Risk Detection Method, Device, Computer-Readable Storage Medium, Electronic Device and Computer Program Product

Abstract

An embodiment of this specification discloses a risk detection method, apparatus, computer-readable storage medium, electronic device, and computer program product. The risk detection method may include: after obtaining the target content that the user needs to publish, obtaining the risk level of the user determined based on the user's historical content publishing behavior in the user historical data, so as to use a risk detection strategy in which the strictness during risk detection is positively correlated with the user's risk level to perform risk detection processing on the target content that the user needs to publish, thereby obtaining the risk detection result of the target content.

Description

Technical Field

[0001] This specification relates to the technical field of risk detection, and particularly to a risk detection method, apparatus, computer-readable storage medium, electronic device, and computer program product. Background Art

[0002] With the increasing demand for interaction and communication among people and the development and progress of science and technology, people have gradually started to use the application programs installed on terminal devices to carry out content creation and digital social networking. Currently, when users publish content in these application programs, service providers usually need to perform relatively comprehensive risk detection on the content that users need to publish, in order to prevent the spread of inappropriate content. However, with the explosive growth of the content that users need to publish, how to improve the efficiency and accuracy of risk detection for the content that users need to publish, so as to reduce the risks brought by the content published by users, has become a technical problem that urgently needs to be solved. Summary of the Invention

[0003] Embodiments of this specification provide a risk detection method, apparatus, computer-readable storage medium, electronic device, and computer program product, which can use a risk detection strategy in which the strictness of risk detection is positively correlated with the risk level of the user to perform risk detection processing on the target content that the user needs to publish, which is beneficial to improving the efficiency and accuracy of risk detection for the target content that the user needs to publish, so as to reduce the risks brought by the target content published by the user.

[0004] A risk detection method provided by an embodiment of this specification, the method includes:

[0005] Obtain the target content that the user needs to publish;

[0006] Obtain the risk level of the user determined according to the user's historical data; wherein, the user's historical data includes data generated based on the user's historical content publishing behavior;

[0007] Perform risk detection processing on the target content according to a risk detection strategy corresponding to the risk level of the user, and obtain a risk detection result of the target content; wherein, the strictness of the risk detection strategy is positively correlated with the risk level of the user.

[0008] An embodiment of this specification also provides a risk detection apparatus, including:

[0009] A first acquisition module, configured to obtain the target content that the user needs to publish;

[0010] A second acquisition module, configured to acquire the risk level of the user determined according to the user's historical data; wherein, the user's historical data includes data generated based on the user's historical content publishing behavior.

[0011] A risk detection module, configured to perform risk detection processing on the target content according to a risk detection strategy corresponding to the risk level of the user, to obtain a risk detection result of the target content; wherein, the strictness of the risk detection strategy is positively correlated with the risk level of the user.

[0012] An embodiment of this specification also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0013] An embodiment of this specification also provides an electronic device, including: a processor and a memory; wherein, the memory stores a computer program, and the computer program is adapted to be loaded and executed by the processor to implement the steps of the above method.

[0014] An embodiment of this specification also provides a computer program product, on which at least one instruction is stored, and when the at least one instruction is executed by a processor, the steps of the above method are implemented.

[0015] In an embodiment of this specification, after obtaining the target content that the user needs to publish, the risk level of the user determined according to the user's historical data generated based on the user's historical content publishing behavior can be obtained; by using a risk detection strategy in which the strictness during risk detection is positively correlated with the risk level of the user, performing risk detection processing on the target content that the user needs to publish, to obtain the risk detection result of the target content, which is beneficial to improving the efficiency and accuracy of risk detection for the target content that the user needs to publish, so as to reduce the risk brought by the target content published by the user. Description of the Drawings

[0016] Figure 1 It is a schematic diagram of an application scenario of a risk detection solution provided by an embodiment of this specification.

[0017] Figure 2 It is a schematic flowchart of a risk detection method provided by an embodiment of this specification.

[0018] Figure 3 It is a schematic structural diagram of a risk detection device provided by an embodiment of this specification.

[0019] Figure 4 It is a schematic structural diagram of an electronic device provided by an embodiment of this specification. Detailed Embodiments

[0020] To make the objectives, technical solutions, and advantages of this specification clearer, the technical solutions of this specification will be clearly and completely described below in conjunction with specific embodiments of this specification and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this specification without creative efforts belong to the scope protected by this specification.

[0021] Before specifically describing the risk detection solution provided by the embodiments of this specification, the relevant technical background will be described first.

[0022] With the rapid development of the content creation and digital social industries, the content published by users at the application has shown explosive growth. This phenomenon not only enriches people's entertainment choices but also brings huge content risk review challenges to service providers. Traditional content risk review methods often rely on manual full-scale review of various risks that may exist in the content to be published by each user according to a unified standard. This content risk review method often consumes a large amount of human resources, resulting in high risk detection costs. In addition, since this content risk review method relies on the subjective judgment of reviewers, some risks may be missed, or too much review time may be spent on some harmless content, thus easily affecting the efficiency and accuracy of risk detection for the content to be published by users.

[0023] Please refer to Figure 1 , which is a schematic diagram of the application scenario of a risk detection solution provided by the embodiments of this specification.

[0024] As Figure 1As shown, the target application program can be installed on the terminal devices of the user group (e.g., Device 102, Device 103, Device 104, etc.), and Server 101 can be the server device for the target application program. If the user performs an operation of publishing target content using the target application program on Device 103, then Device 103 can send the target content that the user needs to publish to Server 101. Server 101 can obtain the risk level of the user determined based on the user's historical data; wherein, the user's historical data can include data generated based on the user's historical content publishing behavior; so as to perform risk detection processing on the target content according to a risk detection strategy having a corresponding relationship with the user's risk level, and obtain the risk detection result of the target content; wherein, the strictness of the risk detection strategy can be positively correlated with the user's risk level. Thus, it is possible to perform risk detection processing with different strictness levels on the content that users with different risk levels need to publish, which is beneficial to improving the efficiency and accuracy of risk detection for the content that users need to publish, and reducing the risks brought by the content published by users.

[0025] Please refer to Figure 2 , which is a schematic flowchart of a risk detection method provided by an embodiment of this specification. The execution subject of this process can be a risk detection device, or an application program installed on the risk detection device. Next, the following Figure 2 shown process will be elaborated in detail. The risk detection method can specifically include the following steps:

[0026] Step 202, obtain the target content that the user needs to publish.

[0027] In the embodiments of this specification, users can usually publish target content on the target application program and set the public scope of the target content so that a specified user group can view and access the target content. In practical applications, with the different content sharing functions supported by the target application program, the types of target content that users need to publish can often be diverse. For example, the target content can include at least one of video content, audio content, image content, and text content. In addition, it can also include links, codes, etc., and no specific limitations are made here.

[0028] For ease of understanding, an example is given here. For example, when the target application is a social network application mainly for sharing pictures and short videos, the target content that the user needs to publish can be the life moments shown by the user through short videos or multiple images. Or, when the target application is an article and blog sharing application mainly for text content, the target content that the user needs to publish can be the long articles written or shared by the user, as well as the news, opinions, and multimedia content expressed through short text content. Or, when the target application is an application providing a live broadcast function, the target content that the user needs to publish can be the content such as the shopping experience and lifestyle shared by the user in real time, so as to facilitate the user group to purchase related products; or, the target content that the user needs to publish can also be the live broadcast content of the user in aspects such as games and entertainment, so as to facilitate the user to interact with the audience in real time. Of course, the target application can also be an application specifically for sharing music, photographic works, open source code, etc., and the target content that the user needs to publish can be content in the form of audio, images, code, links, etc. No specific limitation is made thereto.

[0029] Step 204, obtain the risk level of the user determined according to the user's historical data; wherein, the user's historical data includes the data generated based on the user's historical content publishing behavior.

[0030] In the embodiments of this specification, the historical content publishing behaviors previously performed by the user often generate relevant user historical data at the target application. Since by analyzing the user's historical data, the user's behavior pattern, content preference, and potential risks can be understood, it can thus be used as an important basis for evaluating the risk level of the user. Based on this, the risk level of the user can be obtained by performing data analysis and processing on the user's historical data generated based on the user's historical content publishing behavior.

[0031] Generally, the higher the risk level of the user, the greater the possibility that the user publishes risky content. Therefore, it is necessary to perform more stringent risk detection on the target content that the user needs to publish to ensure the comprehensiveness and accuracy of the risk detection result generated for the target content, thereby effectively controlling the risks brought by the content published by the user. And the lower the risk level of the user, the smaller the possibility that the user publishes risky content. Therefore, the control strictness for risk detection of the target content that the user needs to publish can be appropriately relaxed, so that on the basis of effectively ensuring the accuracy and effectiveness of the risk detection result generated for the target content, the risk detection efficiency can be improved and the resources consumed for risk detection can be reduced.

[0032] Step 206: Perform a risk detection process on the target content according to a risk detection policy corresponding to the risk level of the user, and obtain a risk detection result of the target content; wherein, the strictness of the risk detection policy is positively correlated with the risk level of the user.

[0033] In the embodiments of this specification, different risk detection policies corresponding to the risk levels of each user can be flexibly configured in advance, and the strictness of the risk detection policy can be positively correlated with the corresponding user risk level, so that a more strict risk detection can be performed on the content to be published by users with a higher risk level, and a relatively loose risk detection can be performed on the content to be published by users with a lower risk level. This can not only ensure the efficiency and accuracy of content risk detection, improve the user experience, but also effectively reduce the risk management cost for the content to be published by users, and has good practicality.

[0034] Figure 2 In the solution, by using a risk detection policy in which the strictness during risk detection is positively correlated with the risk level of the user, a risk detection process is performed on the target content to be published by the user to obtain a risk detection result of the target content, which is beneficial to improving the efficiency and accuracy of risk detection for the target content to be published by the user, so as to reduce the risk and risk detection cost brought by the target content published by the user.

[0035] Based on Figure 2 the method in, the embodiments of this specification also provide some specific implementation schemes of this method, which will be described below.

[0036] In the embodiments of this specification, step 204: Obtaining the risk level of the user determined according to the user historical data may include:

[0037] Obtain preset types of user historical data generated based on the historical content publishing behavior of the user.

[0038] Use a risk assessment model constructed based on a machine learning algorithm to process the user historical data, and obtain a risk assessment result output by the risk assessment model for reflecting the risk level of the user.

[0039] In the embodiments of this specification, by using training samples in advance to train and tune the risk assessment model built based on the machine learning algorithm, the risk assessment model can be enabled to extract useful feature data and the correlation between feature data from a large amount of user historical data to accurately assess the user's risk level. Since the risk assessment model built based on the machine learning algorithm can achieve automated, efficient and accurate user risk level assessment, user historical data samples can be set in advance according to actual needs to train the risk assessment model, and then the user's user historical data can be input into the risk assessment model to obtain the risk assessment result output by the risk assessment model to reflect the user's risk level.

[0040] In actual applications, the specific content of the risk assessment result can be pre-set according to actual needs. For example, the risk assessment result may include information reflecting the user's risk level (for example, high, medium, low, etc.); or, the risk assessment result may also include information reflecting the user's risk score (for example, a risk score based on a percentage, a risk score based on a ten-point scale, etc.); in addition, the risk assessment result may also specifically include information such as the user's risk level / risk score in one or more preset risk areas, wherein the preset risk area is not specifically limited.

[0041] In the embodiment of the present specification, the obtaining of preset types of user history data generated based on the user's historical content publishing behavior may include:

[0042] Obtain at least one of the following data: the historical published content of the user, the interaction data of the user group on the historical published content, and the violation record data of the historical published content to obtain the user historical data. Or,

[0043] According to a preset data analysis strategy, data analysis is performed on at least one of the historical published content of the user, the interaction data of the user group on the historical published content, and the violation record data of the historical published content to obtain the user historical data.

[0044] In the embodiments of this specification, the user's historical published content may include the content that the user has successfully published at the target application, and of course, may also include the content that the user has attempted to publish but has not successfully published. Since the user's preferred content topics (e.g., current affairs, health, entertainment, etc.) can be understood by analyzing the user's historical published content, this can to a certain extent reflect whether the content that the user wants to publish is likely to cause controversy or may violate regulations. Therefore, the user's risk level can be assessed in combination with the user's historical published content.

[0045] The interaction data of the user group with respect to the historical published content may include: the comment content of the user group on the historical published content of the user, the collection behavior data, the forwarding behavior data, the like behavior data, etc. Since by analyzing the interaction behavior and communication data between the user group and the user, the social attitude of the user (e.g., positive or negative) and whether there is an attack situation against other users can be analyzed, therefore, the risk level of the user can be evaluated by combining the interaction data of the user group with respect to the historical published content of the user.

[0046] The violation record data for the historical published content of the user may include: information such as the violation type, the violation time, and the punishment measures (e.g., content deletion, account restricted use, etc.) of the historical published content of the user that have been determined. Since there is an obvious correlation between the violation record data of the user's historical published content and the risk level of the user. And the user's historical violation behaviors, violation frequencies, violation severities, time factors, and the user's compliance repair ability, etc., will all affect the risk assessment result of the user. Therefore, the risk level of the user can be evaluated more accurately by combining the violation record data of the user's historical published content.

[0047] In practical applications, since the data volume of the historical published content of some users and the interaction data of the user group with respect to the historical published content, etc., is relatively large. Therefore, in order to ensure that the risk assessment model constructed based on machine learning algorithms can efficiently and accurately complete the analysis and processing of the user's historical data, so as to improve the accuracy of the risk assessment result output by the risk assessment model for reflecting the risk level of the user, data analysis and processing can be performed on at least one of the historical published content of the user, the interaction data of the user group with respect to the historical published content, and the violation record data of the historical published content according to a preset data analysis strategy, so that the user-related information of a preset type obtained from the data analysis is used as the user's historical data required to be input into the risk assessment model.

[0048] For the sake of easy understanding, an example is given for this. For example, the publishing time of the historical published content of the user can be analyzed to obtain the frequency data and active time period data of the user's published content. Or, the historical published content of the user and the interaction data of the user group with respect to the historical published content can be analyzed collaboratively to determine whether the user has mood swings or sudden high-frequency content publishing for hot events. Or, the violation record data of the historical published content of the user can be analyzed to determine the review passing rate or violation rate of the user's published content. Or, the user group interacting with the historical published content of the user can also be analyzed to identify the influence and potential negative impact of the user in the social circle, and whether the social circle where it is located belongs to an extreme or controversial circle, etc.; no specific limitations are made on this.

[0049] By performing data analysis and processing on the user's historical published content, the interaction data of the user group with respect to the historical published content, and the violation record data of the historical published content, etc., according to a preset data analysis strategy, a more diverse type of user historical data can be obtained, which is beneficial to a more comprehensive understanding of the risk status of individual users, and thus can improve the accuracy and reliability of the risk level of the user generated by the risk identification model based on the user historical data.

[0050] In the embodiments of this specification, performing risk detection processing on the target content according to a risk detection strategy corresponding to the risk level of the user to obtain the risk detection result of the target content may include:

[0051] Determine a risk detection strategy corresponding to the risk level of the user; wherein, the risk detection strategy is used to control the preset quantity ratio of the content that needs to be risk-detected in the user's target content, and at least one of the preset risk types to be detected for the target content, and both the preset quantity ratio and the quantity of the preset risk types are positively correlated with the risk level of the user.

[0052] According to the risk detection strategy, perform risk detection processing on the specified content with the preset quantity ratio included in the target content to obtain the risk detection result of the target content. Or,

[0053] According to the risk detection strategy, perform risk detection processing on whether there is a risk of the preset risk type in the target content to obtain the risk detection result of the target content. Or,

[0054] According to the risk detection strategy, perform risk detection processing on whether there is a risk of the preset risk type in the specified content with the preset quantity ratio included in the target content to obtain the risk detection result of the target content.

[0055] In the embodiments of this specification, the strictness of the risk detection strategy can be reflected in aspects such as the preset quantity ratio of the content that needs to be risk-detected in the user's target content, and the quantity of the preset risk types to be detected for the target content. Generally, the larger the preset quantity ratio, the more content that actually needs to be risk-detected in the target content, so that a more comprehensive risk detection of the target content can be performed to effectively identify the risks carried by the target content. Therefore, the larger the preset quantity ratio, the higher the strictness of the risk detection strategy, and the smaller the preset quantity ratio, the lower the strictness of the risk detection strategy.

[0056] Moreover, generally speaking, the larger the number of the preset risk types, the more types of the preset risks that may exist in the target content to be detected, so that a more comprehensive risk detection of the target content can be carried out to effectively identify the risks that may be carried in the target content. Therefore, the larger the number of the preset risk types, the higher the strictness degree of the risk detection strategy, and the smaller the number of the preset risk types, the lower the strictness degree of the risk detection strategy.

[0057] It can be understood that when the risk detection strategy is only used to control the preset quantity ratio, the risk types to be identified for the content published by users with different risk levels can be consistent; when the risk detection strategy is only used to control the preset risk types, the quantity ratio of the content to be subject to risk detection in the content published by users with different risk levels can be consistent. When the risk detection strategy is used to control the preset quantity ratio and the preset risk types, the risk types to be identified for the content published by users with different risk levels may be different, and moreover, the quantity ratio of the content to be subject to risk detection may also be different, with good flexibility, and no specific limitation is made thereto.

[0058] In practical applications, when performing risk detection processing on the target content according to the risk detection strategy corresponding to the risk level of the user, if the risk detection strategy is used to control the preset quantity ratio, when the target content is video content, the preset quantity ratio of the image frames can be randomly or at equal intervals extracted from the image frames constituting the video content, so as to generate the risk detection result of the target content in combination with the extracted image frames. When the target content is audio content or text content, if multiple segments of audio content with short durations and not connected to each other are intercepted, or multiple short words not connected to each other are intercepted, it is often impossible to effectively identify the content expressed by the user. Therefore, a minimum unit data volume (for example, minimum audio duration, minimum number of words, etc.) can be set to extract audio segments or text segments larger than the minimum unit data volume from the target content until the quantity ratio of the extracted audio content or text content reaches the preset quantity ratio, which is beneficial to improving the accuracy and effectiveness of the risk detection result of the target content generated based on the above-mentioned extracted audio / text content.

[0059] In practical applications, if the risk detection strategy is used to control the preset risk type, the preset risk type can either be a risk type with a relatively high detection priority screened in advance, or, when the risk levels of a user in each preset risk area can be determined, the risk type with a relatively high priority in the preset risk area with a relatively large risk level of the user can be used as the preset risk type that needs to be detected as indicated by the risk detection strategy, which is beneficial to improving the effectiveness and accuracy of the risk detection of the target content to be published for the user.

[0060] In the embodiments of this specification, the risk detection process for the specified content with the preset quantity ratio included in the target content to obtain the risk detection result of the target content may include:

[0061] Using a risk detection model constructed based on a neural network algorithm, perform risk detection processing on the specified content with the preset quantity ratio included in the target content to obtain the initial risk detection result of the target content.

[0062] Correspondingly, the risk detection process for the risk of whether the preset risk type exists in the target content to obtain the risk detection result of the target content may include:

[0063] Using a risk detection model constructed based on a neural network algorithm, perform risk detection processing on the risk of whether the preset risk type exists in the target content to obtain the initial risk detection result of the target content.

[0064] Correspondingly, the risk detection process for the risk of whether the preset risk type exists in the specified content with the preset quantity ratio included in the target content to obtain the risk detection result of the target content may include:

[0065] Using a risk detection model constructed based on a neural network algorithm, perform risk detection processing on the risk of whether the preset risk type exists in the specified content with the preset quantity ratio included in the target content to obtain the initial risk detection result of the target content.

[0066] In the embodiments of this specification, by pre-training and optimizing a risk detection model constructed based on a neural network algorithm using training samples, the risk detection model can extract useful feature data and the correlations between the feature data from the content that the user needs to publish, so as to accurately evaluate various preset risks that may exist in the content that the user needs to publish. Based on this, a user content sample can be set in advance according to actual needs to train a risk detection model, and then the target content that needs to be published currently can be input into the risk detection model to obtain an initial risk detection result output by the risk detection model, which is used to reflect whether the target content carries a preset risk, and is convenient and fast.

[0067] In practical applications, the types of preset risks that the risk detection model can detect can be set according to actual needs. For example, it can check whether the user content has the risk of containing inappropriate content, whether there is a risk of using unauthorized images, videos, or music in the user content, and whether there is a risk of leaking other people's personal information or sensitive data in the user content; in addition, it can also check whether there is a risk of grammar / spelling errors and poor clarity of images, videos, audio, etc. in the user content, and no specific limitations are made in this regard.

[0068] In the embodiments of this specification, the risk detection strategy can also be used to indicate the preset conditions that need to be met for the manual review of the user's target content, so as to control the probability of manual review of the user's target content, and the probability of manual review can be positively correlated with the risk level of the user.

[0069] Correspondingly, the risk detection processing of the target content according to the risk detection strategy having a corresponding relationship with the risk level of the user to obtain the risk detection result of the target content may further include:

[0070] According to the initial risk detection result, it is judged whether the preset condition is met to obtain a first judgment result.

[0071] If the first judgment result indicates that the preset condition is met, the manual risk review result for the target content is obtained to obtain the risk detection result of the target content.

[0072] If the first judgment result indicates that the preset condition is not met, the initial risk detection result is determined as the risk detection result of the target content.

[0073] In the embodiments of this specification, in addition to using a risk detection model constructed based on a neural network algorithm to detect the risk of the target content to be published by a user, manual review is usually an essential part of the user content risk review process. By comprehensively assessing the risk of the target content to be published by a reviewer, the spread of inappropriate content can be further effectively prevented, which helps to improve the quality and compliance of the content successfully published by the user.

[0074] Based on this, in order to save human resources and make full use of the advantages of manual review to control the risk carried by the target content to be published by a user, a risk detection strategy that can indicate the preset conditions required for manual review of the target content for the user can be pre-configured. Moreover, the higher the risk level of the content to be published by the user, the greater the probability of manual review, and the lower the risk level of the content to be published by the user, the smaller the probability of manual review, which is beneficial to improving the efficiency and accuracy of risk detection and reducing the risk detection cost. Among them, the preset conditions can be set according to actual needs. For example, the preset conditions can include: the probability that the target content generated by using a risk detection model constructed based on a neural network algorithm has a risk reaches a preset threshold, or the target content has a specified risk, or the target content is content in a specified field, etc., and no specific limitation is made thereto.

[0075] Therefore, when performing risk detection processing on the target content according to a risk detection strategy corresponding to the risk level of the user, after generating an initial risk detection result of the target content by using a risk detection model constructed based on a neural network algorithm, the initial risk detection result of the target content can be combined to determine whether the preset conditions required for manual review of the target content to be published by the user are met. If so, the target content can be submitted to the reviewer so that the obtained manual risk review result for the target content can be used as the risk detection result of the target content. If not, the initial risk detection result of the target content generated by the risk detection model can be directly used as the risk detection result of the target content, which has good flexibility.

[0076] In the embodiments of this specification, after performing risk detection processing on the target content according to a risk detection strategy corresponding to the risk level of the user and obtaining the risk detection result of the target content, it may further include:

[0077] According to the risk detection result of the target content, determine whether the target content meets the preset content publication conditions to obtain a second determination result.

[0078] If the second determination result indicates that the target content meets the preset content publication conditions, a publication instruction for the target content is generated.

[0079] If the second judgment result indicates that the target content does not meet the preset content release condition, a prohibition release instruction for the target content is generated.

[0080] In the embodiments of this specification, the preset content release conditions for user content can usually be preconfigured according to actual needs; for example, the preset content release conditions may include: the user content does not carry a preset risk, or the risk level of the user content is lower than a threshold, or the quality score of the user content reaches a threshold, etc., which are not specifically limited herein. By generating a release instruction for the target content that meets the preset content release condition to externally release and disseminate the user content that meets the preset content release condition; and generating a prohibition release instruction for the target content that does not meet the preset content release condition to refuse to externally release and disseminate the user content that does not meet the preset content release condition, the quality and compliance of the user content released and disseminated externally can be effectively controlled.

[0081] In the embodiments of this specification, if the second judgment result indicates that the target content meets the preset content release condition, the method may further include:

[0082] Determine a content release strategy corresponding to the risk level of the user; wherein, the content release strategy is used to control the dissemination rate of the target content among the user group, and the dissemination rate is negatively correlated with the risk level of the user.

[0083] Correspondingly, the generation of the release instruction for the target content may include: generating a release instruction for the target content according to the content release strategy.

[0084] In the embodiments of this specification, although detecting the risk of the target content required to be released by the user and making a decision on whether to externally release the target content according to the risk detection result of the target content can effectively control the risk brought by the target content. However, there may still be some new risks or relatively hidden risks during the dissemination process of the target content, thus affecting the security and compliance of the dissemination process of the user content. Based on this, a content release strategy can also be preset according to actual needs, so that the dissemination rate of the target content released by users with a higher risk level is relatively low among the user group, and the dissemination rate of the target content released by users with a lower risk level is relatively high among the user group. So that after the target content is successfully externally released and disseminated, once it is detected that the dissemination of the target content has triggered negative public opinion or social repercussions, measures such as silencing, deleting, or reducing the visibility of the content can be quickly taken to prevent the spread of risky content. This is not only conducive to reducing the immediate impact of the spread of risky content, but also can gain time for subsequent corrective measures to effectively reduce the possibility of risk spread.

[0085] In practical applications, the dissemination rate of user content may be affected by various factors. Therefore, there can be various types of the content publishing strategy. For example, the content publishing strategy can be used to control the weights in the content recommendation algorithm to control the exposure rate, exposure time period, etc. of the user content. Or, the content publishing strategy can be used to control the number of users to whom the user content is pushed, the characteristics of the pushed user group, the push frequency, the number of push times, etc., and no specific limitation is made thereto. By generating a publishing instruction for the target content according to the content publishing strategy, when responding to the publishing instruction to publish and disseminate the target content, the dissemination rate of the target content can be indirectly controlled by controlling the exposure rate, exposure time period, number of users to whom the target content is pushed, the characteristics of the pushed user group, the push frequency, the number of push times, etc., and the flexibility is good.

[0086] In the embodiments of this specification, Figure 2 the method described may further include:

[0087] Storing the target content in a preset database to obtain the updated historical published content of the user. Or,

[0088] Storing the interaction data of the user group for the target content in a preset database to obtain the updated interaction data of the user group for the historical published content of the user. Or,

[0089] Storing the violation record data of the target content in a preset database or a preset blockchain network to obtain the updated violation record data of the historical published content of the user.

[0090] In the embodiments of this specification, whether or not the target content of the user is successfully published, the historical published content of the user can be updated using the target content, which is beneficial to ensuring the comprehensiveness and integrity of the updated historical published content of the user; and the violation record data of the historical published content of the user can also be updated using the violation record data of the target content, which is beneficial to ensuring the comprehensiveness and integrity of the updated violation record data of the historical published content of the user. After the target content of the user is successfully published, if there is interaction data of the user group for the target content, the interaction data of the historical published content of the user can also be updated using the interaction data of the user group for the target content to ensure the comprehensiveness and integrity of the updated interaction data of the historical published content of the user.

[0091] In practical applications, since the data stored in the blockchain network usually cannot be modified or deleted, and the decentralized feature of the blockchain network enables its data management process not to rely on a single manager or platform, the probability that the data stored in the blockchain network is unilaterally controlled or manipulated can be reduced. Based on this, a service provider can store the violation record data of the target content (and historical published content) of a user in a preset blockchain network to ensure the integrity and credibility of the relevant violation record data of the user, which is conducive to enhancing the trust between the user and the service provider. Of course, other types of user historical data of the user can also be stored in the preset blockchain network, and no specific limitation is made thereto.

[0092] In the embodiments of this specification, Figure 2 the method described in

[0093] may further include:

[0094] Determine whether a risk level update condition for the user is satisfied to obtain a third determination result.

[0095] If the third determination result indicates that the risk level update condition for the user is satisfied, obtain the latest user historical data generated based on the historical content publishing behavior of the user; wherein, the latest user historical data includes at least one of the updated historical published content, the updated interaction data, and the updated violation record data.

[0096] In the embodiments of this specification, the risk level update condition for the user can be configured in advance according to actual needs. For example, the risk level update condition can be that the time interval since the last update of the user's risk level reaches a threshold, or the number of contents published by the user in a recent preset time period reaches a threshold, or the number of violation records generated by the user in a recent preset time period reaches a threshold, etc., and no specific limitation is made thereto.

[0097] When the condition for updating the risk level for the user is met, the latest user historical data generated based on the user's historical content publishing behavior can be combined to evaluate the user's current risk level, which serves as the updated risk level of the user. Subsequently, when the user needs to publish content, the updated risk level of the user can be combined to determine the corresponding risk detection strategy and / or content publishing strategy, so as to perform risk detection on the content to be published by the user in combination with the risk detection strategy, and to control the dissemination rate of the content in combination with the content publishing strategy.

[0098] In the embodiments of the present specification, by combining the latest user historical data of the user to update the user's risk level, and being able to perform risk detection control and content publishing control on the content to be published by the user subsequently in combination with the updated risk level of the user, it is beneficial to encourage users to comply with content publishing norms, promote users' participation in content risk governance, thereby reducing the risks brought by users' published content at the source, contributing to creating a healthy and safe content ecological environment, and more effectively ensuring the quality and compliance of user content.

[0099] Based on the same idea, the embodiments of the present specification also provide a device corresponding to the above method. Please refer to Figure 3 , which is a schematic structural diagram of a risk detection device provided by the embodiments of the present specification. As Figure 3 shown, the risk detection device 3 can be implemented as all or part of an electronic device through software, hardware, or a combination of both. According to some embodiments, the risk detection device 3 may include a first acquisition module 31, a second acquisition module 32, and a risk detection module 33, where:

[0100] The first acquisition module 31 is configured to acquire the target content to be published by the user.

[0101] The second acquisition module 32 is configured to acquire the risk level of the user determined according to the user historical data; wherein, the user historical data includes data generated based on the user's historical content publishing behavior.

[0102] The risk detection module 33 is configured to perform risk detection processing on the target content according to a risk detection strategy having a corresponding relationship with the risk level of the user, and obtain a risk detection result of the target content; wherein, the strictness of the risk detection strategy is positively correlated with the risk level of the user.

[0103] Optionally, the second acquisition module 32 may include:

[0104] The first acquisition unit is configured to acquire preset types of user historical data generated based on the user's historical content publishing behavior.

[0105] A risk level assessment unit is used to process the user historical data by using a risk assessment model constructed based on a machine learning algorithm, and obtain a risk assessment result output by the risk assessment model, which is used to reflect the risk level of the user.

[0106] Optionally, the first acquisition unit may specifically be used for:

[0107] Acquire at least one of the historical published content of the user, the interaction data of the user group for the historical published content, and the violation record data of the historical published content, to obtain the user historical data; or,

[0108] According to a preset data analysis strategy, perform data analysis and processing on at least one of the historical published content of the user, the interaction data of the user group for the historical published content, and the violation record data of the historical published content, to obtain the user historical data.

[0109] Optionally, the risk detection module 33 may include:

[0110] A first determination unit is used to determine a risk detection strategy corresponding to the risk level of the user; wherein, the risk detection strategy is used to control the preset quantity ratio of the content that needs to be risk-detected in the target content of the user, and at least one of the preset risk types that need to be detected for the target content, and both the preset quantity ratio and the quantity of the preset risk types are positively correlated with the risk level of the user.

[0111] A first risk detection unit is used to perform risk detection processing on the specified content with the preset quantity ratio included in the target content according to the risk detection strategy, to obtain a risk detection result of the target content; or,

[0112] A second risk detection unit is used to perform risk detection processing on whether there is a risk of the preset risk type in the target content according to the risk detection strategy, to obtain a risk detection result of the target content; or,

[0113] A third risk detection unit is used to perform risk detection processing on whether there is a risk of the preset risk type in the specified content with the preset quantity ratio included in the target content according to the risk detection strategy, to obtain a risk detection result of the target content.

[0114] Optionally, the first risk detection unit may specifically be used for: using a risk detection model constructed based on a neural network algorithm, performing risk detection processing on the specified content with the preset quantity ratio included in the target content, to obtain an initial risk detection result of the target content.

[0115] Optionally, the second risk detection unit may specifically be used to: use a risk detection model constructed based on a neural network algorithm to perform risk detection processing on whether there is a risk of the preset risk type in the target content, and obtain an initial risk detection result of the target content.

[0116] Optionally, the third risk detection unit may specifically be used to: use a risk detection model constructed based on a neural network algorithm to perform risk detection processing on whether there is a risk of the preset risk type in the specified content with the preset quantity ratio included in the target content, and obtain an initial risk detection result of the target content.

[0117] Optionally, the risk detection strategy may also be used to indicate preset conditions that need to be met for manual review of the target content of the user, so as to control the probability of manual review of the target content of the user, and the probability of manual review is positively correlated with the risk level of the user.

[0118] Optionally, the risk detection module 33 may further include:

[0119] A judgment unit, configured to judge whether the preset conditions are met according to the initial risk detection result, and obtain a first judgment result.

[0120] A second acquisition unit, configured to obtain a manual risk review result for the target content and obtain a risk detection result of the target content if the first judgment result indicates that the preset conditions are met.

[0121] A second determination unit, configured to determine the initial risk detection result as the risk detection result of the target content if the first judgment result indicates that the preset conditions are not met.

[0122] Optionally, the target content may include at least one of video content, audio content, image content, and text content.

[0123] Optionally, Figure 3 the device in

[0124] A first condition judgment module, configured to judge whether the target content meets preset content release conditions according to the risk detection result of the target content, and obtain a second judgment result.

[0125] A first generation module, configured to generate a release instruction for the target content if the second judgment result indicates that the target content meets the preset content release conditions.

[0126] A second generation module, configured to generate a prohibition instruction for the target content if the second judgment result indicates that the target content does not meet the preset content release condition.

[0127] Optionally, Figure 3 the device in

[0128] A determination module, configured to determine a content release strategy corresponding to the risk level of the user if the second judgment result indicates that the target content meets the preset content release condition; wherein, the content release strategy is used to control the dissemination rate of the target content among the user group, and the dissemination rate is negatively correlated with the risk level of the user.

[0129] The first generation module may specifically be configured to: generate a release instruction for the target content according to the content release strategy.

[0130] Optionally, Figure 3 the device in

[0131] A first storage module, configured to store the target content in a preset database to obtain the updated historical release content of the user; or,

[0132] A second storage module, configured to store the interaction data of the user group for the target content in a preset database to obtain the updated interaction data of the user group for the historical release content of the user; or,

[0133] A third storage module, configured to store the violation record data of the target content in a preset database or a preset blockchain network to obtain the updated violation record data of the historical release content of the user.

[0134] Optionally, Figure 3 the device in

[0135] A second condition judgment module, configured to judge whether a risk level update condition for the user is met to obtain a third judgment result.

[0136] A third acquisition module, configured to obtain the latest user historical data generated based on the historical content release behavior of the user if the third judgment result indicates that the risk level update condition for the user is met; wherein, the latest user historical data includes at least one of the updated historical release content, the updated interaction data, and the updated violation record data.

[0137] An update module is configured to process the latest user historical data by using a risk assessment model constructed based on a machine learning algorithm, so as to obtain a risk assessment result output by the risk assessment model, which is used to reflect the updated risk level of the user. Wherein, the updated risk level of the user is used to determine a risk detection strategy required for risk detection and control of the content to be published for the user within a preset time period in the future, or is used to determine a content publishing strategy required for content publishing control of the content to be published for the user within a preset time period in the future.

[0138] The above device embodiment corresponds to the method embodiment. For specific descriptions, reference can be made to the description in the method embodiment part, which will not be elaborated here. The device embodiment is obtained based on the corresponding method embodiment and has the same technical effect as the corresponding method embodiment. For specific descriptions, reference can be made to the corresponding method embodiment.

[0139] This specification embodiment also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the risk detection method as described above Figure 2 as shown. For the specific execution process, reference can be made to the specific descriptions in the related embodiments of the risk detection method, which will not be elaborated here.

[0140] This specification also provides a computer program product, which stores at least one instruction. The at least one instruction is loaded and executed by the processor to implement the risk detection method as described above Figure 2 as shown. For the specific execution process, reference can be made to the specific descriptions in the related embodiments of the risk detection method, which will not be elaborated here.

[0141] This specification embodiment also provides Figure 4 a schematic structural diagram of the electronic device as shown. As Figure 4 shown, at the hardware level, the electronic device may include a processor, an internal bus, a network interface, a memory, and a non-volatile memory. Of course, it may also include other hardware required for other services. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to implement the risk detection method as described above Figure 2 as shown. For the specific execution process, reference can be made to the specific descriptions in the related embodiments of the risk detection method, which will not be elaborated here.

[0142] Of course, in addition to the software implementation, this specification does not exclude other implementation manners, such as a logic device or a combination of software and hardware. That is to say, the execution subject of the following processing flow is not limited to each logic unit, and may also be hardware or a logic device.

[0143] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other, and the differences between each embodiment and other embodiments are emphasized. In particular, for the computer-readable storage medium, computer program product, and Figure 4 electronic device shown, since it is basically similar to the method embodiment, the description is relatively simple, and for the relevant parts, reference can be made to the partial description of the method embodiment.

[0144] In the 1990s, improvements to a technology could be clearly distinguished as either hardware improvements (e.g., improvements to circuit structures such as diodes, transistors, switches, etc.) or software improvements (improvements to method flows). However, with the development of technology, many method flow improvements today can be regarded as direct improvements to hardware circuit structures. Almost all designers obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement to a method flow cannot be implemented using a hardware entity module. For example, a Programmable Logic Device (PLD) (such as a Field Programmable Gate Array (FPGA)) is an integrated circuit whose logic function is determined by the user programming the device. Designers can program themselves to "integrate" a digital system onto a single PLD, without having to ask a chip manufacturer to design and fabricate a dedicated integrated circuit chip. Moreover, nowadays, instead of manually fabricating integrated circuit chips, this programming is mostly implemented using "logic compiler" software, which is similar to the software compilers used in program development and writing. The original code before compilation also has to be written in a specific programming language, which is called a Hardware Description Language (HDL), and there is not just one type of HDL, but many types, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones currently are VHDL (Very-High-Speed Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also be aware that by simply performing a little logical programming on the method flow using the above-mentioned several hardware description languages and programming it into an integrated circuit, it is easy to obtain the hardware circuit that implements the logical method flow.

[0145] The controller can be implemented in any suitable manner. For example, the controller can take the form of, for example, a microprocessor or a processor and a computer-readable medium storing computer-readable program code (such as software or firmware) executable by the (micro)processor, logic gates, switches, an application specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of the controller include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that in addition to implementing the controller in the form of pure computer-readable program code, it is entirely possible to logically program the method steps to enable the controller to be implemented in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers, etc. to achieve the same function. Therefore, such a controller can be considered a hardware component, and the devices included therein for implementing various functions can also be regarded as the structures within the hardware component. Or even, the devices for implementing various functions can be regarded as either software modules for implementing the method or structures within the hardware component.

[0146] The systems, devices, modules, or units illustrated in the above embodiments can be specifically implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or any combination of these devices.

[0147] For the convenience of description, when describing the above devices, they are described separately as various units according to their functions. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0148] Those skilled in the art should understand that the embodiments of this specification can be provided as a method, a system, or a computer program product. Therefore, this specification can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program code.

[0149] This specification is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the specification. It should be understood that each flow and / or block in the flowchart and / or block diagram, and combinations of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general purpose computers, special purpose computers, embedded processors, or other programmable data processing devices to produce a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices produce means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0150] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufacture including instruction means that implement the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0151] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0152] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and memory.

[0153] The memory may include non-permanent memory in the form of computer-readable media, random access memory (RAM), and / or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of computer-readable media.

[0154] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0155] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0156] It should be understood by those skilled in the art that the embodiments of this specification may be provided as methods, systems or computer program products. Therefore, this specification may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0157] This specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0158] The various embodiments in this specification are described in a progressive manner. For the parts that are the same or similar among the various embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the system embodiments, since they are basically similar to the method embodiments, the description is relatively simple. For the related parts, reference can be made to the partial description of the method embodiments.

[0159] The above description is only for the embodiments of this specification and is not intended to limit this specification. For those skilled in the art, various changes and modifications can be made to this specification. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of this specification shall be included within the scope of the claims of this specification.

Claims

1. A risk detection method, the method comprising: Obtaining target content to be published by a user; Obtaining the risk level of the user determined according to the user's historical data; wherein, the user's historical data includes data generated based on the user's historical content publishing behavior; Performing risk detection processing on the target content according to a risk detection strategy corresponding to the risk level of the user, to obtain a risk detection result of the target content; wherein, the strictness of the risk detection strategy is positively correlated with the risk level of the user; the risk detection strategy is used to control the preset risk types that need to be detected for the target content, and is used to indicate the preset conditions that need to be met for manual review of the target content, so as to control the manual review probability for the target content, and both the number of the preset risk types and the manual review probability are positively correlated with the risk level of the user; the preset conditions include: the probability that the target content generated by the risk detection model has a risk reaches a preset threshold, or the target content has a specified risk, or the target content is content in a specified field.

2. The method according to claim 1, wherein the obtaining the risk level of the user determined according to the user's historical data comprises: Obtaining preset types of user historical data generated based on the user's historical content publishing behavior; Using a risk assessment model constructed based on a machine learning algorithm to process the user historical data, to obtain a risk assessment result output by the risk assessment model for reflecting the risk level of the user.

3. The method according to claim 2, wherein the obtaining preset types of user historical data generated based on the user's historical content publishing behavior comprises: Obtaining at least one of the user's historical published content, the interaction data of the user group for the historical published content, and the violation record data of the historical published content, to obtain the user historical data; or, Performing data analysis processing on at least one of the user's historical published content, the interaction data of the user group for the historical published content, and the violation record data of the historical published content according to a preset data analysis strategy, to obtain the user historical data.

4. The method according to claim 1, wherein the performing risk detection processing on the target content according to a risk detection strategy corresponding to the risk level of the user, to obtain a risk detection result of the target content, comprises: Determining a risk detection strategy corresponding to the risk level of the user; wherein, the risk detection strategy is used to control the preset quantity proportion of the content that needs to be risk-detected in the target content of the user, and at least one of the preset risk types that need to be detected for the target content, and both the preset quantity proportion and the number of the preset risk types are positively correlated with the risk level of the user; According to the risk detection strategy, risk detection processing is performed on the specified content with the preset quantity ratio included in the target content to obtain the risk detection result of the target content; or, According to the risk detection strategy, risk detection processing is performed on whether there is a risk of the preset risk type in the target content to obtain the risk detection result of the target content; or, According to the risk detection strategy, risk detection processing is performed on whether there is a risk of the preset risk type in the specified content with the preset quantity ratio included in the target content to obtain the risk detection result of the target content.

5. The method according to claim 4, wherein the performing risk detection processing on the specified content with the preset quantity ratio included in the target content to obtain the risk detection result of the target content comprises: Using a risk detection model constructed based on a neural network algorithm, performing risk detection processing on the specified content with the preset quantity ratio included in the target content to obtain an initial risk detection result of the target content; The performing risk detection processing on whether there is a risk of the preset risk type in the target content to obtain the risk detection result of the target content comprises: Using a risk detection model constructed based on a neural network algorithm, performing risk detection processing on whether there is a risk of the preset risk type in the target content to obtain an initial risk detection result of the target content; The performing risk detection processing on whether there is a risk of the preset risk type in the specified content with the preset quantity ratio included in the target content to obtain the risk detection result of the target content comprises: Using a risk detection model constructed based on a neural network algorithm, performing risk detection processing on whether there is a risk of the preset risk type in the specified content with the preset quantity ratio included in the target content to obtain an initial risk detection result of the target content.

6. The method according to claim 5, wherein the risk detection strategy is further used to indicate preset conditions required for manual review of the target content of the user, so as to control the manual review probability of the target content of the user, and the manual review probability is positively correlated with the risk level of the user; The performing risk detection processing on the target content according to the risk detection strategy having a corresponding relationship with the risk level of the user to obtain the risk detection result of the target content further comprises: According to the initial risk detection result, judging whether the preset conditions are met to obtain a first judgment result; If the first judgment result indicates that the preset conditions are met, obtaining a manual risk review result for the target content to obtain the risk detection result of the target content; If the first judgment result indicates that the preset conditions are not met, determining the initial risk detection result as the risk detection result of the target content.

7. The method according to claim 6, wherein the target content includes: At least one of video content, audio content, image content, and text content.

8. According to the method described in any one of claims 1-7, after performing risk detection processing on the target content according to the risk detection strategy corresponding to the risk level of the user to obtain the risk detection result of the target content, the method further includes: Judging whether the target content meets the preset content release condition according to the risk detection result of the target content to obtain a second judgment result; If the second judgment result indicates that the target content meets the preset content release condition, generating a release instruction for the target content; If the second judgment result indicates that the target content does not meet the preset content release condition, generating a prohibited release instruction for the target content.

9. According to the method described in claim 8, if the second judgment result indicates that the target content meets the preset content release condition, the method further includes: Determining a content release strategy corresponding to the risk level of the user; wherein, the content release strategy is used to control the dissemination rate of the target content among the user group, and the dissemination rate is negatively correlated with the risk level of the user; The generating a release instruction for the target content includes: Generating a release instruction for the target content according to the content release strategy.

10. According to the method described in claim 9, the method further includes: Storing the target content in a preset database to obtain the updated historical release content of the user; Or, Storing the interaction data of the user group for the target content in a preset database to obtain the updated interaction data of the user group for the historical release content of the user; or, Storing the violation record data of the target content in a preset database or a preset blockchain network to obtain the updated violation record data of the historical release content of the user.

11. According to the method described in claim 10, the method further includes: Judging whether the risk level update condition for the user is met to obtain a third judgment result; If the third judgment result indicates that the risk level update condition for the user is met, obtaining the latest user historical data generated based on the historical content release behavior of the user; wherein, the latest user historical data includes at least one of the updated historical release content, the updated interaction data, and the updated violation record data; Processing the latest user historical data by using a risk assessment model constructed based on a machine learning algorithm to obtain a risk assessment result output by the risk assessment model for reflecting the updated risk level of the user; wherein, the updated risk level of the user is used to determine the risk detection strategy required for risk detection control of the content to be released by the user within a future preset time period, or to determine the content release strategy required for content release control of the content to be released by the user within a future preset time period.

12. A risk detection device, comprising: A first acquisition module, configured to acquire the target content to be released by the user; A second acquisition module, configured to acquire the risk level of the user determined according to the user's historical data; wherein, the user's historical data includes data generated based on the user's historical content publishing behavior; A risk detection module, configured to perform risk detection processing on the target content according to a risk detection strategy corresponding to the risk level of the user, to obtain a risk detection result of the target content; wherein, the strictness of the risk detection strategy is positively correlated with the risk level of the user; the risk detection strategy is used to control the preset risk types to be detected for the target content, and is used to indicate the preset conditions that need to be met for manual review of the target content, so as to control the probability of manual review for the target content, and both the number of the preset risk types and the probability of manual review are positively correlated with the risk level of the user; the preset conditions include: the probability that the target content generated by the risk detection model has a risk reaches a preset threshold, or the target content has a specified risk, or the target content is content in a specified field.

13. A computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 11 are implemented.

14. An electronic device, comprising: A processor and a memory; wherein, the memory stores a computer program, and the computer program is adapted to be loaded and executed by the processor to perform the steps of the method according to any one of claims 1 to 11.

15. A computer program product, on which at least one instruction is stored, and when the at least one instruction is executed by a processor, the steps of the method according to any one of claims 1 to 11 are implemented.

Images