Model training method and device, identity recognition method and device, medium and equipment

Abstract

The invention provides a model training method and device, an identity recognition method and device, a medium and equipment. The method comprises the following steps: acquiring a plurality of groups of training data; each group of training data comprises at least one entity relation graph, and each entity relation graph represents information that a user operates through at least one virtual identity in a network; and training a graph neural network model based on the training data to obtain a target model, the target model being used for identifying whether a plurality of entity relationship graphs point to the same user, and determining a unique identifier of each user. Through the method, the target model can be trained, so that whether a plurality of virtual identities are operated by the same user or not can be determined according to the target model.

Description

Technical Field

[0001] This specification relates to the technical field of deep learning, and particularly to a model training, identity recognition method, apparatus, medium, and device. Background Art

[0002] The same user in the physical world can have multiple identities in the cyber world. For example, a user can register or log in to multiple accounts simultaneously in an application system (such as a shopping platform, a payment platform).

[0003] Other users or back-end service personnel will consider multiple accounts as multiple different "users". This enables malicious users to use multiple identities on the network to disguise their true identities, thus resulting in violations.

[0004] Therefore, for network security and personal privacy protection, it is very important to know the different identities of the same person on the network. Summary of the Invention

[0005] To overcome the problems in the related art, this specification provides a model training, identity recognition method, apparatus, medium, and device.

[0006] According to the first aspect of the embodiments of this specification, a model training method is provided. The method includes:

[0007] Obtain a plurality of groups of training data; each group of training data includes at least one entity relationship graph, and each of the entity relationship graphs represents information of a user operating through at least one virtual identity in the network;

[0008] Train a graph neural network model based on the training data to obtain a target model, where the target model is used to identify whether a plurality of entity relationship graphs point to the same user, and to determine the unique identifier of each user.

[0009] According to the second aspect of the embodiments of this specification, a model training method is provided. The method includes:

[0010] Obtain a plurality of entity relationship graphs; each of the entity relationship graphs represents information of a user operating through at least one virtual identity in the network;

[0011] Input the plurality of entity relationship graphs into the target model to obtain user information output by the target model; the user information includes: the unique identifier of the user pointed to by each of the entity relationship graphs;

[0012] Wherein, the target model is trained by the model training method of the first aspect.

[0013] According to a third aspect of the embodiments of the present specification, a model training device is provided, the device includes:

[0014] A first acquisition module, configured to acquire several groups of training data; each group of training data includes at least one entity relationship graph, and each of the entity relationship graphs represents information of a user operating through at least one virtual identity in a network;

[0015] A training module, configured to train a graph neural network model based on the training data to obtain a target model, where the target model is used to identify whether multiple entity relationship graphs point to the same user, and to determine a unique identifier of each user.

[0016] According to a fourth aspect of the embodiments of the present specification, an identity recognition device is provided, the device includes:

[0017] A second acquisition module, configured to acquire multiple entity relationship graphs; each of the entity relationship graphs represents information of a user operating through at least one virtual identity in a network;

[0018] An input module, configured to input the multiple entity relationship graphs into the target model to obtain user information output by the target model; the user information includes: a unique identifier of the user pointed to by each of the entity relationship graphs;

[0019] Wherein, the target model is trained by the model training device of the third aspect.

[0020] According to a fifth aspect of the embodiments of the present specification, a computer-readable storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the steps of any of the methods provided in the embodiments of the first aspect or the second aspect are implemented.

[0021] According to a sixth aspect of the embodiments of the present specification, a computer device is provided, including a memory, a processor, and a computer program stored on the memory and executable on the processor; wherein, the processor is configured to execute the steps of any of the methods provided in the embodiments of the first aspect or the second aspect.

[0022] The technical solutions provided by the embodiments of the present specification may include the following beneficial effects:

[0023] In the embodiments of the present specification, the entity relationship graph is used as training data to train the graph neural network model, and the trained target model is obtained. Since the entity relationship graph not only includes key information about virtual identities, but also includes the association relationships between various key information, this helps the model extract more features and obtain more accurate training results.

[0024] In this way, it helps to determine whether multiple virtual identities corresponding to multiple entity relationship diagrams are operated by the same user through the trained target model.

[0025] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit this specification. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings herein are incorporated into and constitute a part of this specification, showing embodiments consistent with this specification, and are used together with the specification to explain the principles of this specification.

[0027] Figure 1 The flowchart of a model training method provided by an embodiment of this specification is shown.

[0028] Figure 2 The entity relationship diagram in a model training method provided by an embodiment of this specification is shown.

[0029] Figure 3 The flowchart of an identity recognition method provided by an embodiment of this specification is shown.

[0030] Figure 4 The structural schematic diagram of a model training device provided by an embodiment of this specification is shown.

[0031] Figure 5 The structural schematic diagram of an identity recognition device provided by an embodiment of this specification is shown.

[0032] Figure 6 It is a hardware structure diagram of a computer device where a device is located shown according to an exemplary embodiment of this specification. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0033] Exemplary embodiments will be described in detail herein, and examples thereof are shown in the accompanying drawings. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this specification. On the contrary, they are merely examples of devices and methods consistent with some aspects of this specification as detailed in the appended claims.

[0034] The terms used in this specification are only for the purpose of describing specific embodiments and are not intended to limit this specification. The singular forms "a", "the", and "said" used in this specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0035] It should be understood that although the terms first, second, third, etc. may be used in this specification to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, without departing from the scope of this specification, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "when" or "while" or "in response to determining".

[0036] The same user in the physical world can have multiple identities in the cyber world. For example, a user can register or log in to multiple accounts simultaneously in a software.

[0037] For other users or back-end service personnel, these are actually two different "users". This enables malicious users to use multiple identities on the network to disguise their true identities, thus resulting in violations.

[0038] For example, a user registers two accounts on a shopping website, namely a buyer account and a seller account. At this time, if the user uses the seller account on the shopping website to post a product, and then switches to the buyer account to claim the coupon issued by the platform to purchase this product, earning the amount of the coupon issued by the platform.

[0039] For another example, a user misappropriates (or steals) someone else's super administrator account, grants super permissions to his own ordinary user account, and then logs in to the ordinary user account to do something harmful to the platform through this super permission.

[0040] These are all caused by the inability to recognize the multiple identities of users in the cyber world. Therefore, the embodiments of this specification implement the recognition of users' multiple identities by providing a model training and identity recognition method, and the embodiments of this specification will be described in detail below.

[0041] Figure 1 The flowchart of a model training method provided by the embodiments of this specification is shown, as Figure 1 shown, the method includes the following steps:

[0042] Step 101, obtain a plurality of groups of training data; each group of training data includes at least one entity relationship graph, and each of the entity relationship graphs represents information of a user operating through at least one virtual identity in the network.

[0043] A virtual identity refers to the "identity" of a user in the physical world in the cyber world, such as an account. Information that a user operates through at least one virtual identity in the network includes, but is not limited to: account information of one or more accounts, identity information representing the user's identity (such as ID number, user name), communication information (phone number, email, other social account information, etc.), login device information when logging in to each account, network address at the time of login, behavioral information after logging in to the platform, and so on.

[0044] Among them, the account information of multiple accounts is associated together through the user's identity information or communication information.

[0045] Example 1: When user A registers for account A, the ID information filled in is ID number A; when user A registers for account B, the ID number filled in is also ID number A. Then account A and account B are regarded as different virtual identities, but different virtual identities can be associated through "ID number A" and thus constructed in the same entity relationship diagram, assumed to be entity relationship diagram A.

[0046] Example 2: On the basis of Example 1 above, if user A registers for account C again and the ID number filled in is ID B (such as the ID number of a parent or a spouse), then from the perspective of the accounts, account C has no association with account B (or account A). Therefore, the entity relationship diagram (such as entity relationship diagram B) constructed from the relevant information of account C (such as ID B, communication information, login device information, etc.) and entity relationship diagram A are two independent entity relationship diagrams.

[0047] Therefore, entity relationship diagram A and entity relationship diagram B can be used as two different sets of training data and input into the model, enabling the model to learn the characteristics of different entity relationship diagrams through unsupervised learning (or supervised learning).

[0048] Exemplarily, obtaining several sets of training data may include the following steps:

[0049] Obtain the account information and login device information of the user on at least one platform; sort out each piece of information and the association relationships between each piece of information in the account information and the login device information to obtain at least one of the entity relationship diagrams.

[0050] Exemplarily, the following information of user A is obtained: super administrator account (iam_id), general administrator account (iam_id), ordinary user account (ram_id), guest account (iam_id), and user ID registered in subsystem A (sysA_user_id), user ID registered in subsystem B (sysB_user_id), registered mobile phone number (mobile), registered email (email), membership number (vip_num), mobile phone serial number IMEI (International Mobile Equipment Identity), computer identification number (mac_addr), MAC address (Media Access Control Address, computer hardware address).

[0051] Among them, the membership number can be an identifier used to characterize the membership characteristics when the account opens the membership. Through the above information, the association relationship between each key information can be further determined. For example:

[0052] The user ID registered in subsystem A (sysA_user_id) is bound to: registered mobile phone number (mobile), registered email (email), membership number (vip_num), and the information of the device it logs in is: computer identification number (mac_addr), logged-in super administrator account (iam_id).

[0053] The user ID registered in subsystem B (sysB_user_id) is bound to: membership number (vip_num), and the information of the device it logs in is: mobile phone serial number IMEI, computer identification number (mac_addr), and logged-in guest account (iam_id).

[0054] At this time, taking each information as the "points" of the entity relationship diagram (which can also be called nodes or entities), and taking the associations between each information as the "edges" (the association relationships between each node or each entity), an entity relationship diagram is constructed. Exemplarily, the entity relationship diagram can be constructed using a graph database GDB (graph database, a type of database that can store graph structures and support querying of graph structure data).

[0055] It should be noted that if there is the same information in the information bound and logged in by the user ID registered in the above-mentioned subsystem A and the user ID registered in subsystem B (for example, the computer identification number of the device logged in by the user ID registered in subsystem A is the same as the computer identification number of the device logged in by the user ID registered in subsystem B), it can be built into the same entity relationship diagram.

[0056] Figure 2The entity relationship diagram in a model training method provided by an embodiment of this specification is shown. As Figure 2 shown, where Figure 2 in the upper figure (the first entity relationship diagram) of Figure 2 , the meanings of each serial number are as follows:

[0057] Serial number 1: Computer identification number (mac_addr), assumed to be M5001.

[0058] Serial number 2: A subsystem registered user ID (sysA_user_id), assumed to be: sysA001.

[0059] Serial number 3: Registered mobile phone number (mobile) 1, assumed to be: 001.

[0060] Serial number 4: Registered mobile phone number (mobile) 2, assumed to be: 002.

[0061] Serial number 5: Membership number (vip_num), assumed to be 620001.

[0062] Serial number 6: B subsystem registered user ID (sysB_user_id), assumed to be sysB901.

[0063] It can be seen that the two virtual identities of "A subsystem registered user ID (sysA_user_id): sysA001" and "B subsystem registered user ID (sysB_user_id): sysB901" are constructed in an entity relationship diagram by using the same "computer identification number (mac_addr): M5001".

[0064] Figure 2 in the lower figure (the second entity relationship diagram) of Figure 2 , the meanings of each serial number are as follows:

[0065] Serial number 10: B subsystem registered user ID (sysB_user_id), assumed to be sysB808

[0066] Serial number 7: Membership number (vip_num), assumed to be 625002.

[0067] Serial number 8: Computer identification number (mac_addr), assumed to be M6661.

[0068] Serial number 9: Mobile phone serial number IMEI, assumed to be imei008.

[0069] At this time, it can be seen that the virtual identity of "B subsystem registered user ID (sysB_user_id)" constructs an entity relationship diagram alone.

[0070] Note that, in order to improve the accuracy of training or recognition, it can be the staff or the background server to confirm that multiple virtual identities point to the same user before constructing them in the same entity relationship graph. Or, the graph database automatically determines whether to construct multiple virtual identities into the same entity relationship graph (for example, by setting weights for different types of nodes).

[0071] Among them, before sorting out the entity relationship graph, the training data can also be preprocessed to facilitate the execution of subsequent processing procedures.

[0072] Step 102, train the graph neural network model based on the training data to obtain a target model, where the target model is used to identify whether multiple entity relationship graphs point to the same user, and determine the unique identifier of each user.

[0073] Among them, multiple entity relationship graphs pointing to the same user means that multiple virtual identities under multiple entity relationship graphs are operated by the same user.

[0074] The number of entity relationship graphs and the number of virtual identities can be the same or different. For example, if an entity relationship graph includes two virtual identities, then the two virtual identities in this entity relationship graph are both operated by the same user; if two entity relationship graphs both point to the same user, and there are multiple virtual identities in the two entity relationship graphs, then these multiple virtual identities are all operated by the same user pointed to by these two entity relationship graphs.

[0075] Exemplarily, the graph neural network model can be GAT (Graph Attention Networks), so that the attention mechanism can be used to process graph-structured data.

[0076] The processing principle of GAT is as follows:

[0077] First, in the initial stage, we need to assign a unique feature vector to each node, which is called h^(0) (that is, the zero power of h). Then, we can generate new features for each node in the next layer by fusing the features of each node with its neighboring nodes, that is, h^(1). This operation can be repeated multiple times, and each time a new set of feature vectors h^(I) will be generated, where I is the number of layers. This process can enable the model to have good generalization ability, that is, even for newly added entities or newly emerging roles or identities in the entity relationship graph, it can be well adapted and make correct predictions. In addition, since each node fuses on the features of its neighboring nodes, the model can better capture the complex relationships in the graph and can more easily discover potential patterns and trends.

[0078] The target model trained in the above manner can well identify whether each virtual identity in the input (one or more) entity relationship diagrams is operated by the same user.

[0079] Optionally, Figure 3 The flowchart of an identity recognition method provided by an embodiment of this specification is shown, as Figure 3 shown, the method includes the following steps:

[0080] Step 301, obtain multiple entity relationship diagrams; each of the entity relationship diagrams represents information about a user's operations in the network through at least one virtual identity.

[0081] When wanting to know whether multiple virtual identities are operated by the same user, it is necessary to obtain the entity relationship diagrams corresponding to the multiple virtual identities, and then input the entity relationship diagrams into the target model to obtain the result output by the target model.

[0082] It should be noted that when an entity relationship diagram includes multiple virtual identities, each virtual identity in the same entity relationship diagram is operated by the same user, and this does not need to be determined additionally through the target model. Therefore, what the target model actually wants to determine is whether the virtual identities in multiple entity relationship diagrams are operated by the same user.

[0083] Exemplarily, multiple entity relationship diagrams can be obtained through at least one of the following methods:

[0084] Method 1: Obtain the entity relationship diagrams pre-organized in the database.

[0085] The database can be the aforementioned graph database GDB. The database stores pre-organized entity relationship diagrams. Exemplarily, if the entity relationship diagram of account A has been stored in the database, then the entity relationship diagram of this account A can be directly obtained.

[0086] And / or Method 2: Organize each piece of information and the association relationships between each piece of information in the account information and login device information of the user on at least one platform to obtain at least one of the entity relationship diagrams.

[0087] When new operation information of the virtual identity is obtained (such as binding a new email address, a new mobile phone number, a new ID number, etc.), or a new virtual identity is obtained (such as a newly registered account), a new entity relationship diagram can be re-organized. The organization method has been described in the foregoing embodiments and will not be elaborated here.

[0088] If there is only one entity relationship diagram to be input into the target model provided by the user (hereinafter referred to as the entity relationship diagram to be recognized), then in fact, the user wants to compare this entity relationship diagram with several entity relationship diagrams in the database to determine whether it points to the same user as one or more entity relationship diagrams in the database. At this time, after obtaining this entity relationship diagram to be recognized, it is also necessary to obtain at least one entity relationship diagram from the database (or compare each entity relationship diagram in the database with this entity relationship diagram to be recognized in turn).

[0089] At this time, the entity relationship diagram is obtained by the above-mentioned method 1 and the above-mentioned method 2 together. In this application, the entity relationship diagram to be recognized is mainly obtained by the above-mentioned method 2.

[0090] Step 302, input the multiple entity relationship diagrams into the target model to obtain the user information output by the target model; the user information includes: the unique identifier of the user pointed to by each entity relationship diagram; wherein, the target model is trained by the aforementioned model training method.

[0091] In one embodiment, the unique identifier can directly use the user's ID number. In another embodiment, considering that the user may not provide their ID number, or it is very likely that the same user uses multiple accounts registered with someone else's ID number (such as using the accounts of their spouse and parents), therefore, to distinguish the actual operating user, a unique identifier needs to be generated for them. At this time, the unique identifier is not the ID number.

[0092] The user information output by the target model includes the unique identifier corresponding to each virtual identity in the input entity relationship diagram.

[0093] Suppose, in the 2 input entity relationship diagrams, the entity relationship Figure 1 (including virtual identity 1 and virtual identity 2) and the entity relationship Figure 2 (including virtual identity 3).

[0094] At this time, since virtual identity 1 and virtual identity 2 are in the same entity relationship diagram, the users they point to are also the same (such as user A). At this time, the unique identifiers corresponding to virtual identity 1 and virtual identity 2 are both the unique identifiers generated for user A.

[0095] The target model can determine whether the entity relationship Figure 1 and the entity relationship Figure 2 point to the same user by analyzing them. If they point to the same user, the unique identifier of virtual identity 3 output is also the unique identifier of user A; otherwise, it is necessary to re-determine the entity relationship Figure 1 and the entity relationship Figure 2 ​Figure 2 The unique identifier corresponding to the virtual identity 3 in

[0096] Optionally, the method further includes the following steps:

[0097] Store the mapping relationship between the entity relationship diagram and the unique identifier in a database.

[0098] In this way, due to the introduction of the database, the target model can call the database to assist in completing the recognition of the entity relationship diagram. At this time, performing step 302 to input the multiple entity relationship diagrams into the target model to obtain the user information output by the target model includes the following steps:

[0099] Step 3021, input the multiple entity relationship diagrams into the target model to obtain the unique identifier output by the target model for each of the entity relationship diagrams, where the unique identifier is determined by the target model based on the mapping relationship stored in the database.

[0100] Among them, inputting the multiple entity relationship diagrams into the target model includes:

[0101] Input the entity relationship diagram and each target entity relationship diagram in the mapping relationship into the target model; the unique identifier is the first identifier or the second identifier. The first identifier is the unique identifier of the target user obtained when the target model determines that the entity relationship diagram and any one of the target entity relationship diagrams point to the same target user; the second identifier is the unique identifier generated for the user pointed to by the entity relationship diagram when the target model determines that the entity relationship diagram and each target entity relationship diagram in the mapping relationship point to different users.

[0102] Exemplarily, the multiple entity relationship diagrams (multiple entity relationship diagrams to be recognized) can be recognized to determine whether the multiple entity relationship diagrams to be recognized point to the same user, and then, for the entity relationship diagrams to be recognized that point to different users, compare them with the entity relationship diagrams in the database.

[0103] If the entity relationship diagram to be recognized is the same as (or the similarity reaches a specified threshold) any one of the target entity relationship diagrams in the database, or is recognized as pointing to the same user, the unique identifier (that is, the first identifier) in the mapping relationship can be directly obtained.

[0104] If the entity relationship diagram to be recognized points to different users from each of the target entity relationship diagrams in the database, directly generate a new unique identifier (that is, the second identifier) for the user pointed to by this entity relationship diagram (entity relationship diagram to be recognized) input into the target model.

[0105] Step 3022: Obtain the user information output by the target model based on the unique identifier of each entity relationship diagram.

[0106] Figure 4 FIG. 4 shows a schematic structural diagram of a model training device provided by an embodiment of the present specification. As Figure 4 shown, the device includes:

[0107] A first acquisition module 401, configured to acquire a plurality of groups of training data; each group of training data includes at least one entity relationship diagram, and each of the entity relationship diagrams represents information of a user operating through at least one virtual identity in a network.

[0108] A training module 402, configured to train a graph neural network model based on the training data to obtain a target model, where the target model is used to identify whether a plurality of entity relationship diagrams point to the same user and determine the unique identifier of each user.

[0109] Optionally, when the first acquisition module 401 is configured to acquire a plurality of groups of training data, it is configured to:

[0110] Acquire the account information and login device information of the user on at least one platform.

[0111] Sort out each piece of information and the association relationship between each piece of information in the account information and the login device information to obtain at least one of the entity relationship diagrams.

[0112] Figure 5 FIG. 5 shows a schematic structural diagram of an identity recognition device provided by an embodiment of the present specification. As Figure 5 shown, the device includes:

[0113] A second acquisition module 501, configured to acquire a plurality of entity relationship diagrams; each of the entity relationship diagrams represents information of a user operating through at least one virtual identity in a network;

[0114] An input module 502, configured to input the plurality of entity relationship diagrams into a target model to obtain the user information output by the target model; the user information includes: the unique identifier of the user pointed to by each of the entity relationship diagrams; wherein, the target model is trained by the foregoing model training device.

[0115] Optionally, when the second acquisition module 501 is configured to acquire a plurality of entity relationship diagrams, it is configured to:

[0116] Sort out each piece of information and the association relationship between each piece of information in the account information and the login device information of the user on at least one platform to obtain at least one of the entity relationship diagrams.

[0117] Optionally, the device further includes:

[0118] A storage module, configured to store the mapping relationship between the entity relationship diagram and the unique identifier in a database.

[0119] When the input module 502 is configured to input the multiple entity relationship diagrams into a target model to obtain user information output by the target model, it is configured to:

[0120] Input the multiple entity relationship diagrams into the target model to obtain a unique identifier output by the target model for each of the entity relationship diagrams, where the unique identifier is determined by the target model based on the mapping relationship stored in the database.

[0121] Based on the unique identifier of each entity relationship diagram, obtain the user information output by the target model.

[0122] Optionally, when the input module 502 is configured to input the multiple entity relationship diagrams into the target model, it is configured to:

[0123] Input each target entity relationship diagram in the entity relationship diagram and the mapping relationship into the target model; the unique identifier is a first identifier or a second identifier, where the first identifier is the unique identifier of the target user obtained when the target model determines that the entity relationship diagram and any one of the target entity relationship diagrams point to the same target user; the second identifier is a unique identifier generated by the target model for the user pointed to by the entity relationship diagram when the target model determines that the entity relationship diagram and each target entity relationship diagram in the mapping relationship point to different users.

[0124] The implementation processes of the functions and roles of each module in the above device are specifically described in detail in the implementation processes of the corresponding steps in the above method, and will not be elaborated here.

[0125] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiments described above are only illustrative. The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution in this specification. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0126] Figure 6This is a hardware structure diagram of a computer device where a device shown in an exemplary embodiment of this specification is located. The device may include: a processor 601, a memory 602, an input / output interface 603, a communication interface 604, and a bus 605. Among them, the processor 601, the memory 602, the input / output interface 603, and the communication interface 604 are communicatively connected to each other inside the device through the bus 605.

[0127] The processor 601 may be implemented in ways such as a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the methods provided in the embodiments of this specification.

[0128] The memory 602 may be implemented in forms such as a ROM (Read Only Memory), a RAM (Random Access Memory), a static storage device, a dynamic storage device, etc. The memory 602 may store an operating system and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 602 and are called and executed by the processor 601.

[0129] The input / output interface 603 is used to connect to an input / output module to implement information input and output. The input / output module may be configured as a component in the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Among them, the input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.

[0130] The communication interface 604 is used to connect to a communication module (not shown in the figure) to implement communication interaction between this device and other devices. Among them, the communication module may implement communication in a wired manner (such as USB, network cable, etc.) or in a wireless manner (such as mobile network, WIFI, Bluetooth, etc.).

[0131] The bus 605 includes a path for transmitting information between various components of the device (such as the processor 601, the memory 602, the input / output interface 603, and the communication interface 604).

[0132] It should be noted that although the above device only shows the processor 601, the memory 602, the input / output interface 603, the communication interface 604, and the bus 605, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, those skilled in the art can understand that the above device may also only include the components necessary to implement the solution of the embodiments of this specification, and does not necessarily include all the components shown in the figure.

[0133] The embodiments of this specification also provide a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the steps of any of the methods provided in the embodiments of this specification.

[0134] Computer-readable media includes both permanent and non-permanent, removable and non-removable media and can store information by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0135] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, commodity or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, commodity or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, commodity or device comprising the element.

[0136] The above description has been made of specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the acts or steps recited in the claims may be performed in a different order than in the embodiments and still achieve the desired result. Additionally, the processes depicted in the figures do not necessarily require the particular order shown or sequential order to achieve the desired result. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A model training method, characterized in that, Including: Obtaining a plurality of sets of training data; Each set of training data includes at least one entity relationship graph, and each of the entity relationship graphs represents information of a user operating through at least one virtual identity in a network; Training a graph neural network model based on the training data to obtain a target model, where the target model is used to identify whether multiple entity relationship graphs point to the same user and to determine the unique identifier of each user.

2. The method according to claim 1, wherein The obtaining a plurality of sets of training data includes: Obtaining the account information and login device information of the user on at least one platform; Sorting out each piece of information in the account information and the login device information and the association relationship between each piece of information to obtain at least one of the entity relationship graphs.

3. An identity recognition method, characterized in that, The method includes: Obtaining a plurality of entity relationship graphs; each of the entity relationship graphs represents information of a user operating through at least one virtual identity in a network; Inputting the plurality of entity relationship graphs into the target model to obtain user information output by the target model; the user information includes: the unique identifier of the user pointed to by each of the entity relationship graphs; Wherein, the target model is trained by the method described in claim 1 or 2.

4. The method according to claim 3, characterized in that, The obtaining a plurality of entity relationship graphs includes: Sorting out each piece of information in the account information and the login device information of the user on at least one platform and the association relationship between each piece of information to obtain at least one of the entity relationship graphs.

5. The method according to claim 3, wherein The method further includes: Storing the mapping relationship between the entity relationship graph and the unique identifier in a database; The inputting the plurality of entity relationship graphs into the target model to obtain user information output by the target model includes: Inputting the plurality of entity relationship graphs into the target model to obtain the unique identifier output by the target model for each of the entity relationship graphs, and the unique identifier is determined by the target model based on the mapping relationship stored in the database; Based on the unique identifier of each entity relationship graph, obtaining the user information output by the target model.

6. The method according to claim 5, characterized in that The inputting the plurality of entity relationship graphs into the target model includes: Inputting each target entity relationship graph in the entity relationship graph and the mapping relationship into the target model; the unique identifier is the first identifier or the second identifier. The first identifier is the unique identifier of the target user obtained when the target model determines that the entity relationship graph and any one of the target entity relationship graphs point to the same target user; the second identifier is the unique identifier generated by the target model for the user pointed to by the entity relationship graph when the target model determines that the entity relationship graph and each target entity relationship graph in the mapping relationship point to different users.

7. A model training device, characterized in that, The apparatus includes: A first obtaining module, configured to obtain a plurality of sets of training data; each set of training data includes at least one entity relationship graph, and each of the entity relationship graphs represents information of a user operating through at least one virtual identity in a network; A training module, configured to train a graph neural network model based on the training data to obtain a target model, where the target model is used to identify whether multiple entity relationship graphs point to the same user and to determine the unique identifier of each user.

8. An identity recognition device, characterized in that, The apparatus includes: A second acquisition module, configured to acquire a plurality of entity relationship graphs; each of the entity relationship graphs characterizes information of a user operating in a network through at least one virtual identity; An input module, configured to input the plurality of entity relationship graphs into a target model, and obtain user information output by the target model; the user information includes: a unique identifier of a user pointed to by each of the entity relationship graphs; Wherein, the target model is trained by the device described in claim 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, the steps of the method according to any one of claims 1-6 are implemented.

10. A computer device, characterized in that, It includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the steps of the method according to any one of claims 1-6 are implemented.

Images