source label system and a second entity in the target label system.

[0173] Specifically, the source label system is a label system of the label output platform itself, which can be constructed according to a preset multimedia resource and does not depend on any other function platform.

[0174] The source label system includes a plurality of first entities with a hierarchical relationship, and the target label system includes a plurality of second entities with a hierarchical relationship, wherein the first entity and the second entity have a mapping relationship constructed based on the label knowledge graph, so that after the target label is determined, the source label matched with the target label can be determined based on the preset relationship.

[0175] In an optional embodiment, the label output platform maintains its own source label system and a plurality of other label systems corresponding to a plurality of function platforms, a mapping table between the source label system and each other label system is constructed in advance to obtain a mapping table corresponding to each other label system, and the mapping table corresponding to each other label system includes a corresponding relationship between each first entity in the source label system and a second entity of the other label system. When the determination instruction is received, the target label system to which the target label belongs is determined, so that the mapping table corresponding to the target label system is obtained, and the target label is searched from the mapping table to obtain the source label corresponding to the target label.

[0176] The above label knowledge graph can be a knowledge graph corresponding to the original label system. Based on the label knowledge graph of the original label, knowledge reasoning is performed, and a mapping relationship between the first entity in the original label system and the second entity in the target label system is obtained.

[0177] It should be noted that the mapping relationship between the first entity and the second entity can be one-to-one, one-to-many, or many-to-one, or combined mapping, that is, a combination of multiple first entities and a combination of multiple second entities have a mapping relationship.

[0178] In the above step, the label output platform starts to perform the labeling of the label in response to the determination instruction. It should be noted that in the scheme of the present application, the label output platform does not directly label the target content based on the target label when performing the labeling of the label, but finds the original label matched with the target label, and then performs the labeling based on the found original label. Such a way can avoid customizing and developing the label output platform for each function platform.

[0179] Step S113, the cloud server determines that the material labeled by the original label in the target content is a material that needs to be labeled by the target label.

[0180] Specifically, the display content labeled by the original label can be obtained by the label output platform by labeling the target content with the label in the original label system.

[0181] For different types of target content, the display content obtained by the original label may be different. For text type target content, the display content obtained by the original label can be text content corresponding to the target label; for video type target content, the display content obtained by the original label can be the start and end time of the video content corresponding to the target label; for target content containing multiple images, the display content obtained by the original label can be the identification of one or more images corresponding to the target label.

[0182] Step S114, the cloud server returns the material that needs to be labeled by the target label to the client.

[0183] According to the above, the cloud server in the above embodiment receives at least one target label and target content input by a user through a client, wherein the target label belongs to a target label system; the cloud server determines a original label matched with the target label based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system; the cloud server determines that the display content corresponding to the original label in the target content is the display content corresponding to the target label; and the cloud server returns the display content corresponding to the target label to the client. The above scheme maintains its own original label system according to the customization demand under different label systems, and then uses the knowledge graph to perform relationship reasoning through its own original label capability, obtains the mapping relationship between the customized label and the self-owned label, and performs label annotation of different label systems based on the mapping relationship, so that it is not necessary to customize the development of each different label system, and the technical problem of low label generation efficiency caused by the different label systems of different platforms and the need for customized development of different label systems is solved.

[0184] Embodiment 6

[0185] According to the embodiments of the application, a label annotation device for performing the label annotation method in Embodiment 5 is further provided, Figure 12 is a schematic diagram of a label annotation device according to Embodiment 6 of the application, as Figure 12 shown, the device 1200 comprises:

[0186] A receiving module 1202 is configured to receive, by a cloud server, at least one target label and target content input by a user through a client, wherein the target label belongs to a target label system.

[0187] A first determining module 1204 is configured to determine, by the cloud server, an original label matched with the target label based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system.

[0188] A second determining module 1206 is configured to determine, by the cloud server, that the material labeled by the original label in the target content is the material that needs to be labeled by the target label.

[0189] A returning module 1208 is configured to return, by the cloud server, the material that needs to be labeled by the target label to the client.

[0190] Embodiment 7

[0191] The application provides a label content annotation method as Figure 13 shown. Figure 13is a flowchart of a method for labeling a label content according to the embodiment 7 of the present application, combined with Figure 13 As shown in the figure, the method comprises the following steps:

[0192] In step S131, at least one target label and a news to be labeled are received by a user through a client, wherein the target label belongs to a target label system of a preset news platform.

[0193] The embodiment can be applied to the field of news, and the steps can be executed by a cloud server of a label output platform. The label output platform is used for labeling labels for various types of multimedia content.

[0194] The label output platform can provide an operation interface on the client for the user to input at least one target label and a news to be labeled. The operation interface can be a human-computer interaction interface provided by the label output platform for the user. The operation interface is used to display the news to be labeled for label labeling, and a label library. The user selects at least one label from the label library to label the news to be labeled. The label library is a label library of a specified function platform, which is determined according to the demand of label labeling. For example, the user needs to apply the labeling result to an e-commerce platform, and the label library is the label library of the e-commerce platform. For another example, the user needs to apply the labeling result to a news platform, and the label library is the label library of the news platform. For another example, the user needs to apply the labeling result to a short video platform, and the label library is the label library of the short video platform.

[0195] Specifically, the label system is composed of entities with a preset hierarchical relationship, and each entity represents a label. The target label system is the label system composed of the label library.

[0196] In step S132, the original label matching the target label is determined based on the label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine the mapping relationship between the first entity in the original label system and the second entity in the target label system.

[0197] Specifically, the original label system is the label system of the label output platform itself, which can be constructed according to the preset multimedia resources, and does not depend on any other function platform.

[0198] The original label system includes a plurality of first entities with a hierarchical relationship, and the target label system includes a plurality of second entities with a hierarchical relationship, wherein the first entity and the second entity have a mapping relationship based on the label knowledge graph, so that after the target label is determined, the original label matching the target label can be determined based on the preset relationship.

[0199] In an optional embodiment, the label output platform maintains its own original label system and a plurality of other label systems corresponding to a plurality of functional platforms, and a mapping table between the original label system and each other label system is constructed in advance to obtain a mapping table corresponding to each other label system, and the mapping table corresponding to each other label system includes the corresponding relationship between each first entity in the original label system and a second entity of the other label system. When the determination instruction is received, the target label is determined to belong to a target label system, so as to obtain the mapping table corresponding to the target label system, and then the target label is searched from the mapping table to obtain the original label corresponding to the target label.

[0200] The above label knowledge graph can be a knowledge graph corresponding to the original label system. Based on the label knowledge graph of the original label, knowledge reasoning is performed to obtain the mapping relationship between the first entity in the original label system and the second entity in the target label system.

[0201] It should be noted that the mapping relationship between the first entity and the second entity can be one-to-one, one-to-many or many-to-one, or combined mapping, that is, a combination of a plurality of first entities and a combination of a plurality of second entities have a mapping relationship.

[0202] In the above step, the label output platform starts to perform the labeling of the label in response to the determination instruction. It should be noted that in the scheme of the present application, the label output platform does not directly label the news to be labeled based on the target label, but finds the original label matching the target label, and then performs label labeling based on the found original label. Such a way can avoid customizing development of the label output platform for each functional platform.

[0203] Step S133, determining that the material labeled by the original label in the news to be labeled is the material that needs to be labeled by the target label.

[0204] Step S134, returning the material that needs to be labeled by the target label to the client.

[0205] In an optional embodiment, taking the target label system as the label system of the news platform as an example, and the target content is a large number of news videos. The target label "fire" is obtained, and a plurality of original labels corresponding to the target label are found, such as

fire situation

fire

fire situation

fire

[0206] According to the above, the above-mentioned embodiments of the present application receive at least one target label and a news to be labeled input by a user through a client, wherein the target label belongs to a target label system of a preset news platform; an original label matched with the target label is determined based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system; it is determined that material labeled by the original label in the news to be labeled is material that needs to be labeled by the target label; and the client is returned with the material that needs to be labeled by the target label. The above-mentioned scheme aims at customized needs under different label systems, maintains its own original label system, and then uses a knowledge graph to perform relationship reasoning through its own original label capability, obtains a mapping relationship between a customized label and a self-owned label, and performs label labeling of different label systems based on the mapping relationship, thereby eliminating the need for customized development for different label systems, and solving the technical problem of low label generation efficiency caused by different label systems of different platforms and the need for customized development for different label systems.

[0207] Embodiment 8

[0208] According to the embodiments of the present application, a label labeling device for performing the label labeling method in embodiment 7 is also provided, Figure 14 is a schematic diagram of a label labeling device according to the embodiments of the present application, as Figure 12 shown, the device 1400 comprises:

[0209] The receiving module 1402 is configured to receive at least one target label and a news to be labeled input by a user through a client, wherein the target label belongs to a target label system of a preset news platform.

[0210] The first determining module 1404 is configured to determine an original label matched with the target label based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system.

[0211] The second determining module 1406 is configured to determine that material labeled by the original label in the news to be labeled is material that needs to be labeled by the target label.

[0212] The returning module 1408 is configured to return the client with the material that needs to be labeled by the target label.

[0213] Embodiment 9

[0214] The present application provides a label content labeling method as Figure 15 shown. Figure 15 is a flowchart of a label content labeling method according to the embodiments of the present application, combined withFigure 15 The method comprises the following steps:

[0215] In step S151, at least one target label and product description information input by a user through a client are received, wherein the target label belongs to a target label system of a preset e-commerce platform.

[0216] The embodiment can be applied to the e-commerce field, and the steps can be executed by a cloud server of a label output platform. The label output platform is used for labeling labels for various types of multimedia content.

[0217] The label output platform can provide an operation interface on the client for the user to input at least one target label and product description information. The operation interface can be a human-computer interaction interface provided by the label output platform for the user. The operation interface is used to display product description information to be labeled and a label library. The user selects at least one label from the label library to label the product description information. The label library is a label library of a specified function platform, which is determined according to the labeling requirements. For example, if the user needs to apply the labeling result to an e-commerce platform, the label library is the label library of the e-commerce platform. For another example, if the user needs to apply the labeling result to a news platform, the label library is the label library of the news platform. For another example, if the user needs to apply the labeling result to a short video platform, the label library is the label library of the short video platform.

[0218] Specifically, the label system is composed of entities with a preset hierarchical relationship, and each entity represents a label. The target label system is the label system composed of the label library.

[0219] In step S152, an original label matching the target label is determined based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine the mapping relationship between a first entity in the original label system and a second entity in the target label system.

[0220] Specifically, the original label system is the label system of the label output platform itself, which can be constructed according to the preset multimedia resources and does not depend on any other function platform.

[0221] The original label system includes a plurality of first entities with a hierarchical relationship, and the target label system includes a plurality of second entities with a hierarchical relationship, wherein the first entity and the second entity have a mapping relationship based on the label knowledge graph, so that after the target label is determined, the original label matching the target label can be determined based on the preset relationship.

[0222] In an optional embodiment, the label output platform maintains its own original label system and a plurality of other label systems corresponding to a plurality of functional platforms, and a mapping table between the original label system and each other label system is constructed in advance to obtain a mapping table corresponding to each other label system, and the mapping table corresponding to each other label system includes the corresponding relationship between each first entity in the original label system and a second entity of the other label system. When the determination instruction is received, the target label is determined to belong to a target label system, so as to obtain the mapping table corresponding to the target label system, and then the target label is searched from the mapping table, so as to obtain the original label corresponding to the target label.

[0223] The above label knowledge graph can be a knowledge graph corresponding to the original label system. Based on the label knowledge graph of the original label, knowledge reasoning is performed, so as to obtain the mapping relationship between the first entity in the original label system and the second entity in the target label system.

[0224] It should be noted that the mapping relationship between the first entity and the second entity can be one-to-one, one-to-many or many-to-one, or combined mapping, that is, a combination of a plurality of first entities and a combination of a plurality of second entities have a mapping relationship.

[0225] In the above step, the label output platform starts to perform the labeling of the label in response to the determination instruction. It should be noted that in the scheme of the present application, the label output platform does not directly label the product description information based on the target label when performing the labeling of the label, but finds the original label matching the target label, and then performs the labeling of the label based on the found original label. Such a way can avoid customizing and developing the label output platform for each functional platform.

[0226] Step S153, determining that the material labeled by the original label in the product description information is the material that needs to be labeled by the target label.

[0227] Step S154, returning the material that needs to be labeled by the target label to the client.

[0228] In an optional embodiment, taking the target label system of an e-commerce platform as an example, the target content is a large number of product images. The target label

women's clothing

dresses

skirts

dresses

skirts

[0229] It can be learned from the above that the above-mentioned embodiment of the present application receives at least one target label and product description information input by a user through a client, wherein the target label belongs to a target label system of a preset e-commerce platform; determines a original label matched with the target label based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system; determines that material annotated by the original label in the product description information is material that needs to be annotated by the target label; and returns the material that needs to be annotated by the target label to the client. The above-mentioned scheme maintains its own original label system for customized needs under different label systems, and then uses a knowledge graph to perform relationship reasoning through its own original label capability, obtains a mapping relationship between a customized label and a self-owned label, and performs label annotation of different label systems based on the mapping relationship, so that it is not necessary to customize development for each different label system, and the technical problem of low label generation efficiency caused by the need for customized development for different label systems due to different label systems of different platforms in the related art is solved.

[0230] Embodiment 10

[0231] According to the embodiments of the present application, a label annotation device for performing the label annotation method in embodiment 9 is further provided, Figure 16 which is a schematic diagram of a label annotation device according to embodiment 10 of the present application, as Figure 16 shown, the device 1600 comprises:

[0232] A receiving module 1602 is configured to receive at least one target label and product description information input by a user through a client, wherein the target label belongs to a target label system of a preset e-commerce platform.

[0233] A first determining module 1604 is configured to determine a original label matched with the target label based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system.

[0234] A second determining module 1606 is configured to determine that material annotated by the original label in the product description information is material that needs to be annotated by the target label.

[0235] A returning module 1608 is configured to return the material that needs to be annotated by the target label to the client.

[0236] Embodiment 11

[0237] The present application provides a label content annotation method as Figure 17 shown. Figure 17is a flowchart of a method for labeling a label content according to the embodiment 11 of the present application, combined with Figure 17 As shown in the figure, the method comprises the following steps:

[0238] In step S171, at least one course theme label and teaching material input by a user through a client are received, wherein the course theme label belongs to a target label system of an online education platform.

[0239] The embodiment can be applied to the field of education, and the steps can be executed by a cloud server of a label output platform. The label output platform is used for labeling a plurality of types of multimedia content.

[0240] The teaching material can be an education courseware, a teaching question bank, etc. The second entity in the target label system of the online education platform can be a course theme, i.e., the course theme label. The course theme label can be a knowledge point in teaching, etc.

[0241] The label output platform can provide an operation interface on the client for the user to input at least one course theme label and teaching material. The operation interface can be a human-computer interaction interface provided by the label output platform for the user. The operation interface is used to display teaching material to be labeled and a label library. The user selects at least one label from the label library to label the teaching material. The label library is a label library of a specified function platform, which is determined according to the labeling requirements. In the above scheme, the label library can be a label library composed of teaching knowledge points.

[0242] Specifically, the label system is composed of entities with a preset hierarchical relationship, and each entity represents a label. The target label system is a label system composed of the label library.

[0243] In step S172, a primary label matching the course theme label is determined based on a label knowledge graph, wherein the primary label belongs to a primary label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the primary label system and a second entity in the target label system.

[0244] Specifically, the primary label system is a label system of the label output platform itself, which can be constructed according to a plurality of multimedia resources, and does not depend on any other function platform.

[0245] The primary label system includes a plurality of first entities with a hierarchical relationship, and the target label system includes a plurality of second entities with a hierarchical relationship, wherein the first entity and the second entity have a mapping relationship constructed based on the label knowledge graph. Therefore, after the course theme label is determined, the primary label matching the course theme label can be determined based on the preset relationship.

[0246] In an optional embodiment, the original label system and a plurality of other label systems corresponding to a plurality of functional platforms are maintained in the label output platform, a mapping table between the original label system and each other label system is constructed in advance, and a mapping table corresponding to each other label system is obtained, the mapping table corresponding to each other label system includes a corresponding relationship between each first entity in the original label system and a second entity of the other label system. When the determination instruction is received, the target label system to which the target label belongs is determined, so as to obtain the mapping table corresponding to the target label system, and the target label is searched from the mapping table, so as to obtain the original label corresponding to the target label.

[0247] The above label knowledge graph can be a knowledge graph corresponding to the original label system. Based on the label knowledge graph of the original label, knowledge reasoning is performed, so as to obtain the mapping relationship between the first entity in the original label system and the second entity in the target label system.

[0248] It should be noted that the mapping relationship between the first entity and the second entity can be one-to-one, one-to-many or many-to-one, or combined mapping, that is, a combination of a plurality of first entities and a combination of a plurality of second entities have a mapping relationship.

[0249] In the above step, the label output platform starts to perform the labeling of the label in response to the determination instruction. It should be noted that in the scheme of the present application, the label output platform does not directly label the teaching material based on the target label when performing the labeling of the label, and finds the original label matched with the target label, and then performs the labeling of the label based on the found original label. Such a way can avoid customizing development of the label output platform for each functional platform.

[0250] Step S173, determining that the original label annotated material in the teaching material is the material that needs to be annotated with the course theme label.

[0251] Step S174, returning the material that needs to be annotated with the course theme label to the client.

[0252] In an optional embodiment, the target label system is the label system of an online education platform, and the teaching material is a large number of test question banks. The course theme label

trigonometric function

sine function

cosine function

sine function

cosine function

[0253] According to the above, the above-mentioned embodiments of the present application receive at least one course theme label and teaching material input by a user through a client, wherein the course theme label belongs to a target label system of an online education platform; an original label matched with the course theme label is determined based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system; teaching material labeled by the original label in the teaching material is determined as material that needs to be labeled by the course theme label; and the client is returned with the material that needs to be labeled by the course theme label. The above-mentioned scheme maintains its own original label system for customized needs under different label systems, and instead uses knowledge graph for relationship reasoning through its own original label capability to obtain a mapping relationship between customized labels and self-owned labels, and label labeling of different label systems is performed based on the mapping relationship, so that it is not necessary to customize development for each different label system, and the technical problem of low label generation efficiency caused by the need for customized development for different label systems of different platforms in the related art is solved.

[0254] Embodiment 12

[0255] According to the embodiments of the present application, a label labeling device for performing the label labeling method in embodiment 11 is also provided, Figure 18 is a schematic diagram of a label labeling device according to embodiment 12 of the present application, as Figure 18 shown, the device 1800 comprises:

[0256] The receiving module 1802 is configured to receive at least one course theme label and teaching material input by a user through a client, wherein the course theme label belongs to a target label system of an online education platform.

[0257] The first determining module 1804 is configured to determine an original label matched with the course theme label based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system.

[0258] The second determining module 1806 is configured to determine teaching material labeled by the original label in the teaching material as material that needs to be labeled by the course theme label.

[0259] The returning module 1808 is configured to return the client with the material that needs to be labeled by the course theme label.

[0260] Embodiment 13

[0261] The present application provides a label labeling method as Figure 19The method for labeling the label content shown. Figure 19 is a flowchart of a method for labeling label content according to embodiment 13 of the present application, in conjunction with Figure 19 The method comprises the following steps:

[0262] In step S191, at least one disease name label and medical record material input by a user through a client are received, wherein the disease name label belongs to a target label system of a medical platform.

[0263] The present embodiment can be applied in the medical field, and the steps can be executed by a cloud server of a label output platform. The label output platform is used for labeling labels for various types of multimedia content.

[0264] The medical record material can be a medical record or medication information, etc., and the disease name label can be a disease name, a drug name, etc.

[0265] The label output platform can provide an operation interface on the client for the user to input at least one disease name label and medical record material. The operation interface can be a human-computer interaction interface provided by the label output platform for the user, which is used to display the medical record material to be labeled and a label library, and the user selects at least one disease name label from the label library to label the medical record material. The label library is a label library of a specified function platform, which is determined according to the requirements of label labeling. In the above scheme, the label library can be a label library composed of disease names.

[0266] Specifically, the label system is composed of entities with a preset hierarchical relationship, and each entity represents a label. The target label system is the label system composed of the label library.

[0267] In step S192, an original label matching the disease name label is determined based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine the mapping relationship between a first entity in the original label system and a second entity in the target label system.

[0268] Specifically, the original label system is the label system of the label output platform itself, which can be constructed according to the preset multimedia resources, and does not depend on any other function platform.

[0269] The original label system includes a plurality of first entities with a hierarchical relationship, and the target label system includes a plurality of second entities with a hierarchical relationship, wherein the first entity and the second entity have a mapping relationship constructed based on the label knowledge graph, so that after the disease name label is determined, the original label matching the disease name label can be determined based on the preset relationship.

[0270] In an optional embodiment, the label output platform maintains its own original label system and a plurality of other label systems corresponding to a plurality of functional platforms, a mapping table between the original label system and each other label system is constructed in advance to obtain a mapping table corresponding to each other label system, and the mapping table corresponding to each other label system includes the corresponding relationship between each first entity in the original label system and a second entity of the other label system. When the determination instruction is received, the target label system to which the target label belongs is determined, so as to obtain the mapping table corresponding to the target label system, and the target label is searched from the mapping table, so as to obtain the original label corresponding to the target label.

[0271] The above label knowledge graph can be a knowledge graph corresponding to the original label system. Based on the label knowledge graph of the original label, knowledge reasoning is performed to obtain the mapping relationship between the first entity in the original label system and the second entity in the target label system.

[0272] It should be noted that the mapping relationship between the first entity and the second entity can be one-to-one, one-to-many or many-to-one, or combined mapping, that is, a combination of a plurality of first entities and a combination of a plurality of second entities have a mapping relationship.

[0273] In the above step, the label output platform starts to perform the labeling of the label in response to the determination instruction. It should be noted that in the scheme of the present application, the label output platform does not directly label the teaching material based on the target label when performing the labeling of the label, and finds the original label matching the target label, and then performs the labeling of the label based on the found original label. Such a way can avoid customizing development of the label output platform for each functional platform.

[0274] Step S193, determining that the original label annotated medical record material in the medical record material is the material that needs to be annotated with the disease name label.

[0275] Step S194, returning the material that needs to be annotated with the disease name label to the client.

[0276] In an optional embodiment, the target label system is the label system of a medical platform, and the case material is a large amount of case information. The disease name label

headache

migraine

neuralgia

migraine

neuralgia

[0277] According to the above, the above-mentioned embodiment of the present application receives at least one disease name label and medical record material input by a user through a client, wherein the disease name label belongs to a target label system of a medical platform; determines a original label matched with the disease name label based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system; determines that medical record material labeled by the original label in the medical record material is material that needs to be labeled by the disease name label; and returns the material that needs to be labeled by the disease name label to the client. The above-mentioned scheme maintains its own original label system according to the customization demand under different label systems, and then uses the knowledge graph to perform relationship reasoning through its own original label capability, obtains the mapping relationship between the customized label and the self-owned label, and performs label labeling of different label systems based on the mapping relationship, so that it is not necessary to customize the development of each different label system, and the technical problem of low label generation efficiency caused by the need to customize the development of different label systems of different platforms in the related art is solved.

[0278] Embodiment 14

[0279] According to the embodiments of the present application, a label labeling device for performing the label labeling method in embodiment 14 is further provided, Figure 20 is a schematic diagram of a label labeling device according to embodiment 14 of the present application, as Figure 20 shown, the device 2000 comprises:

[0280] A receiving module 2002 is configured to receive at least one disease name label and medical record material input by a user through a client, wherein the disease name label belongs to a target label system of a medical platform.

[0281] A first determining module 2004 is configured to determine a original label matched with the disease name label based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system.

[0282] A second determining module 2006 is configured to determine that medical record material labeled by the original label in the medical record material is material that needs to be labeled by the disease name label.

[0283] A returning module 2008 is configured to return the material that needs to be labeled by the disease name label to the client.

[0284] Embodiment 15

[0285] The present application provides a label labeling method as Figure 21A labeling method of label content. Figure 21 is a flowchart of a labeling method of label content according to Embodiment 15 of the present application, combined with Figure 21 As shown in the figure, the method comprises the following steps.

[0286] In step S211, at least one target label and multimedia content to be labeled are received by a user through a client, wherein the target label belongs to a target label system of an entertainment platform, and the multimedia content comprises at least one of the following: long video, short video and song audio.

[0287] The embodiment can be applied to the field of entertainment, and the steps can be executed by a cloud server of a label output platform. The label output platform is used for label labeling of various types of multimedia content.

[0288] The multimedia content can be video content (including long video and short video), audio content such as song or cross-talk, etc. The target label can be song name, movie name, star name, name in a movie or TV series, scene description in a movie or TV series, etc.

[0289] The label output platform can provide an operation interface on the client for the user to input at least one target label and multimedia content. The operation interface can be a human-computer interaction interface provided by the label output platform for the user, which is used to display multimedia content to be labeled and a label library, and the user selects at least one target label from the label library to label the multimedia content. The label library is a label library of a specified function platform, which is determined according to the needs of label labeling. In the above scheme, the label library can be a label library composed of entertainment materials.

[0290] Specifically, the label system is composed of entities with a preset hierarchical relationship, and each entity represents a label. The target label system is a label system composed of the label library.

[0291] In step S212, a target label matching the target label is determined based on a label knowledge graph, wherein the target label belongs to an original label system, and the label knowledge graph is used to determine the mapping relationship between a first entity in the original label system and a second entity in the target label system.

[0292] Specifically, the original label system is the label system of the label output platform itself, which can be constructed according to the preset multimedia resources, and does not depend on any other function platform.

[0293] The original label system includes a plurality of first entities having a hierarchical relationship, and the target label system includes a plurality of second entities having a hierarchical relationship. The first entities and the second entities have a mapping relationship based on the label knowledge graph. Therefore, after determining the disease name label, the original label matched with the disease name label can be determined based on the preset relationship.

[0294] In an optional embodiment, the label output platform maintains its own original label system and a plurality of other label systems corresponding to a plurality of functional platforms. A mapping table between the original label system and each other label system is constructed in advance to obtain a mapping table corresponding to each other label system. The mapping table corresponding to each other label system includes a corresponding relationship between each first entity in the original label system and a second entity of the other label system. When the determination instruction is received, the target label system to which the target label belongs is determined, so as to obtain the mapping table corresponding to the target label system. The target label can be obtained by searching the mapping table.

[0295] The label knowledge graph can be a knowledge graph corresponding to the original label system. Knowledge reasoning based on the label knowledge graph of the original label can obtain the mapping relationship between the first entity in the original label system and the second entity in the target label system.

[0296] It should be noted that the mapping relationship between the first entity and the second entity can be one-to-one, one-to-many, or many-to-one. Alternatively, the mapping relationship can be a combined mapping, i.e., a combination of a plurality of first entities and a combination of a plurality of second entities.

[0297] In the above step, the label output platform starts to perform the labeling of the label in response to the determination instruction. It should be noted that in the scheme of the present application, the label output platform does not directly label the multimedia content based on the target label when performing the labeling of the label. Instead, the original label matched with the target label is searched, and then the original label is used to perform the labeling. Such a way can avoid customizing and developing the label output platform for each functional platform.

[0298] In step S213, it is determined that the material labeled by the original label in the multimedia content is a material that needs to be labeled by the target label.

[0299] In step S214, the material that needs to be labeled by the target label is returned to the client.

[0300] In an optional embodiment, the target tag system is a tag system of an entertainment platform, and the multimedia content is a large amount of video and television content. A target tag is acquired, a plurality of original tags corresponding to the target tag are found based on the target tag, and the video and television content is tagged based on the original tag system of the tag output platform itself, so that the video and television content corresponding to the original tags such as and is output as the tagging result of the target tag.

[0301] It can be learned from the above that the above-mentioned embodiments of the present application receive at least one target tag and multimedia content to be tagged input by a user through a client, wherein the target tag belongs to a target tag system of an entertainment platform, and the multimedia content includes at least one of a long video, a short video, and a song audio; an original tag matching the target tag is determined based on a tag knowledge graph, wherein the original tag belongs to an original tag system, and the tag knowledge graph is used to determine a mapping relationship between a first entity in the original tag system and a second entity in the target tag system; material in the multimedia content tagged by the original tag is determined as material needing to be tagged by the target tag; and the material needing to be tagged by the target tag is returned to the client. The above-mentioned scheme maintains its own original tag system for customized needs under different tag systems, and then uses the knowledge graph to perform relationship reasoning through its own original tag capability to obtain a mapping relationship between a customized tag and a self-owned tag, and performs tag tagging of different tag systems based on the mapping relationship, so that it is not necessary to customize and develop each different tag system, thereby solving the technical problem of low tag generation efficiency caused by the need to customize and develop different tag systems for different platforms in the related art.

[0302] As an optional embodiment, the method further includes determining a mapping relationship between a first entity in the original tag system and a second entity in the target tag system, wherein the step includes: acquiring the original tag system, and acquiring a corresponding tag knowledge graph based on the original tag system; and determining the mapping relationship based on the tag knowledge graph and a plurality of second entities in the target tag system through a preset neural network model.

[0303] Embodiment 14

[0304] According to the embodiments of the present application, a tag tagging device for performing the tag tagging method in embodiment 14 is further provided, Figure 22 is a schematic diagram of a tag tagging device according to embodiment 14 of the present application, as Figure 22 shown, the device 2200 includes:

[0305] The receiving module 2202 is configured to receive at least one target label and multimedia content to be labeled input by a user through a client, wherein the target label belongs to a target label system of an entertainment platform, and the multimedia content includes at least one of the following: a long video, a short video and a song audio;

[0306] The first determining module 2204 is configured to determine a primary label matched with the target label based on a label knowledge graph, wherein the primary label belongs to a primary label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the primary label system and a second entity in the target label system.

[0307] The second determining module 2206 is configured to determine that a material labeled by the primary label in the multimedia content is a material that needs to be labeled by the target label.

[0308] The returning module 2208 is configured to return the material that needs to be labeled by the target label to the client.

[0309] As an optional embodiment, the apparatus further includes a third determining module configured to determine a mapping relationship between a first entity in the primary label system and a second entity in the target label system, wherein the third determining module includes: an obtaining sub-module configured to obtain the primary label system and obtain a corresponding label knowledge graph based on the primary label system; and a determining sub-module configured to determine the mapping relationship by using a preset neural network model based on the label knowledge graph and a plurality of second entities in the target label system.

[0310] Embodiment 15

[0311] The embodiments of the present application can provide a computing device, which can be any one of the computing devices in the computing device group. Alternatively, in the present embodiment, the computing device can also be replaced by a terminal device such as a mobile terminal.

[0312] Alternatively, in the present embodiment, the computing device can be located in at least one network device of a plurality of network devices of a computer network.

[0313] In the embodiment, the computing device can execute program codes of the following steps in the vulnerability detection method of the application: displaying target content to be labeled on an operation interface; if the operation interface detects that at least one target label in the label library is selected, generating a determination instruction, wherein the determination instruction carries the at least one target label, and the target label belongs to a target label system; in response to the determination instruction, determining a original label matched with the target label based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system; and determining that material labeled by the original label in the target content is material that needs to be labeled by the target label.

[0314] Optionally, Figure 23 is a structural block diagram of a computing device according to embodiment 15 of the application. As shown in the figure, the computing device A can include one or more (only one is shown in the figure) processors 232, a memory 234, and a peripheral interface 236. Figure 23

[0315] The memory can be used to store software programs and modules, such as program instructions / modules corresponding to the security vulnerability detection method in the embodiments of the application. The processor executes various functions and data processing by running the software programs and modules stored in the memory, that is, implements the system vulnerability attack detection method described above. The memory can include a high-speed random access memory, and can also include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some examples, the memory can further include a memory remotely arranged with respect to the processor, which can be connected to the terminal A through a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0316] The processor can call information and application programs stored in the memory through the transmission device to execute the following steps: displaying target content to be labeled on an operation interface; if the operation interface detects that at least one target label in the label library is selected, generating a determination instruction, wherein the determination instruction carries the at least one target label, and the target label belongs to a target label system; in response to the determination instruction, determining a original label matched with the target label based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system; and determining that material labeled by the original label in the target content is material that needs to be labeled by the target label.

[0317] ​Optionally, the processor can further execute program codes of the following steps: after determining that the material labeled by the original label in the target content is the material that needs to be labeled by the target label, displaying a labeling result of the target content on the operation interface, wherein the labeling result is a content labeling result of the material labeled by the original label using the target label; if the labeling result is an incorrect labeling result, calling an editing interface and displaying the editing interface on the operation interface; modifying the labeling result in response to an editing instruction; and updating a mapping relationship between a first entity in the original label system and a second entity in the target label system based on the modified labeling result.

[0318] Optionally, the processor can further execute program codes of the following steps: updating the labeling result by inputting description information on the editing interface, and saving the updated target label to the target label system.

[0319] Optionally, the processor can further execute program codes of the following steps: performing feature combination on the original label system entity based on the label knowledge graph, generating new label information, and synchronizing the new label information to the original label system.

[0320] Optionally, the processor can further execute program codes of the following steps: if it is detected that the similarity between the new label information and the existing label information in the original label system exceeds a threshold, terminating the synchronization of the new label information to the original label system.

[0321] By adopting the embodiment of the present application, a labeling scheme of label content is provided. The scheme displays target content to be labeled on an operation interface; if the operation interface detects that at least one target label in a label library is selected, a determination instruction is generated, wherein the determination instruction carries the at least one target label, and the target label belongs to a target label system; in response to the determination instruction, an original label that matches the target label is determined based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system; and it is determined that the material labeled by the original label in the target content is the material that needs to be labeled by the target label. The above scheme maintains its own original label system according to the customization demand of different label systems, and then uses the knowledge graph to perform relationship reasoning through its own original label capability, obtains a mapping relationship between a customized label and a self-owned label, and performs label labeling of different label systems based on the mapping relationship, so that it is not necessary to customize and develop each different label system, and the technical problem of low label generation efficiency caused by the difference between label systems of different platforms and the need for customization development of different label systems is solved.

[0322] Those skilled in the art can understand that, Figure 23 The structure shown is only schematic, and the computing device can also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a palm computer, a Mobile Internet Device (MID), a PAD, or the like. Figure 23 It does not limit the structure of the electronic device. For example, the computing device A can further include more or fewer components (such as a network interface, a display device, etc.) than those shown in the figure, or have a different configuration from that shown in the figure. Figure 23 Figure 23 It does not limit the structure of the electronic device. For example, the computing device A can further include more or fewer components (such as a network interface, a display device, etc.) than those shown in the figure, or have a different configuration from that shown in the figure.

[0323] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device by a program, and the program can be stored in a computer readable storage medium, which can include a flash disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, etc.

[0324] Embodiment 16

[0325] The embodiments of the present application also provide a storage medium. Optionally, in the present embodiment, the above-mentioned storage medium can be used to save the program code executed by the labeling method of the label content provided in Embodiment 1.

[0326] Optionally, in the present embodiment, the above-mentioned storage medium can be located in any one of the computing devices in the computing device group in the computer network, or in any one of the mobile terminals in the mobile terminal group.

[0327] Optionally, in the present embodiment, the storage medium is configured to store program code for performing the following steps: displaying target content to be labeled on an operation interface; if the operation interface detects that at least one target label in the label library is selected, a determination instruction is generated, wherein the determination instruction carries the at least one target label, and the target label belongs to a target label system; in response to the determination instruction, determining a original label matched with the target label based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine the mapping relationship between a first entity in the original label system and a second entity in the target label system; determining that the material labeled by the original label in the target content is the material that needs to be labeled by the target label.

[0328] The above-mentioned serial numbers of the embodiments of the present application are only for description, and do not represent the advantages and disadvantages of the embodiments.​

[0329] In the above-described embodiments of the present application, the description of each embodiment focuses on different aspects, and the parts not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0330] In several embodiments provided in the present application, it should be understood that the disclosed technical contents can be implemented by other ways. Among them, the above-described device embodiments are only schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division way, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units or modules shown or discussed can be indirect coupling or communication connection through some interfaces, units or modules, which can be electrical or other forms.

[0331] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or they can be distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0332] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The above integrated unit can be realized in the form of hardware or in the form of software functional unit.

[0333] The integrated unit, if realized in the form of software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the essential part or all or part of the technical solutions that make contributions to the prior art can be embodied in the form of software product, which is stored in a storage medium and includes a plurality of instructions for making a computer device (which can be a personal computer, a server or a network device, etc.) execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and various program code storage media.

[0334] The above merely describes the preferred embodiments of the present application, and it should be pointed out that, for those skilled in the art, several improvements and refinements can be made without departing from the principles of the present application, and these improvements and refinements should also be considered as falling within the protection scope of the present application.

Claims

1. A method of annotating content of a label, characterized by, The method comprises the following steps: displaying target content to be labeled on an operation interface; if the operation interface detects that at least one target label in a label library is selected, a determination instruction is generated, wherein the determination instruction carries the at least one target label, and the target label belongs to a target label system; in response to the determination instruction, an original label matched with the target label is determined based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system; determining that material labeled by the original label in the target content is material that needs to be labeled by the target label; wherein the original label system comprises a plurality of first entities having a hierarchical relationship, and the target label system comprises a plurality of second entities having a hierarchical relationship.

2. The method of claim 1, wherein, After determining that the material labeled by the original label in the target content is the material that needs to be labeled by the target label, the method further comprises: displaying a labeling result of the target content on the operation interface, wherein the labeling result is content labeling of the material labeled by the original label using the target label; if the labeling result is an incorrect labeling, an editing interface is called, and the editing interface is displayed on the operation interface; in response to an editing instruction, the labeling result is modified; the mapping relationship between the first entity in the original label system and the second entity in the target label system is updated based on the modified labeling result.

3. The method of claim 2, wherein, The labeling result is updated by inputting description information on the editing interface, and the updated target label is saved to the target label system.

4. The method of claim 2, wherein, Based on the label knowledge graph, the original label system entity is combined to generate new label information, and the new label information is synchronized to the original label system.

5. The method of claim 4, wherein, If it is detected that the similarity between the new label information and the existing label information in the original label system exceeds a threshold value, the synchronization of the new label information to the original label system is terminated.

6. A method of annotating content of a label, characterized by, The method comprises the following steps: obtaining target content to be labeled; calling a selected target label from a label library, wherein the target label belongs to a target label system; determining an original label matched with the target label based on a label knowledge graph, wherein the original label belongs to an entity of an original label system, and the label knowledge graph is used to record a mapping relationship between a first entity in the original label system and a second entity in the target label system; based on the original label matched with the target label, labeling material labeled by the original label from the target content; content labeling of the material labeled by the original label using the target label; wherein the original label system comprises a plurality of first entities having a hierarchical relationship, and the target label system comprises a plurality of second entities having a hierarchical relationship.

7. A method of annotating a label, characterized by, The method comprises the following steps: a cloud server receives at least one target label and target content input by a user through a client, wherein the target label belongs to a target label system; The cloud server determines an original label matched with the target label based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system; The cloud server determines that the material labeled by the original label in the target content is a material that needs to be labeled by the target label; The cloud server returns the material that needs to be labeled by the target label to the client; The original label system includes a plurality of first entities having a hierarchical relationship, and the target label system includes a plurality of second entities having a hierarchical relationship.

8. A method of annotating a label, characterized by, Comprise: Receiving at least one target label and a news to be labeled input by a user through a client, wherein the target label belongs to a target label system of a preset news platform; Determine an original label matched with the target label based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system; Determine that the material labeled by the original label in the news to be labeled is a material that needs to be labeled by the target label; Return the material that needs to be labeled by the target label to the client; The original label system includes a plurality of first entities having a hierarchical relationship, and the target label system includes a plurality of second entities having a hierarchical relationship.

9. A method of annotating a label, characterized by, Comprise: Receiving at least one target label and product description information input by a user through a client, wherein the target label belongs to a target label system of a preset e-commerce platform; Determine an original label matched with the target label based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system; Determine that the material labeled by the original label in the product description information is a material that needs to be labeled by the target label; Return the material that needs to be labeled by the target label to the client; The original label system includes a plurality of first entities having a hierarchical relationship, and the target label system includes a plurality of second entities having a hierarchical relationship.

10. A method of annotating a label, characterized by, Comprise: Receiving at least one course theme label and teaching material input by a user through a client, wherein the course theme label belongs to a target label system of an online education platform; Determine an original label matched with the course theme label based on a label knowledge graph, wherein the original label belongs to an original label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the original label system and a second entity in the target label system; Determine that the teaching material labeled by the original label in the teaching material is a material that needs to be labeled by the course theme label; Return the material that needs to be labeled by the course theme label to the client; The original label system includes a plurality of first entities having a hierarchical relationship, and the target label system includes a plurality of second entities having a hierarchical relationship.

11. A method of annotating a label, characterized by, The method comprises the following steps: receiving at least one disease name label and medical record material input by a user through a client, wherein the disease name label belongs to a target label system of a medical platform; determining a primary label matched with the disease name label based on a label knowledge graph, wherein the primary label belongs to a primary label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the primary label system and a second entity in the target label system; determining that medical record material labeled by the primary label in the medical record material is material that needs to be labeled by the disease name label; returning the material that needs to be labeled by the disease name label to the client; wherein the primary label system comprises a plurality of first entities having a hierarchical relationship, and the target label system comprises a plurality of second entities having a hierarchical relationship.

12. A method of annotating a label, characterized by, The method comprises the following steps: receiving at least one target label and multimedia content to be labeled input by a user through a client, wherein the target label belongs to a target label system of an entertainment platform, and the multimedia content comprises at least one of the following: long video, short video and song audio; determining a primary label matched with the target label based on a label knowledge graph, wherein the primary label belongs to a primary label system, and the label knowledge graph is used to determine a mapping relationship between a first entity in the primary label system and a second entity in the target label system; determining that material labeled by the primary label in the multimedia content is material that needs to be labeled by the target label; returning the material that needs to be labeled by the target label to the client; wherein the primary label system comprises a plurality of first entities having a hierarchical relationship, and the target label system comprises a plurality of second entities having a hierarchical relationship.

Images