E-commerce Information Security Detection Method and Its Device, Equipment, Medium, Product

Abstract

The present application discloses an e-commerce information security detection method, its device, equipment, medium, and product. The method includes: responding to a security detection request, obtaining the storage address carried by the request, and obtaining the commodity information to be detected according to the storage address; constructing an index item, which includes the version information of the currently enabled security classification model and the encoded mapping information of the storage address; obtaining the detection result mapped to the index item from the buffer area. When the acquisition fails, calling the currently enabled security classification model to perform information security detection on the commodity information to obtain the detection result, and storing the detection result mapped to the index item in the buffer area; pushing the detection result to answer the security detection request. The present application can quickly call the information security detection result of the commodity information published by the independent station, facilitating the independent station to decide whether to publish the commodity information.

Description

Technical Field

[0001] The present application relates to the field of e-commerce information technology, and in particular to an e-commerce information security detection method and its corresponding device, computer equipment, computer-readable storage medium, and computer program product. Background Art

[0002] In e-commerce platforms, especially those based on independent websites, each independent website usually runs an instance of an online store. The store management user can enter the product description information through the background of the online store to publish new products, or modify the product description information of the online products and resubmit them for publication. The corresponding product description information is regarded as an independent product unit by the background, and the corresponding information is stored in the product database.

[0003] The technical support for independent sites is implemented by the e-commerce platform, but the deployment of independent sites may be distributed on servers in various countries and regions. Each country and region has its own information security management laws and regulations. Based on this, the e-commerce platform needs to provide relevant detection service support to determine whether the product description information posted by merchant users violates the regulations. Generally, this is done by providing a neural network model that has been pre-trained to a convergence state to classify and detect certain specific product information in the product description information, so as to determine the corresponding security category of the product information, and control the release of the product description information based on whether the security category belongs to the violation type.

[0004] In order to achieve technical upgrades, the e-commerce platform will continue to update the neural network model so that the corresponding detection service can continuously and timely correct and improve the accuracy of judging security categories. The detection service is generally implemented to independently and concurrently serve the calls of each independent station, and independently manage the various data used and generated by the model. Therefore, this service uses different versions of neural network models to concurrently process a large amount of commodity information for a long time, which will generate a large amount of relevant data. How to use this data will affect whether the operating resources of the backend computer equipment of the e-commerce platform can be efficiently scheduled, and also affect whether the timeliness of information security detection can be guaranteed for each independent station. Summary of the invention

[0005] The primary purpose of the present application is to solve at least one of the above problems and to provide an e-commerce information security detection method and its corresponding device, computer equipment, computer-readable storage medium, and computer program product.

[0006] In order to meet the various objectives of this application, this application adopts the following technical solutions:

[0007] An e-commerce information security detection method provided to meet one of the purposes of this application includes the following steps:

[0008] In response to a security detection request, obtain the storage address carried by the request, and obtain the product information to be detected according to the storage address;

[0009] Construct an index entry, which includes the version information of the currently enabled security classification model and the encoded mapping information of the storage address;

[0010] Obtain the detection result mapped to the index entry from the buffer. When the acquisition fails, call the currently enabled security classification model to perform information security detection on the product information to obtain the detection result, and store the detection result mapped to the index entry in the buffer;

[0011] Push the detection result to answer the security detection request.

[0012] In some deepened embodiments, constructing an index entry includes the following steps:

[0013] Obtain the version number of the currently enabled security classification model as its version information;

[0014] Apply a digital digest algorithm to calculate the hash value of the storage address as the encoded mapping information of the storage address;

[0015] Concatenate the version information and the encoded mapping information into an index entry of the product information to be detected.

[0016] In some deepened embodiments, obtaining the detection result mapped to the index entry from the buffer includes the following steps:

[0017] Query whether there is a target key-value pair containing the index entry in the key-value pair mapping table in the buffer. When the target key-value pair exists, obtain the detection result in the value range of the target key-value pair;

[0018] When the target key-value pair does not exist, access the storage address to call the product information, input the product information into the currently enabled security classification model to perform information security detection, and obtain the corresponding detection result;

[0019] Construct a key-value pair corresponding to the product information with the index entry as the key domain and the detection result output by the security classification model as the value domain, and store it in the key-value pair mapping table in the buffer.

[0020] In some extended embodiments, before the step of responding to the security detection request, the following steps are included:

[0021] Receive and distribute and store the product description information submitted by independent site editors, where the product description information includes picture information and / or text information for describing product objects;

[0022] Determine the picture information and / or text information in the commodity description information as the commodity information to be detected according to the input parameters defined by the currently enabled security classification model, and obtain the storage address corresponding to the commodity information;

[0023] Trigger a security detection request to obtain the detection result of the commodity information, and include the storage address of the commodity information to be detected in the request.

[0024] In an extended partial embodiment, after the step of pushing the detection result to answer the security detection request, the following steps are included:

[0025] According to the confidence levels corresponding to multiple security categories included in the detection result, determine the security category with the highest confidence level as the detection category corresponding to the commodity information;

[0026] Judge whether the detection category belongs to a preset allowed category. When it is an allowed category, release the commodity object to the online store.

[0027] In a deepened partial embodiment, the process of the security classification model performing information security detection on the commodity information includes the following steps:

[0028] Preprocess the commodity information;

[0029] Use a semantic feature extraction model to extract deep semantic features from the preprocessed commodity information to obtain the semantic feature information of the commodity information;

[0030] Map the semantic feature information to the classification space to obtain the confidence levels of each security category of the commodity information corresponding to the classification space, and form the detection result of the commodity information.

[0031] In an extended partial embodiment, after the step of pushing the detection result to answer the security detection request, the following steps are further included:

[0032] Respond to the arrival event of the timing task, and obtain the version information of the currently enabled security classification model;

[0033] Retrieve the index item in the buffer area that does not contain the version information, and clear the key-value pair where the index item is located from the buffer area.

[0034] An e-commerce information security detection device provided to meet one of the purposes of this application includes a request response module, an index construction module, a result acquisition module, and a push response module, where: the request response module is used to respond to a security detection request, obtain the storage address carried by the request, and obtain the product information to be detected according to the storage address; the index construction module is used to construct an index item, which includes the version information of the currently enabled security classification model and the encoding mapping information of the storage address; the result acquisition module is used to obtain the detection result mapped to the index item from the buffer area. When the acquisition fails, call the currently enabled security classification model to perform information security detection on the product information to obtain the detection result, and map and store the detection result with the index item in the buffer area; the push response module is used to push the detection result to respond to the security detection request.

[0035] In some further embodiments, the index construction module includes: a version query unit, which is used to obtain the version number of the currently enabled security classification model as its version information; an encoding mapping unit, which is used to apply a digital digest algorithm to calculate the hash value of the storage address as the encoding mapping information of the storage address; a splicing index unit, which is used to splice the version information and the encoding mapping information into an index item of the product information to be detected.

[0036] In some further embodiments, the result acquisition module includes: a cache call unit, which is used to query whether there is a target key-value pair containing the index item in the key-value pair mapping table in the buffer area. When the target key-value pair exists, obtain the detection result in the value range of the target key-value pair; a model call unit, which is used to when the target key-value pair does not exist, access the storage address to call the product information, input the product information into the currently enabled security classification model to perform information security detection, and obtain the corresponding detection result; a result cache unit, which is used to construct a key-value pair corresponding to the product information with the index item as the key domain and the detection result output by the security classification model as the value domain, and store it in the key-value pair mapping table in the buffer area.

[0037] In an extended partial embodiment, the e-commerce information security detection device of the present application further includes the following modules that run prior to the request response module: a submission processing module, configured to receive and distributively store the product description information submitted by an independent site editor, where the product description information includes picture information and / or text information for describing a product object; an information invocation module, configured to determine the picture information and / or text information in the product description information as the product information to be detected according to the input parameters defined by the currently enabled security classification model, and obtain the storage address corresponding to the product information; a request trigger module, configured to trigger a security detection request to obtain the detection result of the product information, where the storage address of the product information to be detected is included in the request.

[0038] In an extended partial embodiment, the e-commerce information security detection device of the present application further includes the following modules that run after the push response module: a category determination module, configured to determine the security category with the highest confidence among the multiple security categories included in the detection result as the detection category corresponding to the product information according to the confidence levels corresponding to the multiple security categories included in the detection result; a product release module, configured to determine whether the detection category belongs to a preset allowed category, and when it is an allowed category, release the product object to an online store.

[0039] In a deepened partial embodiment, the security classification model includes: a preprocessing unit, configured to preprocess the product information; a feature extraction unit, configured to extract deep semantic features from the preprocessed product information using a semantic feature extraction model to obtain the semantic feature information of the product information; a classification mapping unit, configured to map the semantic feature information to a classification space to obtain the confidence levels of each security category corresponding to the product information in the classification space, and form the detection result of the product information.

[0040] In an extended partial embodiment, the e-commerce information security detection device of the present application further includes the following modules that run after the push response module: a task response module, configured to respond to the arrival event of a timing task and obtain the version information of the currently enabled security classification model; a cache cleaning module, configured to retrieve the index items in the buffer area that do not contain the version information, and clear the key-value pairs where the index items are located from the buffer area.

[0041] A computer device provided to meet one of the purposes of the present application includes a central processing unit and a memory, where the central processing unit is configured to call and run a computer program stored in the memory to execute the steps of the e-commerce information security detection method of the present application.

[0042] A computer-readable storage medium provided for another object of the present application stores a computer program implemented according to the e-commerce information security detection method in the form of computer-readable instructions. When the computer program is called and run by a computer, it executes the steps included in the method.

[0043] A computer program product provided for another object of the present application includes a computer program / instructions. When the computer program / instructions are executed by a processor, they implement the steps of the method described in any embodiment of the present application.

[0044] Compared with the prior art, the technical solution of the present application has multiple technical advantages, including but not limited to the following aspects:

[0045] First of all, the present application realizes a commodity information security detection service. After a security detection request is triggered at an independent station, in response to the security detection request, the storage address of the commodity information to be detected is determined. Then, according to the version information of the currently enabled security classification model and the encoding mapping information of the storage address, an index item is constructed. As the name implies, the index item has the function of uniqueness. According to this index item, the system buffer is queried, and the detection results generated by the security classification model during the historical service process in the buffer are preferentially utilized. Only when the above-mentioned detection results cannot be obtained from the buffer, the security classification model is called to perform a security detection on the commodity information, and the corresponding detection results are determined, and the detection results are associated with the index item and stored in the buffer for subsequent preferential invocation. Accordingly, for the same commodity information, the detection results of the security classification model of the same version can be reused from the buffer. That is, even if the user modifies the commodity information of the same commodity multiple times, only the security classification model needs to be called once for its commodity information to determine its detection results. For a computer device that needs to concurrently respond to security detection requests from multiple independent stations, it can greatly save system overhead and improve the immediacy of returning detection results.

[0046] Secondly, since the version information of the security classification model is associated in the index item, it can be ensured that once the security classification model is replaced by its new version, the constructed index item cannot find the corresponding detection results in the buffer as the version information is replaced. Then, the new version of the security classification model is automatically called to re-obtain the detection results of the commodity information, so as to ensure that it can automatically adapt to the upgrade and replacement of the security classification model and always provide the latest version of the detection results for the security detection of commodity information.

[0047] In addition, the service implemented by the technical solution of the present application, as a security detection service relatively independent of the independent website, has relatively concentrated functions. It can obtain product information according to the storage address of the product information, can concurrently respond to a large number of security detection requests, and can use a caching mechanism to minimize the invocation of the security classification model. The combination of these means enables the security detection service of the e-commerce platform to be deployed at the lowest cost and provide the service with the greatest economic value. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The above and / or additional aspects and advantages of the present application will become apparent and be readily understood from the following description of the embodiments in conjunction with the drawings, where:

[0049] Figure 1 is a flowchart of a typical embodiment of the e-commerce information security detection method of the present application;

[0050] Figure 2 is a flowchart of the process of constructing index items in an embodiment of the present application;

[0051] Figure 3 is a flowchart of the process of obtaining the detection result of product information in an embodiment of the present application;

[0052] Figure 4 is a flowchart of the process of triggering a security detection request in an embodiment of the present application;

[0053] Figure 5 is a flowchart of the working process of the security classification model in an embodiment of the present application;

[0054] Figure 6 is a flowchart of the process of further cleaning the buffer area in an embodiment of the present application;

[0055] Figure 7 is a schematic block diagram of the principle of the e-commerce information security detection device of the present application;

[0056] Figure 8 is a schematic structural diagram of a computer device adopted by the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] The embodiments of the present application will be described in detail below. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals denote the same or similar elements or elements with the same or similar functions throughout. The embodiments described below with reference to the drawings are exemplary and are only used to explain the present application and should not be construed as limiting the present application.

[0058] Those skilled in the art can understand that, unless specifically stated otherwise, the singular forms "a", "an", "the" and "said" used herein may also include the plural forms. It should be further understood that the term "comprising" used in the specification of this application means the presence of the stated features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their groups. It should be understood that when we say that an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The phrase "and / or" used herein includes all or any unit and all combinations of one or more of the associated listed items.

[0059] Those skilled in the art can understand that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as the general understanding of those of ordinary skill in the art to which this application belongs. It should also be understood that terms such as those defined in a general dictionary should be understood to have a meaning consistent with the meaning in the context of the prior art, and will not be interpreted in an idealized or overly formal sense unless specifically defined as here.

[0060] Those skilled in the art can understand that the "client", "terminal", and "terminal device" used herein include both devices with wireless signal receivers that only have the ability to receive and no ability to transmit, and devices with both receiving and transmitting hardware that can perform two-way communication on a two-way communication link. Such devices can include: cellular or other communication devices such as personal computers and tablet computers, which have a single-line display or a multi-line display or a cellular or other communication device without a multi-line display; PCS (Personal Communications Service), which can combine voice, data processing, fax, and / or data communication capabilities; PDA (Personal Digital Assistant), which can include a radio frequency receiver, a pager, Internet / intranet access, a web browser, a notepad, a calendar, and / or a GPS (Global Positioning System) receiver; conventional laptop and / or palm computers or other devices, which are conventional laptop and / or palm computers or other devices with and / or including a radio frequency receiver. The "client", "terminal", and "terminal device" used herein can be portable, transportable, installed in a vehicle (air, sea, and / or land), or suitable for and / or configured to run locally, and / or run in a distributed manner at any other location on the earth and / or in space. The "client", "terminal", and "terminal device" used herein can also be a communication terminal, an Internet access terminal, a music / video playback terminal, such as a PDA, a MID (Mobile Internet Device), and / or a mobile phone with music / video playback functions, or can also be devices such as a smart TV and a set-top box.

[0061] The hardware referred to by names such as "server", "client", and "service node" in this application is essentially an electronic device with the equivalent capabilities of a personal computer, and is a hardware device with the necessary components disclosed by the von Neumann principle, including a central processing unit (including an arithmetic unit and a controller), a memory, an input device, and an output device. The computer program is stored in its memory, and the central processing unit loads the program stored in the external memory into the memory for execution, executes the instructions in the program, and interacts with the input / output devices to complete specific functions.

[0062] It should be noted that the concept of "server" in this application can similarly be extended to the case of server clusters. According to the network deployment principles understood by those skilled in the art, the various servers should be logically divided. Physically, these servers can either be independent of each other but can be invoked through interfaces, or be integrated into a single physical computer or a set of computer clusters. Those skilled in the art should understand this flexibility and should not be restricted by this when implementing the network deployment method of this application.

[0063] One or several technical features of this application, unless expressly specified, can either be deployed on the server and accessed by the client remotely invoking the online service interface provided by the server, or be directly deployed and run on the client for access.

[0064] The neural network models cited or possibly cited in this application, unless expressly specified, can either be deployed on a remote server and remotely invoked on the client, or be deployed on a client with sufficient device capabilities for direct invocation. In some embodiments, when it runs on the client, its corresponding intelligence can be obtained through transfer learning to reduce the requirements for the client's hardware operating resources and avoid over-occupying the client's hardware operating resources.

[0065] All kinds of data involved in this application, unless expressly specified, can either be remotely stored on the server or stored on the local terminal device, as long as it is suitable for being invoked by the technical solution of this application.

[0066] Those skilled in the art should be aware that although the various methods of this application are described based on the same concept and thus show commonality with each other, unless otherwise specified, these methods can be executed independently. Similarly, for the various embodiments disclosed in this application, they are all proposed based on the same inventive concept. Therefore, for concepts with the same expression, as well as concepts that are only appropriately transformed for convenience although the concept expressions are different, they should be equivalently understood.

[0067] For the various embodiments to be disclosed in this application, unless expressly stated to be mutually exclusive, the relevant technical features involved in each embodiment can be cross-combined to flexibly construct new embodiments, as long as such combination does not deviate from the creative spirit of this application and can meet the requirements in the prior art or solve certain deficiencies in the prior art. Those skilled in the art should be aware of this flexibility.

[0068] An e-commerce information security detection method of the present application can be programmed as a computer program product and implemented by running on a client or a server. For example, in the application scenario of the e-commerce platform of the present application, it is generally deployed on the server for implementation. Thus, by accessing the interface opened after the computer program product runs, human-computer interaction can be carried out with the process of the computer program product through a graphical user interface to execute this method.

[0069] Please refer to Figure 1 , the e-commerce information security detection method of the present application can be implemented as an online service with relatively independent functions, serving the information security detection of each independent site of the e-commerce platform and concurrently responding to the calls of each independent site. Therefore, in its typical embodiment, it includes the following steps:

[0070] Step S1100: Respond to the security detection request, obtain the storage address carried by the request, and obtain the commodity information to be detected according to the storage address:

[0071] When any independent site calls the online service implemented by the present application, that is, the information security detection service, a security detection request can be triggered and sent to the detection service, and the storage address of the commodity information uploaded by the management user of the online store is carried in the detection request.

[0072] The commodity information belongs to the commodity description information corresponding to the commodity published by the user. The commodities published on the e-commerce platform are defined by the above-mentioned commodity description information, so as to define a commodity object at the data level, so that corresponding calls can be made to each commodity with the commodity object. The above-mentioned commodity description information includes various picture information and text information. The picture information is usually the relevant pictures used to display the appearance of the commodity, including the main picture of the commodity and each detail and environmental picture in the commodity details, etc.; the text information can include any text information such as commodity title, commodity details, commodity introduction, and attribute data of the commodity.

[0073] The commodity information pointed to by the storage address carried by the security detection request can be the picture information, text information or a combination of both in the commodity description information, and can be flexibly adapted according to the input parameter requirements of the security classification model adopted by the present application, that is, the basic information for its classification.

[0074] An independent site can submit the product description information corresponding to a new product, or edit and submit the product description information for a product that has been published online. After each submission of the product description information, the corresponding product description information is transmitted to the database engine for storage, so as to obtain the storage address corresponding to each specific product description information, such as the storage address of a product picture, or the storage address corresponding to a product text, etc. Thus, as long as a certain picture information and / or text information is called, the respective storage addresses corresponding to the picture information and / or text information can be obtained accordingly. Conversely, according to each storage address, the binary data of the picture information or text information pointed to by each storage address can be obtained accordingly. Considering the fact that there is a huge amount of product description information on the e-commerce platform, the platform generally uses a distributed storage system to store the product description information. Therefore, when it is necessary to transmit the corresponding product information to the information security detection service of the present application, by carrying the storage address of the corresponding product information in the security detection request for transmission, the transmission bandwidth between service clusters can be saved, and the corresponding service can call the corresponding product information to be detected according to the storage address by itself.

[0075] Step S1200, construct an index item, which includes the version information of the currently enabled security classification model and the encoding mapping information of the storage address:

[0076] In the present application, the detection result obtained by performing security detection on the product information by the security classification model is pre-stored in the system buffer area for efficient reuse, avoiding repeated calls to the security classification model to perform repeated classification and identification on the same product information, thereby saving system overhead and improving the detection result acquisition efficiency through the caching mechanism.

[0077] In the buffer area, for example, the Redis mechanism can be adopted to store the mapping relationship between the product information and its detection result. For this purpose, an index item will be constructed for each product information, and a unique correspondence with the specific product information is established through this index item, playing the role of a unique feature, and then the index item is associated and stored with the corresponding detection result. Based on the Redis key-value pair storage mechanism, the index item can be stored as the data in the key field of the key-value pair, and the detection result can be stored as the data in the value field of the key-value pair. Thus, the buffer area actually stores a key-value pair mapping table, and according to the index item of a product information, its corresponding detection result can be obtained.

[0078] The index item corresponds one by one in units of storage addresses. For example, when the storage address points to a product picture, the index item corresponds to the product picture; when the storage address points to a part of the text stored centrally, the index item corresponds to the text.

[0079] The index item mentioned above includes two parts of information. The first part is the version information of the currently enabled security classification model, and the second part is the encoded mapping information obtained by encoding according to the storage address. It is not difficult to understand that the version information in the first part is set corresponding to different versions of the security classification model. Generally, the version number of the security classification model or other feature information that can indicate different versions, such as the MD5 hash value of the digital signature, can be directly used, which has the effect of distinguishing the detection results generated by different versions of the security classification model. The encoded mapping information in the second part is generated by encoding according to the storage address and has the effect of uniquely corresponding to specific commodity information. Therefore, the one-to-one correspondence between the index item and the specific commodity information is specifically reflected as the one-to-one correspondence between the encoded mapping information in the second part and the storage address of the commodity information.

[0080] According to the above principle, when constructing the index item of the storage address in the security detection request, first convert the storage address into the corresponding encoded mapping information according to a preset encoding algorithm, and then combine the version information of the currently enabled security classification model with the encoded mapping information into an index item. Due to the combined effect of the encoded mapping information and the version information, this index item has the function of uniqueness and can be used to determine the detection result obtained by a certain version of the security classification model for a specific commodity information pointed to by a certain storage address.

[0081] Step S1300: Obtain the detection result mapped to the index item from the buffer. When the acquisition fails, call the currently enabled security classification model to perform information security detection on the commodity information to obtain the detection result, and map the detection result to the index item and store it in the buffer.

[0082] After constructing the index item corresponding to the storage address in the security detection request, the key-value pair corresponding to the index item can be queried from the buffer, and then the detection result mapped to the index item can be determined from the key-value pair. If the commodity information pointed to by the storage address has been detected by the currently enabled security classification model beforehand, the corresponding detection result will be pre-stored in the buffer. Thus, there will be such a key-value pair, and the corresponding detection result can be obtained from the key-value pair. Otherwise, if there is no key-value pair corresponding to the index item, it indicates that there is no corresponding detection result in the buffer, and the acquisition of the detection result corresponding to the index item from the buffer has failed, which also means that the commodity information pointed to by the storage address in the security detection request has not been detected by the currently enabled security classification model. Therefore, it is necessary to call the currently enabled security classification model to perform information security detection on this commodity information.

[0083] When invoking the currently enabled security classification model to detect the product information pointed to by the storage address in the security detection request, first obtain the corresponding product information from the database according to the storage address, then, according to the requirements of the security classification model, perform appropriate preprocessing on the product information, and input it into the security classification model for detection to obtain the corresponding detection result.

[0084] The security classification model described above can be a preset model constructed based on rules, machine learning, or deep learning, and it is only required to be suitable for outputting corresponding detection results according to the corresponding product information. It can be implemented by those skilled in the art according to actual needs and can be continuously upgraded and iterated without affecting the manifestation of the creative spirit of this application.

[0085] The security classification model can be suitable for separately processing picture information or text information, or jointly processing picture information and text information to obtain a detection result, which specifically depends on how those skilled in the art construct the model and does not affect the manifestation of the creative spirit of this application. Correspondingly, when providing the storage address of the product information in the security detection request, one or more storage addresses corresponding to the specific product information should be provided according to the input required by the construction of the security classification model.

[0086] After the security classification model determines the detection result for the product information pointed to by the storage address, it can be reused subsequently. Therefore, it can be stored in the buffer area. For this purpose, according to the implementation principle of the buffer area disclosed above, the detection result determined by the security classification model based on the product information pointed to by the storage address can be associated with the index item corresponding to the storage address, and constructed into mapping relationship data, that is, the key-value pair, and stored in the key-value pair mapping table of the buffer area.

[0087] When the security classification model only separately determines the detection result for a single picture information or a single text information pointed to by a single storage address, the mapping relationship data between the single storage address and the detection result represents the detection result of the product information pointed to by the storage address.

[0088] When the security classification model obtains the detection result through comprehensive detection of the combination of a single picture information and a single text information, independent corresponding storage addresses are provided for the single picture information and the single text information respectively in the security detection request. Therefore, two index entries can be obtained. At this time, each index entry and the detection result can be constructed into a key-value pair, that is, two key-value pairs are obtained and stored in the key-value pair mapping table in the buffer area. Subsequently, if a new security detection request is triggered and the two storage addresses are provided, the two key-value pairs can be called correspondingly. According to whether the detection results of the two key-value pairs are the same, when the two detection results are the same, it can be confirmed as the real detection result. If the two detection results are different, the security classification model can be called again to re-determine the detection result. Finally, similarly here, the latest detection result can be cached.

[0089] The latest detection result saved in the buffer area is associated with the index entry. Therefore, when retrieving according to the same index entry later, the corresponding detection result can be obtained without calling the security classification model again for repeated detection.

[0090] When the security classification model is upgraded and the version number is changed, according to the index entry constructed in the previous step, due to the change of version information, it is different from that before the security classification model is not upgraded. Therefore, the corresponding key-value pair cannot be found in the buffer area. In this case, only the latest version of the security classification model can be called to re-determine the detection result. Therefore, in fact, through the construction mechanism of the index entry, a mechanism for adaptively upgrading the security classification model and updating the detection result in the buffer area is realized.

[0091] Step S1400, push the detection result to respond to the security detection request:

[0092] According to the foregoing steps, it can be seen that after responding to the security detection request, according to the index entry of the storage address, the detection result of the commodity information corresponding to the storage address is obtained from the buffer area, or the security classification model is called to obtain the corresponding detection result when the detection result does not exist in the buffer area. In any case, the detection result expected by the security detection request can be obtained. Therefore, the detection result can be used to respond to the security detection request and returned to the corresponding requester, and the corresponding requester can implement further business processes and control whether the corresponding commodity information can be published according to the detection result.

[0093] It can be seen that based on the ability of the present application to implement its technical solution as an online service, a large number of concurrent security detection requests can be responded to, and a large number of detection results corresponding to the large number of security detection requests can be cached and managed to achieve efficient scheduling.

[0094] According to the typical embodiments of the present application and their variant embodiments, it can be known that, compared with the prior art, the technical solution of the present application has multiple technical advantages, including but not limited to the following aspects:

[0095] First of all, the present application realizes a commodity information security detection service. After a security detection request is triggered on an independent website, in response to the security detection request, the storage address of the commodity information to be detected is determined. Then, according to the version information of the currently enabled security classification model and the coding mapping information of the storage address, an index item is constructed. As the name implies, the index item has the function of uniqueness. According to this index item, the system buffer is queried, and the detection results generated by the security classification model during the historical service process in the buffer are preferentially utilized. Only when the above-mentioned detection results cannot be obtained from the buffer, the security classification model is called to perform a security detection on the commodity information, and the corresponding detection results are determined, and the detection results are associated with the index item and stored in the buffer for subsequent preferential invocation. Accordingly, for the same commodity information, the detection results of the same version of the security classification model can be reused from the buffer. That is, even if the user modifies the commodity information of the same commodity multiple times, only the security classification model needs to be called once for its commodity information to determine its detection results. For a computer device that needs to concurrently respond to security detection requests from multiple independent websites, it can greatly save system overhead and improve the immediacy of returning detection results.

[0096] Secondly, since the version information of the security classification model is associated in the index item, it can be ensured that once the security classification model is replaced by its new version, the constructed index item cannot find the corresponding detection results in the buffer as the version information is replaced. Then, the new version of the security classification model is automatically called to re-obtain the detection results of the commodity information, so as to ensure automatic adaptation to the upgrade and replacement of the security classification model and ensure that the latest version of the detection results is always provided for the security detection of commodity information.

[0097] In addition, the service implemented by the technical solution of the present application, as a security detection service relatively independent of the independent website, has relatively concentrated functions. It can obtain commodity information according to the storage address of the commodity information, can concurrently respond to a large number of security detection requests, and can use the cache mechanism to minimize the invocation of the security classification model. The combination of these means makes the security detection service of the e-commerce platform realize the service with the greatest economic value deployed at the lowest cost.

[0098] Please refer to Figure 2 , in some deepened embodiments, the step S1200 of constructing an index item includes the following steps:

[0099] Step S1210: Obtain the version number of the currently enabled security classification model as its version information:

[0100] In this embodiment, when constructing an index entry, first detect the version number of the currently enabled security classification model. The currently enabled security classification model is the security classification model that has been configured for use in the security detection service of this application. It can be obtained by upgrading the original security classification model after the security classification model is retrained and converges. Correspondingly, its version number is updated, so it can be called here.

[0101] Step S1220: Apply a digital digest algorithm to calculate the hash value of the storage address as the encoded mapping information of the storage address:

[0102] Subsequently, a preset digital digest algorithm, such as the MD5 algorithm, can be used to calculate the hash value for each storage address provided in the security detection request. The hash value is generally a 256-bit binary value or a 32-bit hexadecimal value, and this hash value is the encoded mapping information corresponding to each storage address. According to the principle of the digital digest algorithm, this hash value generally has uniqueness. Similarly, other digital digest algorithms can also be used for implementation, such as MD4, etc., which can be flexibly implemented by those skilled in the art.

[0103] Step S1230: Concatenate the version information and the encoded mapping information into an index entry of the product information to be detected:

[0104] Furthermore, concatenate the version information corresponding to the security classification model, that is, its version number, with the encoded mapping information, that is, the hash value, in sequence to obtain a string, which can be used as the index entry corresponding to the storage address.

[0105] In this embodiment, the version number of the security classification model and the hash value of the storage address are used to construct the corresponding index entry. The version number can be used to distinguish different versions of the security classification model, and the hash value can be used to distinguish different product information. Therefore, when querying the key-value pair in the cache area according to the index entry later, whether it is a change in the version of the security classification model or a change caused by different product information, it can be reflected by this index entry. Thus, the security detection service of this application can adapt to these information changes and flexibly decide whether to call the corresponding detection result in the cache area or call the security classification model to re-obtain the detection result.

[0106] Please refer to Figure 3 , in some in-depth embodiments, step S1300: Obtain the detection result mapped to the index entry from the cache area, including the following steps:

[0107] Step S1310: Query whether there is a target key-value pair containing the index term in the key-value pair mapping table in the buffer. When there is such a target key-value pair, obtain the detection result in the value range of the target key-value pair:

[0108] After the construction of the index term for a storage address is completed, it is necessary to query the corresponding key-value pair in the buffer. Accordingly, first query in the buffer whether there is a key that is consistent with the index term. When there is, it is confirmed that there is a corresponding key-value pair, so that the detection result mapped to the index term can be directly called from the key-value pair.

[0109] Step S1320: When there is no such target key-value pair, access the storage address to call the commodity information, input the commodity information into the currently enabled security classification model for information security detection, and obtain the corresponding detection result:

[0110] If the corresponding key-value pair of the index term is not found in the buffer in the previous step, it means that the acquisition of the detection result corresponding to the index term from the buffer fails. Accordingly, it is necessary to call the currently enabled security classification model to re-determine the detection result of the commodity information corresponding to the index term.

[0111] To implement the security detection of the commodity information, first obtain the corresponding commodity information from the database according to the storage address, such as the picture information or text information mentioned above. Then, after preprocessing the commodity information according to the requirements of the security classification model, input it into the security classification model for detection, and finally obtain the detection result output by the model. Depending on the structure of the security classification model, the content included in the detection result is also correspondingly different, which does not affect the manifestation of the inventive concept of this application.

[0112] Step S1330: Construct a key-value pair corresponding to the commodity information with the index term as the key domain and the detection result output by the security classification model as the value domain, and store it in the key-value pair mapping table in the buffer:

[0113] To implement the caching of the latest detection result obtained by the security classification model, according to the construction principle of the buffer, use the index term corresponding to the storage address as the key and its corresponding detection result as the value to construct a new key-value pair, and store it in the key-value pair mapping table in the buffer.

[0114] This embodiment reflects the process of storing and invoking the mapping relationship between the index item and the detection result according to the key-value pair mapping table in the buffer. According to this process, it can be known that the detection results in the buffer are placed in the order of priority invocation, so as to ensure that the security detection request can quickly obtain the corresponding detection result, and avoid invoking the security classification model for detection for each request, which can save system overhead. When there is no detection result corresponding to the index item in the buffer, it can adaptively invoke the security classification model to perform immediate detection to obtain the detection result. Therefore, it can adapt to the upgrade and replacement of the security classification model and timely and adaptively update the detection results in the buffer to ensure that the security detection request can obtain the latest detection service.

[0115] Please refer to Figure 4 , in some extended embodiments, before the step S1100, the step of responding to the security detection request, the following steps are included:

[0116] Step S1010: Receive and distributively store the product description information submitted by the independent site editor. The product description information includes picture information and / or text information for describing the product object:

[0117] For an independent site, a service for publishing and maintaining the product description information of the product object is provided for the independent site. After receiving the product description information submitted by each independent site for the purpose of publishing or updating a certain product object, it can be distributively stored in the data warehouse, thereby generating the storage address corresponding to the product description information. Specifically, the storage addresses of each specific product information in the product description information can be generated, such as the storage addresses of each product picture therein, or the storage addresses corresponding to a certain part of the text unit therein.

[0118] Generally, whether it is when a product is first published or when the product description information of a published product is edited, multiple product information is processed in batches. For example, for a product object, its product description information will include its product title, main product picture, product details text, product detail pictures, etc. These product information are generally edited uniformly through the same page and can be submitted uniformly after editing. Thus, when the product description information is submitted and stored uniformly each time, the storage addresses corresponding to each specific product information therein can be determined, regardless of whether the corresponding product information is updated in this submission.

[0119] Generally speaking, for the commodity pictures described above, they are stored independently as single pictures. Therefore, each commodity picture has a corresponding storage address. For the text information described above, generally, all the text information corresponding to a commodity object can be regarded as a group, and a single storage address can be provided for calling, or it can also be divided into multiple groups and multiple storage addresses can be provided accordingly. In this regard, those skilled in the art can implement it flexibly.

[0120] Step S1020: Determine the picture information and / or text information in the commodity description information as the commodity information to be detected according to the input parameters defined by the currently enabled security classification model, and obtain the storage address corresponding to the commodity information:

[0121] In theory, information security detection needs to be performed on each specific commodity information in the commodity description information. Therefore, each commodity information can be used as a unit, and according to the input parameter requirements of the security classification model, determine what kind of commodity information to be detected is organized, provide the storage addresses of these commodity information, and encapsulate them as the data carried by the security detection request.

[0122] In one embodiment, the security classification model is suitable for performing security detection on a single commodity picture. Correspondingly, it defines the input of a single commodity picture through input parameters. Accordingly, each commodity picture can be regarded as independent picture information, and the storage addresses can be provided for each commodity picture respectively to call the security detection service of the present application for detection.

[0123] In another embodiment, the security classification model is suitable for performing security detection on a single group of text information. Correspondingly, it defines the input of a single group of text information through input parameters. Accordingly, each group of text information can be regarded as independent text information, and the storage addresses can be provided for each group respectively to call the security detection service of the present application for detection.

[0124] In still another embodiment, the security classification model is implemented to simultaneously use picture information and text information for information security detection. Thus, through the definition of its input parameters, it is required to provide the storage addresses corresponding to the picture information and text information at the same time, usually the storage address of a single commodity picture and the storage address of a single group of text information. Therefore, the two storage addresses are used as parameters to call the security detection service of the present application for detection.

[0125] Step S1030: Trigger a security detection request to obtain the detection result of the commodity information, and the storage address of the commodity information to be detected is included in the request:

[0126] After constructing the storage address to be transmitted corresponding to the described security classification model, the security detection request can be invoked, and the storage address is used as a parameter carried by the request and transmitted to the security detection service of the present application. The security detection service of the present application can execute the business logic disclosed in the foregoing embodiments of the present application according to the storage address therein, so as to obtain corresponding detection results.

[0127] This embodiment further discloses determining the storage address provided to the security detection service of the present application by correspondingly processing the product description information submitted by the user according to the parameters defined by the security classification model. It can be seen that for the security detection service of the present application, as long as the corresponding interfaces are standardized, and then when the e-commerce platform needs to implement information security detection, the storage address of the corresponding product information is provided correspondingly, the compatible docking can be achieved, and the expected detection results can be obtained according to the innovative business logic of the present application.

[0128] In some extended embodiments, on the basis of the previous embodiment, after step S1400, the step of pushing the detection result to answer the security detection request, the following steps are included:

[0129] Step S1511: Determine the security category with the highest confidence among the multiple security categories corresponding to the confidence levels included in the detection result as the detection category corresponding to the product information:

[0130] In this embodiment, the described security classification model is a neural network model implemented based on deep learning. It extracts deep semantic features from the product information input therein, and then performs classification mapping on the basis of the deep semantic features to obtain the confidence levels corresponding to each specific security category mapped to a preset classification space, that is, the classification probabilities of belonging to each security category. The respective security categories and their confidence levels corresponding to the classification space constitute the detection result obtained by the security classification model detection.

[0131] After invoking the security detection service of the present application to obtain the described detection result, the detection result can be parsed to obtain the confidence levels corresponding to each security category therein. Then, the security category with the highest confidence can be determined therefrom as the detection category corresponding to the product information detected by the security detection service.

[0132] Each security category included in the classification space is a category pre-planned in accordance with the laws and regulations of the country or region where the independent station is located. Corresponding meanings can be assigned to these categories in advance, such as "pornography-related", "drug-related", "gambling-related", "health-related", etc. Specifically, those skilled in the art can flexibly preset according to actual needs.

[0133] Step S1512: Determine whether the detection category belongs to a preset permitted category. When it is a permitted category, publish the product object to the online store:

[0134] Business logics for processing different security categories can be preset according to different countries and regions. For example, after determining the detection category corresponding to the product information, first determine whether the detected category belongs to a preset permitted category. For example, the "health category" belongs to the permitted category. When a product information is detected as this category, the product description information containing this product information can be published to the online store of the independent station to make it visible to the public. Otherwise, if it does not belong to the permitted category, the user can be reminded to edit and modify it again.

[0135] This embodiment further improves the business process of publishing the product description information corresponding to the product object in the online store, realizing a business closed-loop. It can be seen from this that with the support of the caching mechanism of the security detection service implemented in this application, when the online store publishes the product description information of its products, it can quickly determine the security of the product information therein with the help of this security detection service, avoiding the need to repeatedly call the security classification model for security detection of each existing specific product information in the product description information every time the product description information of the same product is edited. Instead, through the flexible use of the cache, efficient and fast detection can be achieved, obtaining the corresponding detection results, and quickly updating the product description information based on the detection results.

[0136] Please refer to Figure 5 , in some further embodiments, the process of the security classification model performing information security detection on the product information includes the following steps:

[0137] Step S2100: Preprocess the product information:

[0138] Depending on the structure of the security classification model, the input data it depends on is different, and different preprocessing may be required. For example:

[0139] When the security classification model depends on image information for detection, conventional image preprocessing can be performed on the image information, including image scaling, cropping, product positioning, etc. Those skilled in the art can select the preprocessing means that can be adopted according to the actual situation and implement it.

[0140] When the security classification model depends on text information for detection, preprocessing such as removing stop words, removing spaces, and word segmentation can be performed on the text information. Similarly, it can also be flexibly implemented by those skilled in the art.

[0141] When the security classification model depends on both image information and text information for comprehensive detection, corresponding preprocessing is performed on these two types of product information respectively.

[0142] Step S2200: Use a semantic feature extraction model to extract deep semantic features from the preprocessed product information, and obtain the semantic feature information of the product information:

[0143] The described security classification model includes a semantic feature extraction model corresponding to its structure. Depending on its different structures, the included semantic feature extraction models can be different. For example:

[0144] When the security classification model is constructed to perform security identification on picture information, the semantic feature extraction model can be implemented using a neural network model based on CNN, such as basic models like FastCNN and Resnet series, to represent the features of the picture information and obtain the corresponding semantic feature information.

[0145] When the security classification model is constructed to identify text information, the semantic feature extraction model can be implemented using neural network models suitable for processing text, such as Bert, TextCNN, and LSTM, to represent the features of the text information and obtain the corresponding semantic feature information.

[0146] When the security classification model is constructed to perform security identification using both picture information and text information simultaneously, it can be in two parallel paths. The models corresponding to the above picture information and text information are respectively adopted, and after each performs feature representation to obtain their respective semantic feature information, the two-way semantic feature information can be combined into the same semantic feature information through a splicing layer.

[0147] Step S2300: Map the semantic feature information to the classification space to obtain the confidence levels of each security category of the product information corresponding to the classification space, and form the detection result of the product information:

[0148] After the security classification model obtains the final semantic feature information corresponding to the input product information, it can map this semantic feature information to a preset classification space through a fully connected layer, thereby obtaining the confidence levels corresponding to each preset security category in the classification space, and thus obtaining the detection result corresponding to the input product information.

[0149] It can be understood that the described security classification model can be iteratively trained using a training data set until it converges to obtain security detection capabilities, thereby realizing its upgrade and replacement. Then it is redeployed to the security detection service of this application and its version number is updated accordingly, thereby providing version information different from the historical version, which will also cause the detection results of the product information already existing in the buffer area to be re-obtained to update the detection results. When the security identification accuracy of the security classification model is improved, the accuracy of the detection results in the buffer area will also inevitably be improved.

[0150] As can be seen from this embodiment, the security classification model of the present application is based on deep learning to acquire the intelligent classification ability of the input commodity information to serve the security category detection of the commodity information. The deep learning model can adapt to the upgrade of the data set and continuously upgrade its security detection ability, thereby improving the service ability of the security detection service of the present application, improving the accuracy of its security identification, making the security detection of the commodity description information on the e-commerce platform more accurate and reliable, automatically filtering unsafe commodity information for the e-commerce platform, and playing a role in maintaining the healthy operation of the e-commerce platform and its independent sites.

[0151] Please refer to Figure 6 , in an extended partial embodiment, after the step S1400, the step of pushing the detection result to respond to the security detection request, the following steps are further included:

[0152] Step S1521, in response to the arrival event of the timing task, obtain the version information of the currently enabled security classification model:

[0153] As mentioned above, in the buffer area of the present application, due to the long-term storage of a large number of concurrent security detection requests, a large number of key-value pairs are accumulated. The security classification model is constantly updated. If the detection results corresponding to the historical versions of the security classification model continue to occupy the buffer area and consume system overhead, it will no longer be practical. Therefore, the buffer area can be periodically cleaned.

[0154] In order to realize the periodic cleaning of the cache, it can be achieved through a timing task. The system runs a timing scheduled task. When the timing arrives, the corresponding timing task arrival event is triggered. Respond to this event and obtain the version information of the currently enabled security classification model to determine the latest security classification model.

[0155] Step S1522, retrieve the index items in the buffer area that do not contain the version information, and clear the key-value pair where the index item is located from the buffer area:

[0156] Furthermore, from the key-value pair mapping table in the buffer area, retrieve all the key-value pairs whose key domain, that is, their index items, do not contain the version information. These key-value pairs are the key-value pairs storing the detection results of the old version of the security classification model. Delete these key-value pairs from the buffer area, and the historical data generated by the old version of the security classification model is cleared, completing the cleaning of the buffer area.

[0157] This embodiment automatically cleans the redundant information in the buffer area through a timing task, which can avoid the ineffective occupation of cache resources, improve the cache access efficiency, save system overhead, thereby improving the system response speed and the service effectiveness of the security detection service of the present application.

[0158] Please refer toFigure 7 , An e-commerce information security detection device provided to meet one of the purposes of this application is a functional embodiment of the e-commerce information security detection method of this application. The device includes a request response module 1100, an index construction module 1200, a result acquisition module 1300, and a push response module 1400, where: The request response module 1100 is used to respond to a security detection request, obtain the storage address carried by the request, and obtain the commodity information to be detected according to the storage address; The index construction module 1200 is used to construct an index item, which includes the version information of the currently enabled security classification model and the encoding mapping information of the storage address; The result acquisition module 1300 is used to obtain the detection result mapped to the index item from the buffer area. When the acquisition fails, call the currently enabled security classification model to perform information security detection on the commodity information to obtain the detection result, and map and store the detection result to the index item in the buffer area; The push response module 1400 is used to push the detection result to respond to the security detection request.

[0159] In some further embodiments, the index construction module 1200 includes: a version query unit, which is used to obtain the version number of the currently enabled security classification model as its version information; an encoding mapping unit, which is used to apply a digital digest algorithm to calculate the hash value of the storage address as the encoding mapping information of the storage address; a splicing index unit, which is used to splice the version information and the encoding mapping information into an index item of the commodity information to be detected.

[0160] In some further embodiments, the result acquisition module 1300 includes: a cache call unit, which is used to query whether there is a target key-value pair containing the index item in the key-value pair mapping table in the buffer area. When the target key-value pair exists, obtain the detection result in the value range of the target key-value pair; a model call unit, which is used to, when the target key-value pair does not exist, access the storage address to call the commodity information, input the commodity information into the currently enabled security classification model to perform information security detection, and obtain the corresponding detection result; a result cache unit, which is used to construct a key-value pair corresponding to the commodity information with the index item as the key domain and the detection result output by the security classification model as the value domain, and store it in the key-value pair mapping table in the buffer area.

[0161] In an extended partial embodiment, the e-commerce information security detection device of the present application further includes the following modules that run prior to the request response module 1100: a submission processing module, configured to receive and distributively store the product description information submitted by an independent site editor, where the product description information includes picture information and / or text information for describing a product object; an information invocation module, configured to determine the picture information and / or text information in the product description information as the product information to be detected according to the input parameters defined by the currently enabled security classification model, and obtain the storage address corresponding to the product information; a request triggering module, configured to trigger a security detection request to obtain the detection result of the product information, and include the storage address of the product information to be detected in the request.

[0162] In an extended partial embodiment, the e-commerce information security detection device of the present application further includes the following modules that run after the push response module 1400: a category determination module, configured to determine, according to the confidence levels corresponding to multiple security categories included in the detection result, the security category with the highest confidence level as the detection category corresponding to the product information; a product release module, configured to determine whether the detection category belongs to a preset allowed category, and when it is an allowed category, release the product object to an online store.

[0163] In a deepened partial embodiment, the security classification model includes: a preprocessing unit, configured to preprocess the product information; a feature extraction unit, configured to use a semantic feature extraction model to extract deep semantic features from the preprocessed product information to obtain the semantic feature information of the product information; a classification mapping unit, configured to map the semantic feature information to a classification space to obtain the confidence levels of each security category of the product information corresponding to the classification space, and form the detection result of the product information.

[0164] In an extended partial embodiment, the e-commerce information security detection device of the present application further includes the following modules that run after the push response module 1400: a task response module, configured to respond to the arrival event of a timing task and obtain the version information of the currently enabled security classification model; a cache cleaning module, configured to retrieve the index items in the buffer area that do not include the version information, and clear the key-value pairs where the index items are located from the buffer area.

[0165] To solve the above technical problems, an embodiment of the present application further provides a computer device. As Figure 8As shown, it is a schematic diagram of the internal structure of a computer device. The computer device includes a processor, a computer-readable storage medium, a memory, and a network interface connected through a system bus. Among them, the computer-readable storage medium of the computer device stores an operating system, a database, and computer-readable instructions. The database can store a control information sequence. When the computer-readable instructions are executed by the processor, the processor can implement an e-commerce information security detection method. The processor of the computer device is used to provide computing and control capabilities to support the operation of the entire computer device. The memory of the computer device can store computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor can execute the e-commerce information security detection method of this application. The network interface of the computer device is used to connect and communicate with a terminal. Those skilled in the art can understand that Figure 8 The structure shown in the figure is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0166] In this embodiment, the processor is used to execute Figure 7 the specific functions of each module and its sub-modules in the figure. The memory stores the program code and various types of data required to execute the above modules or sub-modules. The network interface is used for data transmission between the user terminal and the server. The memory in this embodiment stores the program code and data required to execute all modules / sub-modules in the e-commerce information security detection device of this application. The server can call the program code and data of the server to execute the functions of all sub-modules.

[0167] This application also provides a storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors are caused to execute the steps of the e-commerce information security detection method of any embodiment of this application.

[0168] This application also provides a computer program product, including a computer program / instructions. When the computer program / instructions are executed by one or more processors, the steps of the method described in any embodiment of this application are implemented.

[0169] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments of the present application can be completed by instructing relevant hardware through a computer program. This computer program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above various methods. Among them, the aforementioned storage medium can be a computer-readable storage medium such as a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc.

[0170] In summary, the present application can quickly call the information security detection results of the product information published on the independent website, facilitating the independent website to decide whether to publish the said product information.

[0171] Those skilled in the art of this technology can understand that the steps, measures, and solutions in the various operations, methods, and processes discussed in the present application can be alternated, changed, combined, or deleted. Further, the other steps, measures, and solutions in the various operations, methods, and processes discussed in the present application can also be alternated, changed, rearranged, decomposed, combined, or deleted. Further, the steps, measures, and solutions in the prior art that are the same as those disclosed in the various operations, methods, and processes in the present application can also be alternated, changed, rearranged, decomposed, combined, or deleted.

[0172] The above are only some implementation manners of the present application. It should be noted that for those of ordinary skill in the art of this technology, without departing from the principle of the present application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present application.

Claims

1. An e-commerce information security detection method, characterized in that, It includes the following steps: In response to a security detection request, obtain the storage address carried by the request, and obtain the product information to be detected according to the storage address; Construct an index item, which includes the version information of the currently enabled security classification model and the encoded mapping information of the storage address; Obtain the detection result mapped to the index item from the buffer. When the acquisition fails, call the currently enabled security classification model to perform information security detection on the product information to obtain the detection result, and store the detection result mapped to the index item in the buffer; Push the detection result to answer the security detection request; Among them, the constructing of the index item includes: Obtain the version number of the currently enabled security classification model as its version information; Apply a digital digest algorithm to calculate the hash value of the storage address as the encoded mapping information of the storage address; Concatenate the version information and the encoded mapping information into an index item of the product information to be detected; The obtaining of the detection result mapped to the index item from the buffer includes: Query whether there is a target key-value pair containing the index item in the key-value pair mapping table in the buffer. When the target key-value pair exists, obtain the detection result in the value range of the target key-value pair; When the target key-value pair does not exist, access the storage address to call the product information, input the product information into the currently enabled security classification model to perform information security detection, and obtain the corresponding detection result; Construct a key-value pair corresponding to the product information with the index item as the key domain and the detection result output by the security classification model as the value domain, and store it in the key-value pair mapping table in the buffer.

2. The e-commerce information security detection method according to claim 1, wherein, Before the step of responding to the security detection request, it includes the following steps: Receive and distribute and store the product description information submitted by the independent site editor. The product description information includes picture information and / or text information for describing the product object; Determine the picture information and / or text information in the product description information as the product information to be detected according to the input parameters defined by the currently enabled security classification model, and obtain the corresponding storage address of the product information; Trigger a security detection request to obtain the detection result of the product information, and include the storage address of the product information to be detected in the request.

3. The e-commerce information security detection method according to claim 2, characterized in that, After the step of pushing the detection result to answer the security detection request, it includes the following steps: According to the confidence levels corresponding to multiple security categories included in the detection result, determine the security category with the highest confidence level as the detection category corresponding to the product information; Judge whether the detection category belongs to a preset allowed category. When it is an allowed category, publish the product object to the online store.

4. The e-commerce information security detection method according to claim 1, wherein The process of the security classification model performing information security detection on the product information includes the following steps: Preprocess the product information; Adopt a semantic feature extraction model to extract deep semantic features from the preprocessed product information to obtain the semantic feature information of the product information; Map the semantic feature information to the classification space to obtain the confidence levels of each security category of the product information corresponding to the classification space, and constitute the detection result of the product information.

5. The e-commerce information security detection method according to any one of claims 1 to 4, characterized in that After the step of pushing the detection result to respond to the security detection request, the following steps are further included: Respond to the arrival event of the timing task, and obtain the version information of the currently enabled security classification model; Retrieve the index item in the buffer that does not contain the version information, and clear the key-value pair where the index item is located from the buffer.

6. An e-commerce information security detection device, characterized in that, Include: A request response module, configured to respond to a security detection request, obtain the storage address carried by the request, and obtain the product information to be detected according to the storage address; An index construction module, configured to construct an index item, where the index item includes the version information of the currently enabled security classification model and the encoded mapping information of the storage address; A result acquisition module, configured to obtain the detection result mapped to the index item from the buffer. When the acquisition fails, call the currently enabled security classification model to perform information security detection on the product information to obtain the detection result, and map and store the detection result to the index item in the buffer; A push response module, configured to push the detection result to respond to the security detection request; Among them, the construction of the index item includes: Obtain the version number of the currently enabled security classification model as its version information; Apply a digital digest algorithm to calculate the hash value of the storage address as the encoded mapping information of the storage address; Concatenate the version information and the encoded mapping information into an index item of the product information to be detected; The obtaining of the detection result mapped to the index item from the buffer includes: Query whether there is a target key-value pair containing the index item in the key-value pair mapping table in the buffer. When the target key-value pair exists, obtain the detection result in the value range of the target key-value pair; When the target key-value pair does not exist, access the storage address to call the product information, input the product information into the currently enabled security classification model to perform information security detection, and obtain the corresponding detection result; Construct a key-value pair corresponding to the product information with the index item as the key domain and the detection result output by the security classification model as the value domain, and store it in the key-value pair mapping table in the buffer.

7. A computer device, comprising a central processing unit and a memory, characterized in that, The central processing unit is used to call and run the computer program stored in the memory to execute the steps of the method described in any one of claims 1 to 4.

8. A computer-readable storage medium, characterized in that, It stores in the form of computer-readable instructions a computer program implemented according to the method described in any one of claims 1 to 4. When the computer program is called and run by the computer, it executes the steps included in the corresponding method.

Images