Identity information security management system and method based on artificial intelligence

Abstract

The invention relates to the field of information security, in particular to an identity information security management system and method based on artificial intelligence, and the system comprises a data exchange module, an information agent module, an AI modeling module, a network convolution module and a verification restoration module, the information agent module is used for extracting data and signing, the AI modeling module is used for carrying out physical modeling, the network convolution module is used for generating a verification function and carrying out high-dimensional network convolution, and the verification restoration module is used for restoring an original data packet. The information security of the database and the identity security of the user are protected, the difficulty and the cost of network attacks are increased, the overall security level of communication is improved, the resource utilization efficiency and the compliance are improved, and the real-time performance, the flexibility and the intelligence of the security management process are ensured.

Description

Technical Field

[0001] The present invention relates to the field of information security, and in particular to an identity information security management system and method based on artificial intelligence. Background Art

[0002] Information security management is a key technology to protect data from unauthorized access, leakage, tampering and destruction. Commonly used security management methods include data encryption management, access control, intrusion detection, security information and event management, etc., which can ensure the security of sensitive information.

[0003] Since some data will generate user logs, public certificates, URL directional links and other identity identifiers during the exchange process, it is much easier to obtain user IDs than to obtain data. Existing security management technologies focus on data tampering and leakage prevention, but lack protection for user personal information. If attackers analyze the identities of both parties, it will not only increase the risk of data forgery, but may also lead to identity theft, financial losses, privacy violations and other problems.

[0004] Since user addresses must be transmitted publicly and identity information is difficult to encrypt, existing technologies avoid identity monitoring by forwarding through proxy servers. However, the connection between the proxy server and the user is easily hijacked. Through reverse address analysis, the user ID and access content can still be identified. At the same time, the management of user identity permissions and access types is also one of the difficulties in identity authentication. Summary of the invention

[0005] The purpose of the present invention is to provide an identity information security management system and method based on artificial intelligence to solve the problems raised in the above background technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: an identity information security management system based on artificial intelligence, comprising: a data exchange module, an information proxy module, an AI modeling module, a network convolution module and a verification and restoration module; The data exchange module is used to set up a user interaction interface on the access end. After the user logs in using an ID, he enters the user authority and operation request, parses the operation request into an instruction pointing to the data end, and establishes independent memory on the access end and the data end to record the operation records of each user, and generates a database entry for each operation of all user IDs; The information proxy module is used to establish a proxy server between the access terminal and the data terminal, receive user instructions, extract data from the data terminal according to the instructions, and store the extracted data packets or the user's modification instructions for the data packets in the server memory, generate a unique signature based on the proxy server IP, and perform data signing when the data packets are uploaded or downloaded; The AI ​​modeling module is used to train a generative AI model in the proxy server, and use the AI ​​model to perform physical modeling for all database entries under each user ID, using the confidentiality level of the data packet as the collision volume of the physical module, using the data volume of the data packet as the collision mass of the physical model, and setting the collision volume to decrease with the number of times the user accesses the same data packet, using the user access level as the maximum collision force of the model, and after receiving the instruction, recording the physical model corresponding to the user ID that issued the instruction as a first-class model, and recording the models corresponding to the remaining user IDs as second-class models; The network convolution module is used to convert the data packet authority involved in the user instruction into a collision force input into each physical model, AI identifies the momentum of the collision model, generates a verification function, and uses each verification function as a node in turn to perform high-dimensional network convolution processing on the acquired data packet, designates the first type of model node as a unique high-dimensional, and the second type of model as a low-dimensional convolution layer, to form a model verification network, and outputs the data packet processed by the network; The verification and restoration module is used to input the processed data packet into the access terminal, and the user or the access terminal provides user permissions and historical operation logs, generates a type of model according to the verification function corresponding to the signature, uses the verification function corresponding to the type of model as the only high-dimensional item to perform inverse convolution decoding on the data packet, and after cleaning all low-dimensional components, restores the original data and activates the modification instructions stored in the server memory.

[0007] Further, the data exchange module includes: an instruction parsing unit and a data management unit; The instruction parsing unit is used to build an exchange interface with the user, obtain user login information, and parse the user's access request to the data end into an instruction pointing to the data end; The data management unit is used to store the operation records of all users, and the operation records include: user ID, operation time and operation instructions.

[0008] Further, the information proxy module includes: a proxy service unit, an information link unit and an identity storage unit; The proxy service unit is used to receive user instructions from the access terminal and extract data packets from the data terminal according to the user instructions, and the extraction method includes hardware extraction and wireless extraction; The information link unit is used to construct a transmission link between the proxy server and the access terminal, and ensure the reliability of the short-distance link between the proxy server and the data terminal; The identity storage unit is used to generate a digital signature according to the proxy server IP and mark it after the data packet.

[0009] Further, the AI ​​modeling module includes: an artificial intelligence unit and a physical construction unit; The artificial intelligence unit is used to train the generative AI model, build the underlying logic of the physical model, and provide computing power support for model migration; The physical construction unit is used to modularize each operation record of the user, and to superimpose modules with the same ID in space to form a physical model.

[0010] Furthermore, the network convolution module includes: a collision input unit, a unique verification unit and a high-dimensional convolution unit; The collision input unit is used to compile the user's instruction into a collision with the physical model, and calculate the collision force to determine the momentum of each module after the collision; The unique verification unit is used to determine the parameters of the base kernel function according to the position and momentum of each module, obtain the verification function, and designate the verification function corresponding to the source ID as a unique high dimension; The high-dimensional convolution unit is used to perform layer-by-layer convolution using high-dimensional and low-dimensional verification functions respectively to obtain a model verification network, and use the model verification network to process data packets.

[0011] Further, the verification and restoration module includes: a digital signature unit and a dimension retrieval unit; The digital signature unit is used to send a data packet to an access terminal, and the access terminal parses the signature of the data packet and generates a decoding record; The dimension retrieval unit is used to obtain the operation information of the current user, use the operation information to decode the high-dimensional convolution layer, and restore the original data packet after layer-by-layer analysis.

[0012] An identity information security management method based on artificial intelligence includes the following steps: Step S1. The user logs in the ID at the access end, issues an access request, compiles the access request into an instruction pointing to the data end, sets a proxy server with a solid connection at the data end, and the proxy server receives the access end instruction and extracts the corresponding data packet in the database; Step S2. The data end obtains the operation records of all user IDs on the sub-library where the data packet is located, trains the generative AI model, and uses the AI ​​model to build the physical space. For each operation record, a corresponding physical module is generated with the operation object and access rights as parameters. All physical modules under the same ID are stacked in the space to form a physical model. Step S3. According to the data packet permissions involved in the instructions, the user instructions are compiled into collisions with the physical model, and the collision force is calculated to determine the momentum of each module after the collision. The parameters of the base kernel function are determined according to the position and momentum of each module after the collision, and the base kernel function model with variable parameters is selected for fitting to obtain the verification function; Step S4. The physical model corresponding to the instruction source user ID is recorded as a first-class model, and the models corresponding to the other user IDs are recorded as second-class models. The data packet uses each verification function as a node in turn, performs high-dimensional network convolution layer by layer, and designates the nodes corresponding to the first-class model as the only high-dimensional one to obtain a model verification network, and uses the model verification network to process the data packet; Step S5. The proxy server signs the processed data packet and sends it to the access end. The access end issues the operation record of the current ID user, reproduces the high-dimensional physical model, obtains the verification function of the high-dimensional node, decodes the high-dimensional convolution layer, and restores the original data packet after deconvolution of the low-dimensional convolution layer.

[0013] Further, step S1 includes: Step S11. A user interaction interface is constructed on the access end. The user provides an ID and password to log in to the interaction interface. When accessing the virtual database in the interface, all user operations are recorded and stored, and the operations are compiled into binary instructions pointing to the data end according to the database protocol; Step S12. The data terminal is solidly connected to the proxy server. The proxy server has a replaceable IP address cluster, can extract information from the database, receive data with a public IP, and send data through any IP. After the proxy server receives the instruction directed to the data terminal, it extracts the data packet from the data terminal according to the data index corresponding to the instruction.

[0014] Further, step S2 includes: Step S21. The same training library is pre-set on the data end and the access end, and the generative AI model is trained using the training library so that the generative AI model has the underlying logic of the physical entity, thereby obtaining the ability to construct a physical space. The generative AI models selected include: physical information neural network, graph neural network, generative adversarial network and variational autoencoder; Step S22. For each operation record, the confidentiality level of the data packet is used as the collision volume of the physical module, and the data volume of the data packet is used as the collision quality of the physical model. The collision volume is set to decrease with the number of times the user accesses the same data packet. All modules under each user ID are stacked in space to form a physical model.

[0015] Further, step S3 includes: Step S31. Analyze the user instruction, obtain the confidentiality level of each data packet involved in the user instruction, and calculate the collision force F on the physical model: ; Wherein, m represents the number of data packets involved in the instruction, Wj represents the data volume of the jth data packet, RE represents the authority level of the user from which the instruction comes, Rj represents the confidentiality level of the jth data packet, a represents the physical amplification factor, and c0 represents the preset basic force; Step S32. Use the collision force to analyze the collision of the physical models under all user IDs, obtain the position and momentum of each model after the collision, and select the base kernel function model with variable parameters to fit the verification function: ; Among them, T(x) is the verification function, x is the input force, G(x) is the selected basis kernel function, n is the number of modules in the physical model, ki represents the mass of the i-th module, σi represents the distance between the position of the i-th module and the initial position, and xi is the momentum of the i-th module.

[0016] Further, step S4 includes: Step S41. The verification function corresponding to the user ID that issues the instruction is used as the only high-dimensional node. If the user issues the instruction for the first time, the preset standard verification function is used as the only high-dimensional node, and the verification functions under the other user IDs are used as low-dimensional nodes; Step S42: Perform convolution layer by layer using the verification functions of high-dimensional and low-dimensional nodes to obtain a model verification network. The processing logic of the model verification network is: ; Where P is the network output, P1 and P2 are the real and imaginary parts of the output respectively, TE is the verification function feature of the high-dimensional node, u is the number of low-dimensional nodes, Tz is the verification function feature of the z-th low-dimensional node, E is the data in the data packet, and b is the offset. When the verification functions of the data end and the access end are consistent, the offset is 1; Step S43: Use the model verification network to process the data packet to obtain an encrypted data packet, and store the encrypted data packet into the proxy server memory.

[0017] Further, step S5 includes: Step S51. After the proxy server signs the data packet, it sends the data packet to the access terminal, and the access terminal parses the signature of the data packet and generates a decoding record; Step S52. The access end obtains all operation records of the current user ID on the database, or requires the user to provide all operation records on the database, restores the verification function feature TE of the high-dimensional node according to the operation record, restores Tz from the imaginary part of the output, and obtains the original data packet after deconvolution layer by layer of data packet analysis.

[0018] Compared with the prior art, the beneficial effects achieved by the present invention are: The present invention records the identity, authority and historical operations of each user in a database, and for each user identity ID, uses artificial intelligence to perform physical modeling for each operation record, determines the physical parameters of the model according to the user identity, and converts the digital key into an unmodifiable physical model. This enables the server to take corresponding protective measures based on real-time threats and risk changes, and monitors and protects the information security of the database and the identity security of users as a whole.

[0019] The present invention converts the permissions involved in user instructions into collision forces input into various physical models, performs high-dimensional network convolution, designates user nodes as unique high-dimensional, generates proxy data packets, and requires attackers to obtain the physical models of all users to decrypt data, thereby increasing the difficulty and cost of network attacks and improving the overall security level of communications.

[0020] The present invention generates a physical model according to the verification function corresponding to the signature through the user authority and historical operation log issued by the access terminal, decodes the data packet using the verification function corresponding to the model and obtains a unique high-dimensional item, and restores the original data after inverse convolution, which can improve resource utilization efficiency and compliance and ensure the real-time, flexibility and intelligence of the security management process. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 It is a structural schematic diagram of an identity information security management system based on artificial intelligence of the present invention; Figure 2 It is a schematic diagram of the steps of an identity information security management method based on artificial intelligence of the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] See also Figure 1 ,The present invention provides a technical solution: an identity information security management system based on artificial intelligence, including: a data exchange module, an information proxy module, an AI modeling module, a network convolution module and a verification and restoration module; The data exchange module is used to set up a user interaction interface on the access end. After the user logs in using an ID, he enters the user authority and operation request, parses the operation request into an instruction pointing to the data end, and establishes independent memory on the access end and the data end to record the operation records of each user, and generates a database entry for each operation of all user IDs; The data exchange module includes: an instruction parsing unit and a data management unit; The instruction parsing unit is used to build an exchange interface with the user, obtain user login information, and parse the user's access request to the data end into an instruction pointing to the data end; The data management unit is used to store the operation records of all users, and the operation records include: user ID, operation time and operation instructions.

[0024] The information proxy module is used to establish a proxy server between the access terminal and the data terminal, receive user instructions, extract data from the data terminal according to the instructions, and store the extracted data packets or the user's modification instructions for the data packets in the server memory, generate a unique signature based on the proxy server IP, and perform data signing when the data packets are uploaded or downloaded; The information proxy module includes: a proxy service unit, an information link unit and an identity storage unit; The proxy service unit is used to receive user instructions from the access terminal and extract data packets from the data terminal according to the user instructions, and the extraction method includes hardware extraction and wireless extraction; The information link unit is used to construct a transmission link between the proxy server and the access terminal, and ensure the reliability of the short-distance link between the proxy server and the data terminal; The identity storage unit is used to generate a digital signature according to the proxy server IP and mark it after the data packet.

[0025] The AI ​​modeling module is used to train a generative AI model in the proxy server, and use the AI ​​model to perform physical modeling for all database entries under each user ID, using the confidentiality level of the data packet as the collision volume of the physical module, using the data volume of the data packet as the collision mass of the physical model, and setting the collision volume to decrease with the number of times the user accesses the same data packet, using the user access level as the maximum collision force of the model, and after receiving the instruction, recording the physical model corresponding to the user ID that issued the instruction as a first-class model, and recording the models corresponding to the remaining user IDs as second-class models; The AI ​​modeling module includes: an artificial intelligence unit and a physical construction unit; The artificial intelligence unit is used to train the generative AI model, build the underlying logic of the physical model, and provide computing power support for model migration; The physical construction unit is used to modularize each operation record of the user, and to superimpose modules with the same ID in space to form a physical model.

[0026] The network convolution module is used to convert the data packet authority involved in the user instruction into a collision force input into each physical model, AI identifies the momentum of the collision model, generates a verification function, and uses each verification function as a node in turn to perform high-dimensional network convolution processing on the acquired data packet, designates the first type of model node as a unique high-dimensional, and the second type of model as a low-dimensional convolution layer, to form a model verification network, and outputs the data packet processed by the network; The network convolution module includes: a collision input unit, a unique verification unit and a high-dimensional convolution unit; The collision input unit is used to compile the user's instruction into a collision with the physical model, and calculate the collision force to determine the momentum of each module after the collision; The unique verification unit is used to determine the parameters of the base kernel function according to the position and momentum of each module, obtain the verification function, and designate the verification function corresponding to the source ID as a unique high dimension; The high-dimensional convolution unit is used to perform layer-by-layer convolution using high-dimensional and low-dimensional verification functions respectively to obtain a model verification network, and use the model verification network to process data packets.

[0027] The verification and restoration module is used to input the processed data packet into the access terminal, and the user or the access terminal provides user permissions and historical operation logs, generates a type of model according to the verification function corresponding to the signature, uses the verification function corresponding to the type of model as the only high-dimensional item to perform inverse convolution decoding on the data packet, and after cleaning all low-dimensional components, restores the original data and activates the modification instructions stored in the server memory.

[0028] The verification and restoration module includes: a digital signature unit and a dimension retrieval unit; The digital signature unit is used to send a data packet to an access terminal, and the access terminal parses the signature of the data packet and generates a decoding record; The dimension retrieval unit is used to obtain the operation information of the current user, use the operation information to decode the high-dimensional convolution layer, and restore the original data packet after layer-by-layer analysis.

[0029] like Figure 2 As shown, an identity information security management method based on artificial intelligence includes the following steps: Step S1. The user logs in the ID at the access end, issues an access request, compiles the access request into an instruction pointing to the data end, sets a proxy server with a solid connection at the data end, and the proxy server receives the access end instruction and extracts the corresponding data packet in the database; Step S1 includes: Step S11. A user interaction interface is constructed on the access end. The user provides an ID and password to log in to the interaction interface. When accessing the virtual database in the interface, all user operations are recorded and stored, and the operations are compiled into binary instructions pointing to the data end according to the database protocol; Step S12. The data terminal is solidly connected to the proxy server. The proxy server has a replaceable IP address cluster, can extract information from the database, receive data with a public IP, and send data through any IP. After the proxy server receives the instruction directed to the data terminal, it extracts the data packet from the data terminal according to the data index corresponding to the instruction.

[0030] Step S2. The data end obtains the operation records of all user IDs on the sub-library where the data packet is located, trains the generative AI model, and uses the AI ​​model to build the physical space. For each operation record, a corresponding physical module is generated with the operation object and access rights as parameters. All physical modules under the same ID are stacked in the space to form a physical model. Step S2 includes: Step S21. The same training library is pre-set on the data end and the access end, and the generative AI model is trained using the training library so that the generative AI model has the underlying logic of the physical entity, thereby obtaining the ability to construct a physical space. The generative AI models selected include: physical information neural network, graph neural network, generative adversarial network and variational autoencoder; Step S22. For each operation record, the confidentiality level of the data packet is used as the collision volume of the physical module, and the data volume of the data packet is used as the collision quality of the physical model. The collision volume is set to decrease with the number of times the user accesses the same data packet. All modules under each user ID are stacked in space to form a physical model.

[0031] Step S3. According to the data packet permissions involved in the instructions, the user instructions are compiled into collisions with the physical model, and the collision force is calculated to determine the momentum of each module after the collision. The parameters of the base kernel function are determined according to the position and momentum of each module after the collision, and the base kernel function model with variable parameters is selected for fitting to obtain the verification function; Step S3 includes: Step S31. Analyze the user instruction, obtain the confidentiality level of each data packet involved in the user instruction, and calculate the collision force F on the physical model: ; Wherein, m represents the number of data packets involved in the instruction, Wj represents the data volume of the jth data packet, RE represents the authority level of the user from which the instruction comes, Rj represents the confidentiality level of the jth data packet, a represents the physical amplification factor, and c0 represents the preset basic force; Step S32. Use the collision force to analyze the collision of the physical models under all user IDs, obtain the position and momentum of each model after the collision, and select the base kernel function model with variable parameters to fit the verification function: ; Among them, T(x) is the verification function, x is the input force, G(x) is the selected basis kernel function, n is the number of modules in the physical model, ki represents the mass of the i-th module, σi represents the distance between the position of the i-th module and the initial position, and xi is the momentum of the i-th module.

[0032] Step S4. The physical model corresponding to the instruction source user ID is recorded as a first-class model, and the models corresponding to the other user IDs are recorded as second-class models. The data packet uses each verification function as a node in turn, performs high-dimensional network convolution layer by layer, and designates the nodes corresponding to the first-class model as the only high-dimensional one to obtain a model verification network, and uses the model verification network to process the data packet; Step S4 includes: Step S41. The verification function corresponding to the user ID that issues the instruction is used as the only high-dimensional node. If the user issues the instruction for the first time, the preset standard verification function is used as the only high-dimensional node, and the verification functions under the other user IDs are used as low-dimensional nodes; Step S42: Perform convolution layer by layer using the verification functions of high-dimensional and low-dimensional nodes to obtain a model verification network. The processing logic of the model verification network is: ; Where P is the network output, P1 and P2 are the real and imaginary parts of the output respectively, TE is the verification function feature of the high-dimensional node, u is the number of low-dimensional nodes, Tz is the verification function feature of the z-th low-dimensional node, E is the data in the data packet, and b is the offset. When the verification functions of the data end and the access end are consistent, the offset is 1; Step S43: Use the model verification network to process the data packet to obtain an encrypted data packet, and store the encrypted data packet into the proxy server memory.

[0033] Step S5. The proxy server signs the processed data packet and sends it to the access end. The access end issues the operation record of the current ID user, reproduces the high-dimensional physical model, obtains the verification function of the high-dimensional node, decodes the high-dimensional convolution layer, and restores the original data packet after deconvolution of the low-dimensional convolution layer.

[0034] Step S5 includes: Step S51. After the proxy server signs the data packet, it sends the data packet to the access terminal, and the access terminal parses the signature of the data packet and generates a decoding record; Step S52. The access end obtains all operation records of the current user ID on the database, or requires the user to provide all operation records on the database, restores the verification function feature TE of the high-dimensional node according to the operation record, restores Tz from the imaginary part of the output, and obtains the original data packet after deconvolution layer by layer of data packet analysis.

[0035] Embodiment: There are three users in total who have permissions for data packet 1, and their permission levels are 2, 1 and 3 respectively. After user 1 logs in, he submits an access instruction to data packet 1. The confidentiality level of the data packet is 1, the data volume is 10Mb, the amplification factor a=0.1, the basic force c0=1N, and the collision force is 2N. The operation records of users 1, 2 and 3 are used to establish physical models, and the collision physical model is simulated with a collision force of 2N to obtain a verification function.

[0036] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0037] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An identity information security management method based on artificial intelligence, characterized in that: The method comprises the following steps: Step S1. The user logs in the ID at the access end, issues an access request, compiles the access request into an instruction pointing to the data end, sets a proxy server with a solid connection at the data end, and the proxy server receives the access end instruction and extracts the corresponding data packet in the database; Step S2. The data end obtains the operation records of all user IDs on the sub-library where the data packet is located, trains the generative AI model, and uses the AI ​​model to build the physical space. For each operation record, a corresponding physical module is generated with the operation object and access rights as parameters. All physical modules under the same ID are stacked in the space to form a physical model. Step S3. According to the data packet permissions involved in the instructions, the user instructions are compiled into collisions with the physical model, and the collision force is calculated to determine the momentum of each module after the collision. The parameters of the base kernel function are determined according to the position and momentum of each module after the collision, and the base kernel function model with variable parameters is selected for fitting to obtain the verification function; Step S4. The physical model corresponding to the instruction source user ID is recorded as a first-class model, and the models corresponding to the other user IDs are recorded as second-class models. The data packet uses each verification function as a node in turn, performs high-dimensional network convolution layer by layer, and designates the nodes corresponding to the first-class model as the only high-dimensional one to obtain a model verification network, and uses the model verification network to process the data packet; Step S5. The proxy server signs the processed data packet and sends it to the access end. The access end issues the operation record of the current ID user, reproduces the high-dimensional physical model, obtains the verification function of the high-dimensional node, decodes the high-dimensional convolution layer, and restores the original data packet after deconvolution of the low-dimensional convolution layer.

2. According to claim 1, the identity information security management method based on artificial intelligence is characterized by: Step S1 includes: Step S11. A user interaction interface is constructed on the access end. The user provides an ID and password to log in to the interaction interface. When accessing the virtual database in the interface, all user operations are recorded and stored, and the operations are compiled into binary instructions pointing to the data end according to the database protocol; Step S12. The data terminal is solidly connected to the proxy server, and the proxy server has a replaceable IP address cluster, which can extract information from the database, receive data with a public IP, and send data through any IP. After receiving the instruction directed to the data terminal, the proxy server extracts the data packet from the data terminal according to the data index corresponding to the instruction; Step S2 includes: Step S21. The same training library is pre-set on the data end and the access end, and the generative AI model is trained using the training library so that the generative AI model has the underlying logic of the physical entity, thereby obtaining the ability to construct a physical space. The generative AI models selected include: physical information neural network, graph neural network, generative adversarial network and variational autoencoder; Step S22. For each operation record, the confidentiality level of the data packet is used as the collision volume of the physical module, and the data volume of the data packet is used as the collision quality of the physical model. The collision volume is set to decrease with the number of times the user accesses the same data packet. All modules under each user ID are stacked in space to form a physical model.

3. The method for secure management of identity information based on artificial intelligence according to claim 2, characterized in that: Step S3 includes: Step S31. Analyze the user instruction, obtain the confidentiality level of each data packet involved in the user instruction, and calculate the collision force F on the physical model: ； Wherein, m represents the number of data packets involved in the instruction, Wj represents the data volume of the jth data packet, RE represents the authority level of the user from which the instruction comes, Rj represents the confidentiality level of the jth data packet, a represents the physical amplification factor, and c0 represents the preset basic force; Step S32. Use the collision force to analyze the collision of the physical models under all user IDs, obtain the position and momentum of each model after the collision, and select the base kernel function model with variable parameters to fit the verification function: ； Among them, T(x) is the verification function, x is the input force, G(x) is the selected basis kernel function, n is the number of modules in the physical model, ki represents the mass of the i-th module, σi represents the distance between the position of the i-th module and the initial position, and xi is the momentum of the i-th module.

4. The method for secure management of identity information based on artificial intelligence according to claim 3, characterized in that: Step S4 includes: Step S41. The verification function corresponding to the user ID that issues the instruction is used as the only high-dimensional node. If the user issues the instruction for the first time, the preset standard verification function is used as the only high-dimensional node, and the verification functions under the other user IDs are used as low-dimensional nodes; Step S42: Perform convolution layer by layer using the verification functions of high-dimensional and low-dimensional nodes to obtain a model verification network. The processing logic of the model verification network is: ； Where P is the network output, P1 and P2 are the real and imaginary parts of the output respectively, TE is the verification function feature of the high-dimensional node, u is the number of low-dimensional nodes, Tz is the verification function feature of the z-th low-dimensional node, E is the data in the data packet, and b is the offset. When the verification functions of the data end and the access end are consistent, the offset is 1; Step S43: Use the model verification network to process the data packet to obtain an encrypted data packet, and store the encrypted data packet into the proxy server memory.

5. The method for secure management of identity information based on artificial intelligence according to claim 4, characterized in that: Step S5 includes: Step S51. After the proxy server signs the data packet, it sends the data packet to the access terminal, and the access terminal parses the signature of the data packet and generates a decoding record; Step S52. The access end obtains all operation records of the current user ID on the database, or requires the user to provide all operation records on the database, restores the verification function feature TE of the high-dimensional node according to the operation record, restores Tz from the imaginary part of the output, and obtains the original data packet after deconvolution layer by layer of data packet analysis.

6. An identity information security management system based on artificial intelligence, characterized in that: The system includes the following modules: a data exchange module, an information proxy module, an AI modeling module, a network convolution module, and a verification and restoration module; The data exchange module is used to set up a user interaction interface on the access end. After the user logs in using an ID, he enters the user authority and operation request, parses the operation request into an instruction pointing to the data end, and establishes independent memory on the access end and the data end to record the operation records of each user, and generates a database entry for each operation of all user IDs; The information proxy module is used to establish a proxy server between the access terminal and the data terminal, receive user instructions, extract data from the data terminal according to the instructions, and store the extracted data packets or the user's modification instructions for the data packets in the server memory, generate a unique signature based on the proxy server IP, and perform data signing when the data packets are uploaded or downloaded; The AI ​​modeling module is used to train a generative AI model in the proxy server, and use the AI ​​model to perform physical modeling for all database entries under each user ID, using the confidentiality level of the data packet as the collision volume of the physical module, using the data volume of the data packet as the collision mass of the physical model, and setting the collision volume to decrease with the number of times the user accesses the same data packet, using the user access level as the maximum collision force of the model, and after receiving the instruction, recording the physical model corresponding to the user ID that issued the instruction as a first-class model, and recording the models corresponding to the remaining user IDs as second-class models; The network convolution module is used to convert the data packet authority involved in the user instruction into a collision force input into each physical model, AI identifies the momentum of the collision model, generates a verification function, and uses each verification function as a node in turn to perform high-dimensional network convolution processing on the acquired data packet, designates the first type of model node as a unique high-dimensional, and the second type of model as a low-dimensional convolution layer, to form a model verification network, and outputs the data packet processed by the network; The verification and restoration module is used to input the processed data packet into the access terminal, and the user or the access terminal provides user permissions and historical operation logs, generates a type of model according to the verification function corresponding to the signature, uses the verification function corresponding to the type of model as the only high-dimensional item to perform inverse convolution decoding on the data packet, and after cleaning all low-dimensional components, restores the original data and activates the modification instructions stored in the server memory.

7. The identity information security management system based on artificial intelligence according to claim 6 is characterized by: The data exchange module includes: an instruction parsing unit and a data management unit; The instruction parsing unit is used to build an exchange interface with the user, obtain user login information, and parse the user's access request to the data end into an instruction pointing to the data end; The data management unit is used to store the operation records of all users, and the operation records include: user ID, operation time and operation instructions.

8. The identity information security management system based on artificial intelligence according to claim 7 is characterized by: The information proxy module includes: a proxy service unit, an information link unit and an identity storage unit; The proxy service unit is used to receive user instructions from the access terminal and extract data packets from the data terminal according to the user instructions, and the extraction method includes hardware extraction and wireless extraction; The information link unit is used to construct a transmission link between the proxy server and the access terminal, and ensure the reliability of the short-distance link between the proxy server and the data terminal; The identity storage unit is used to generate a digital signature according to the proxy server IP and mark it after the data packet.

9. The identity information security management system based on artificial intelligence according to claim 8 is characterized by: The AI ​​modeling module includes: an artificial intelligence unit and a physical construction unit; The artificial intelligence unit is used to train the generative AI model, build the underlying logic of the physical model, and provide computing power support for model migration; The physical construction unit is used to modularize each operation record of the user, and to superimpose modules with the same ID in space to form a physical model.

10. The identity information security management system based on artificial intelligence according to claim 9 is characterized in that: The AI ​​modeling module includes: an artificial intelligence unit and a physical construction unit; The artificial intelligence unit is used to train the generative AI model, build the underlying logic of the physical model, and provide computing power support for model migration; The physical construction unit is used to modularize each operation record of the user, and to superimpose modules with the same ID in space to form a physical model; The network convolution module includes: a collision input unit, a unique verification unit and a high-dimensional convolution unit; The collision input unit is used to compile the user's instruction into a collision with the physical model, and calculate the collision force to determine the momentum of each module after the collision; The unique verification unit is used to determine the parameters of the base kernel function according to the position and momentum of each module, obtain the verification function, and designate the verification function corresponding to the source ID as a unique high dimension; The high-dimensional convolution unit is used to perform layer-by-layer convolution using high-dimensional and low-dimensional verification functions respectively to obtain a model verification network, and use the model verification network to process data packets.

Images