Audit method, device and electronic equipment based on content understanding

Through the review method based on content understanding, the management and control needs are determined and the corresponding audit model is called to review the content of different business parties, which solves the problems of inflexible audit methods and limited coverage in the existing technology, and achieves flexibility and broadness Sex improvement.

CN113469732BActive Publication Date: 2024-11-19BEIJING BAIDU NETCOM SCI & TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202110652248.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-06-11
Publication Date
2024-11-19
Estimated Expiration
2041-06-11

AI Technical Summary

Technical Problem

The existing technology is difficult to apply to content reviews of different business parties, resulting in inflexible audit methods and limited coverage.

Method used

Through the review method based on content understanding, we determine the management and control needs, obtain the corresponding content tags, configure the content audit system, and call the audit model to review the objects to be reviewed, output the audit results, and realize content audit of different business parties.

Benefits of technology

It realizes content auditing without the need to configure the audit system separately for different business parties and is suitable for different management and control needs, improving the flexibility and coverage of the audit method.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113469732B_ABST
    Figure CN113469732B_ABST
Patent Text Reader

Abstract

The present disclosure discloses an audit method, device and electronic device based on content understanding, which relates to the field of computer technology, and in particular to the field of information processing technology. The specific implementation scheme is: determine the control requirements; obtain at least one content tag corresponding to the control requirements, and configure the content audit system based on the at least one content tag, wherein any one of the at least one content tag corresponds to at least one audit model; when the object to be audited is obtained, the audit model corresponding to the at least one content tag is called based on the content audit system to perform content audit on the object to be audited; and output the audit result, which is used to characterize the content of the object to be audited.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, in particular to the field of information processing technology, and specifically to an audit method, device and electronic device based on content understanding. Background Art

[0002] With the development of mobile Internet, content consumption has been upgraded, and the content presentation forms include text, images, videos, etc. Content needs to be reviewed before it is published and users obtain information to ensure that high-quality content is pushed to users. The main purpose of the review is to filter out risky content. By establishing different risk identification models for different risk categories, risks can be identified and filtered through the established risk identification models when they occur again. Summary of the invention

[0003] The present disclosure provides a content understanding-based audit method, device, and electronic device.

[0004] According to a first aspect of the present disclosure, there is provided a review method based on content understanding, comprising:

[0005] Determine control needs;

[0006] Acquire at least one content tag corresponding to the control requirement, and configure a content review system based on the at least one content tag, wherein any one of the at least one content tag corresponds to at least one review model;

[0007] When the object to be reviewed is obtained, the review model corresponding to the at least one content tag is called based on the content review system to perform content review on the object to be reviewed;

[0008] The audit result is output, where the audit result is used to characterize the content of the object to be audited.

[0009] According to a second aspect of the present disclosure, there is provided a review device based on content understanding, comprising:

[0010] Determination module, used to determine control requirements;

[0011] a configuration module, configured to obtain at least one content tag corresponding to the control requirement, and configure a content review system based on the at least one content tag, wherein any one of the at least one content tag corresponds to at least one review model;

[0012] An audit module, configured to, when an object to be audited is acquired, perform content audit on the object to be audited by calling the audit model corresponding to the at least one content tag based on the content audit system;

[0013] The output module is used to output the audit result, and the audit result is used to characterize the content of the object to be audited.

[0014] According to a third aspect of the present disclosure, there is provided an electronic device, including:

[0015] at least one processor; and

[0016] a memory communicatively connected to the at least one processor; wherein,

[0017] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described in the first aspect.

[0018] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to execute the method according to the first aspect.

[0019] According to a fifth aspect of the present disclosure, a computer program product is provided, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the method according to the first aspect.

[0020] The solution provided by the present disclosure does not require the configuration of different audit systems for different business parties. The audit system based on content understanding can be applied to the content audit of the audit objects of business parties with different management and control requirements, making the audit method of electronic devices more flexible and covering a wider range of audit objects.

[0021] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0023] Figure 1 is a flow chart of a content understanding-based audit method provided according to an embodiment of the present disclosure;

[0024] Figure 2 It is a scenario diagram applicable to a content understanding-based audit method provided by an embodiment of the present disclosure;

[0025] Figure 3 is a structural diagram of a content understanding-based auditing device provided according to an embodiment of the present disclosure;

[0026] Figure 4is a block diagram of an electronic device for implementing the content understanding-based audit method of the embodiment of the present disclosure; DETAILED DESCRIPTION

[0027] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0028] The disclosed embodiment provides an audit method based on content understanding.

[0029] Please refer to Figure 1 , Figure 1 is a flow chart of a content understanding-based audit method provided by an embodiment of the present disclosure. Figure 1 As shown, the audit method based on content understanding includes the following steps:

[0030] Step S101: Determine management and control requirements.

[0031] In the embodiments of the present disclosure, the management and control requirements may refer to the management and control requirements of the business party, and the business party may refer to a business platform or business software that provides content consumption for users, such as search software, video playback platform, online shopping software, social software, etc.

[0032] The control requirements of the business party may refer to the control requirements of the business party for the content displayed on its own platform or software, such as not allowing the displayed content to contain uncivilized language, etc. It is understandable that the control requirements of different business parties may be different.

[0033] It should be noted that the method provided in the embodiment of the present disclosure may be applied to an electronic device, and the electronic device may be installed with a business platform or business software of a business party. Thus, the electronic device can also obtain the management and control requirements of the business party on which it is installed.

[0034] Step S102: Obtain at least one content tag corresponding to the control requirement, and configure a content review system based on the at least one content tag.

[0035] Among them, any one of the at least one content label corresponds to at least one audit model. For example, if there are three content labels corresponding to the management and control requirements, each of the three content labels can correspond to at least one audit model. The audit model can be a pre-built neural network model that can audit the audited object. The audit model can be trained based on a large number of training samples. The training samples can be training samples related to the content labels. The training method of the neural network model can refer to the principles and implementation methods in the relevant technology, and the embodiments of the present disclosure will not be described in detail.

[0036] In the disclosed embodiment, the electronic device is installed with an audit system based on content understanding. The audit system based on content understanding may be a pre-built audit system. The audit system may include a requirement layer, a configuration layer and a label layer. The requirement layer is provided with different management and control requirements. The configuration layer may be configured with labels according to the management and control requirements of the business party. The label layer includes various pre-set content labels. The configuration layer configures corresponding labels for the management and control requirements based on the content labels in the label layer. One content label corresponds to at least one audit model.

[0037] like Figure 2 As shown, the demand layer of the audit system includes sensitive information control, low-quality content control, qualification access control, compliance control, etc.; the label layer includes three types of labels: pictures, texts, and landing pages, wherein the picture type labels include household appliances, characters, underwear, alcohol, flags, and underwear, the text type labels include hints, public opinion descriptions, health care, and gossip descriptions, and the landing page type labels include illegal agents and web page tampering; if the configuration layer needs to configure labels for sensitive information control, the configuration layer can be based on the three types of labels: picture type, text type, and landing page type to select corresponding content labels, such as character content labels and flag content labels for picture type labels, public opinion description content labels and gossip description content labels for text type labels, and illegal agent content labels and web page tampering content labels for landing page content labels. In this way, based on the business party's control requirements for sensitive information, it is possible to obtain content labels corresponding to the sensitive information control requirements, and then based on these corresponding content labels, it is possible to configure the business party's content audit system, and the content audit system also includes all audit models for the corresponding content labels.

[0038] Step S103: When the object to be reviewed is obtained, the review model corresponding to the at least one content tag is called based on the content review system to perform content review on the object to be reviewed.

[0039] Among them, the object to be reviewed may refer to a business object applied to the business party. For example, the business party is an advertising delivery platform, and the business object of the business party is the advertisement that needs to be delivered on the platform. It can be understood that if the electronic device is installed with the advertising delivery platform, when the advertisement to be delivered that needs to be delivered on the advertising delivery platform is obtained, the advertisement to be delivered is also the object to be reviewed, and it is necessary to review the content of the advertisement to be delivered to determine whether its content meets the management and control requirements of the advertising delivery platform.

[0040] In the disclosed embodiment, after determining the management and control requirements of the business party, the electronic device can obtain at least one content tag corresponding to the management and control requirements of the business party from the audit system based on content understanding, configure the content audit system of the business party based on these content tags, and then when the object to be audited of the business party is obtained, it can call the audit model corresponding to the at least one content tag based on the configured content audit system to perform content audit on the object to be audited.

[0041] It should be noted that if there are three content tags corresponding to the business party's management and control requirements, one content tag corresponds to one audit model, and the other two content tags each correspond to two audit models, then the content audit system configured for the business party corresponds to five audit models; when the object to be audited of the business party is obtained, these five audit models are called to perform content audit on the object to be audited.

[0042] Step S104: outputting the audit result, wherein the audit result is used to characterize the content of the object to be audited.

[0043] In the embodiment of the present disclosure, the electronic device can be based on the content review system configured for the business party, call the corresponding review model to perform content review on the business party's object to be reviewed, and then output the review result, which is used to characterize the content of the object to be reviewed, or to describe the content of the object to be reviewed.

[0044] Among them, the content audit of the audited object based on the audit model corresponding to the content audit system can be that the audited object is used as the input of each audit model, and the content audit of the audited object is performed based on the audit model, for example, text extraction or image recognition can be performed on the audited object to output the audit result used to describe the content of the audited object. For example, if the audited object is an image, and the image includes "xx Building", the output audit result can be "xx Building", that is, the audit result is used to describe the image content in the image.

[0045] According to the solution provided by the embodiments of the present disclosure, after determining the management and control requirements of the business party, the electronic device can obtain at least one content tag corresponding to the management and control requirements of the business party from the audit system based on content understanding, configure the content audit system of the business party based on these content tags, and then when the object to be audited of the business party is obtained, the corresponding audit model can be called based on the configured content audit system to perform content audit on the object to be audited, and an audit result for characterizing the content of the object to be audited can be output, and the user can decide whether the object to be audited meets the management and control requirements based on the audit result. In this way, the electronic device does not need to configure different audit systems for different business parties separately. The audit system based on content understanding can be applied to the content audit of the audit objects of business parties with different management and control requirements, making the audit method of the electronic device more flexible and covering a wider range of audit objects.

[0046] Optionally, the step S103 may include:

[0047] In the case of obtaining the object to be reviewed, determining the type of the object to be reviewed, the type including at least one of text and image;

[0048] In the case where the type of the object to be reviewed includes text, calling a text review model in the at least one review model to review the text of the object to be reviewed to determine whether the text of the object to be reviewed matches a preset text;

[0049] In the case that the type of the object to be reviewed includes an image, an image review model in the at least one review model is called to identify the image of the object to be reviewed to determine whether the image of the object to be reviewed matches a preset image.

[0050] For example, the business party is an advertising delivery platform, and the object to be reviewed is an advertisement to be reviewed that needs to be delivered on the platform, and the advertisement includes text and images. When the electronic device reviews the advertisement to be reviewed, it can call a text review model in a corresponding review model based on the configured content review system to review the text of the advertisement to be reviewed. For example, it can extract keywords from the text included in the advertisement to be reviewed to determine whether the extracted keywords match the preset text; at the same time, the image review model in the corresponding review model is called based on the configured content review system to identify the image of the advertisement to be reviewed to determine whether the identified image matches the preset image.

[0051] In the disclosed embodiment, based on the type of the object to be audited, the text audit and / or image audit are performed respectively through the audit model corresponding to the type, and then the text of the object to be audited matches the preset text, and / or the image of the object to be audited matches the preset image, so as to output the corresponding audit result. In this way, the corresponding audit model is specifically called to perform the audit based on the type of the object to be audited, so that the output audit result can be more appropriate to the type of the object to be audited, and the accuracy of the audit result is effectively improved.

[0052] Further, the step S104 may include:

[0053] When the text of the object to be audited matches the preset text, outputting the audit result including the preset text;

[0054] In the case where the image of the object to be reviewed matches the preset image, a review result including text information for describing the preset image is output.

[0055] It should be noted that the text audit model can be obtained by training a large number of text training samples, and the text audit model can include at least one preset text, and the text audit model corresponds to the content label. For example, the content label is gossip description, and the preset text included in the text audit model corresponding to the content label can include gossip, scandal, revelations, hot news, etc. If the configured content audit system includes a text audit model corresponding to the gossip description content label, when the text audit model is called to audit the text of the audit object, it is to determine whether the text matches these preset texts; if it matches, the audit result including the preset text is output; for example, if the text of the audit object includes "scandal", it means that the text matches the preset text "scandal", and the audit result output for the audit object also includes the word "scandal". Furthermore, based on the audit result, the user can determine that the audit object involves gossip description, which can provide a more intuitive reference for the user to determine whether the audit object meets the management and control requirements.

[0056] Optionally, the image review model is also obtained by training with a large number of image training samples. The image review model may include at least one preset image, and the image review model corresponds to the content label. For example, if the content label is a flag, and the configured content review system includes an image review model corresponding to the flag content label, when the image review model is called to review the image of the object to be reviewed, it is to determine whether the image included in the object to be reviewed matches the preset image corresponding to the content label; if it matches, the review result including text information describing the preset image is output. Furthermore, it can provide a more intuitive reference for users to determine whether the object to be reviewed meets the management and control requirements. For example, if the business party is an advertising delivery platform, if it does not meet the requirements of the Advertising Law, then based on the review result, the user can more intuitively and accurately know whether the object to be reviewed does not meet the management and control requirements of the business party, and can provide a reference for users to improve the object to be reviewed.

[0057] In the disclosed embodiment, based on the type of the object to be audited, the object to be audited is subjected to text audit and / or image audit respectively through the audit model corresponding to the type, and the audit result outputted includes preset text and / or text information for describing the preset image, so that the user can more intuitively know the content included in the object to be audited based on the audit result, and provide a more intuitive reference for the user to determine whether the object to be audited meets the management and control requirements of the business party. In addition, when the object to be audited does not meet the management and control requirements, the user can know what content does not meet the management and control requirements, so as to provide suggestions for the user to improve the object to be audited, which is more beneficial to help the user obtain the object to be audited that meets the management and control requirements of the business party.

[0058] It should be noted that in the embodiments of the present disclosure, the review system based on content understanding may be a preset review system in which the electronic device has been configured with content tags, or a general review system; further, the user may also configure the review system based on content understanding according to the management and control requirements of the business party.

[0059] Optionally, in the embodiment of the present disclosure, the method further includes:

[0060] In the case where the business party includes a target management and control requirement, a target content tag corresponding to the target management and control requirement is obtained from a content tag database; wherein the at least one content tag includes the target content tag.

[0061] It should be noted that the electronic device may be pre-set with a content tag database, and the content tag database may include content tags corresponding to different management and control requirements, or may include content tags corresponding to all management and control requirements of the business party that may be applied.

[0062] When the target control requirements of the business party are obtained, if the target content tag corresponding to the target control requirement is not included in the audit system based on content understanding, or is not included in the preset audit system, the electronic device can obtain the target content tag corresponding to the target control requirement from the content tag database based on user operation, and then the electronic device can obtain the target content tag corresponding to the target control requirement after determining the target control requirements of the business party. It should be noted that the number of the target content tags can be at least one, and each target content tag corresponds to at least one audit model. Furthermore, the electronic device can configure a content audit system corresponding to the target control requirements of the business party based on the target content tag, and when the object to be audited is obtained, it can call the audit model corresponding to the target content tag based on the configured content audit system to perform content audit on the object to be audited.

[0063] In the disclosed embodiment, when the business party includes target management and control requirements, the electronic device can obtain the target content label corresponding to the target management and control requirements from the content label database, thereby enabling the electronic device to flexibly configure the content label of the business party's management and control requirements according to the business party's management and control requirements, thereby meeting the different management and control requirements of different business parties, making the method more applicable.

[0064] It should be noted that the electronic device can be configured to generate a directed acyclic graph (DAG) according to the management and control requirements of the business party to build a bridge between the management and control requirements of the business party and the content label. The DAG can be used to calculate the corresponding content label return according to the management and control requirements. Optionally, the DAG can be automatically generated, and each operator maintains a sub-dag separately. A dag is equivalent to a content label. According to the operator set determined by the management and control requirements, an overall DAG will be automatically generated. The algorithm is as follows:

[0065] DAG=(dag0∪dag1…∪dag n );

[0066] Among them, DAG is the overall dag, n is the nth operator, dag n is the dag of the nth operator. For details, reference may be made to the relevant technology, and this disclosure will not go into detail.

[0067] Furthermore, before acquiring the target content tag corresponding to the target management and control requirement from the content tag database, the method further includes:

[0068] If the target content tag is not included in the content tag database, obtaining input information;

[0069] constructing the target content tag based on the input information, and constructing at least one audit model corresponding to the target content tag;

[0070] The target content tag is added into the content tag database.

[0071] In the disclosed embodiment, if the content tag database does not include the target content tag corresponding to the target control requirement, the target content tag can be constructed by obtaining the user's input information, and at least one audit model corresponding to the target content tag is constructed, and the constructed target content tag is added to the content tag database, which expands the number and range of content tags in the content tag database, so that the content tag database can be applied to more business parties and different control requirements. In this way, for the content tags that do not exist, the user only needs to construct the corresponding content tags, and there is no need to configure a separate audit system for the target control requirements of the business party.

[0072] Optionally, when the business party's management and control needs no longer meet the current risk management requirements, or the business party's management and control needs need to be upgraded, the content tag corresponding to the management and control need can be found and the content tag corresponding to the management and control need can be upgraded. For example, the audit model of the content tag can be extended, such as the preset text or preset image in the extended audit model; or content tags and corresponding audit models that match the management and control needs can be added.

[0073] For example, if the banner content label in the image content label corresponding to the sensitive information control demand does not meet the current risk control requirements, only the banner content label can be optimized, such as adding a preset image in the banner content label review model, so that the optimized banner content label can meet the current risk control requirements. There is no need to upgrade and optimize all content labels corresponding to the entire sensitive information control demand. This can effectively simplify the upgrade and optimization of the business party's control requirements and reduce the upgrade cost of the review system based on content understanding.

[0074] The solution provided by the embodiments of the present disclosure can configure a universal content review system for different business parties, and can be applied to content review of review objects of business parties with different management and control requirements, thereby eliminating the need to configure different review systems separately for different business parties; or it can also be based on the management and control requirements of the business parties to flexibly configure corresponding content tags, thereby also being able to meet the needs of business parties with different management and control requirements, so as to meet the review of the business parties' objects to be reviewed, thereby improving the scalability of the review system and being able to adapt to the review needs of different business parties.

[0075] The disclosed embodiment also provides an audit device based on content understanding.

[0076] Please refer to Figure 3 , Figure 3 is a structural diagram of a content understanding-based auditing device provided by an embodiment of the present disclosure. Figure 3 As shown, the content understanding-based auditing device 300 includes:

[0077] Determination module 301, used to determine the control requirements;

[0078] A configuration module 302 is used to obtain at least one content tag corresponding to the control requirement, and configure a content review system based on the at least one content tag, wherein any one of the at least one content tag corresponds to at least one review model;

[0079] The review module 303 is used to perform content review on the object to be reviewed by calling the review model corresponding to the at least one content tag based on the content review system when the object to be reviewed is obtained;

[0080] The output module 304 is used to output the audit result, where the audit result is used to represent the content of the object to be audited.

[0081] Optionally, the content understanding-based auditing device 300 further includes:

[0082] A first acquisition module is used to acquire a target content tag corresponding to the target management requirement from a content tag database when the business party includes a target management requirement;

[0083] The at least one content tag includes the target content tag.

[0084] Optionally, the content understanding-based auditing device 300 further includes:

[0085] A second acquisition module, configured to acquire input information when the target content tag is not included in the content tag database;

[0086] A construction module, configured to construct the target content tag based on the input information, and to construct at least one audit model corresponding to the target content tag;

[0087] An adding module is used to add the target content tag into the content tag database.

[0088] Optionally, the review module 303 is further used to:

[0089] In the case of obtaining the object to be reviewed, determining the type of the object to be reviewed, the type including at least one of text and image;

[0090] In the case where the type of the object to be reviewed includes text, calling a text review model in the at least one review model to review the text of the object to be reviewed to determine whether the text of the object to be reviewed matches a preset text;

[0091] In the case that the type of the object to be reviewed includes an image, an image review model in the at least one review model is called to identify the image of the object to be reviewed to determine whether the image of the object to be reviewed matches a preset image.

[0092] Optionally, the output module 304 is further used for:

[0093] When the text of the object to be audited matches the preset text, outputting the audit result including the preset text;

[0094] In the case where the image of the object to be reviewed matches the preset image, a review result including text information for describing the preset image is output.

[0095] The content understanding-based audit device 300 provided by the present disclosure can configure a common content audit system for different business parties, and can be applicable to content audit of audit objects of business parties with different management and control requirements, thereby eliminating the need to configure different audit systems separately for different business parties; or it can also be based on the management and control requirements of the business party to flexibly configure corresponding content tags, thereby also being able to meet the needs of business parties with different management and control requirements, so as to meet the audit of the business party's audited objects, thereby improving the scalability of the audit system and being applicable to the audit needs of different business parties.

[0096] It should be noted that the content understanding-based audit device 300 provided in this embodiment can implement all the technical solutions of the above-mentioned content understanding-based audit method embodiment, and therefore can at least achieve all the above-mentioned technical effects, which will not be repeated here.

[0097] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0098] Figure 4A schematic block diagram of an example electronic device 400 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0099] like Figure 4 As shown, the electronic device 400 includes a computing unit 401, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 402 or a computer program loaded from a storage unit 408 into a random access memory (RAM) 403. In the RAM 403, various programs and data required for the operation of the device 400 can also be stored. The computing unit 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0100] A number of components in the device 400 are connected to the I / O interface 405, including: an input unit 406, such as a keyboard, a mouse, etc.; an output unit 407, such as various types of displays, speakers, etc.; a storage unit 408, such as a disk, an optical disk, etc.; and a communication unit 409, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 409 allows the device 400 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0101] The computing unit 401 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 401 performs the various methods and processes described above, such as the above-mentioned audit method based on content understanding. For example, in some embodiments, the audit method based on content understanding may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 408. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 400 via the ROM 402 and / or the communication unit 409. When the computer program is loaded into the RAM 403 and executed by the computing unit 401, one or more steps of the audit method based on content understanding described above may be performed. Alternatively, in other embodiments, the computing unit 401 may be configured to perform the audit method based on content understanding in any other appropriate manner (eg, by means of firmware).

[0102] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0103] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0104] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0105] To provide interaction with a user, the systems and techniques described herein may be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0106] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0107] A computer system may include clients and servers. Clients and servers are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship to each other.

[0108] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0109] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A review method based on content understanding, including: Determine control needs; Acquire at least one content tag corresponding to the control requirement, and configure a content review system based on the at least one content tag, wherein any one of the at least one content tag corresponds to at least one review model; When the object to be reviewed is obtained, the review model corresponding to the at least one content tag is called based on the content review system to perform content review on the object to be reviewed; Outputting the audit result, wherein the audit result is used to characterize the content of the object to be audited; The method further includes: in the case where the business party includes a target management and control requirement, acquiring a target content tag corresponding to the target management and control requirement from a content tag database; Wherein, the at least one content tag includes the target content tag; Before acquiring the target content tag corresponding to the target management and control requirement from the content tag database, the method further includes: If the target content tag is not included in the content tag database, obtaining input information; constructing the target content tag based on the input information, and constructing at least one audit model corresponding to the target content tag; Adding the target content tag into the content tag database; The method further comprises: When the management and control requirements of the business party need to be upgraded, the content tags corresponding to the management and control requirements are upgraded; The acquiring of at least one content tag corresponding to the control requirement specifically includes: According to the control requirement configuration, a directed acyclic graph DAG is generated to build a bridge between the control requirement and the content label. The corresponding content label is returned by calculating the control requirement through the DAG. The DAG is automatically generated, and each operator maintains a sub-dag separately. A dag is equivalent to a content label. According to the operator set determined by the control requirement, an overall DAG will be automatically generated, as shown in the following formula: DAY=(day0∪day1…∪day n ); Among them, DAG is the overall dag, n is the nth operator, dag n is the dag of the nth operator.

2. The method according to claim 1, wherein: When the object to be reviewed is obtained, the at least one review model is called based on the content review system to review the object to be reviewed, including: In the case of obtaining the object to be reviewed, determining the type of the object to be reviewed, the type including at least one of text and image; In the case where the type of the object to be reviewed includes text, calling a text review model in the at least one review model to review the text of the object to be reviewed to determine whether the text of the object to be reviewed matches a preset text; In the case that the type of the object to be reviewed includes an image, an image review model in the at least one review model is called to identify the image of the object to be reviewed to determine whether the image of the object to be reviewed matches a preset image.

3. The method according to claim 2, wherein: The output audit results include: When the text of the object to be audited matches the preset text, outputting the audit result including the preset text; In the case where the image of the object to be reviewed matches the preset image, a review result including text information for describing the preset image is output.

4. A review device based on content understanding, comprising: Determination module, used to determine control requirements; a configuration module, configured to obtain at least one content tag corresponding to the control requirement, and configure a content review system based on the at least one content tag, wherein any one of the at least one content tag corresponds to at least one review model; An audit module, configured to, when an object to be audited is acquired, perform content audit on the object to be audited by calling the audit model corresponding to the at least one content tag based on the content audit system; An output module, used to output the audit result, wherein the audit result is used to characterize the content of the object to be audited; The device also includes: A first acquisition module is used to acquire a target content tag corresponding to the target management requirement from a content tag database when the business party includes a target management requirement; Wherein, the at least one content tag includes the target content tag; The device also includes: A second acquisition module, configured to acquire input information when the target content tag is not included in the content tag database; A construction module, configured to construct the target content tag based on the input information, and to construct at least one audit model corresponding to the target content tag; An adding module, used for adding the target content tag into the content tag database; The device also includes: An upgrade module, used to upgrade the content tags corresponding to the management and control requirements when the management and control requirements of the business party need to be upgraded; Wherein, the configuration module is specifically used for: According to the control requirement configuration, a directed acyclic graph DAG is generated to build a bridge between the control requirement and the content label. The corresponding content label is returned by calculating the control requirement through the DAG. The DAG is automatically generated, and each operator maintains a sub-dag separately. A dag is equivalent to a content label. According to the operator set determined by the control requirement, an overall DAG will be automatically generated, as shown in the following formula: DAY=(day0∪day1…∪day n ); Among them, DAG is the overall dag, n is the nth operator, dag n is the dag of the nth operator.

5. The device according to claim 4, wherein: The audit module is also used to: In the case of obtaining the object to be reviewed, determining the type of the object to be reviewed, the type including at least one of text and image; In the case where the type of the object to be reviewed includes text, calling a text review model in the at least one review model to review the text of the object to be reviewed to determine whether the text of the object to be reviewed matches a preset text; In the case that the type of the object to be reviewed includes an image, an image review model in the at least one review model is called to identify the image of the object to be reviewed to determine whether the image of the object to be reviewed matches a preset image.

6. The device according to claim 5, wherein: The output module is also used for: When the text of the object to be audited matches the preset text, outputting the audit result including the preset text; In the case where the image of the object to be reviewed matches the preset image, a review result including text information for describing the preset image is output.

7. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 3.

8. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-3.

9. A computer program product, comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 3.