Account login verification method and device based on artificial intelligence, medium and equipment
By using an AI-based account login verification method, which trains a model using historical login information to calculate login confidence, the problem of easily cracked CAPTCHAs in existing technologies is solved, achieving higher account security and user protection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGDONG MECHANICAL & ELECTRICAL COLLEGE
- Filing Date
- 2026-04-24
- Publication Date
- 2026-06-19
AI Technical Summary
Existing account login verification methods rely on CAPTCHAs, which pose significant security risks. These CAPTCHAs can be easily obtained by criminals through data interception or fraud, resulting in insufficient account security.
An AI-based account login verification method is adopted. The account login verification model is trained by collecting historical login information. The login confidence is calculated using current and recent login information. If the confidence is low, an abnormal alarm is sent and the account is locked. If the confidence is high, the account functions are restricted or the recycle bin is displayed or hidden, thus realizing intelligent verification of the account.
It improves the security of online account logins, prevents unauthorized login attempts, protects user information and funds, and reduces losses caused by account theft.
Smart Images
Figure CN122247739A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the technical field of account login security verification, specifically to an artificial intelligence-based account login verification method, apparatus, medium, and device. Background Technology
[0002] As society continues to progress and people's living standards continue to improve, communication network technology is gradually being widely applied to all sectors of society. These advanced communication network technologies play an extremely important role in social production and family life, creating substantial wealth for overall social development and improved family lives, and greatly propelling society towards greater modernization and intelligence.
[0003] However, with the rapid development of computer technology and the increasing popularity of the internet, some inherent shortcomings of the internet have created opportunities for criminals. These criminals often use various viruses, Trojan horses, or other malicious programs to steal other people's online accounts. Once they successfully log into someone's online account, they will wantonly steal the data, use these accounts for fraudulent activities, or even spread viruses. Their illegal activities cause extremely serious losses to users, potentially leading to the leakage of users' personal privacy and huge economic losses.
[0004] In the security protection system of online accounts, the login verification process is undoubtedly a crucial line of defense. Most existing online account verification technologies rely solely on CAPTCHAs. However, it's important to note that this seemingly simple CAPTCHA verification method still harbors significant security vulnerabilities. CAPTCHAs can easily be obtained by criminals through data interception, or they can be tricked into handing over CAPTCHAs to users via telecommunications fraud. Therefore, existing account login verification methods still have significant shortcomings in terms of security and urgently need further improvement and refinement to address the increasingly complex cybersecurity landscape. Summary of the Invention
[0005] The purpose of this application is to overcome the shortcomings and deficiencies in the prior art and provide an account login verification method, apparatus, medium and device based on artificial intelligence.
[0006] The first aspect of this application provides an artificial intelligence-based account login verification method, including:
[0007] Collect historical login information of network accounts; the historical login information includes login location, login method and login time;
[0008] The historical login information is used as training samples to train an account login verification model corresponding to the network account.
[0009] In response to the current login operation of the network account, the current login information and recent login information of the network account are input into the account login verification model to obtain the current login confidence of the network account;
[0010] If the current login confidence level is less than or equal to the first preset threshold, an account login anomaly alarm is sent to the designated terminal device, and the network account is locked for a preset protection time.
[0011] As one implementation, after the step of sending an account login anomaly alarm to a designated terminal device and locking the network account for a preset protection time, the method further includes:
[0012] If a login consent instruction is received from the terminal device before the protection period ends, the network account will be unlocked and the current login operation of the network account will be approved.
[0013] As one implementation method, after inputting the current login information and recent login information of the network account into the account login verification model to obtain the current login confidence level of the network account, the method further includes:
[0014] If the current login confidence level is greater than the first preset threshold and less than or equal to the second preset threshold, the current login operation is approved.
[0015] Restrict the account functions of the currently logged-in network account, including: limiting the number of messages that the network account can send while currently logged in, prohibiting the network account from sending mass messages while currently logged in, and limiting the amount of funds that the network account can spend while currently logged in.
[0016] As one implementation method, the step of restricting the account functions of the currently logged-in network account further includes:
[0017] Deleted information or emails from a network account while it is currently logged in are stored in a hidden recycle bin; wherein, the hidden recycle bin is not displayed on the operation interface of a network account that is logged in when the login confidence level is greater than the first preset threshold and less than or equal to the second preset threshold.
[0018] When the login confidence level of the network account is greater than the second preset threshold, the hidden recycle bin and the information or emails stored in the hidden recycle bin are displayed on the network account's operation interface.
[0019] As one implementation method, the step of using the historical login information as training samples to train an account login verification model corresponding to the network account includes:
[0020] Multiple first login information training samples are obtained from the historical login information; wherein each first login information training sample includes a target login information sample corresponding to a target time node, and a recent login information sample whose time is earlier than the target login information sample.
[0021] Cluster analysis was performed on the multiple first login information training samples to obtain several training sample groups;
[0022] From each training sample group, several first login information training samples are randomly selected for manual calibration of login confidence, resulting in several second login information training samples corresponding to each training sample group.
[0023] The artificial intelligence model is trained based on the aforementioned training samples of second login information to obtain the initial model for account login verification;
[0024] Based on the account login verification initial model, the login confidence of the multiple first login information training samples is predicted and calibrated to obtain multiple third login information training samples.
[0025] The initial model for account login verification is trained based on the multiple training samples of third-party login information to obtain the account login verification model.
[0026] As one implementation method, the step of training an initial account login verification model based on the plurality of third login information training samples to obtain the account login verification model includes:
[0027] The normalized confidence difference is obtained by comparing the confidence scores of each third login information training sample with the confidence scores of the second login information training samples in the same training sample group.
[0028] The information difference between each third login information training sample and the second login information training sample of the same training sample group is used to obtain the normalized information difference.
[0029] If the difference between the normalized confidence difference and the normalized information difference is greater than a preset difference threshold, the login confidence of the corresponding third login information training sample is canceled to restore it to the first login information training sample.
[0030] The login confidence of the recovered first login information training sample is manually calibrated to obtain the fourth login information training sample;
[0031] The initial model for account login verification is trained based on the multiple training samples of third login information and the training samples of fourth login information to obtain the account login verification model.
[0032] As one implementation, the account login verification model includes a backbone network layer, a feature convolution processing module, several feature convolution attention processing modules, a feature fusion layer, and a classifier;
[0033] The backbone network layer is used to perform feature extraction processing on the current login information and the recent login information to obtain information features;
[0034] The feature convolution processing module is used to perform convolution processing on the information features to obtain information convolution features;
[0035] The feature convolutional attention processing module is used to perform convolutional fusion and channel attention processing on the information convolutional features to obtain information convolutional attention features;
[0036] The feature fusion layer is used to fuse the information convolutional features and the information convolutional attention features to obtain fused features;
[0037] The classifier is used to output the current login confidence score based on the fusion features.
[0038] Compared to related technologies, the AI-based account login verification method of this application pre-trains an account login verification model corresponding to the network account using the network account as a training sample. Then, each time the network account logs in, in response to the current login operation, the current login information and recent login information of the network account are input into the account login verification model to obtain the current login confidence level of the network account. If the current login confidence level is less than or equal to a first preset threshold, an account login anomaly alarm is sent to a designated terminal device, and the network account is locked for a preset protection time. The account login verification model based on AI technology can improve the login security of network accounts.
[0039] A second aspect of this application provides an artificial intelligence-based account login verification device, comprising:
[0040] The historical login information collection module is used to collect historical login information of network accounts; the historical login information includes login location, login method and login time;
[0041] The model training module is used to train the account login verification model corresponding to the network account by using the historical login information as training samples.
[0042] The login confidence acquisition module is used to respond to the current login operation of the network account by inputting the current login information and recent login information of the network account into the account login verification model to obtain the current login confidence of the network account.
[0043] The alarm module is used to send an account login anomaly alarm to a designated terminal device if the current login confidence level is less than or equal to a first preset threshold, and to lock the network account for a preset protection time.
[0044] Compared to related technologies, the AI-based account login verification device of this application pre-trains an account login verification model corresponding to the network account using the network account as a training sample. Then, each time the network account logs in, in response to the current login operation, the current login information and recent login information of the network account are input into the account login verification model to obtain the current login confidence level of the network account. If the current login confidence level is less than or equal to a first preset threshold, an account login anomaly alarm is sent to a designated terminal device, and the network account is locked for a preset protection time. The account login verification model based on AI technology can improve the login security of network accounts.
[0045] A third aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the artificial intelligence-based account login verification method described above.
[0046] A fourth aspect of this application provides a computer device including a storage device, a processor, and a computer program stored in the storage device and executable by the processor, wherein the processor executes the computer program to implement the steps of the artificial intelligence-based account login verification method described above.
[0047] To provide a clearer understanding of this application, the specific embodiments of this application will be described below in conjunction with the accompanying drawings. Attached Figure Description
[0048] Figure 1 This is a flowchart of an AI-based account login verification method according to an embodiment of this application.
[0049] Figure 2 This is a flowchart illustrating the training process of an account login verification model according to one embodiment of this application.
[0050] Figure 3 This is a schematic diagram of the module connections of an AI-based account login verification device according to an embodiment of this application.
[0051] 100. Account login verification device; 101. Historical login information collection module; 102. Model training module; 103. Login confidence acquisition module; 104. Alarm module. Detailed Implementation
[0052] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.
[0053] It should be understood that the described embodiments are merely some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of the embodiments of this application.
[0054] In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. In the description of this application, it should be understood that the terms "first," "second," "third," etc., are used only to distinguish similar objects and are not necessarily used to describe a specific order or sequence, nor should they be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances. The singular forms "a," "the," and "the" used in this application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise. The word "if" as used herein can be interpreted as "when," "when," or "in response to determination."
[0055] Furthermore, in the description of this application, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.
[0056] Please see Figure 1 This is a flowchart of an AI-based account login verification method according to the first embodiment of this application. The method includes:
[0057] S1: Collect historical login information of network accounts; the historical login information includes login location, login method and login time.
[0058] The login location can be the location information at the time of login. Login methods include web login, computer program login, mobile program login, and mini-program login within an application. Login time includes the time zone corresponding to the location information at the time of login.
[0059] S2: Using the historical login information as training samples, train the account login verification model corresponding to the network account.
[0060] S3: In response to the current login operation of the network account, input the current login information and recent login information of the network account into the account login verification model to obtain the current login confidence of the network account.
[0061] Recent login information refers to login information from several previous login attempts or login information within a preset time period, such as the previous login, the previous five login attempts, login information from the past week, login information from the past month, etc.
[0062] S4: If the current login confidence level is less than or equal to the first preset threshold, send an account login anomaly alarm to the designated terminal device and lock the network account for a preset protection time.
[0063] By locking the network account, it is possible to prevent the network account from being successfully logged in when the current login confidence level is less than or equal to a first preset threshold. At the same time, an account login anomaly alarm is sent to notify the user of the login anomaly and to take actions such as changing the password.
[0064] The specified terminal device can be a specified email address, a specified phone number, a specified chat software account, or a communication device with a specified network address. The communication device can be a mobile phone, computer, tablet computer, etc.
[0065] Artificial intelligence (AI) is the broadest concept, aiming to enable machines to simulate human intelligent behavior, such as understanding language, recognizing images, and making decisions and reasoning. Machine learning is one of the main methods for achieving AI; it trains models with data to give machines the ability to "learn." Deep learning is an advanced branch of machine learning, and its core is the use of multi-layered neural networks for automatic feature extraction and pattern recognition. Therefore, deep learning training (i.e., the process of training deep neural networks with large amounts of data) is one of the technical means of artificial intelligence.
[0066] Compared to related technologies, the AI-based account login verification method of this application pre-trains an account login verification model corresponding to the network account using the network account as a training sample. Then, each time the network account logs in, in response to the current login operation, the current login information and recent login information of the network account are input into the account login verification model to obtain the current login confidence level of the network account. If the current login confidence level is less than or equal to a first preset threshold, an account login anomaly alarm is sent to a designated terminal device, and the network account is locked for a preset protection time. The account login verification model based on AI technology can improve the login security of network accounts.
[0067] In a feasible embodiment, after step S4: sending an account login anomaly alarm to a designated terminal device and locking the network account for a preset protection time, the method further includes:
[0068] S41: If a login consent instruction is received from the terminal device before the protection time expires, the network account is unlocked and the current login operation of the network account is consented to.
[0069] In this embodiment, receiving a login consent instruction from the terminal device indicates that the user agrees to the login operation, thus agreeing to the current login operation of the network account to facilitate normal login and use of the network account.
[0070] In a feasible embodiment, after step S3: inputting the current login information and recent login information of the network account into the account login verification model to obtain the current login confidence of the network account, the method further includes:
[0071] S31: If the current login confidence level is greater than the first preset threshold and less than or equal to the second preset threshold, agree to the current login operation;
[0072] S32: Restrict the account functions of the currently logged-in network account, including: limiting the number of messages that the network account can send while currently logged in, prohibiting the network account from sending mass messages while currently logged in, and limiting the amount of funds that the network account can spend while currently logged in.
[0073] In this embodiment, if the current login confidence level is greater than the first preset threshold and less than or equal to the second preset threshold, it indicates that the current login operation has a high probability of being a normal login, but there is still a small risk of abnormal login. Therefore, it is necessary to restrict the account functions of the currently logged-in network account in order to protect the information interaction security and financial security of the network account.
[0074] In one feasible embodiment, step S32: restricting the account functions of the currently logged-in network account, further includes:
[0075] S321: Store the information or emails deleted by the network account in the current login state in a hidden recycle bin; wherein, the hidden recycle bin is not displayed on the operation interface of the network account when the login confidence is greater than the first preset threshold and less than or equal to the second preset threshold;
[0076] S322: When the login confidence level of the network account is greater than the second preset threshold, the hidden recycle bin and the information or emails stored in the hidden recycle bin are displayed on the operation interface of the network account.
[0077] In this embodiment, when the login confidence level is greater than a first preset threshold and less than or equal to a second preset threshold, the information or emails deleted under the current login status are stored in a hidden recycle bin. Then, when the login confidence level of the network account is greater than the second preset threshold, the hidden recycle bin and the information or emails stored in the hidden recycle bin are displayed. This helps protect the information of the network account and prevents the information of the network account from being directly deleted when the login confidence level is greater than the first preset threshold and less than or equal to the second preset threshold.
[0078] Please see Figure 2 In one feasible embodiment, step S2: using the historical login information as training samples to train an account login verification model corresponding to the network account, includes:
[0079] S21: Obtain multiple first login information training samples from the historical login information; wherein each first login information training sample includes a target login information sample corresponding to the target time node, and a recent login information sample whose time is earlier than the target login information sample;
[0080] S22: Perform cluster analysis on the multiple first login information training samples to obtain several training sample groups;
[0081] Cluster analysis is an unsupervised learning method whose core objective is to automatically group similar objects into the same cluster based on the similarity between data, while making objects in different clusters as dissimilar as possible. It does not require pre-labeled data; instead, it starts from the data itself to discover hidden structures or patterns.
[0082] S23: Randomly select several first login information training samples from each training sample group to manually calibrate the login confidence, and obtain several second login information training samples corresponding to each training sample group.
[0083] S24: Train the artificial intelligence model based on the aforementioned training samples of the second login information to obtain the initial model for account login verification;
[0084] Among them, the artificial intelligence model is a deep learning model.
[0085] S25: Based on the account login verification initial model, predict and calibrate the login confidence of the multiple first login information training samples to obtain multiple third login information training samples;
[0086] S26: Train the initial model for account login verification based on the multiple training samples of third login information to obtain the account login verification model.
[0087] In this embodiment, multiple first login information training samples are divided into several training sample groups through cluster analysis. Several first login information training samples are randomly selected for manual calibration of login confidence to train the initial model for account login verification. Then, based on the initial model for account login verification, the login confidence of multiple uncalibrated first login information training samples is predicted and calibrated before being used to continue training the initial model for account login verification. This helps to improve the efficiency of login confidence calibration of login information training samples and the training efficiency of the model.
[0088] In a feasible embodiment, S26: The step of training the initial account login verification model based on the plurality of third login information training samples to obtain the account login verification model includes:
[0089] S261: The normalized confidence difference is obtained by taking the confidence difference between the login confidence of each third login information training sample and the login confidence of the second login information training sample in the same training sample group.
[0090] S262: The information difference between each third login information training sample and the second login information training sample of the same training sample group is used to obtain the normalized information difference.
[0091] S263: If the difference between the normalized confidence difference and the normalized information difference is greater than a preset difference threshold, cancel the login confidence of the corresponding third login information training sample to restore it to the first login information training sample.
[0092] S264: Manually calibrate the login confidence of the recovered first login information training sample to obtain the fourth login information training sample;
[0093] S265: Train the initial model for account login verification based on the multiple training samples of third login information and the training samples of fourth login information to obtain the account login verification model.
[0094] In this embodiment, since the login information training samples in the same training sample group have certain data similarity, the difference in login confidence corresponding to the login information training samples in the same training sample group will not be too large under normal circumstances. By combining the normalized information difference and normalized confidence difference between the login information training samples in the same training sample group, it is possible to quickly determine whether the login confidence calibration of the third login information training sample is normal. This is beneficial for screening out the third login information training samples with abnormal calibration, recalibrating them manually, and then training the initial account login verification model in combination with the normally calibrated third login information training samples, thereby improving the accuracy of the trained account login verification model.
[0095] In one feasible embodiment, the account login verification model includes a backbone network layer, a feature convolution processing module, several feature convolution attention processing modules, a feature fusion layer, and a classifier;
[0096] The backbone network layer is used to perform feature extraction processing on the current login information and the recent login information to obtain information features;
[0097] The feature convolution processing module is used to perform convolution processing on the information features to obtain information convolution features;
[0098] The feature convolutional attention processing module is used to perform convolutional fusion and channel attention processing on the information convolutional features to obtain information convolutional attention features;
[0099] The feature fusion layer is used to fuse the information convolutional features and the information convolutional attention features to obtain fused features;
[0100] The classifier is used to output the current login confidence score based on the fusion features.
[0101] Please see Figure 3 The second embodiment of this application provides an artificial intelligence-based account login verification device 100, comprising:
[0102] The historical login information collection module 101 is used to collect historical login information of network accounts; the historical login information includes login location, login method and login time.
[0103] Model training module 102 is used to train an account login verification model corresponding to the network account by using the historical login information as training samples.
[0104] The login confidence acquisition module 103 is used to respond to the current login operation of the network account by inputting the current login information and recent login information of the network account into the account login verification model to obtain the current login confidence of the network account.
[0105] The alarm module 104 is used to send an account login anomaly alarm to a designated terminal device if the current login confidence level is less than or equal to a first preset threshold, and to lock the network account for a preset protection time.
[0106] It should be noted that the AI-based account login verification device 100 provided in the second embodiment of this application is only illustrated by the above-described division of functional modules when executing the AI-based account login verification method. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. Furthermore, the AI-based account login verification device 100 provided in the second embodiment of this application and the AI-based account login verification method of the first embodiment of this application belong to the same concept, and its implementation process is detailed in the method embodiment, which will not be repeated here.
[0107] A third aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the artificial intelligence-based account login verification method described above.
[0108] A fourth aspect of this application provides a computer device including a storage device, a processor, and a computer program stored in the storage device and executable by the processor, wherein the processor executes the computer program to implement the steps of the artificial intelligence-based account login verification method described above.
[0109] The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without any inventive effort.
[0110] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0111] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function selected in one or more boxes.
[0112] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function selected in one or more boxes.
[0113] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0114] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.
[0115] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.
[0116] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.
[0117] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.
Claims
1. An artificial intelligence-based account login verification method, characterized by, include: Collect historical login information of network accounts; The historical login information includes login location, login method, and login time; The historical login information is used as training samples to train an account login verification model corresponding to the network account. In response to the current login operation of the network account, the current login information and recent login information of the network account are input into the account login verification model to obtain the current login confidence of the network account; If the current login confidence level is less than or equal to the first preset threshold, an account login anomaly alarm is sent to the designated terminal device, and the network account is locked for a preset protection time. 2.The AI-based account login verification method of claim 1, wherein, After the steps of sending an account login anomaly alarm to a designated terminal device and locking the network account for a preset protection time, the method further includes: If a login consent instruction is received from the terminal device before the protection period ends, the network account will be unlocked and the current login operation of the network account will be approved. 3.The AI-based account login verification method of claim 1, wherein, After inputting the current login information and recent login information of the network account into the account login verification model to obtain the current login confidence of the network account, the method further includes: If the current login confidence level is greater than the first preset threshold and less than or equal to the second preset threshold, the current login operation is approved. Restrict the account functions of the currently logged-in network account, including: limiting the number of messages that the network account can send while currently logged in, prohibiting the network account from sending mass messages while currently logged in, and limiting the amount of funds that the network account can spend while currently logged in. 4.The AI-based account login verification method of claim 1, wherein, The step of restricting the account functions of the currently logged-in network account further includes: Deleted information or emails from a network account while it is currently logged in are stored in a hidden recycle bin; wherein, the hidden recycle bin is not displayed on the operation interface of a network account that is logged in when the login confidence level is greater than the first preset threshold and less than or equal to the second preset threshold. When the login confidence level of the network account is greater than the second preset threshold, the hidden recycle bin and the information or emails stored in the hidden recycle bin are displayed on the network account's operation interface. 5.The AI-based account login verification method of claim 1, wherein, The step of using the historical login information as training samples to train an account login verification model corresponding to the network account includes: Multiple first login information training samples are obtained from the historical login information; wherein each first login information training sample includes a target login information sample corresponding to a target time node, and a recent login information sample whose time is earlier than the target login information sample. Cluster analysis was performed on the multiple first login information training samples to obtain several training sample groups; From each training sample group, several first login information training samples are randomly selected for manual calibration of login confidence, resulting in several second login information training samples corresponding to each training sample group. The artificial intelligence model is trained based on the aforementioned training samples of second login information to obtain the initial model for account login verification; Based on the account login verification initial model, the login confidence of the multiple first login information training samples is predicted and calibrated to obtain multiple third login information training samples. The initial model for account login verification is trained based on the multiple training samples of third-party login information to obtain the account login verification model. 6.The AI-based account login verification method of claim 5, wherein, The steps for training the initial account login verification model based on the multiple third-party login information training samples to obtain the account login verification model include: The normalized confidence difference is obtained by comparing the confidence scores of each third login information training sample with the confidence scores of the second login information training samples in the same training sample group. The information difference between each third login information training sample and the second login information training sample of the same training sample group is used to obtain the normalized information difference. If the difference between the normalized confidence difference and the normalized information difference is greater than a preset difference threshold, the login confidence of the corresponding third login information training sample is canceled to restore it to the first login information training sample. The login confidence of the recovered first login information training sample is manually calibrated to obtain the fourth login information training sample; The initial model for account login verification is trained based on the multiple training samples of third login information and the training samples of fourth login information to obtain the account login verification model.
7. The artificial intelligence-based account login verification method according to any one of claims 1-6, characterized in that, The account login verification model includes a backbone network layer, a feature convolution processing module, several feature convolution attention processing modules, a feature fusion layer, and a classifier; The backbone network layer is used to perform feature extraction processing on the current login information and the recent login information to obtain information features; The feature convolution processing module is used to perform convolution processing on the information features to obtain information convolution features; The feature convolutional attention processing module is used to perform convolutional fusion and channel attention processing on the information convolutional features to obtain information convolutional attention features; The feature fusion layer is used to fuse the information convolutional features and the information convolutional attention features to obtain fused features; The classifier is used to output the current login confidence score based on the fusion features.
8. An account login verification device based on artificial intelligence, characterized in that, include: The historical login information collection module is used to collect historical login information of network accounts; The historical login information includes login location, login method, and login time; The model training module is used to train the account login verification model corresponding to the network account by using the historical login information as training samples. The login confidence acquisition module is used to respond to the current login operation of the network account by inputting the current login information and recent login information of the network account into the account login verification model to obtain the current login confidence of the network account. The alarm module is used to send an account login anomaly alarm to a designated terminal device if the current login confidence level is less than or equal to a first preset threshold, and to lock the network account for a preset protection time.
9. A computer readable storage medium, the computer readable storage medium storing a computer program, characterized in that: When the computer program is executed by the processor, it implements the steps of the AI-based account login verification method as described in any one of claims 1 to 7.
10. A computer device, comprising: The device includes a storage device, a processor, and a computer program stored in the storage device and executable by the processor, wherein the processor executes the computer program to implement the steps of the artificial intelligence-based account login verification method as described in any one of claims 1 to 7.