Webpage violation alarm method and device based on large model, equipment and medium

Abstract

The invention provides a webpage violation warning method and device based on a large model, equipment and a medium, and relates to the technical field of artificial intelligence such as large models, the method comprises the following steps: obtaining a to-be-detected target webpage, and identifying whether the target webpage meets a violation detection triggering condition or not; in response to the fact that the target webpage is recognized to meet the violation detection triggering condition, extracting a detection text object and a detection image object in the target webpage to obtain a detection object set of the target webpage; inputting the detection object set into a pre-acquired large model, performing violation detection on the detection object set through the model capability of the large model, and identifying whether a violation object exists in the detection object set; and in response to the fact that the illegal object exists in the detection object set, obtaining a webpage user of the target webpage, and sending a webpage illegal alarm to the webpage user.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing, and in particular to the field of artificial intelligence technologies such as large models. Background Art

[0002] With the development of technology, the Internet is becoming more and more important in people's work and life. However, with the development of network technology, the illegal information in the network environment has caused a certain degree of impact on people's work and life. In this scenario, it is necessary to detect and process the illegal information in the network environment.

[0003] In the related art, illegal information on web pages can be checked and detected manually, but the manual reliance is high and the efficiency is low. Summary of the invention

[0004] The present disclosure proposes a web page violation warning method, device, equipment and medium based on a large model.

[0005] According to a first aspect of the present disclosure, a web page violation alarm method based on a large model is proposed, comprising: acquiring a target web page to be detected, and identifying whether the target web page satisfies a violation detection trigger condition; in response to identifying that the target web page satisfies the violation detection trigger condition, extracting a detection text object and a detection image object in the target web page to obtain a detection object set of the target web page; inputting the detection object set into a pre-acquired large model, performing violation detection on the detection object set through the model capability of the large model, and identifying whether there are violation objects in the detection object set; in response to identifying that there are violation objects in the detection object set, acquiring a web page user of the target web page, and issuing a web page violation alarm to the web page user.

[0006] According to a second aspect of the present disclosure, a web page violation alarm device based on a large model is proposed, comprising: an identification module, used to obtain a target web page to be detected, and identify whether the target web page satisfies a violation detection trigger condition; an extraction module, used to extract detection text objects and detection image objects in the target web page in response to identifying that the target web page satisfies the violation detection trigger condition, and obtain a detection object set of the target web page; a detection module, used to input the detection object set into a pre-acquired large model, perform violation detection on the detection object set through the model capability of the large model, and identify whether there are violation objects in the detection object set; an alarm module, used to obtain a web page user of the target web page in response to identifying that there are violation objects in the detection object set, and issue a web page violation alarm to the web page user.

[0007] According to a third aspect of the present disclosure, an electronic device is proposed, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the large model-based web page violation alarm method proposed in the first aspect above.

[0008] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is proposed, wherein the computer instructions are used to enable the computer to execute the large model-based web page violation alarm method proposed in the first aspect above.

[0009] According to a fifth aspect of the present disclosure, a computer program product is proposed, including a computer program, which, when executed by a processor, implements the large model-based web page violation warning method proposed in the first aspect.

[0010] 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

[0011] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0012] Figure 1 A schematic diagram of a flow chart of a web page violation warning method based on a large model according to an embodiment of the present disclosure;

[0013] Figure 2 A schematic diagram of a flow chart of a web page violation warning method based on a large model according to another embodiment of the present disclosure;

[0014] Figure 3 A schematic diagram of a flow chart of a web page violation warning method based on a large model according to another embodiment of the present disclosure;

[0015] Figure 4 A schematic diagram of a flow chart of a web page violation warning method based on a large model according to another embodiment of the present disclosure;

[0016] Figure 5 This is a structural diagram of a web page violation warning device based on a large model according to an embodiment of the present disclosure;

[0017] Figure 6 A schematic block diagram of an electronic device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0018] 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.

[0019] Data processing is a basic part of system engineering and automatic control. Data is a form of expression of facts, concepts or instructions, which can be processed by manual or automatic devices. After data is interpreted and given a certain meaning, it becomes information. Data processing is the collection, storage, retrieval, processing, transformation and transmission of data. The basic purpose of data processing is to extract and derive valuable and meaningful data for certain specific people from a large amount of data that may be disorganized and difficult to understand.

[0020] A large model refers to a machine learning model with large-scale parameters and complex computational structures. It is usually built from deep neural networks and has billions or even hundreds of billions of parameters. The purpose of designing a large model is to improve the model's expressiveness and predictive performance, and to be able to handle more complex tasks and data. Large models are widely used in various fields, including natural language processing, computer vision, speech recognition, and recommendation systems.

[0021] The LLM in the big model is a type of natural language processing model based on deep learning. Its main features are huge model parameters and complex neural network structure, strong language understanding, context perception and language generation capabilities. It can automatically learn useful feature representations from input data and generate relevant text.

[0022] Artificial Intelligence (AI) is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. Artificial Intelligence is a branch of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing and expert systems. Since the birth of artificial intelligence, the theory and technology of artificial intelligence have become increasingly mature, and the application field has continued to expand. It can be imagined that the technological products brought by artificial intelligence in the future will be the "container" of human wisdom. Artificial Intelligence can simulate the information process of human consciousness and thinking.

[0023] Figure 1FIG. 1 is a flow chart of a webpage violation warning method based on a large model according to an embodiment of the present disclosure. Figure 1 As shown, the method includes:

[0024] S101, obtaining a target web page to be detected, and identifying whether the target web page meets a violation detection trigger condition.

[0025] In implementation, people can obtain information by browsing web pages. In this scenario, illegal information may appear in the web pages that people browse. The illegal information may include information that violates relevant laws and regulations, which is not specifically limited here.

[0026] In the disclosed embodiment, a web page that needs to be detected for violations may be marked as a target web page to be detected. Before the target web page is detected for violations, it may be identified whether there is a need to detect violation information on the target web page.

[0027] Optionally, a preset identification method corresponding to a suspected illegal web page can be obtained, and based on the method, it can be identified whether the current target web page meets the preset violation detection trigger condition, and then the web page needs to be detected for illegal information. It can be understood that based on this identification step, the web pages in the network environment can be preliminarily screened, and web pages that may contain illegal information can be screened out as target web pages to be detected.

[0028] As an example, Figure 2 As shown, it can be Figure 2 The browser software development kit (Browser Software Development Kit, browser SDK) shown collects web page information related to the target web page and transmits the collected information to Figure 2 The trigger shown is Figure 2 The trigger shown identifies whether the target webpage meets the violation detection trigger condition. When it is identified that the target webpage meets the violation detection trigger condition, Figure 2 The trigger shown initiates the violation detection process for the target web page.

[0029] in, Figure 2 The trigger shown may be a Hypertext Transfer Protocol (http) trigger, or may be another type of trigger, which is not specifically limited here.

[0030] S102 , in response to identifying that the target webpage meets the violation detection triggering condition, extracting detection text objects and detection image objects in the target webpage to obtain a detection object set of the target webpage.

[0031] In the disclosed embodiment, when it is identified that a target webpage satisfies a violation detection trigger condition, a violation detection process for the target webpage can be started, and webpage information extraction and violation detection for the target webpage can be started based on the start of the process.

[0032] Optionally, the text information and image information in the target web page can be extracted respectively, and then based on the extracted text information and image information, text objects and image objects required for illegal information detection are generated and marked as detection text objects and detection image objects respectively.

[0033] As an example, Figure 2 As shown, it can be Figure 2 The detection service module shown extracts text information and image information from the target web page, and integrates the extracted text information and image information based on a preset information integration algorithm, thereby generating a detection text object and a detection image object for violation detection.

[0034] Furthermore, a set consisting of the detected text objects and the detected image objects is marked as a detection object set of the target web page.

[0035] S103, inputting the detection object set into the pre-acquired large model, performing violation detection on the detection object set through the model capability of the large model, and identifying whether there are any violation objects in the detection object set.

[0036] In the disclosed embodiment, violation detection of the target web page can be performed through the model capability of the large model, wherein the acquired detection object set can be input into the pre-acquired large model, and the detection text objects and detection image objects in the detection object set can be respectively detected for violations through the large model, and then the corresponding violation detection results can be output.

[0037] Among them, the objects indicating the presence of violation information in the violation detection result can be marked as violation objects in the detection object set.

[0038] Optionally, for violation detection of the detected text object, the model capability of the large model can be used to detect and identify violation keywords of the detected text object, and determine whether the detected text object is a violation object based on the recognition result of the violation keywords.

[0039] Optionally, for violation detection of a detected image object, the image elements in the detected image object may be extracted through the model capability of the large model, and whether the detected image object is a violation object may be identified based on the extracted image elements.

[0040] Further, when the detected text object is identified as a violation object and / or the detected image object is identified as a violation object, it can be determined that there is a violation object in the detected object set.

[0041] S104, in response to identifying that there is a violation object in the detection object set, obtaining a web page user of the target web page, and issuing a web page violation warning to the web page user.

[0042] In the disclosed embodiment, when it is identified that the target webpage contains illegal information, it is necessary to issue an alarm to the browser of the webpage. In this scenario, the browser of the target webpage can be marked as the webpage user of the target webpage.

[0043] Optionally, relevant violation information of the violation object in the target webpage may be obtained, and according to the warning information normally corresponding to the violation information, a webpage violation warning of the target webpage may be issued to the webpage user based on a preset warning method.

[0044] Among them, the web page violation alarm can be issued when the web page user logs in to the target web page, and the alarm information of the web page violation alarm of the target web page can also be sent through the communication method reserved by the web page user, which is not specifically limited here.

[0045] The webpage violation warning method based on the large model proposed in the present disclosure obtains the target webpage to be detected and identifies whether the target webpage meets the violation detection triggering condition. After identifying that the target webpage meets the violation detection triggering condition, the violation detection process of the target webpage is started, wherein the detection text object and the detection image object in the target webpage can be extracted to obtain the corresponding detection object set, and the detection object set is subjected to violation detection through the model capability of the large model, and whether there is a violation object in the detection object set is identified, and when the violation object is identified in the detection object set, a webpage violation warning is issued to the webpage user of the target webpage. In the present disclosure, the violation detection of the detection object set is performed through the model capability of the large model, which reduces the degree of dependence of the violation detection of the target webpage on manual work, improves the detection efficiency and accuracy of the text violation detection and image violation detection of the target webpage, and when the violation object is identified in the detection object set of the target webpage, a webpage violation warning is issued to the webpage user of the target webpage, which improves the vigilance of the webpage user, reduces the possibility of the webpage user being polluted by the violation information, reduces the possibility of the interests of the webpage user being damaged, optimizes the security and stability of the network environment, and improves the network usage experience.

[0046] In the above embodiment, the extraction of the detection object set can be combined with Figure 3 Further understanding, Figure 3 FIG. 1 is a flow chart of a webpage violation warning method based on a large model according to another embodiment of the present disclosure. Figure 3 As shown, the method includes:

[0047] S301, obtaining a target web page to be detected, and identifying whether the target web page meets a violation detection trigger condition.

[0048] Optionally, a violation detection request is generated for the target web page, and whether the target web page satisfies a violation detection trigger condition is identified based on the violation detection request, wherein, in response to identifying that the target web page satisfies the violation detection trigger condition, a preset cloud computing module is accessed, and violation detection is performed on the target web page through the cloud computing service provided by the cloud computing module.

[0049] As an example, Figure 2 As shown, it can be Figure 2 The browser software development kit shown extracts the web page information of the target web page, and generates a corresponding violation detection request based on the extracted web page information, and transmits it to Figure 2 The trigger module shown in this example can call a preset violation detection trigger condition and compare the received web page information with the violation detection trigger condition. When the comparison result indicates that the web page information meets the violation detection trigger condition, it can be determined that the target web page can start to detect violation information on the web page.

[0050] Optionally, when the trigger identifies that the target web page meets the violation detection trigger condition, Figure 2 The access service shown accesses the cloud computing module, and performs violation detection on the target web page by calling the cloud computing service provided by the cloud computing module.

[0051] S302, obtaining web page information and web page screenshots of the target web page.

[0052] In the disclosed embodiment, the text information related to the target webpage may be marked as webpage information, wherein the webpage information may include the name of the target webpage and a web link (URL), etc.

[0053] As an example, Figure 2 As shown, it can be Figure 2 The browser software development kit shown realizes the extraction of web page information of a target web page.

[0054] Furthermore, the target web page can be captured by a preset screenshot tool to obtain a screenshot of the target web page. For example, Figure 2 The browser software development kit shown implements image capture processing of a target web page, thereby obtaining a web page screenshot of the target web page.

[0055] S303, performing text scanning on the target web page, extracting suspected illegal sentences in the target web page, and obtaining a detection text object according to the web page information and the suspected illegal sentences.

[0056] Optionally, a text scan is performed on the target web page to obtain web page text information of the target web page, and suspected illegal sentences are extracted from the web page text information.

[0057] In the disclosed embodiment, a text scan can be performed on the target web page based on a text scanning and extraction algorithm in the related technology, and then the web page statements of the target web page can be extracted according to the scanning processing results, and some sentences that may contain illegal information can be identified from the scanned web page statements through a violation statement identification strategy, and marked as suspected illegal statements of the target web page.

[0058] Optionally, the violation type to which the suspected violation statement belongs is obtained, vector concatenation is performed on the suspected violation statement, the violation type and the web page information to obtain a concatenated first detection vector, and the first detection vector is determined as a detection text object of the target web page.

[0059] In the disclosed embodiment, the violation information that violates relevant laws and regulations that appears on a web page may include multiple violation types. In this scenario, for the suspected violation statement on the target web page, identification conditions of the multiple violation types can be obtained, and the suspected violation statement can be compared with each identification condition respectively. The violation type to which the suspected violation statement belongs can be determined from the multiple violation types through the comparison result.

[0060] Optionally, according to the vector conversion algorithm in the relevant technology, the suspected illegal statements, the illegal types and the web page information of the target web page are respectively converted into vectors, and according to the vector splicing method in the relevant technology, the three vectors obtained after the conversion are spliced ​​to obtain the corresponding splicing vector as the first detection vector, wherein the first detection vector can be marked as the detection text object in the target web page.

[0061] As an example, Figure 2 As shown, it can be Figure 2 The detection service shown extracts suspected illegal sentences in the target web page, determines the illegal type of the suspected illegal sentence, and performs vector concatenation on the suspected illegal sentence, illegal type and web page information to obtain a first detection vector, and uses it as a detection text object for illegal detection.

[0062] In this example, Figure 2 The detection service shown can convert the suspected violation statement, violation type and web page information from text to vector by calling the embedding service in the cloud computing service provided by the cloud computing module, thereby obtaining the corresponding first detection vector.

[0063] Optionally, by Figure 2The call of the first vector database service (Bes service) shown can store the first detection vector in a preset vector library. In this example, the stored vector data of the first detection vector may include <vector of suspected violation statement, identification (Identification, id) of suspected violation statement, text content of suspected violation statement, violation type of suspected violation statement, database storage type of suspected violation statement>.

[0064] S304, performing image scanning on the target webpage, extracting suspected illegal images in the target webpage, and obtaining a detection image object according to the webpage information and the suspected illegal images.

[0065] Optionally, an image scan is performed on the target web page to obtain web page image information of the target web page, and suspected illegal images are extracted from the web page image information.

[0066] In the disclosed embodiments, a text scan can be performed on a target web page based on an image scanning algorithm in the related art, and then web page image information of the target web page can be extracted based on the scanning processing result. Furthermore, the extracted web page image information can be scanned for illegal image elements, and based on the scanning results, some images that may contain illegal information can be screened out from all the web page image information and marked as suspected illegal images of the target web page.

[0067] As an example, you can use Figure 2 The image storage service (MySQL service) shown is called to store the suspected illegal images of the target webpage in a preset image database, wherein the storage data of the suspected illegal images in the database may include <the identification of the suspected illegal images, the illegal image feature information of the suspected illegal images, and the illegal type of the suspected illegal images>

[0068] Optionally, the violation type to which the suspected violation image belongs is obtained, vector splicing is performed on the suspected violation image, the violation type and the webpage screenshot to obtain a spliced ​​second detection vector, and the second detection vector is determined as the detection image object of the target webpage.

[0069] In the disclosed embodiment, for the suspected illegal image on the target web page, identification conditions of each of the multiple illegal types can be obtained, and the suspected illegal image can be compared with each identification condition respectively, and the illegal type to which the suspected illegal image belongs can be determined from the multiple illegal types through the comparison result.

[0070] Optionally, according to the vector conversion algorithm in the relevant technology, vector conversion is performed on the suspected illegal image, illegal type and web page screenshot of the target web page respectively, and according to the vector splicing method in the relevant technology, the three vectors obtained after the conversion are spliced ​​to obtain the corresponding splicing vector as the second detection vector, wherein the second detection vector can be marked as the detection image object in the target web page.

[0071] As an example, Figure 2 As shown, it can be Figure 2 The detection service shown extracts suspected illegal images from the target web page, determines the illegal type of the suspected illegal images, and performs vector splicing on the suspected illegal sentences, illegal types and web page screenshots to obtain a second detection vector, and uses it as a detection image object for illegal detection.

[0072] In this example, Figure 2 The detection service shown can convert suspected violation images, violation types, and web page screenshots from images to vectors by calling the image embedding service (Image to Vector, Img2Vec) in the cloud computing service provided by the cloud computing module, thereby obtaining a corresponding second detection vector.

[0073] Optionally, by Figure 2 The call of the second vector database service (Bes service) shown can store the second detection vector in a preset vector library. In this example, the stored vector data of the second detection vector can include <vector of suspected illegal image, identification (Identification, id) of suspected illegal image, database storage type of suspected illegal image>.

[0074] S305: Obtain a detection object set of the target web page based on the detection of text objects and the detection of image objects.

[0075] Optionally, an algorithm may be constructed based on a data set in related technologies, and algorithm processing may be performed on detected text objects and detected image objects, and a detection object set consisting of detected text objects and detected image objects may be obtained according to the results of the algorithm processing.

[0076] The web page violation alarm method based on a large model proposed in the present disclosure, when identifying that a target web page meets the violation detection triggering condition, is connected to the cloud computing module, and the target web page is detected for violation through the computing service provided by the cloud computing module. There is no need to develop and train a separate model, thereby reducing the resource consumption cost of violation detection. Through the detection service provided by the cloud computing module, the detection text object and the detection image object of the target web page are extracted, thereby improving the accuracy and efficiency of the extraction of the detection text object and the detection image object, and providing accurate data support for subsequent violation detection of the target web page.

[0077] In the above embodiment, the violation detection of the target web page can be combined with Figure 4 understand, Figure 4 FIG. 1 is a flow chart of a webpage violation warning method based on a large model according to another embodiment of the present disclosure. Figure 4 As shown, the method includes:

[0078] S401, obtaining a preset violation detection strategy library, and obtaining a first target detection strategy corresponding to a detection text object in a detection object set and a second target detection strategy corresponding to a detection image object from the violation detection strategy library.

[0079] In the disclosed embodiment, multiple violation types that may occur in the current network environment can be obtained, and corresponding violation detection strategies can be set for each violation type, thereby obtaining violation detection strategies for each violation type, and marking the database storing the violation detection strategies as a violation detection strategy library.

[0080] Optionally, for the detection text object of the target web page, the violation type to which the detection text object belongs can be obtained, and the violation detection strategy associated with the violation type can be obtained from the violation detection strategy library as the strategy used when performing violation detection on the detection text object, and marked as the first target detection strategy for the detection text object.

[0081] And, obtaining the violation type to which the detected image object belongs, and obtaining the associated strategy of the violation type from the violation detection strategy library as the second target detection strategy used by the detected image object when performing violation detection.

[0082] Optionally, according to the matching algorithm in the relevant technology, the matching degree between each violation detection strategy and the detection text object can be calculated, and the matching degree between each violation detection strategy and the detection image object can be calculated, and then according to the calculation results, the violation detection strategy for matching the detection text object is obtained from the violation detection strategy library as the first target detection strategy, and according to the calculation results, the violation detection strategy for matching the detection image object is obtained from the violation detection strategy library as the second target detection strategy.

[0083] S402, calling the first target detection strategy through the large model to perform violation detection on the detection text object to obtain a first violation detection result of the detection text object.

[0084] In the embodiment of the present disclosure, the violation detection of the detected text object can be performed by calling the model capability of the large model, wherein the detection Figure 2 The model service shown calls the model capabilities of the large model to implement violation detection of text objects.

[0085] Optionally, the first target detection strategy for detecting text objects can be called through the model capability of the large model, and the detection text objects input into the large model can be subjected to violation detection through the called first target detection strategy.

[0086] Among them, in the scenario where the first target detection strategy is built based on illegal keywords, the model capabilities of the large model can be used to perform illegal keyword detection on the detection text object based on the reference illegal keywords carried in the first target detection strategy, and the corresponding detection results can be obtained and output.

[0087] Furthermore, in the scenario where the first target detection strategy is built based on semantic analysis, the model capabilities of the large model can be used to perform semantic analysis detection on the detection text object based on the reference violation semantics defined by the first target detection strategy, and the corresponding detection results can be obtained and output.

[0088] Furthermore, the output result of the large model is marked as the first violation detection result of detecting the text object.

[0089] S403, calling the second target detection strategy through the large model to perform violation detection on the detection image object to obtain a second violation detection strategy for the detection image object.

[0090] In the embodiment of the present disclosure, the specific violation detection process for detecting image objects can be understood by referring to the violation detection process for detecting text objects proposed in the above step S402, which will not be repeated here.

[0091] S404: Identify whether there is a violation object in the detection object set according to the first violation detection result and the second violation detection result.

[0092] In the disclosed embodiment, a second review and verification may be performed on the first violation detection result and the second violation detection result output by the large model, wherein the first violation detection result and the second violation detection result may be subjected to a violation review and verification to obtain a first violation verification result of the first violation detection result and a second violation verification result of the second violation detection result.

[0093] Among them, the secondary audit verification can be marked as a violation audit verification, and the first violation detection result and the second violation detection result can be identified through the violation audit verification to perform a secondary audit confirmation, thereby obtaining a corresponding audit confirmation result.

[0094] Among them, the verification result obtained by performing violation review verification on the first violation detection result can be marked as the first violation verification result, and the verification result obtained by performing violation review verification on the second violation detection result can be marked as the second violation verification result.

[0095] In some implementations, a preset secondary violation audit and verification strategy can be obtained, and the first violation detection result and the second violation detection result can be audited and confirmed based on the strategy, thereby obtaining a first violation verification result corresponding to the first violation detection result and a second violation verification result corresponding to the second violation detection result.

[0096] In other implementations, the first violation detection result and the second violation detection result can be manually reviewed and verified for violations, thereby obtaining a first violation verification result corresponding to the first violation detection result and a second violation verification result corresponding to the second violation detection result.

[0097] As an example, Figure 2 As shown, it can be Figure 2 The alarm service shown is through Figure 2 The communication methods shown, such as SMS, phone call, and email, are used to send a secondary violation review and verification audit notification to the staff. In this scenario, upon receiving the audit notification, the staff can start a secondary violation review and verification of the first violation detection result and the second violation detection result, and obtain a first violation verification result corresponding to the first violation detection result, and a second violation verification result corresponding to the second violation detection result.

[0098] Optionally, in response to the first violation detection result indicating that the detected text object violates the rule, and the first violation verification result indicating that the verification passes, it is determined that there is a violation object in the detected object set, and the detected text object is determined as a violation object.

[0099] In the disclosed embodiment, based on the first violation detection result and the first violation verification result, whether the detected text object is a violation object can be identified and confirmed. It can be understood that when the first violation verification result indicates that the first violation detection result verification has passed, it can be determined that the current violation detection result of the detected text object is correct.

[0100] In this scenario, when the first violation detection result indicates that the detected text object violates the rules, it can be determined that the detected text object is a violation object in the detection object set of the target webpage.

[0101] Optionally, in response to the second violation detection result indicating that the detected image object violates the rules and the second violation verification result indicating that the verification is passed, it is determined that there is a violation object in the detected object set, and the detected image object is determined as a violation object.

[0102] In the disclosed embodiment, based on the second violation detection result and the second violation verification result, whether the detected image object is a violation object can be identified and confirmed. It can be understood that when the second violation verification result indicates that the second violation detection result verification has passed, it can be determined that the current violation detection result of the detected image object is correct.

[0103] In this scenario, when the second violation detection result indicates that the detected image object violates the rules, it can be determined that the detected image object is a violation object in the detection object set of the target webpage.

[0104] It should be noted that the illegal objects in the detection object set may be detection text objects, may be detection image objects, and may also include detection text objects and detection image objects.

[0105] S405 , in response to identifying that there is a violation object in the detection object set, obtaining a web page user of the target web page, and issuing a web page violation warning to the web page user.

[0106] Optionally, violation information corresponding to the violation object is obtained, and corresponding pop-up window warning information is generated according to the violation information, and a pop-up window warning is issued to the web page user based on the pop-up window warning information to issue a web page violation warning.

[0107] In the embodiment of the present disclosure, when it is identified that there is an illegal object in the detection object set, the target web page can be determined to be an illegal web page. In this scenario, relevant alarm processing needs to be performed on the illegal web page.

[0108] Optionally, violation information corresponding to the violating object can be obtained, which may include the violation type to which the violating object belongs, etc., and the violation information can be algorithmically processed based on a preset alarm information generation algorithm, and corresponding alarm information can be generated according to the result of the algorithm processing as pop-up alarm information of the target web page.

[0109] In this scenario, when the target web page is browsed, the pop-up warning information can pop up when the target web page is displayed to inform the web page user of the violation information of the target web page, thereby achieving the purpose of pop-up warning to the web page user, thereby realizing the web page violation alarm of the target web page.

[0110] In the implementation, there is a certain degree of correlation between the dissemination scope of the target web page and the degree of illegal information pollution caused by the target web page to people. In this scenario, the historical number of visits to the target web page can be obtained, and the violation level of the target web page can be determined based on the historical number of visits. The violation handling strategy of the target web page can be determined based on the violation level, and the target web page can be corrected according to the violation handling strategy.

[0111] As an example, Figure 2 As shown, it can be Figure 2 The counting service shown counts the number of visits to the target web page within a set historical time range, wherein the set historical time range is from the past 1 hour to 24 hours. In this example, Figure 2The counting service shown counts the number of visits to the target webpage in the past 1 hour to 24 hours as the historical number of visits to the target webpage.

[0112] In the disclosed embodiment, there is a correlation between the number of visits and the violation level of the target web page. It can be understood that, for two target web pages, the scope of dissemination of the violation information of the target web page with a higher number of visits is greater than the scope of dissemination of the violation information of the target web page with a lower number of visits. Therefore, it can be determined that the violation level of the target web page with a higher number of visits is higher than the violation level of the target web page with a lower number of visits.

[0113] In this scenario, multiple preset violation levels may be obtained, and a violation level matching the historical visit count of the target webpage may be identified from the multiple violation levels, and the violation level is the violation level of the target webpage.

[0114] Furthermore, a preset violation handling strategy set is obtained, and a violation handling strategy corresponding to the target webpage is determined from the violation handling strategy set according to the violation level of the target webpage, and violation correction processing is performed on the target webpage based on execution of the strategy.

[0115] Among them, the violation correction process may include closing the target web page, or may include rectifying and deleting the illegal information in the target web page, or may include other processing methods, which are not specifically limited here.

[0116] In the disclosed embodiment, when a target web page is identified as an illegal web page, the illegal web page may be stored to provide a reference for subsequent web page violation detection.

[0117] Optionally, a preset list of illegal web pages can be obtained, and in response to identifying the presence of illegal objects in the detection object set, the target web page is determined to be an illegal web page, the web page information and violation information of the illegal web page are obtained, and the illegal web page table is updated based on the web page information and violation information to obtain an updated new list of illegal web pages.

[0118] In the disclosed embodiment, a record list of related web page information of historical illegal web pages can be marked as an illegal web page list, and illegal information detection of web pages can be implemented through reference data provided in the illegal web page list.

[0119] As an example, Figure 2 As shown, it can be Figure 2 The data storage module shown stores the illegal web pages to Figure 2 The audit database shown is updated to obtain a new list of illegal web pages.

[0120] Optionally, when it is determined that the target web page is an illegal web page with an illegal object, the relevant information of the target web page can be stored and recorded in the preset illegal web page list, wherein the web page information of the target web page and the relevant illegal information of the illegal object can be obtained, and the illegal web page list is recorded based on the obtained web page information and the illegal information, thereby realizing the update of the illegal web page list and obtaining a new illegal web page list.

[0121] The web page violation alarm method based on a large model proposed in the present invention calls the first target detection strategy through the large model to perform violation detection on the text object, and calls the second target detection strategy through the large model to perform violation detection on the image object, thereby improving the violation detection efficiency of the text object and the image object, performing a secondary audit and verification on the first violation detection result and the second violation detection result, thereby improving the violation detection accuracy of the text object and the image object, updating the list of illegal web pages based on the web page information and violation information of the illegal web pages, providing reference data for subsequent web page violation detection, reducing the detection complexity of subsequent illegal web pages of the same type, determining the violation level of the illegal web page through the historical number of visits to obtain the corresponding violation processing strategy, reducing the possibility of excessive correction processing of the target web page, thereby reducing the possibility of wasting web page resources, and issuing pop-up warnings to web page users, reducing the possibility of web page users being polluted by violation information, reducing the possibility of web page users' interests being damaged, optimizing the security and stability of the network environment, and improving the network usage experience.

[0122] An embodiment of the present disclosure further proposes a web page violation alarm device based on a big model. Since the web page violation alarm device based on a big model proposed in the embodiment of the present disclosure corresponds to the web page violation alarm methods based on a big model proposed in the above-mentioned embodiments, the implementation methods of the above-mentioned web page violation alarm methods based on a big model are also applicable to the web page violation alarm device based on a big model proposed in the embodiment of the present disclosure, and will not be described in detail in the following embodiments.

[0123] Figure 5 FIG. 1 is a schematic diagram of a webpage violation warning device based on a large model according to an embodiment of the present disclosure. Figure 5 As shown, a webpage violation warning device 500 based on a large model includes an identification module 51, an extraction module 52, a detection module 53 and an alarm module 54, wherein:

[0124] The identification module 51 is used to obtain the target web page to be detected and identify whether the target web page meets the violation detection trigger condition;

[0125] An extraction module 52, configured to extract detection text objects and detection image objects in the target webpage in response to identifying that the target webpage satisfies the violation detection trigger condition, and obtain a detection object set of the target webpage;

[0126] A detection module 53 is used to input the detection object set into the pre-acquired large model, perform violation detection on the detection object set through the model capability of the large model, and identify whether there is a violation object in the detection object set;

[0127] The alarm module 54 is used to obtain the web page user of the target web page in response to identifying the existence of a violation object in the detection object set, and issue a web page violation alarm to the web page user.

[0128] In the disclosed embodiment, the extraction module 52 is further used to: obtain web page information and web page screenshots of the target web page; perform text scanning on the target web page, extract suspected illegal sentences in the target web page, and obtain detection text objects based on the web page information and the suspected illegal sentences; perform image scanning on the target web page, extract suspected illegal images in the target web page, and obtain detection image objects based on the web page screenshots and the suspected illegal images; obtain a detection object set of the target web page based on the detection text objects and the detection image objects.

[0129] In the disclosed embodiment, the extraction module 52 is further used to: perform text scanning on the target web page, obtain the web page text information of the target web page, and extract the suspected illegal sentences from the web page text information; obtain the violation type to which the suspected illegal sentences belong; perform vector splicing on the suspected illegal sentences, the violation type and the web page information to obtain a first detection vector after splicing, and determine the first detection vector as the detection text object of the target web page.

[0130] In the disclosed embodiment, the extraction module 52 is further used to: perform image scanning on the target web page, obtain web page image information of the target web page, and extract suspected illegal images from the web page image information; obtain the violation type to which the suspected illegal images belong; perform vector splicing on the suspected illegal images, violation types, and web page screenshots to obtain a spliced ​​second detection vector, and determine the second detection vector as the detection image object of the target web page.

[0131] In the disclosed embodiment, the detection module 53 is further used to: obtain a preset violation detection strategy library, and obtain a first target detection strategy corresponding to a detection text object in a detection object set, and a second target detection strategy for a detection image object from the violation detection strategy library; call the first target detection strategy through a large model to perform violation detection on the detection text object, and obtain a first violation detection result for the detection text object; call the second target detection strategy through a large model to perform violation detection on the detection image object, and obtain a second violation detection strategy for the detection image object; and identify whether there is a violation object in the detection object set based on the first violation detection result and the second violation detection result.

[0132] In the disclosed embodiment, the detection module 53 is further used to: perform a violation review and verification on the first violation detection result and the second violation detection result to obtain a first violation verification result of the first violation detection result and a second violation verification result of the second violation detection result; in response to the first violation detection result indicating that the detected text object is in violation, and the first violation verification result indicating that the verification is passed, determine that a violation object exists in the detection object set, and determine the detected text object as a violation object; in response to the second violation detection result indicating that the detected image object is in violation, and the second violation verification result indicating that the verification is passed, determine that a violation object exists in the detection object set, and determine the detected image object as a violation object.

[0133] In the disclosed embodiment, the alarm module 54 is also used to: obtain violation information corresponding to the violation object, and generate corresponding pop-up alarm information according to the violation information; and issue a pop-up warning to the web page user based on the pop-up alarm information to issue a web page violation alarm.

[0134] In the disclosed embodiment, the alarm module 54 is further used to: obtain the historical number of visits to the target web page, and determine the violation level of the target web page based on the historical number of visits; determine the violation handling strategy of the target web page based on the violation level, and perform violation correction processing on the target web page based on the violation handling strategy.

[0135] In the disclosed embodiment, the identification module 51 is also used to: generate a violation detection request for the target web page, and identify whether the target web page meets the violation detection trigger condition based on the violation detection request; in response to identifying that the target web page meets the violation detection trigger condition, access a preset cloud computing module, and perform violation detection on the target web page through the cloud computing service provided by the cloud computing module.

[0136] In the disclosed embodiment, the device also includes an update module, which is used to: obtain a preset list of illegal web pages; in response to identifying the presence of illegal objects in the detection object set, determine that the target web page is an illegal web page, obtain web page information and violation information of the illegal web page, and update the illegal web page table based on the web page information and violation information to obtain an updated new illegal web page list.

[0137] The webpage violation warning device based on the large model proposed in the present disclosure obtains the target webpage to be detected and identifies whether the target webpage meets the violation detection trigger condition. After identifying that the target webpage meets the violation detection trigger condition, the violation detection process of the target webpage is started, wherein the detection text object and the detection image object in the target webpage can be extracted to obtain the corresponding detection object set, and the detection object set is subjected to violation detection through the model capability of the large model, and whether there is a violation object in the detection object set is identified, and when the violation object is identified in the detection object set, a webpage violation warning is issued to the webpage user of the target webpage. In the present disclosure, the violation detection of the detection object set is performed through the model capability of the large model, which reduces the degree of dependence of the violation detection of the target webpage on manual work, improves the detection efficiency and accuracy of the text violation detection and image violation detection of the target webpage, and when the violation object is identified in the detection object set of the target webpage, a webpage violation warning is issued to the webpage user of the target webpage, which improves the vigilance of the webpage user, reduces the possibility of the webpage user being polluted by the violation information, reduces the possibility of the interests of the webpage user being damaged, optimizes the security and stability of the network environment, and improves the network usage experience.

[0138] According to an embodiment of the present disclosure, the present disclosure also proposes an electronic device, a readable storage medium, and a computer program product.

[0139] Figure 6 A schematic block diagram of an example electronic device 600 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.

[0140] like Figure 6 As shown, the device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 609 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0141] A number of components in the device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 606, such as various types of displays, speakers, etc.; a storage unit 609, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0142] The computing unit 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 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 601 performs the various methods and processes described above, such as a web page violation alarm method based on a large model. For example, in some embodiments, a web page violation alarm method based on a large model may be implemented as a computer software program, which is tangibly included in a machine-readable medium, such as a storage unit 609. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 600 via the ROM 602 and / or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the web page violation alarm method based on the large model described above may be executed. Alternatively, in other embodiments, the computing unit 601 may be configured to execute the large model-based web page violation warning method in any other appropriate manner (for example, by means of firmware).

[0143] 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.

[0144] 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 presented 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, implements the functions / operations specified in the flow chart and / or block diagram. 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.

[0145] 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.

[0146] To propose interactions with a user account, 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 account; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user account can propose input to the computer. Other types of devices may also be used to propose interactions with the user account; for example, feedback proposed to the user account may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user account may be received in any form (including acoustic input, voice input, or tactile input).

[0147] 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 account computer with a graphical user account interface or a web browser through which a user account 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.

[0148] A computer system may include a client and a server. The client and the server 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 with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0149] 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.

[0150] 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 web page violation warning method based on a large model, wherein: The method comprises: Obtaining a target web page to be detected, and identifying whether the target web page meets a violation detection trigger condition; In response to identifying that the target webpage satisfies the violation detection trigger condition, extracting detection text objects and detection image objects in the target webpage to obtain a detection object set of the target webpage; Inputting the detection object set into the pre-acquired large model, performing violation detection on the detection object set by using the model capability of the large model, and identifying whether there are any violation objects in the detection object set; In response to identifying that there is a violation object in the detection object set, a web page user of the target web page is obtained, and a web page violation alarm is issued to the web page user.

2. The method according to claim 1, wherein: In response to identifying that the target webpage satisfies the violation detection trigger condition, extracting the detection text object and the detection image object in the target webpage to obtain the detection object set of the target webpage includes: Obtaining webpage information and webpage screenshots of the target webpage; Performing text scanning on the target webpage, extracting suspected illegal sentences in the target webpage, and obtaining the detection text object according to the webpage information and the suspected illegal sentences; Performing image scanning on the target webpage, extracting suspected illegal images in the target webpage, and obtaining the detection image object according to the webpage screenshot and the suspected illegal images; Based on the detected text objects and the detected image objects, the detected object set of the target webpage is obtained.

3. The method according to claim 2, wherein: The step of performing text scanning on the target webpage, extracting suspected illegal sentences in the target webpage, and obtaining the detection text object according to the webpage information and the suspected illegal sentences includes: Performing text scanning on the target webpage to obtain webpage text information of the target webpage, and extracting the suspected illegal sentence from the webpage text information; Obtaining the violation type to which the suspected violation statement belongs; Vector splicing is performed on the suspected illegal sentence, the illegal type and the web page information to obtain a spliced ​​first detection vector, and the first detection vector is determined as the detection text object of the target web page.

4. The method according to claim 2, wherein: The performing image scanning on the target webpage, extracting the suspected illegal images in the target webpage, and obtaining the detection image object according to the webpage screenshot and the suspected illegal images, comprises: Performing image scanning on the target webpage to obtain webpage image information of the target webpage, and extracting the suspected illegal image from the webpage image information; Obtaining the violation type to which the suspected violation image belongs; Vector splicing is performed on the suspected illegal image, the illegal type, and the webpage screenshot to obtain a spliced ​​second detection vector, and the second detection vector is determined as the detection image object of the target webpage.

5. The method according to claim 1, wherein: The step of inputting the detection object set into the pre-acquired large model, performing violation detection on the detection object set by using the model capability of the large model, and identifying whether there are any violation objects in the detection object set includes: Obtain a preset violation detection strategy library, and obtain a first target detection strategy corresponding to the detection text object in the detection object set and a second target detection strategy corresponding to the detection image object from the violation detection strategy library; Calling the first target detection strategy through the large model to perform violation detection on the detection text object to obtain a first violation detection result of the detection text object; Calling the second target detection strategy through the large model to perform violation detection on the detection image object to obtain a second violation detection strategy for the detection image object; According to the first violation detection result and the second violation detection result, it is identified whether the violation object exists in the detection object set.

6. The method according to claim 5, wherein: The step of identifying whether the violating object exists in the detection object set according to the first violation detection result and the second violation detection result includes: Performing violation review verification on the first violation detection result and the second violation detection result to obtain a first violation verification result of the first violation detection result and a second violation verification result of the second violation detection result; In response to the first violation detection result indicating that the detected text object violates the rules, and the first violation verification result indicating that the verification is passed, determining that the violating object exists in the detected object set, and determining the detected text object as the violating object; In response to the second violation detection result indicating that the detected image object violates the rules and the second violation verification result indicating that the verification is passed, it is determined that the violating object exists in the detected object set, and the detected image object is determined as the violating object.

7. The method according to claim 1, wherein: In response to identifying that there is a violation object in the detection object set, obtaining a web page user of the target web page, and issuing a web page violation warning to the web page user, includes: Obtaining violation information corresponding to the violation object, and generating corresponding pop-up window warning information according to the violation information; A pop-up window warning is provided to the web page user based on the pop-up window warning information to issue a web page violation warning.

8. The method according to claim 7, wherein: The method further comprises: Obtaining the historical number of visits to the target webpage, and determining the violation level of the target webpage according to the historical number of visits; A violation handling strategy for the target webpage is determined according to the violation level, and violation correction processing is performed on the target webpage according to the violation handling strategy.

9. The method according to claim 1, wherein: The method further comprises: generating a violation detection request for the target webpage, and identifying whether the target webpage satisfies the violation detection trigger condition according to the violation detection request; In response to identifying that the target webpage satisfies the violation detection triggering condition, a preset cloud computing module is accessed, and violation detection is performed on the target webpage through a cloud computing service provided by the cloud computing module.

10. The method according to any one of claims 1 to 9, wherein: The method further comprises: Get a preset list of illegal web pages; In response to identifying that there are illegal objects in the detection object set, the target web page is determined to be an illegal web page, web page information and illegal information of the illegal web page are obtained, and the illegal web page table is updated based on the web page information and the illegal information to obtain an updated new illegal web page list.

11. A web page violation warning device based on a large model, wherein: The device comprises: An identification module, used to obtain a target web page to be detected and identify whether the target web page meets a violation detection trigger condition; An extraction module, configured to extract detection text objects and detection image objects in the target webpage in response to identifying that the target webpage satisfies the violation detection trigger condition, and obtain a detection object set of the target webpage; A detection module, used for inputting the detection object set into a pre-acquired large model, performing violation detection on the detection object set through the model capability of the large model, and identifying whether there are any violation objects in the detection object set; The alarm module is used for obtaining the web page user of the target web page in response to identifying the existence of a violation object in the detection object set, and issuing a web page violation alarm to the web page user.

12. The device according to claim 11, wherein The extraction module is further used for: Obtaining webpage information and webpage screenshots of the target webpage; Performing text scanning on the target webpage, extracting suspected illegal sentences in the target webpage, and obtaining the detection text object according to the webpage information and the suspected illegal sentences; Performing image scanning on the target webpage, extracting suspected illegal images in the target webpage, and obtaining the detection image object according to the webpage screenshot and the suspected illegal images; Based on the detected text objects and the detected image objects, the detected object set of the target webpage is obtained.

13. The device according to claim 12, wherein: The extraction module is further used for: Performing text scanning on the target webpage to obtain webpage text information of the target webpage, and extracting the suspected illegal sentence from the webpage text information; Obtaining the violation type to which the suspected violation statement belongs; Vector splicing is performed on the suspected illegal sentence, the illegal type and the web page information to obtain a spliced ​​first detection vector, and the first detection vector is determined as the detection text object of the target web page.

14. The device according to claim 12, wherein: The extraction module is further used for: Performing image scanning on the target webpage to obtain webpage image information of the target webpage, and extracting the suspected illegal image from the webpage image information; Obtaining the violation type to which the suspected violation image belongs; Vector splicing is performed on the suspected illegal image, the illegal type, and the webpage screenshot to obtain a spliced ​​second detection vector, and the second detection vector is determined as the detection image object of the target webpage.

15. The device according to claim 11, wherein The detection module is further used for: Obtain a preset violation detection strategy library, and obtain a first target detection strategy corresponding to the detection text object in the detection object set and a second target detection strategy corresponding to the detection image object from the violation detection strategy library; Calling the first target detection strategy through the large model to perform violation detection on the detection text object to obtain a first violation detection result of the detection text object; Calling the second target detection strategy through the large model to perform violation detection on the detection image object to obtain a second violation detection strategy for the detection image object; According to the first violation detection result and the second violation detection result, it is identified whether the violation object exists in the detection object set.

16. The device according to claim 15, wherein: The detection module is further used for: Performing violation review verification on the first violation detection result and the second violation detection result to obtain a first violation verification result of the first violation detection result and a second violation verification result of the second violation detection result; In response to the first violation detection result indicating that the detected text object violates the rules, and the first violation verification result indicating that the verification is passed, determining that the violating object exists in the detected object set, and determining the detected text object as the violating object; In response to the second violation detection result indicating that the detected image object violates the rules and the second violation verification result indicating that the verification is passed, it is determined that the violating object exists in the detected object set, and the detected image object is determined as the violating object.

17. The device according to claim 11, wherein: The alarm module is also used for: Obtaining violation information corresponding to the violation object, and generating corresponding pop-up window warning information according to the violation information; A pop-up window warning is provided to the web page user based on the pop-up window warning information to issue a web page violation warning.

18. The device according to claim 17, wherein: The alarm module is also used for: Obtaining the historical number of visits to the target webpage, and determining the violation level of the target webpage according to the historical number of visits; A violation handling strategy for the target webpage is determined according to the violation level, and violation correction processing is performed on the target webpage according to the violation handling strategy.

19. The device according to claim 11, wherein: The identification module is further used for: generating a violation detection request for the target webpage, and identifying whether the target webpage satisfies the violation detection trigger condition according to the violation detection request; In response to identifying that the target webpage satisfies the violation detection triggering condition, a preset cloud computing module is accessed, and violation detection is performed on the target webpage through a cloud computing service provided by the cloud computing module.

20. The device according to any one of claims 11 to 19, wherein: The device also includes an updating module, which is used to: Get a preset list of illegal web pages; In response to identifying that there are illegal objects in the detection object set, the target web page is determined to be an illegal web page, web page information and illegal information of the illegal web page are obtained, and the illegal web page table is updated based on the web page information and the illegal information to obtain an updated new illegal web page list.

21. 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 10.

22. 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 1-10.

23. 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 10.

Images