WO2020211247A1 - 账户信息的登录方法、装置、计算机设备及计算机存储介质 - Google Patents

账户信息的登录方法、装置、计算机设备及计算机存储介质 Download PDF

Info

Publication number
WO2020211247A1
WO2020211247A1 PCT/CN2019/103197 CN2019103197W WO2020211247A1 WO 2020211247 A1 WO2020211247 A1 WO 2020211247A1 CN 2019103197 W CN2019103197 W CN 2019103197W WO 2020211247 A1 WO2020211247 A1 WO 2020211247A1
Authority
WO
WIPO (PCT)
Prior art keywords
login
account information
user
risk level
account
Prior art date
Application number
PCT/CN2019/103197
Other languages
English (en)
French (fr)
Inventor
朱坤
Original Assignee
平安科技（深圳）有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 平安科技（深圳）有限公司 filed Critical 平安科技（深圳）有限公司
Publication of WO2020211247A1 publication Critical patent/WO2020211247A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00—Network architectures or network communication protocols for network security
    • H04L63/08—Network architectures or network communication protocols for network security for authentication of entities
    • H04L63/0815—Network architectures or network communication protocols for network security for authentication of entities providing single-sign-on or federations
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00—Network architectures or network communication protocols for network security
    • H04L63/08—Network architectures or network communication protocols for network security for authentication of entities
    • H04L63/083—Network architectures or network communication protocols for network security for authentication of entities using passwords
    • H04L63/0838—Network architectures or network communication protocols for network security for authentication of entities using passwords using one-time-passwords
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00—Network architectures or network communication protocols for network security
    • H04L63/08—Network architectures or network communication protocols for network security for authentication of entities
    • H04L63/0861—Network architectures or network communication protocols for network security for authentication of entities using biometrical features, e.g. fingerprint, retina-scan
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/08—Protocols specially adapted for terminal emulation, e.g. Telnet

Definitions

  • This application relates to the field of information technology, in particular to account information login methods, devices, computer equipment and computer non-volatile readable storage media.
  • the system usually has multiple login methods. For example, when a user enters the Alipay application for the first time, he needs to enter a user name and password, because the Alipay client uses the login method of frequent login. When the user enters the Alipay application again, he can directly enter the Alipay interface without verification. In addition, when the user uses mobile banking to transfer or pay, he needs to log in through the dynamic password authentication of the client, while 58.com and Ganji. The client will use the login method of face verification.
  • the login method recommended by the APP client is not fixed, and the flexibility is relatively low.
  • the following situations may exist.
  • the risk level of the user account is not high, the more complex login method of face verification is recommended , which makes the login operation cumbersome, and when the risk level of the user account is high, the relatively simple login method of long log verification is recommended, which cannot guarantee the security of the user account.
  • this application provides an account information login method, device, computer equipment, and computer non-volatile readable storage medium, the main purpose is to solve the current account login method recommended by the user in the account information login process The problem of low flexibility.
  • a method for logging in account information including:
  • the characteristic data of the multiple groups of user login accounts are input as training data into the deep learning model for training to construct a login risk identification model, which records the characteristic data of the user login account and the risk level of the login account information
  • a login risk identification model which records the characteristic data of the user login account and the risk level of the login account information
  • a login method of the account information that matches the risk level of the login account information is selected, and the account information is logged in.
  • a device for logging in account information comprising:
  • the acquiring unit is configured to acquire characteristic data of multiple groups of user login accounts, and the characteristic data carries a risk level label suitable for the login account information;
  • the construction unit is used to input the characteristic data of the multiple groups of user login accounts as training data into the deep learning model for training, and construct a login risk identification model.
  • the login risk identification model records the characteristic data of the user login account and login The mapping relationship between the risk levels of account information;
  • the identification unit is configured to, when a login request for account information is received, input the characteristic data of the user login account corresponding to the login request into the login risk identification model to identify the risk level of the login account information;
  • the login unit is configured to select a login method of account information matching the risk level of the login account information according to the risk level of the login account information, and log in the account information.
  • a computer non-volatile readable storage medium having computer readable instructions stored thereon, and when the computer readable instructions are executed by a processor, the following steps are implemented: Obtain multiple groups of users The characteristic data of the login account, the characteristic data carries a risk level label applicable to the login account information; the characteristic data of the multiple groups of user login accounts are input as training data into the deep learning model for training, and the login risk identification is constructed
  • the login risk identification model records the mapping relationship between the characteristic data of the user login account and the risk level of the login account information; when a login request for account information is received, the login request corresponds to the characteristics of the user login account Data is input into the login risk identification model to identify the risk level of the login account information; according to the risk level of the login account information, select the login method of the account information that matches the risk level of the login account information, Information to log in.
  • a computer device including a memory, a processor, and computer-readable instructions stored on the memory and executable on the processor, and the processor executes the computer-readable instructions
  • the processor executes the computer-readable instructions
  • the model is trained to construct a login risk identification model.
  • the login risk identification model records the mapping relationship between the characteristic data of the user's login account and the risk level of the login account information; when a login request for account information is received, it will The characteristic data of the user login account corresponding to the login request is input to the login risk identification model to identify the risk level of the login account information; according to the risk level of the login account information, select the risk level that matches the risk level of the login account information The login method of the account information is to log in the account information.
  • this application provides a method and device for logging in account information, which collects multiple sets of user log-in account characteristic data as training data and inputs them into a deep learning model for training, and builds a login risk identification model.
  • the risk level of the login account information is identified through the login risk identification model, and then the login method of the account information that matches the risk level of the login account is selected, the account information is logged in, and the flexibility of account information login is improved .
  • this application uses a login risk identification model to identify the risk level of login account information, and flexibly recommends users to use different based on the risk level of account information
  • the account login method is used to log in the account information. While ensuring the security of the account information login process, the security of the user account is guaranteed during the account information login process, so that the user does not need to manually modify the login method and improve The flexibility of the account information login process is improved.
  • FIG. 1 shows a schematic flowchart of a method for logging in account information provided by an embodiment of the present application
  • FIG. 2 shows a schematic flowchart of another method for logging in account information provided by an embodiment of the present application
  • FIG. 3 shows a schematic structural diagram of an apparatus for logging in account information provided by an embodiment of the present application
  • FIG. 4 shows a schematic structural diagram of another device for logging in account information provided by an embodiment of the present application.
  • the embodiment of the application provides a login method for account information, which can flexibly recommend users to use different account login methods based on the risk level of the account information, thereby improving the flexibility of the account information login process.
  • the method includes:
  • the characteristic data of the user's login account includes but is not limited to the user's login time, operating location, device type, whether the account is tradable, user risk tolerance, identity information, account asset limit and other characteristic data, specifically when the user logs in to the application service for the first time Accounts are registered through personal identification information, such as account name, phone number, gender, etc., as well as settings for application service operations, such as payment deduction order, login method, and login status settings.
  • the user will set account information such as the account name and account login password during the process of registering account information, so that the user can log in to the application service again through the set account information.
  • account information such as the account name and account login password
  • security verification is usually set during login and transaction.
  • the risk is higher when transaction authentication is involved in the use process.
  • long login status is set after login, and security verification is set during transaction.
  • security verification is not set when logging in, and security verification is set when trading. Therefore, the acquired characteristic data carries a risk level label applicable to the login account information.
  • the login risk identification model records the mapping relationship between the characteristic data of the user's login account and the risk level of the login account information. Since the deep learning model has the function of the mapping relationship between the training data, by constructing the login risk identification model, different The characteristic data of the user's login account and the mapping relationship obtained through the training of the login risk identification model can obtain the risk level of the login account information corresponding to the characteristic data of the user's login account.
  • the risk level labels can be set to three levels, from high to low, respectively, level A, level B, and level C.
  • level A is set for feature data involving large-value transactions
  • level A is set for feature data modified by personal information
  • level B is set for feature data involving medium-value transactions, etc.
  • each level can also be set
  • the range of feature data is not limited here, so that the feature data of multiple groups of user login accounts are trained according to the risk level applicable to the login account information.
  • the deep learning model here may be a convolutional neural network model.
  • the network structure of the login risk identification model can be constructed by repeatedly training multiple sets of user login account characteristic data, and the network structure can train the input data , And give the correct input-output relationship, which is equivalent to the mapping relationship between the characteristic data of the user's login account and the risk level of the login account information.
  • the structure of the specific convolutional neural network model can be realized through the convolutional layer, fully connected layer, pooling layer and classification layer structure.
  • the convolutional layer here is equivalent to the hidden layer of the convolutional neural network, which can be a multi-layer structure.
  • the pooling layer is often inserted in the continuous convolutional layer; here the full connection
  • the layer is similar to the convolutional layer.
  • the neurons of the convolutional layer are connected to the output local area of the previous layer. Of course, in order to reduce the output feature vector too much, two fully connected layers can be set. After the training data is trained through several convolutional layers Integrate the feature data of training output.
  • the characteristic data of the login request corresponding to the user's login account is characteristic data with no risk level marked, and the risk level of the login account information mapped to the characteristic data can be identified through the login risk identification model.
  • setting the risk level of login account information can include low-level risk, intermediate risk and high-level risk. If the transaction amount involved in the characteristic data input to the login risk level is greater than 10 yuan, the risk level of the login account information is identified as medium risk.
  • the risk level of the login account information select a login method of the account information that matches the risk level of the login account information, and log in the account information.
  • the risk level of the login account information may include, but is not limited to, face verification, identity verification, dynamic password verification, verification code login, and secret login, etc.
  • the risk level of the login account information indicates the security risk of the account information during the login process
  • the corresponding log-in account information has a lower risk level, indicating that the account information security risk is low, and the user can choose to log in without password or Other login methods that do not require verification.
  • the risk level of the login account is medium, indicating that the account information still has a certain security risk. You can choose the login method of dynamic password verification.
  • the risk level for the login account is higher, indicating that the account information security risk is relatively high. High, you can choose the login method of face verification or identity verification.
  • the embodiment of the application provides a method for logging in to account information.
  • a login risk identification model is constructed.
  • Identify the risk level of the login account information through the login risk identification model, and then select the login method of the account information that matches the risk level of the login account, log the account information, and improve the flexibility of account information login.
  • this application uses a login risk identification model to identify the risk level of login account information, and flexibly recommends users to use different based on the risk level of account information
  • the account login method is used to log in the account information. While ensuring the security of the account information login process, the security of the user account is guaranteed during the account information login process, so that the user does not need to manually modify the login method and improve The flexibility of the account information login process is improved.
  • the embodiment of the application provides another method for logging in account information, which can provide different account login methods based on the user's risk level to ensure the security of user account information during the account login process.
  • the method include:
  • the identity data of the user's registered account may include but is not limited to the user name, gender, contact number, device IP address and other data that can indicate the user's identity.
  • the historical operation data of the user's login account may include but is not limited to user transaction data, User browsing data, etc. can indicate the user's operating habits. Calling the identity data of the user's registered application account through the preset interface can obtain the user's personal identity information data, and calling the historical operation data of the user's login account information through the preset interface can obtain the user's operating habits data.
  • the characteristic data is equivalent to the personal information of the user registered application account and the user login account.
  • Historical operation data of information for example, for network services involving financials, security verification is usually set during login and transaction, such as login account passwords, transaction passwords, etc.
  • security verification is set during transaction.
  • the personal information of the user registered application account and the historical operation data of the user login account information are further set Can extract the characteristic data of the user's login account.
  • the characteristic data of each group of user login accounts mark the characteristic data of each group of user login accounts, and obtain the risk level label corresponding to the login account information of each group of characteristic data.
  • the characteristic data of the user's login account will involve different data ranges. For example, if the characteristic data involves the user's personal information data is changed, the account information may have a security risk, and the personal information data involves multiple Information fields. Not every information field being changed will cause the account information to have a security risk. If the changed personal information data is within the pre-set range of the dangerous information field, it will cause the account information to have a security risk. Similarly, for features Data involving changes in the account amount will lead to security risks in the account information, and different data ranges of the changed amount values will have different levels of security risks. Therefore, according to the data range of the characteristic data of the user's login account, the characteristic data can be marked with a risk level, and the risk level label corresponding to the login account information of the characteristic data can be obtained.
  • a set of feature data each time the user logs into the application account can be used as an N-dimensional vector, and each feature data It is expressed as a value in an N-dimensional vector, and after normalizing the characteristic data, the distribution range of each characteristic data is [-1,1], and the characteristic data not set by the user is expressed as a value of 0, In order to obtain training data in a standard format.
  • the deep learning model may be a convolutional neural network, each layer of structure has different input and output parameters to achieve different functions, through the convolutional neural network to train multiple groups of user login account feature data to obtain users
  • the mapping relationship between the characteristic data of the login account and the risk level of the login account information is equivalent to the login risk identification model.
  • the specific convolutional neural network includes a multi-layer structure.
  • the characteristics of the characteristic data of the user's login account can be extracted through the convolutional layer of the convolutional neural network model to obtain the characteristic parameters of the user's login account at each risk level; through the convolutional neural network model
  • the pooling layer performs dimensionality reduction processing on the characteristic parameters of the user login account at each risk level, and obtains the characteristic parameters of the user login account at each risk level after the dimensionality reduction process;
  • the fully connected layer of the convolutional neural network model summarizes the reduction After dimension processing, the characteristic parameters of the user login account at each risk level are obtained, and the weight values of the characteristic data of the user login account at different risk levels are obtained;
  • the classification layer of the convolutional neural network model is at different risks according to the characteristic data of the user login account
  • the weight value on the level generates a mapping relationship between the characteristic data of the user's login account and the risk level of the login account information, and builds a login risk identification model.
  • a convolutional neural network includes 13 convolutional layers and 3 fully connected layers.
  • 1 pooling layer is connected, and the above 13 convolutional layers and 3 fully connected layers are all processed with a nonlinear activation function.
  • the above convolutional layer, fully connected layer and pooling layer The number of layers is not limited, and can be selected according to actual conditions. Similarly, the activation function selected in each layer is also not limited.
  • the mapping relationship between the characteristic data of the user login account and the risk level of the login account information is output, which is equivalent to the classification results of the characteristic data of the user login account at each risk level.
  • the risk level of the login account information identified by the login risk identification model is only based on the characteristic data of the user login account corresponding to the login request to preliminarily determine the risk of the user login, and further improve the user login account information safety.
  • the risk level of the login account information select a login method of the account information that matches the risk level of the login account information, and log in the account information.
  • the risk level of the login account information in order to facilitate the user to log in account information, after identifying the risk level of the login account information, you can understand the risk of the current account login in real time. If the risk level is high, it indicates the security risk of the login account. If the risk level is low, it means that the security risk of the login account is low, and it does not require user verification to log in. For example, login without password, select further The account information login method that matches the risk level of the login account information is used to log in the account information, thereby ensuring the security of the account information login process.
  • the account information login method is adjusted according to the risk level of the login account information obtained by re-identification, and the account information is logged in .
  • the characteristic data of the user login account information is constantly changing. As the characteristic data of the user login account information changes, if the characteristic data of multiple groups of user login accounts changes, the risk identification model is constructed to identify the data again. The risk level of the login account information will also change, and the login method of the account information is adjusted according to the risk level of the login account information obtained by re-identification, and the account information is logged in.
  • the login time in the characteristic data of the user's login account indicates that the user has been logged in more than one week since the last login.
  • the risk level of the identified user's login account has increased, or the personal information in the characteristic data of the user's login account has been changed.
  • the risk level of the login account is increased, and the account information is further logged in according to the login method corresponding to the increased risk level.
  • an embodiment of the present application provides a device for logging in account information.
  • the device includes: an acquisition unit 31, a construction unit 32, an identification unit 33, Login unit 34.
  • the acquiring unit 31 may be used to acquire characteristic data of multiple groups of user login accounts, where the characteristic data carries risk level labels suitable for login account information;
  • the constructing unit 32 may be used to input the characteristic data of the multiple groups of user login accounts as training data into the deep learning model for training, and construct a login risk identification model.
  • the login risk identification model records the characteristic data of the user login account. Mapping relationship with the risk level of login account information;
  • the identification unit 33 may be used to input the characteristic data of the user login account corresponding to the login request into the login risk identification model when a login request for account information is received, to identify the risk level of the login account information;
  • the login unit 34 may be configured to select a login method of account information matching the risk level of the login account information according to the risk level of the login account information, and log in the account information.
  • the account information login device collects the characteristic data of multiple groups of user login accounts as training data and inputs them into a deep learning model for training to construct a login risk identification model.
  • the risk level of the login account information is identified through the login risk identification model, and then the login method of the account information that matches the risk level of the login account is selected to log in the account information and improve the flexibility of account information login.
  • this application uses a login risk identification model to identify the risk level of login account information, and flexibly recommends users to use different based on the risk level of account information
  • the account login method is used to log in the account information. While ensuring the security of the account information login process, the security of the user account is guaranteed during the account information login process, so that the user does not need to manually modify the login method and improve The flexibility of the account information login process is improved.
  • FIG. 4 is a schematic structural diagram of another device for logging in account information according to an embodiment of the present application. As shown in FIG. 4, the device further includes:
  • the marking unit 35 may be used to mark the characteristic data of each group of user login accounts according to the data range of the characteristic data of each group of user login accounts after acquiring the characteristic data of multiple groups of user login accounts, Obtain the risk level label of each group of characteristic data corresponding to the login account information;
  • the adjustment unit 36 may be used to select a login method of account information matching the risk level of the login account information according to the risk level of the login account information, and after logging in the account information, if the risk level The characteristic data of the group user's login account changes.
  • the account information login method is adjusted according to the risk level of the login account information obtained by re-identification, and the account information is logged in.
  • the marking unit 35 includes:
  • the comparison module 351 can be used to extract the characteristic range of the characteristic information of each group of user login accounts, and compare the characteristic range of the characteristic information of each group of user login accounts with the characteristic ranges corresponding to the preset risk levels Perform comparison and get the comparison result;
  • the marking module 352 may be used to mark the risk level of the characteristic data of each group of user login accounts according to the comparison result, and obtain the risk level label of each group of characteristic data corresponding to the login account information.
  • the acquiring unit 31 includes:
  • the calling module 311 can be used to call the identity data of the user's registered application account and the historical operation data of the user's login account information through a preset interface;
  • the extraction module 312 may be used to extract characteristic data of the user's login account based on the personal information of the user's registered application account and the historical operation data of the user's login account information.
  • the deep learning model is a convolutional neural network
  • the convolutional neural network includes a multi-layer structure
  • the construction unit 32 includes:
  • the extraction module 321 may be used to extract the characteristics of the characteristic data of the user login account through the convolutional layer of the deep learning model to obtain the characteristic parameters of the user login account at various risk levels;
  • the dimensionality reduction module 322 may be used to perform dimensionality reduction processing on the characteristic parameters of the user login account at each risk level through the pooling layer of the deep learning model, and obtain the user login account at each risk level after the dimensionality reduction processing.
  • Characteristic parameters
  • the summary module 323 may be used to summarize the characteristic parameters of the user login account at each risk level after the dimensionality reduction processing through the fully connected layer of the deep learning model, and obtain the weight of the characteristic data of the user login account at different risk levels value;
  • the construction module 324 may be used to generate the difference between the characteristic data of the user login account and the risk level of the login account information according to the weight values of the characteristic data of the user login account at different risk levels through the classification layer of the deep learning model. Mapping relationships to build a login risk identification model.
  • the risk level of the login account includes a low risk level, an intermediate risk level, and a high risk level
  • the login unit 34 includes:
  • the first login module 341 can be used to log in the account information when the risk level of the login account is a low-level risk level, and then select a regular log-in verification or non-secret verification login method that matches the low-level risk level;
  • the second login module 342 can be used for when the risk level of the login account is an intermediate risk level, select a dynamic password authentication login method that matches the intermediate risk level to log in account information;
  • the third login module 343 may be used for when the risk level of the login account is an advanced risk level, select a face verification login method that matches the advanced risk level to log in account information.
  • this embodiment also provides a non-volatile readable storage medium on which a computer readable instruction is stored, and when the readable instruction is executed by a processor
  • the login method of account information as shown in Figure 1 and Figure 2 is realized.
  • the technical solution of this application can be embodied in the form of a software product, which can be stored in a non-volatile readable storage medium (which can be a CD-ROM, U disk, mobile hard disk, etc.), It includes several instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each implementation scenario of this application.
  • a non-volatile readable storage medium which can be a CD-ROM, U disk, mobile hard disk, etc.
  • a computer device which can be a personal computer, a server, or a network device, etc.
  • an embodiment of the present application also provides a computer device, which may be a personal computer, Server, network device, etc.
  • the physical device includes a nonvolatile readable storage medium and a processor; a nonvolatile readable storage medium for storing computer readable instructions; a processor for executing computer readable instructions to The login method of account information as shown in Figure 1 and Figure 2 is realized.
  • the computer device may also include a user interface, a network interface, a camera, a radio frequency (RF) circuit, a sensor, an audio circuit, a WI-FI module, and so on.
  • the user interface may include a display screen (Display), an input unit such as a keyboard (Keyboard), etc., and the optional user interface may also include a USB interface, a card reader interface, and the like.
  • the network interface can optionally include a standard wired interface, a wireless interface (such as a Bluetooth interface, a WI-FI interface), etc.
  • the physical device structure of the device for logging in account information does not constitute a limitation on the physical device, and may include more or fewer components, or a combination of certain components, or different components. Component arrangement.
  • the non-volatile readable storage medium may also include an operating system and a network communication module.
  • the operating system is a program that manages the hardware and software resources of the above-mentioned computer equipment, and supports the operation of information processing programs and other software and/or programs.
  • the network communication module is used to implement communication between various components in the non-volatile readable storage medium and communication with other hardware and software in the physical device.
  • this application can be implemented by means of software plus a necessary general hardware platform, or by hardware.
  • this application uses a login risk identification model to identify the risk level of login account information, and flexibly recommends users to use different account login methods based on the risk level of account information.
  • the account information is logged in, while ensuring the security of the account information login process, so as to ensure the security of the user account during the account information login process, so that the user does not need to manually modify the login method, and improves the account information login process. flexibility.

Landscapes

  • Engineering & Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Computing Systems (AREA)
  • Computer Hardware Design (AREA)
  • Computer Security & Cryptography (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Health & Medical Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Artificial Intelligence (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Molecular Biology (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Information Transfer Between Computers (AREA)

Abstract

本申请公开了一种账户信息的登录方法、装置及计算机非易失性可读存储介质，涉及信息技术领域，可以基于账户信息的风险级别灵活推荐用户使用不同的账户登录方式，从而在账户信息的登录过程中保证用户账户的安全性。所述方法包括：获取多组用户登录账户的特征数据；将所述多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型；当接收到账户信息的登陆请求时，将所述登录请求对应用户登录账户的特征数据输入至所述登录风险识别模型，识别得到登录账户信息的风险级别；根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录。

Description

账户信息的登录方法、装置、计算机设备及计算机存储介质
本申请要求于2019年4月19日提交中国专利局、申请号为201910319436.X、申请名称为“账户信息的登录方法、装置、计算机设备及计算机存储介质”的中国专利申请的优先权，其全部内容通过引用结合在申请中。
技术领域
本申请涉及信息技术领域，尤其是涉及到账户信息的登录方法、装置、计算机设备及计算机非易失性可读存储介质。
背景技术
随着智能手机和ipad等移动终端设备的普及，人们逐渐习惯了使用APP客户端上网的方式，人类的社交活动渐渐地由传统的登门拜访、聚会派对演变为互联网上的虚拟活动，日常购物由传统的商场选购演变为互联网上的购物，例如，用户可以通过论坛、微博、网络游戏等与他人进行互动，还可以通过淘宝、京东商城选购日常用品，而账户就是用户在每个APP中进行相应活动的身份。
在现有的账户信息在登录过程中，系统通常会存在多种登录方式，例如，用户在首次进入支付宝应用时，需要输入用户名和密码的登录方式，由于支付宝客户端采用了常登陆的登录方式，用户再次进入支付宝应用时，可以直接进入支付宝界面，无需经过验证，此外，用户在使用手机银行进行过转账或者支付时，需要通过客户端的动态口令认证的登录方式，而58同城和赶集网的客户端会采用人脸验证的登录方式。
但在实际操作过程中，APP客户端推荐的登录方式不固定，灵活度相对较低，可能存在以下的情况，当用户账号的风险级别不高时，推荐人脸验证这种比较复杂的登录方式，使得登录操作变得繁琐，而当用户账号的风险级别较高时，推荐长登验证这种比较简单的登录方式，无法保证用户账户的安全性。
发明内容
有鉴于此，本申请提供了一种账户信息的登录方法、装置、计算机设备及计算机非易失性可读存储介质，主要目的在于解决目前用户在账户信息的登录过程中所推荐的账户登录方式灵活度较低的问题。
依据本申请一个方面，提供了一种账户信息的登录方法，该方法包括：
获取多组用户登录账户的特征数据，所述特征数据中携带有适用于登录账户信息的风 险级别标签；
将所述多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型，所述登录风险识别模型记录有用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系；
当接收到账户信息的登陆请求时，将所述登录请求对应用户登录账户的特征数据输入至所述登录风险识别模型，识别得到登录账户信息的风险级别；
根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录。
依据本申请另一个方面，提供了一种账户信息的登录装置，所述装置包括：
获取单元，用于获取多组用户登录账户的特征数据，所述特征数据中携带有适用于登录账户信息的风险级别标签；
构建单元，用于将所述多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型，所述登录风险识别模型记录有用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系；
识别单元，用于当接收到账户信息的登陆请求时，将所述登录请求对应用户登录账户的特征数据输入至所述登录风险识别模型，识别得到登录账户信息的风险级别；
登录单元，用于根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录。
根据本申请实施例的第三方面，提供一种计算机非易失性可读存储介质，其上存储有计算机可读指令，该计算机可读指令被处理器执行时实现以下步骤：获取多组用户登录账户的特征数据，所述特征数据中携带有适用于登录账户信息的风险级别标签；将所述多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型，所述登录风险识别模型记录有用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系；当接收到账户信息的登陆请求时，将所述登录请求对应用户登录账户的特征数据输入至所述登录风险识别模型，识别得到登录账户信息的风险级别；根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录。
根据本申请实施例的第四方面，提供一种计算机设备，包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机可读指令，所述处理器执行所述计算机可读指令时实现以下步骤：获取多组用户登录账户的特征数据，所述特征数据中携带有适用于登录账户信息的风险级别标签；将所述多组用户登录账户的特征数据作为训练数据输入至深度学 习模型中进行训练，构建登录风险识别模型，所述登录风险识别模型记录有用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系；当接收到账户信息的登陆请求时，将所述登录请求对应用户登录账户的特征数据输入至所述登录风险识别模型，识别得到登录账户信息的风险级别；根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录。
借由上述技术方案，本申请提供一种账户信息的登录方法及装置，通过收集多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型，当接收到账户信息的登录请求时，通过登录风险识别模型识别登录账户信息的风险级别，进而选取与登录账户的风险级别相匹配的账户信息的登录方式，对账户信息进行登录，提高账户信息登录的灵活性。与现有技术中采用APP客户端推荐的登录方式实现账户信息的登录方法相比，本申请通过采用登录风险识别模型来识别登录账户信息的风险级别，基于账户信息的风险级别灵活推荐用户使用不同的账户登录方式，从而对账户信息进行登录，在保证账户信息登录过程的安全性的同时，从而在账户信息的登录过程中保证用户账户的安全性，使得用户无需手动操作去修改登录方式，提高了账户信息登录过程的灵活性。
附图说明
通过阅读下文优选实施方式的详细描述，各种其他的优点和益处对于本领域普通技术人员将变得清楚明了。附图仅用于示出优选实施方式的目的，而并不认为是对本申请的限制。而且在整个附图中，用相同的参考符号表示相同的部件。在附图中：
图1示出了本申请实施例提供的一种账户信息的登录方法流程示意图；
图2示出了本申请实施例提供的另一种账户信息的登录方法流程示意图；
图3示出了本申请实施例提供的一种账户信息的登录装置的结构示意图；
图4示出了本申请实施例提供的另一种账户信息的登录装置的结构示意图。
具体实施方式
下面将参照附图更详细地描述本公开的示例性实施例。虽然附图中显示了本公开的示例性实施例，然而应当理解，可以以各种形式实现本公开而不应被这里阐述的实施例所限制。相反，提供这些实施例是为了能够更透彻地理解本公开，并且能够将本公开的范围完整的传达给本领域的技术人员。
本申请实施例提供了一种账户信息的登录方法，可以基于账户信息的风险级别灵活推 荐用户使用不同的账户登录方式，提高账户信息登录过程的灵活性，如图1所示，该方法包括：
101、获取多组用户登录账户的特征数据。
其中，用户登录账户的特征数据包括但不限于用户登录时间、操作地点、设备类型、账号是否可交易、用户风险承受能力、身份信息、账户资产额度等特征数据，具体当用户首次登录应用服务的账户时，都会通过个人身份信息进行注册，例如，账户名称、电话号码、性别等，以及会对应用服务的操作进行设置，例如，支付扣款顺序、登录方式、登录状态设置等。
可以理解的是，为了保证账户安全，用户在注册账户信息的过程中，都会设置账户名称以及账户登录密码等账户信息，以便于用户后续通过设置的账户信息再次登录应用服务。针对不同的应用服务的特征，适用于登录账户的特征数据的风险级别有所不同，例如，针对涉及财务类的应用服务风在使用过程中风险较高，通常在登录和交易时会设置安全验证，如登录账户密码，交易密码等，针对涉及社交类的应用服务在使用过程中涉及交易认证时风险较高，通常在登录后设置长登录状态，交易时设置安全验证，针对购物类的应用服务在使用过程中涉及交易认证时风险较高，通常在登录时不会设置安全验证，交易时设置安全验证。所以，获取的特征数据中携带有适用于登录账户信息的风险级别标签。
102、将所述多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型。
其中，登录风险识别模型记录有用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系，由于深度学习模型具有训练数据之间映射关系的作用，通过构建登录风险识别模型，对于不同用户登录账户的特征数据，通过登录风险识别模型训练得到的映射关系可以获取与用户登录账户的特征数据相对应的登录账户信息的风险级别。
由于特征数据中携带有适用于登录账户信息的风险级别标签，例如，可以将风险级别标签设置为三个等级，由高至低分别为等级A、等级B、等级C，当然还可以设置更多等级，具体地，对于涉及大额交易的特征数据设置等级A，对于个人信息修改的特征数据设置等级A，对于涉及中等金额交易的特征数据设置为等级B等等，还可以对每个等级设置特征数据的范围，这里不进行限定，从而根据适用于登录账户信息的风险等级对多组用户登录账户的特征数据进行训练。
对于本申请实施例，这里深度学习模型可以为卷积神经网络模型，具体可以通过反复训练多组用户登录账户的特征数据构建登录风险识别模型的网络结构，该网络结构可以对输入的数据进行训练，并给出正确的输入-输出关系，相当于用户登录账户的特征数据与 登录账户信息的风险级别之间的映射关系。
具体卷积神经网络模型的结构可以通过卷积层、全连接层、池化层以及分类层结构实现，这里的卷积层相当于卷积神经网络的隐含层，可以为多层结构，用于提取更深层次的用户登录账户在各个风险级别上的特征参数；在卷积神经网络模型中，为了减小参数，减低计算，常常在连续卷积层中间隔插入池化层；这里的全连接层与卷积层相似，卷积层的神经元和上一层输出局部区域相连，当然为了减少输出特征向量过多，可以设置两个全连接层，在训练数据通过若干个卷积层训练后对训练输出的特征数据进行整合。
103、当接收到账户信息的登陆请求时，将所述登录请求对应用户登录账户的特征数据输入至所述登录风险识别模型，识别得到登录账户信息的风险级别。
其中，登录请求对应用户登录账户的特征数据为未标记风险级别的特征数据，通过登录风险识别模型可以识别出与特征数据相映射的登录账户信息的风险级别。
例如，设置登录账户信息的风险级别可以包括低级风险、中级风险和高级风险，如果输入至登录风险级别的特征数据涉及的交易金额大于10元，则识别得到登录账户信息的风险级别为中级风险。
104、根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录。
其中，登录账户信息的风险级别可以包括但不局限于人脸验证、身份验证、动态口令验证、验证码登录以及无密登录等。
对于本申请实施例，由于登录账户信息的风险级别说明账户信息在登录过程中存在的安全风险，对应登录账户信息的风险级别较低，说明账户信息安全风险较低，可以选取用户无密登录或其他无需验证的登录方式，对于登录账户的风险级别中等，说明账户信息还存在一定的安全风险，可以选取动态口令的验证的登录方式，对于登录账户的风险级别较高，说明账户信息安全风险较高，可以选取人脸验证或身份验证的登录方式。
本申请实施例提供一种账户信息的登录方法，通过收集多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型，当接收到账户信息的登录请求时，通过登录风险识别模型识别登录账户信息的风险级别，进而选取与登录账户的风险级别相匹配的账户信息的登录方式，对账户信息进行登录，提高账户信息登录的灵活性。与现有技术中采用APP客户端推荐的登录方式实现账户信息的登录方法相比，本申请通过采用登录风险识别模型来识别登录账户信息的风险级别，基于账户信息的风险级别灵活推荐用户使用不同的账户登录方式，从而对账户信息进行登录，在保证账户信息登录过程的安全性的同时，从而在账户信息的登录过程中保证用户账户的安全性，使得用 户无需手动操作去修改登录方式，提高了账户信息登录过程的灵活性。
本申请实施例提供了另一种账户信息的登录方法，可以基于用户的风险级别提供不同的账户登录方式，保证在账户登录过程中用户账户信息的安全性，如图2所示，所述方法包括：
201、通过预设接口调用用户注册应用账户的身份数据以及用户登录账户信息的历史操作数据。
其中，用户注册账户的身份数据可以包括但不局限于用户名称、性别、联系电话、设备IP地址等可以表明用户身份的数据，用户登录账户的历史操作数据可以包括但不局限于用户交易数据、用户浏览数据等可以表明用户从操作习惯的数据。通过预设接口调用用户注册应用账户的身份数据可以获取用户个人身份信息数据，通过预设接口调用用户登录账户信息的历史操作数据可以获取用户操作习惯数据。
202、根据所述用户注册应用账户的个人信息以及用户登录账户信息的历史操作数据，提取用户登录账户的特征数据。
对于本申请实施例，用户注册应用账户后，为了便于操作，针对应用账户的特性都会设置用于登录应用账户的多个特征数据，该特征数据相当于用户注册应用账户的个人信息以及用户登录账户信息的历史操作数据，例如，针对涉及财务类的网络服务，通常在登录和交易时会设置安全验证，如登录账户密码，交易密码等，针对涉及社交类的网络服务，通常在登录后设置长登录状态，交易时设置安全验证，针对购物类的网络服务，通常在登录时不会设置安全验证，交易时设置安全验证，进一步从用户注册应用账户的个人信息以及用户登录账户信息的历史操作数据中可以提取用户登录账户的特征数据。
203、根据每组用户登录账户的特征数据所处的数据范围，对每组用户登录账户的特征数据进行风险等级标记，获取每组特征数据对应登录账户信息的风险级别标签。
对于本申请实施例，由于用户登录账户的特征数据都会涉及不同的数据范围，例如，对于特征数据涉及用户的个人信息数据被更改，可能导致账户信息存在安全风险，而个人信息数据中涉及多个信息字段，并不是每个信息字段被更改都会使得账户信息存在安全风险，如果被更改的个人信息数据处于预先设置的危险信息字段范围内，才会导致账户信息存在安全风险，同理，对于特征数据涉及账户金额变更，都会导致账户信息存在安全风险，而变更金额数值的数据范围不同会存在不同程度的安全风险。所以，针对用户登录账户的特征数据所处的数据范围，可以对特征数据进行风险等级标记，获取特征数据对应登录账户信息的风险级别标签。
204、将所述多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行 训练，构建登录风险识别模型。
具体在将多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练的过程中，可以将用户每次登录应用账户时的一组特征数据作为一个N维向量，每个特征数据表示为N维向量中的一个数值，并且通过对特征数据进行归一化处理后，使得每个特征数据的分布范围为[-1,1]，对于用户未设置的特征数据表示为数值0，从而得到标准格式的训练数据。
对于本申请实施例，深度学习模型可以为卷积神经网络，每层结构具有不同的输入输出参数以实现不同的功能，通过卷积神经网络对多组用户登录账户的特征数据进行训练，得到用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系，该映射关系相当于登录风险识别模型。
具体卷积神经网络包括多层结构，可以通过卷积神经网络模型的卷积层提取用户登录账户的特征数据的特征，得到用户登录账户在各个风险级别上的特征参数；通过卷积神经网络模型的池化层对用户登录账户在各个风险级别上的特征参数进行降维处理，得到降维处理后用户登录账户在各个风险级别上的特征参数；通过卷积神经网络模型的全连接层汇总降维处理后用户登录账户在各个风险级别上的特征参数，得到用户登录账户的特征数据在不同风险级别上的权重值；通过卷积神经网络模型的分类层根据用户登录账户的特征数据在不同风险级别上的权重值生成用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系，构建登录风险识别模型。
例如，卷积神经网络包括13个卷积层，3个全连接层，这里可以设置每个卷积层的卷积核个数分别为64、64、128、128、256、256、512、512、512、512、512、512，并且第2个卷积层与第3个卷积层之间、第4个卷积层与第5个卷积层之间、第6个卷积层与第7个卷积层之间、第8个卷积层与第9个卷积层之间、第10个卷积层与第11个卷积层之间、第13个卷积层与第1个全连接层之间，均连接1个池化层，并且上述13个卷积层和3个全连接层均用非线性激活函数进行处理，这里对上述卷积层、全连接层以及池化层的层数不进行限定，具体可以根据实际情况进行选取，同理，对于每层内所选择的激活函数也不进行限定。当输入多组用户登录账户的特征数据，输出用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系，相当于用户登录账户的特征数据在各个风险级别上的分类结果。
205、当接收到账户信息的登陆请求时，将所述登录请求对应用户登录账户的特征数据输入至所述登录风险识别模型，识别得到登录账户信息的风险级别。
对于本申请实施例，通过登录风险识别模型所识别得到的登录账户信息的风险级别只 是根据登录请求对应用户登录账户的特征数据来初步判断用户登录时存在的风险情况，进一步提高用户登录账户信息的安全性。
206、根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录。
需要说明的是，为了便于用户进行账户信息的登录，在识别得到登录账户信息的风险级别后，可以实时了解当前账户登录时存在的风险，如果风险级别较高，则说明登录账户存在的安全风险较高，需要经过用户验证登录，例如，人脸验证、短信验证等，如果风险级别较低，则说明登录账户存在的安全风险较低，无需经过用户验证登录，例如，无密登录，进一步选取与登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录，从而保证账户信息登录过程中的安全性。
207、若所述多组用户登录账户的特征数据产生变化，当再次接收到账户信息的登录请求时，根据再次识别得到的登录账户信息的风险级别调整账户信息的登录方式，对账户信息进行登录。
可以理解的是，用户登录账户信息的特征数据是不断发生变化的，随着用户登录账户信息的特征数据发生变化，若多组用户登录账户的特征数据产生变化，通过构建风险识别模型再次识别得到的登录账户信息的风险级别也会发生改变，进一步根据再次识别得到的登录账户信息的风险级别调整账户信息的登录方式，对账户信息进行登录。
例如，用户登录账户的特征数据中的登录时间表明用户距离上次登录时间超过一周，识别得到的用户登录账户的风险级别提升，或者用户登录账户的特征数据中的个人信息被更改了，识别得到的登录账户的风险级别提升，进一步根据提升后风险级别对应的登录方式，对账户信息进行登录。
进一步地，作为图1所述方法的具体实现，本申请实施例提供了一种账户信息的登录装置，如图3所示，所述装置包括：获取单元31、构建单元32、识别单元33、登录单元34。
获取单元31，可以用于获取多组用户登录账户的特征数据，所述特征数据中携带有适用于登录账户信息的风险级别标签；
构建单元32，可以用于将所述多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型，所述登录风险识别模型记录有用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系；
识别单元33，可以用于当接收到账户信息的登陆请求时，将所述登录请求对应用户登录账户的特征数据输入至所述登录风险识别模型，识别得到登录账户信息的风险级别；
登录单元34，可以用于根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录。
本申请提供的一种账户信息的登录装置，通过收集多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型，当接收到账户信息的登录请求时，通过登录风险识别模型识别登录账户信息的风险级别，进而选取与登录账户的风险级别相匹配的账户信息的登录方式，对账户信息进行登录，提高账户信息登录的灵活性。与现有技术中采用APP客户端推荐的登录方式实现账户信息的登录方法相比，本申请通过采用登录风险识别模型来识别登录账户信息的风险级别，基于账户信息的风险级别灵活推荐用户使用不同的账户登录方式，从而对账户信息进行登录，在保证账户信息登录过程的安全性的同时，从而在账户信息的登录过程中保证用户账户的安全性，使得用户无需手动操作去修改登录方式，提高了账户信息登录过程的灵活性。
作为图3中所示账户信息的登录装置的进一步说明，图4是根据本申请实施例另一种账户信息的登录装置的结构示意图，如图4所示，所述装置还包括：
标记单元35，可以用于在所述获取多组用户登录账户的特征数据之后，根据每组用户登录账户的特征数据所处的数据范围，对每组用户登录账户的特征数据进行风险等级标记，获取每组特征数据对应登录账户信息的风险级别标签；
调整单元36，可以用于在所述根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录之后，若所述多组用户登录账户的特征数据产生变化，当再次接收到账户信息的登录请求时，根据再次识别得到的登录账户信息的风险级别调整账户信息的登录方式，对账户信息进行登录。
进一步地，所述标记单元35包括：
比对模块351，可以用于提取每组用户登录账户的特征信息所处的特征范围，将所述每组用户登录账户的特征信息所处的特征范围与预先设置的各个风险等级对应的特征范围进行比对，得到比对结果；
标记模块352，可以用于根据所述比对结果对每组用户登录账户的特征数据进行风险等级标记，获取每组特征数据对应登录账户信息的风险级别标签。
进一步地，所述获取单元31包括：
调用模块311，可以用于通过预设接口调用用户注册应用账户的身份数据以及用户登录账户信息的历史操作数据；
提取模块312，可以用于根据所述用户注册应用账户的个人信息以及用户登录账户信息的历史操作数据，提取用户登录账户的特征数据。
进一步地，所述深度学习模型为卷积神经网络，所述卷积神经网络包括多层结构，所述构建单元32包括：
提取模块321，可以用于通过所述深度学习模型的卷积层提取所述用户登录账户的特征数据的特征，得到用户登录账户在各个风险级别上的特征参数；
降维模块322，可以用于通过所述深度学习模型的池化层对所述用户登录账户在各个风险级别上的特征参数进行降维处理，得到降维处理后用户登录账户在各个风险级别上的特征参数；
汇总模块323，可以用于通过所述深度学习模型的全连接层汇总所述降维处理后用户登录账户在各个风险级别上的特征参数，得到用户登录账户的特征数据在不同风险级别上的权重值；
构建模块324，可以用于通过所述深度学习模型的分类层根据所述用户登录账户的特征数据在不同风险级别上的权重值生成用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系，构建登录风险识别模型。
进一步地，所述登录账户的风险级别包括低级风险级别、中级风险级别和高级风险级别，所述登录单元34包括：
第一登录模块341，可以用于当所述登录账户的风险级别为低级风险级别，则选取与低级风险级别相匹配的常登验证或无密验证登录方式，对账户信息进行登录；
第二登录模块342，可以用于当所述登录账户的风险级别为中级风险级别，则选取与中级风险级别相匹配的动态口令认证登录方式，对账户信息进行登录；
第三登录模块343，可以用于当所述登录账户的风险级别为高级风险级别，则选取与高级风险级别相匹配的人脸验证登录方式，对账户信息进行登录。
需要说明的是，本实施例提供的一种账户信息的登录装置所涉及各功能单元的其他相应描述，可以参考图1和图2中的对应描述，在此不再赘述。
基于上述如图1和图2所示方法，相应的，本实施例还提供了一种非易失性可读存储介质，其上存储有计算机可读指令，该可读指令被处理器执行时实现上述如图1和图2所示的账户信息的登录方法。
基于这样的理解，本申请的技术方案可以以软件产品的形式体现出来，该软件产品可以存储在一个非易失性可读存储介质(可以是CD-ROM，U盘，移动硬盘等)中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施场景所述的方法。
基于上述如图1、图2所示的方法，以及图3、图4所示的虚拟装置实施例，为了实 现上述目的，本申请实施例还提供了一种计算机设备，具体可以为个人计算机、服务器、网络设备等，该实体设备包括非易失性可读存储介质和处理器；非易失性可读存储介质，用于存储计算机可读指令；处理器，用于执行计算机可读指令以实现上述如图1和图2所示的账户信息的登录方法。
可选地，该计算机设备还可以包括用户接口、网络接口、摄像头、射频(Radio Frequency，RF)电路，传感器、音频电路、WI-FI模块等等。用户接口可以包括显示屏(Display)、输入单元比如键盘(Keyboard)等，可选用户接口还可以包括USB接口、读卡器接口等。网络接口可选的可以包括标准的有线接口、无线接口(如蓝牙接口、WI-FI接口)等。
本领域技术人员可以理解，本实施例提供的账户信息的登录装置的实体设备结构并不构成对该实体设备的限定，可以包括更多或更少的部件，或者组合某些部件，或者不同的部件布置。
非易失性可读存储介质中还可以包括操作系统、网络通信模块。操作系统是管理上述计算机设备硬件和软件资源的程序，支持信息处理程序以及其它软件和/或程序的运行。网络通信模块用于实现非易失性可读存储介质内部各组件之间的通信，以及与该实体设备中其它硬件和软件之间通信。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到本申请可以借助软件加必要的通用硬件平台的方式来实现，也可以通过硬件实现。通过应用本申请的技术方案，与目前现有技术相比，本申请通过采用登录风险识别模型来识别登录账户信息的风险级别，基于账户信息的风险级别灵活推荐用户使用不同的账户登录方式，从而对账户信息进行登录，在保证账户信息登录过程的安全性的同时，从而在账户信息的登录过程中保证用户账户的安全性，使得用户无需手动操作去修改登录方式，提高了账户信息登录过程的灵活性。
本领域技术人员可以理解附图只是一个优选实施场景的示意图，附图中的模块或流程并不一定是实施本申请所必须的。本领域技术人员可以理解实施场景中的装置中的模块可以按照实施场景描述进行分布于实施场景的装置中，也可以进行相应变化位于不同于本实施场景的一个或多个装置中。上述实施场景的模块可以合并为一个模块，也可以进一步拆分成多个子模块。
上述本申请序号仅仅为了描述，不代表实施场景的优劣。以上公开的仅为本申请的几个具体实施场景，但是，本申请并非局限于此，任何本领域的技术人员能思之的变化都应落入本申请的保护范围。

Claims (20)

  1. 一种账户信息的登录方法，其特征在于，所述方法包括：
    获取多组用户登录账户的特征数据，所述特征数据中携带有适用于登录账户信息的风险级别标签；
    将所述多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型，所述登录风险识别模型记录有用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系；
    当接收到账户信息的登陆请求时，将所述登录请求对应用户登录账户的特征数据输入至所述登录风险识别模型，识别得到登录账户信息的风险级别；
    根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录。
  2. 根据权利要求1所述的方法，其特征在于，所述深度学习模型为卷积神经网络，所述卷积神经网络包括多层结构，所述将所述多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型包括：
    通过所述深度学习模型的卷积层提取所述用户登录账户的特征数据的特征，得到用户登录账户在各个风险级别上的特征参数；
    通过所述深度学习模型的池化层对所述用户登录账户在各个风险级别上的特征参数进行降维处理，得到降维处理后用户登录账户在各个风险级别上的特征参数；
    通过所述深度学习模型的全连接层汇总所述降维处理后用户登录账户在各个风险级别上的特征参数，得到用户登录账户的特征数据在不同风险级别上的权重值；
    通过所述深度学习模型的分类层根据所述用户登录账户的特征数据在不同风险级别上的权重值生成用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系，构建登录风险识别模型。
  3. 根据权利要求1所述的方法，其特征在于，所述获取多组用户登录账户的特征数据包括：
    通过预设接口调用用户注册应用账户的身份数据以及用户登录账户信息的历史操作数据；
    根据所述用户注册应用账户的个人信息以及用户登录账户信息的历史操作数据，提取用户登录账户的特征数据。
  4. 根据权利要求1所述的方法，其特征在于，在所述获取多组用户登录账户的特征数据之后，所述方法还包括：
    根据每组用户登录账户的特征数据所处的数据范围，对每组用户登录账户的特征数据进行风险等级标记，获取每组特征数据对应登录账户信息的风险级别标签。
  5. 根据权利要求4所述的方法，其特征在于，所述根据每组用户登录账户的特征数据所处的数据范围，对每组用户登录账户的特征数据进行风险等级标记，获取每组特征数据对应登录账户信息的风险级别标签包括：
    提取每组用户登录账户的特征信息所处的特征范围，将所述每组用户登录账户的特征信息所处的特征范围与预先设置的各个风险等级对应的特征范围进行比对，得到比对结果；
    根据所述比对结果对每组用户登录账户的特征数据进行风险等级标记，获取每组特征数据对应登录账户信息的风险级别标签。
  6. 根据权利要求1所述的方法，其特征在于，所述登录账户的风险级别包括低级风险级别、中级风险级别和高级风险级别，所述根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录包括：
    当所述登录账户的风险级别为低级风险级别，则选取与低级风险级别相匹配的常登验证或无密验证登录方式，对账户信息进行登录；
    当所述登录账户的风险级别为中级风险级别，则选取与中级风险级别相匹配的动态口令认证登录方式，对账户信息进行登录；
    当所述登录账户的风险级别为高级风险级别，则选取与高级风险级别相匹配的人脸验证登录方式，对账户信息进行登录。
  7. 根据权利1-6中任一项所述的方法，其特征在于，在所述根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录之后，所述方法还包括：
    若所述多组用户登录账户的特征数据产生变化，当再次接收到账户信息的登录请求时，根据再次识别得到的登录账户信息的风险级别调整账户信息的登录方式，对账户信息进行登录。
  8. 一种账户信息的登录装置，其特征在于，所述装置包括：
    获取单元，用于获取多组用户登录账户的特征数据，所述特征数据中携带有适用于登录账户信息的风险级别标签；
    构建单元，用于将所述多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型，所述登录风险识别模型记录有用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系；
    识别单元，用于当接收到账户信息的登陆请求时，将所述登录请求对应用户登录账户的特征数据输入至所述登录风险识别模型，识别得到登录账户信息的风险级别；
    登录单元，用于根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录。
  9. 根据权利要求8所述的装置，其特征在于，所述深度学习模型为卷积神经网络，所述卷积神经网络包括多层结构，所述构建单元包括：
    提取模块，用于通过所述深度学习模型的卷积层提取所述用户登录账户的特征数据的特征，得到用户登录账户在各个风险级别上的特征参数；
    降维模块，用于通过所述深度学习模型的池化层对所述用户登录账户在各个风险级别上的特征参数进行降维处理，得到降维处理后用户登录账户在各个风险级别上的特征参数；
    汇总模块，用于通过所述深度学习模型的全连接层汇总所述降维处理后用户登录账户在各个风险级别上的特征参数，得到用户登录账户的特征数据在不同风险级别上的权重值；
    构建模块，用于通过所述深度学习模型的分类层根据所述用户登录账户的特征数据在不同风险级别上的权重值生成用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系，构建登录风险识别模型。
  10. 根据权利要求8所述的装置，其特征在于，所述获取单元包括：
    调用模块，用于通过预设接口调用用户注册应用账户的身份数据以及用户登录账户信息的历史操作数据；
    提取模块，用于根据所述用户注册应用账户的个人信息以及用户登录账户信息的历史操作数据，提取用户登录账户的特征数据。
  11. 根据权利要求8所述的装置，其特征在于，所述装置还包括：
    标记单元，用于在所述获取多组用户登录账户的特征数据之后，根据每组用户登录账户的特征数据所处的数据范围，对每组用户登录账户的特征数据进行风险等级标记，获取每组特征数据对应登录账户信息的风险级别标签。
  12. 根据权利要求11所述的装置，其特征在于，所述标记单元包括：
    比对模块，用于提取每组用户登录账户的特征信息所处的特征范围，将所述每组用户登录账户的特征信息所处的特征范围与预先设置的各个风险等级对应的特征范围进行比对，得到比对结果；
    标记模块，用于根据所述比对结果对每组用户登录账户的特征数据进行风险等级标记，获取每组特征数据对应登录账户信息的风险级别标签。
  13. 根据权利要求8所述的装置，其特征在于，所述登录账户的风险级别包括低级风险级别、中级风险级别和高级风险级别，所述登录单元包括：
    第一登录模块，用于当所述登录账户的风险级别为低级风险级别，则选取与低级风险级别相匹配的常登验证或无密验证登录方式，对账户信息进行登录；
    第二登录模块，用于当所述登录账户的风险级别为中级风险级别，则选取与中级风险级别 相匹配的动态口令认证登录方式，对账户信息进行登录；
    第三登录模块，用于当所述登录账户的风险级别为高级风险级别，则选取与高级风险级别相匹配的人脸验证登录方式，对账户信息进行登录。
  14. 根据权利要求8-13中任一项所述的装置，其特征在于，所述装置还包括：
    调整单元，用于在所述根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录之后，若所述多组用户登录账户的特征数据产生变化，当再次接收到账户信息的登录请求时，根据再次识别得到的登录账户信息的风险级别调整账户信息的登录方式，对账户信息进行登录。
  15. 一种计算机非易失性可读存储介质，其上存储有计算机可读指令，其特征在于，所述计算机可读指令被处理器执行时实现账户信息的登录方法，包括：
    获取多组用户登录账户的特征数据，所述特征数据中携带有适用于登录账户信息的风险级别标签；将所述多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型，所述登录风险识别模型记录有用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系；当接收到账户信息的登陆请求时，将所述登录请求对应用户登录账户的特征数据输入至所述登录风险识别模型，识别得到登录账户信息的风险级别；根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录。
  16. 根据权利要求15所述的计算机非易失性可读存储介质，其特征在于，所述计算机可读指令被处理器执行时实现所述深度学习模型为卷积神经网络，所述卷积神经网络包括多层结构，所述将所述多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型包括：
    通过所述深度学习模型的卷积层提取所述用户登录账户的特征数据的特征，得到用户登录账户在各个风险级别上的特征参数；通过所述深度学习模型的池化层对所述用户登录账户在各个风险级别上的特征参数进行降维处理，得到降维处理后用户登录账户在各个风险级别上的特征参数；通过所述深度学习模型的全连接层汇总所述降维处理后用户登录账户在各个风险级别上的特征参数，得到用户登录账户的特征数据在不同风险级别上的权重值；通过所述深度学习模型的分类层根据所述用户登录账户的特征数据在不同风险级别上的权重值生成用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系，构建登录风险识别模型。
  17. 根据权利要求15所述的计算机非易失性可读存储介质，其特征在于，所述计算机可读指令被处理器执行时实现所述获取多组用户登录账户的特征数据包括：
    通过预设接口调用用户注册应用账户的身份数据以及用户登录账户信息的历史操作数据； 根据所述用户注册应用账户的个人信息以及用户登录账户信息的历史操作数据，提取用户登录账户的特征数据。
  18. 一种计算机设备，包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机可读指令，其特征在于，所述处理器执行所述计算机可读指令时实现账户信息的登录方法，包括：
    获取多组用户登录账户的特征数据，所述特征数据中携带有适用于登录账户信息的风险级别标签；将所述多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型，所述登录风险识别模型记录有用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系；当接收到账户信息的登陆请求时，将所述登录请求对应用户登录账户的特征数据输入至所述登录风险识别模型，识别得到登录账户信息的风险级别；根据所述登录账户信息的风险级别，选取与所述登录账户信息的风险级别相匹配的账户信息的登录方式，对账户信息进行登录。
  19. 根据权利要求18所述的计算机设备，其特征在于，所述处理器执行所述计算机可读指令时实现所述深度学习模型为卷积神经网络，所述卷积神经网络包括多层结构，所述将所述多组用户登录账户的特征数据作为训练数据输入至深度学习模型中进行训练，构建登录风险识别模型包括：
    通过所述深度学习模型的卷积层提取所述用户登录账户的特征数据的特征，得到用户登录账户在各个风险级别上的特征参数；通过所述深度学习模型的池化层对所述用户登录账户在各个风险级别上的特征参数进行降维处理，得到降维处理后用户登录账户在各个风险级别上的特征参数；通过所述深度学习模型的全连接层汇总所述降维处理后用户登录账户在各个风险级别上的特征参数，得到用户登录账户的特征数据在不同风险级别上的权重值；通过所述深度学习模型的分类层根据所述用户登录账户的特征数据在不同风险级别上的权重值生成用户登录账户的特征数据与登录账户信息的风险级别之间的映射关系，构建登录风险识别模型。
  20. 根据权利要求18所述的计算机设备，其特征在于，所述处理器执行所述计算机可读指令时实现所述获取多组用户登录账户的特征数据包括：
    通过预设接口调用用户注册应用账户的身份数据以及用户登录账户信息的历史操作数据；根据所述用户注册应用账户的个人信息以及用户登录账户信息的历史操作数据，提取用户登录账户的特征数据。
PCT/CN2019/103197 2019-04-19 2019-08-29 账户信息的登录方法、装置、计算机设备及计算机存储介质 WO2020211247A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910319436.X 2019-04-19
CN201910319436.XA CN110166438B (zh) 2019-04-19 2019-04-19 账户信息的登录方法、装置、计算机设备及计算机存储介质

Publications (1)

Publication Number Publication Date
WO2020211247A1 true WO2020211247A1 (zh) 2020-10-22

Family

ID=67639663

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/103197 WO2020211247A1 (zh) 2019-04-19 2019-08-29 账户信息的登录方法、装置、计算机设备及计算机存储介质

Country Status (2)

Country Link
CN (1) CN110166438B (zh)
WO (1) WO2020211247A1 (zh)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112487296A (zh) * 2020-12-07 2021-03-12 南京酷沃智行科技有限公司 基于账号系统的个性化智能车机系统、实现方法及车辆
CN112581259A (zh) * 2020-12-16 2021-03-30 同盾控股有限公司 账户风险识别方法及装置、存储介质、电子设备
CN113468510A (zh) * 2021-07-15 2021-10-01 中国银行股份有限公司 异常登录行为数据检测方法及装置
CN115065512A (zh) * 2022-05-31 2022-09-16 北京奇艺世纪科技有限公司 一种账号登录方法、系统、装置、电子设备以及存储介质
CN116257840A (zh) * 2022-12-28 2023-06-13 南京邮电大学盐城大数据研究院有限公司 一种基于大数据的登录信息查询管理系统及方法

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110166438B (zh) * 2019-04-19 2022-03-18 平安科技（深圳）有限公司 账户信息的登录方法、装置、计算机设备及计算机存储介质
CN110647738B (zh) * 2019-09-29 2021-09-03 武汉极意网络科技有限公司 业务风控适配方法、装置、设备及存储介质
CN111447221B (zh) * 2020-03-26 2022-07-19 支付宝(杭州)信息技术有限公司 使用生物特征核实身份的方法和系统
CN112165379B (zh) * 2020-09-28 2022-08-05 武汉虹信技术服务有限责任公司 一种用户安全登录方法、装置及终端设备
CN113194079B (zh) * 2021-04-23 2022-09-09 平安科技（深圳）有限公司 登录验证方法、装置、设备及存储介质
CN113779521B (zh) * 2021-09-09 2024-05-24 北京安天网络安全技术有限公司 身份认证方法、装置、存储介质及电子设备
CN117521042B (zh) * 2024-01-05 2024-05-14 创旗技术有限公司 基于集成学习的高危授权用户识别方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105303442A (zh) * 2015-11-04 2016-02-03 中国民生银行股份有限公司 网上银行开户账号检测方法和装置
US20160088000A1 (en) * 2014-09-18 2016-03-24 Microsoft Corporation Lateral movement detection
CN108269187A (zh) * 2018-01-29 2018-07-10 深圳壹账通智能科技有限公司 金融业务的验证方法、装置、设备和计算机存储介质
CN109753783A (zh) * 2018-11-28 2019-05-14 北京友信科技有限公司 一种基于机器学习的单点登录方法、装置及计算机可读存储介质
CN110166438A (zh) * 2019-04-19 2019-08-23 平安科技（深圳）有限公司 账户信息的登录方法、装置、计算机设备及计算机存储介质

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103685169B (zh) * 2012-09-06 2019-02-12 盛趣信息技术（上海）有限公司 登录方法及系统
US20140282978A1 (en) * 2013-03-15 2014-09-18 Sergio Demian LERNER Method and apparatus for secure interaction with a computer service provider
CN104093141B (zh) * 2014-06-27 2016-09-28 北京奇虎科技有限公司 终端应用的登录方法、装置、客户端及电子设备
CN105915490A (zh) * 2015-11-18 2016-08-31 乐视致新电子科技（天津）有限公司 一种安全问题交互方法和系统
CN107864112B (zh) * 2016-09-28 2021-01-26 平安科技（深圳）有限公司 登录安全验证方法和装置
CN107239680B (zh) * 2017-05-22 2019-09-06 微梦创科网络科技（中国）有限公司 一种对用户登录进行风险评估的方法及装置
CN108650226B (zh) * 2018-03-30 2019-10-29 平安科技（深圳）有限公司 一种登录验证方法、装置、终端设备及存储介质
CN108551443B (zh) * 2018-03-30 2021-07-23 平安科技（深圳）有限公司 一种应用登录方法、装置、终端设备及存储介质
CN108537028A (zh) * 2018-04-17 2018-09-14 西安电子科技大学 一种计算机身份识别系统及方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160088000A1 (en) * 2014-09-18 2016-03-24 Microsoft Corporation Lateral movement detection
CN105303442A (zh) * 2015-11-04 2016-02-03 中国民生银行股份有限公司 网上银行开户账号检测方法和装置
CN108269187A (zh) * 2018-01-29 2018-07-10 深圳壹账通智能科技有限公司 金融业务的验证方法、装置、设备和计算机存储介质
CN109753783A (zh) * 2018-11-28 2019-05-14 北京友信科技有限公司 一种基于机器学习的单点登录方法、装置及计算机可读存储介质
CN110166438A (zh) * 2019-04-19 2019-08-23 平安科技（深圳）有限公司 账户信息的登录方法、装置、计算机设备及计算机存储介质

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112487296A (zh) * 2020-12-07 2021-03-12 南京酷沃智行科技有限公司 基于账号系统的个性化智能车机系统、实现方法及车辆
CN112581259A (zh) * 2020-12-16 2021-03-30 同盾控股有限公司 账户风险识别方法及装置、存储介质、电子设备
CN112581259B (zh) * 2020-12-16 2023-09-19 同盾控股有限公司 账户风险识别方法及装置、存储介质、电子设备
CN113468510A (zh) * 2021-07-15 2021-10-01 中国银行股份有限公司 异常登录行为数据检测方法及装置
CN115065512A (zh) * 2022-05-31 2022-09-16 北京奇艺世纪科技有限公司 一种账号登录方法、系统、装置、电子设备以及存储介质
CN115065512B (zh) * 2022-05-31 2024-03-15 北京奇艺世纪科技有限公司 一种账号登录方法、系统、装置、电子设备以及存储介质
CN116257840A (zh) * 2022-12-28 2023-06-13 南京邮电大学盐城大数据研究院有限公司 一种基于大数据的登录信息查询管理系统及方法
CN116257840B (zh) * 2022-12-28 2023-10-20 南京邮电大学盐城大数据研究院有限公司 一种基于大数据的登录信息查询管理系统及方法

Also Published As

Publication number Publication date
CN110166438B (zh) 2022-03-18
CN110166438A (zh) 2019-08-23

Similar Documents

Publication Publication Date Title
WO2020211247A1 (zh) 账户信息的登录方法、装置、计算机设备及计算机存储介质
US11138300B2 (en) Multi-factor profile and security fingerprint analysis
US11588824B2 (en) Systems and methods for proximity identity verification
US11068575B2 (en) Authentication system
US10762508B2 (en) Detecting fraudulent mobile payments
US10659439B2 (en) Device identification scoring
EP2748781B1 (en) Multi-factor identity fingerprinting with user behavior
CN103856472B (zh) 一种账户登录的方法及装置
CN104303483B (zh) 用于社交网络的基于用户的识别系统
US20150220933A1 (en) Methods and systems for making secure online payments
WO2016033698A1 (en) Method and system for real-time authentication of user access to a resource
CN110311891B (zh) 账户管理方法、装置、计算机设备及存储介质
US11823197B2 (en) Authenticating based on user behavioral transaction patterns
CN111598162A (zh) 一种黄牛风险监测方法、终端设备和存储介质
CN107656959B (zh) 一种留言方法、装置及留言设备
US10713538B2 (en) System and method for learning from the images of raw data
EP2896005A1 (en) Multi-factor profile and security fingerprint analysis
CN105988998B (zh) 关系网络构建方法及装置
CN107679865B (zh) 一种基于触压面积的身份验证方法和装置
CN106302367A (zh) 事务处理方法和系统
US20220414193A1 (en) Systems and methods for secure adaptive illustrations
CN111539742B (zh) 信息处理方法、装置、电子设备及存储介质
US11947643B2 (en) Fraud detection system, fraud detection method, and program
CN110210205B (zh) 登录状态的确定方法、装置、计算机设备及计算机存储介质
CN109165328A (zh) 一种用户身份认证方法及装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19925299

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19925299

Country of ref document: EP

Kind code of ref document: A1