Administrator face recognition data export protection method based on deep learning

Abstract

The invention discloses an administrator face recognition data export protection method based on deep learning. The method comprises the following steps: S1, starting multi-modal biological feature verification; s2, dynamically adjusting verification strength and complexity; s3, detecting whether abnormal behaviors or potential attacks exist or not; s4, generating a potential attack behavior sample, and training and optimizing an anomaly detection module in combination with an imitation learning technology; s5, when the system detects an abnormal behavior or a potential attack, starting an enhanced active cheating module, and confusing an attacker to enter a virtual environment through an induced data operation and hierarchical virtualization technology; s6, in the virtual environment, the system generates and optimizes an induction strategy through combination of a near-end strategy optimization algorithm and a meta-strategy gradient algorithm; and S7, recording all operation logs by the background system, and triggering multi-level security policy adjustment according to abnormal conditions. According to the method, deep learning and multi-modal biological feature verification are combined, an enhanced active cheating module is introduced, and the data export security is enhanced.

Description

Technical Field

[0001] The present invention relates to the field of information security technology, and in particular to a method for protecting administrator face recognition data export based on deep learning. Background Art

[0002] With the development of informatization and digital transformation, businesses and organizations have accumulated vast amounts of sensitive business data and user privacy data. This data is not only crucial to business operations but also poses a significant threat to hackers, malicious attackers, and insiders. Data export, in particular, involves significant data migration and flow, posing a significant risk of data leakage. Illegal data export can result in significant financial losses and damage to brand reputation. Therefore, ensuring the security and legality of data during export has become a crucial issue in the field of information security.

[0003] Existing data export protection technologies typically rely on a single authentication method or permission management strategy. For example, common authentication methods include password verification, biometric recognition technologies such as fingerprint recognition and facial recognition. While these authentication methods offer a certain degree of security, they are relatively vulnerable to complex attacks such as replay attacks and biometric forgery. Furthermore, many systems employ static permission management, meaning that once an administrator passes identity authentication, they gain access to all sensitive data for which they have permission. These systems lack the ability to dynamically adjust authentication strength based on data sensitivity, resulting in insufficient protection against potential attacks or unauthorized operations.

[0004] To detect abnormal operational behavior, existing technologies mostly use rule-based detection systems, using pre-set behavioral rules to determine whether an operation is normal. However, this approach is limited in effectiveness against complex and diverse attack vectors. Attackers can circumvent rule detection by mimicking normal operating patterns, further increasing system risk. Furthermore, existing anomaly detection technologies often only provide post-event alerts and block operations, lacking proactive mechanisms to induce and confuse attackers. Once attackers break through initial defenses, they could potentially access real data, causing irreversible damage.

[0005] The introduction of technologies such as generative adversarial networks (GANs) has significantly enhanced the capabilities of anomaly detection systems. GANs have demonstrated significant potential in generating forged samples and conducting adversarial training. However, current applications still face limitations. For example, forged samples generated by GANs struggle to capture complex attack behavior patterns, especially those associated with different environmental conditions. Therefore, further leveraging the unique characteristics of GANs, combined with attackers' historical behavior patterns, to generate more realistic and effective forged samples remains a major challenge in existing technologies.

[0006] Existing systems generally lack proactive defense mechanisms after detecting an attack. Most protection measures rely on passive blocking or triggering alarms. This passive defense strategy easily alerts attackers, potentially leading them to employ more covert methods to circumvent system detection, posing a greater threat to data security. Even in systems with robust security protections, which rely primarily on operational log audits and multiple verifications, these systems often fail to effectively prevent further attacks when attackers intentionally evade detection.

[0007] Furthermore, the flexibility and adaptability of permissions management is another major flaw of existing systems. In real-world applications, the sensitivity of data and the administrator's permission requirements may change dynamically depending on the environment and tasks. Existing permission management systems are mostly static and pre-configured, lacking the ability to adaptively adjust administrator permissions and data sensitivity. This can lead to a mismatch between verification strength and data sensitivity in some cases, potentially causing unnecessary verification redundancy or data leakage risks due to overly broad permissions. Existing dynamic permission management mechanisms still rely on relatively simple conditional judgments, making it difficult to adapt to complex operating environments and potential threats in real time.

[0008] Therefore, how to provide a deep learning-based administrator face recognition data export protection method is a problem that technicians in this field urgently need to solve. Summary of the Invention

[0009] One purpose of the present invention is to propose a method for protecting administrator facial recognition data export based on deep learning. The present invention combines multimodal biometric verification, generative adversarial networks, and active induction defense strategies to provide comprehensive protection against security risks in the data export process. The system uses deep learning algorithms for identity authentication and dynamic permission management to ensure that the verification strength matches the data sensitivity. At the same time, it confuses potential attackers by generating false data and operating interfaces, and combines reinforcement learning to dynamically optimize the induction strategy. This method has the advantages of accurate identity authentication, flexible permission management, efficient anomaly detection, and comprehensive active protection, significantly improving the security and robustness of the data export process.

[0010] According to an embodiment of the present invention, a method for protecting administrator face recognition data export based on deep learning includes the following steps:

[0011] S1. When the administrator requests data export, the system initiates multimodal biometric verification and uses a multi-task learning neural network to verify the administrator's identity;

[0012] S2. After identity verification is passed, the system dynamically adjusts the verification strength and complexity based on the administrator's permissions and the sensitivity of the requested data;

[0013] S3. The system monitors the administrator's operational behavior and uses the Transformer model to analyze the operation sequence to detect abnormal behavior or potential attacks.

[0014] S4. Generate potential attack behavior samples through generative adversarial networks, and combine imitation learning technology to learn historical attacker behavior patterns to train and optimize the anomaly detection module;

[0015] S5. When the system detects abnormal behavior or potential attacks, it activates the enhanced active deception module, generates multi-layered false data and false operation interfaces, and uses inductive data manipulation and hierarchical virtualization technology to confuse attackers into entering the virtual environment.

[0016] S6. In a virtual environment, the system generates and optimizes the induction strategy through a proximal policy optimization algorithm combined with a meta-policy gradient algorithm. It dynamically adjusts the induction method based on the attacker's behavior pattern and records all the attacker's operations.

[0017] S7. The backend system records all operation logs and triggers multi-level security policy adjustments based on abnormal situations, including increasing verification strength, locking accounts, or restricting data export permissions, and retaining complete operation records.

[0018] Optionally, the S1 specifically includes:

[0019] S11. When the system starts, it receives an authentication request from the administrator and collects the administrator's multimodal biometric data in real time, including facial images, fingerprints, irises, and voice data;

[0020] S12, inputting the collected multimodal biometric data into a multi-task learning neural network model, wherein the multi-task learning neural network includes a feature extraction layer and a task branching layer;

[0021] S13, the feature extraction layer uses a convolutional neural network to extract features of each biological feature and generate a feature vector;

[0022] S14, the task branch layer performs classification tasks based on the feature vectors of the biological features, the face feature vectors are classified and judged using the fully connected layer, the fingerprint feature vectors are matched using the support vector machine model, the iris feature vectors are compared after adjusting the weights using the attention mechanism, and the voice feature vectors are processed by the long short-term memory network to generate a matching result;

[0023] S15. The biometric verification results are comprehensively judged using a weighted fusion method. The weights are dynamically adjusted based on current environmental factors, including light intensity, sound background, and fingerprint clarity.

[0024] R=w1·r1+w2·r2+…+w n ·r n ;

[0025] Among them, R represents the final authentication result, w i represents the weight of the i-th biometric feature, r i represents the classification result of the i-th biometric feature;

[0026] S16. When the final authentication result R is greater than the threshold T set by the system r When the administrator identity verification is passed; if R≤T r , the verification fails, the administrator's operation request is rejected, and the security alert mechanism is triggered.

[0027] Optionally, the S2 specifically includes:

[0028] S21, the system receives the authority information after the administrator passes the identity authentication, and the authority information includes the administrator's role, authority level, historical operation records and risk assessment of the current operation. The administrator authority matrix is ​​P a ={p ij}, where p ij Indicates that the administrator is in the i-th system operation module at time t j The permission level at the time;

[0029] S22, the system reads the data sensitivity information requested by the administrator to be exported. The data sensitivity changes dynamically according to the data category, encryption level and business importance. The requested data sensitivity matrix is ​​D s ={d kl}, where d kl Indicates that the k-th data at time t l sensitivity;

[0030] S23, the system according to the administrator authority matrix P a and the request data sensitivity matrix D s The multi-dimensional mapping relationship between them is used to calculate the comprehensive adjustment coefficient C:

[0031]

[0032] Among them, γ ik (t j ,t l ) represents the dynamic weight adjustment coefficient between administrator privileges and data sensitivity, p ij Indicates the administrator's authority level in the i-th system module at time t j The permission value when d kl Indicates the sensitivity of the kth category of data requested by the administrator at time t l α represents the nonlinear adjustment coefficient of the permission value, β represents the nonlinear adjustment coefficient of the sensitivity value, λ represents the time attenuation coefficient, n represents the number of system permission modules, and m represents the number of data types.

[0033] S24. Verify the strength based on the comprehensive adjustment coefficient C:

[0034]

[0035] Where V(C) represents the strength verification function, V1 represents low strength verification, V2 represents medium strength verification, V3 represents high strength verification, V4 represents the highest strength verification, and T1, T2, and T3 represent system safety thresholds.

[0036] S25. The system dynamically adjusts the complexity of the verification process based on the dynamic changes in the operating environment over time, and ultimately determines the administrator's data export range;

[0037] S26. When the comprehensive adjustment coefficient C is lower than the set minimum threshold T3, the system rejects the administrator's data export request, triggers an alarm, and records the abnormal operation behavior.

[0038] Optionally, the S3 specifically includes:

[0039] S31. The system receives and processes the operation sequence log generated by the administrator during the data export process. The operation sequence log includes operation time, data access frequency, data type and operation device information. The operation sequence is O={o1, o2, ..., o N}, where o i represents the administrator’s i-th operation record, and N represents the total number of operations;

[0040] S32. Preprocess the operation sequence and extract the time interval of each operation:

[0041] Δt i =t i+1 -t i ;

[0042] Where Δt i Indicates the time interval between the i-th operation and the next operation, t i Indicates the timestamp of the administrator’s i-th operation, t i+1 Indicates the timestamp of the administrator's i+1th operation;

[0043] S33. The system inputs the preprocessed operation sequence into the Transformer-based anomaly detection model. The Transformer model uses a multi-head self-attention mechanism to capture the temporal dependencies in the operation sequence and calculate the attention weights between operations:

[0044]

[0045] in, represents the attention weight between the i-th operation and the j-th operation under the g-th attention head, represents the query vector of the i-th operation under the g-th attention head, represents the key vector of the jth operation under the gth attention head, d q represents the dimension of the vector, η ij Indicates dynamic adjustment of bias term, softmax indicates normalization function;

[0046] S34. Aggregate the operation dependencies under all attention heads into an operation feature vector. The model merges the outputs of multiple heads through a weighted multi-head attention mechanism to generate the final operation feature vector:

[0047]

[0048] Among them, Y i represents the final aggregated feature vector of the i-th operation, G represents the number of attention heads, and V g represents the learnable weight of each attention head, Represents the value vector under the g-th attention head;

[0049] S35, according to Y i and Δt i , the system generates an abnormal behavior score for the operation:

[0050]

[0051] Among them, S op represents the abnormal behavior score of the entire operation sequence, B ij represents the attention weight between operations, κ i represents the adaptive adjustment factor of the i-th operation, ζ represents the time attenuation coefficient, t i and t jIndicates the timestamp of the operation;

[0052] S36, when abnormal behavior score S op Exceeds the set threshold T op When the score is lower than the threshold, the system determines that the operation is abnormal, triggers the protection mechanism and records the abnormal operation log; if the score is lower than the threshold, the operation is considered normal and the system continues to allow the data export operation.

[0053] Optionally, the S4 specifically includes:

[0054] S41, the system extracts normal operation and attack behavior features based on the administrator's operation log and historical attack behavior data. The normal behavior feature set is N = {n1, n2, ..., n p}, where n i represents the normal behavior feature of the i-th type, and the attack behavior feature set is A={a1,a2,…,a q}, where a j represents the characteristics of the j-th attack behavior;

[0055] S42. The system inputs the attack behavior features into the Generative Adversarial Network. The Generative Adversarial Network consists of a generator G and a discriminator D. The generator generates potential attack behavior samples G(z) based on random noise z0:

[0056] G(z)=G(θ g ,z0);

[0057] Among them, θ g Represents the learnable parameters of the generator, z0 represents the random noise vector input to the generator;

[0058] S43, the imitation learning part uses the historical attack behavior dataset P mimic The characteristic operation mode M={m1,m2,…,m r} to train the generator so that its output G(z) is close to the real attack pattern:

[0059]

[0060] Among them, L G represents the loss function of the generator, E represents the mathematical expectation, P z represents the random noise distribution, D(G(z)) represents the output probability of the discriminator for the generated sample, λ g Represents the adjustment coefficient of imitation learning, ∥G(z)-m∥ 2 represents the difference between the generated sample and the imitated attack behavior, and m represents the historical attack behavior pattern;

[0061] S44, the discriminator D is used to distinguish between real attack behaviors and fake attack behaviors. The optimization goal is to maximize the recognition probability of real attack samples:

[0062]

[0063] Among them, L D represents the loss function of the discriminator, P data represents the distribution of real attack behaviors, and x represents real attack samples;

[0064] S45. By optimizing the generation of potential attack behavior samples through a generative adversarial network, the system continuously trains the anomaly detection module. The anomaly detection module improves its ability to detect complex and covert attack behaviors by learning from the generated attack samples. When a potential attack behavior is detected, the protection mechanism is triggered, the attack operation is recorded, and the model is updated.

[0065] S46. Based on the continuously updated operation log, the system continuously optimizes the generator's generation capabilities by combining generative adversarial networks with imitation learning technology.

[0066] Optionally, the S5 specifically includes:

[0067] S51. When the system detects abnormal behavior or potential attack, it starts the enhanced active deception module to generate multi-level false data and false operation interface. The false data feature set is F = {f1, f2, ..., f m}, where f i Represents the characteristics of the i-th type of false data:

[0068] f i =g(d i ,θ f );

[0069] Among them, d i represents the true data feature of the i-th category, θ f represents the transformation parameters for generating false data, g(·) represents the data transformation function;

[0070] S52, the system generates a fake operation interface through virtualization technology to induce the attacker to operate. The set of virtual operation interfaces is V = {v1, v2, ..., v k}, where v j Represents the jth virtual operation interface:

[0071] v j =h(u j ,θ v );

[0072] Among them, u j represents the operation behavior of the jth attacker, θ vrepresents the transformation parameter for generating the fake interface, h(·) represents the mapping function between the attacker's operation and the fake interface;

[0073] S53, the system dynamically generates induced operations through the induced data operation mechanism, and the induced operation set is I={i1,i2,…,i e}, where i l Indicates the first induction operation and calculates the ineffectiveness score of the induction operation:

[0074]

[0075] Among them, S i represents the invalidity score of the induced operation, e represents the total number of induced operations, α l represents the weight of the first induction operation, Indicates the result of the attacker's first fake operation in a fake environment. Indicates the first real operation result, μ l Represents the adaptive adjustment coefficient, Δt l Indicates the time interval between the first operation and the last operation, γ l represents the time attenuation coefficient;

[0076] S54, the system scores S based on the ineffectiveness of the induced operation i Dynamically adjust the complexity of fake data and fake interfaces, and dynamically optimize the fake environment as the score increases:

[0077]

[0078] Among them, C f represents the complexity of the false environment, m represents the total number of false data categories, ρ i represents the weight of the false data of type i, δ i represents the adaptive complexity adjustment coefficient, and T represents the scoring threshold set by the system;

[0079] S55, the system dynamically adjusts the characteristics of the fake environment according to the attacker's behavior and operation records, and records the operation logs, and compares the attacker's fake operation records. and corresponding real operation records Calculate the matching score M r :

[0080]

[0081] Among them, θ q represents the weight of the qth operation.

[0082] Optionally, the S6 specifically includes:

[0083] S61. The system monitors the attacker's operational behavior in a virtual environment and collects behavioral data in real time. The system initializes the induction strategy based on the operation sequence and uses a proximal strategy optimization algorithm to generate the initial strategy.

[0084] S62. The system dynamically optimizes the initial strategy through the meta-policy gradient algorithm. The meta-policy gradient algorithm adjusts the induced strategy in real time according to the attacker's operation mode:

[0085]

[0086] Among them, L Meta (θ) represents the loss function of the meta-policy gradient algorithm, θ k represents the parameters of the k-th strategy, represents the kth strategy, Represents the parameter θ k The gradient, E represents the mathematical expectation, R t represents the reward function at time step t, and T represents the total length of the time step;

[0087] S63. The system continuously optimizes the inducement strategy based on the feedback from the attacker's operation behavior:

[0088]

[0089] Among them, F t represents the feedback value at time step t, r i represents the reward value of the attacker's ith operation in the fake environment, λ3 represents the time interval sensitivity coefficient, γ f represents the discount factor, t i -t i-1 Indicates the time difference between two operations;

[0090] S64. Based on the feedback function and the results of meta-policy gradient optimization, the system dynamically adjusts the induction strategy:

[0091]

[0092] Among them, θ′ represents the updated policy parameters, θ represents the policy parameters before the update, and τ represents the learning rate;

[0093] S65. The system dynamically adjusts the inducement operation in the fake environment based on the optimized policy parameters. According to the attacker's operation mode, the system generates a new inducement interface and data.

[0094] The beneficial effects of the present invention are:

[0095] First, the present invention significantly improves the security and accuracy of identity authentication through multimodal biometric verification. Traditional identity authentication typically relies on a single biometric feature, which is vulnerable to counterfeit attacks or specific environmental restrictions. However, the present invention uses a multi-task learning neural network to simultaneously verify multiple biometric features of the administrator, such as face, fingerprint, iris, and voice, and dynamically adjusts the weight of each biometric feature based on changes in the current environment. In this way, the system can still ensure efficient and accurate identity authentication under adverse conditions such as low light, sound interference, or blurred fingerprints, enhancing the robustness and flexibility of the verification process.

[0096] Secondly, the dynamic permission management mechanism of the present invention further ensures security during the data export process. Compared with the static, pre-set permission management mode in the prior art, the present invention uses a deep learning algorithm to calculate the verification strength and complexity in real time based on the administrator's permission information and the sensitivity of the requested data, ensuring that the administrator's operating permissions match the sensitivity level of the data. This method of dynamically adjusting the verification strength can effectively prevent administrators from unauthorized access to highly sensitive data, thereby reducing the risk of data leakage. At the same time, the system can flexibly adjust the permission verification process according to changes in the operating environment, so that security and operational convenience are well balanced.

[0097] In terms of anomaly detection, the present invention significantly improves the ability to detect potential attack behaviors by introducing generative adversarial networks (GANs) and imitation learning technology. Traditional anomaly detection often relies on preset rules or simple behavioral analysis models, which makes it difficult to deal with complex and covert attack methods. The present invention generates forged samples of potential attack behaviors through generative adversarial networks, and further optimizes the training process of the anomaly detection module by imitating the operating modes of historical attackers. As a result, the system can more accurately identify attack behaviors that attempt to imitate normal operations, ensuring that any potential threats can be identified and processed in a timely manner during the data export process.

[0098] In addition, a notable feature of the present invention is the introduction of an enhanced active deception module. When the system detects that an attacker has potential abnormal behavior, it does not simply block or alarm. The system will generate multi-level false data and false operation interfaces to induce the attacker to enter the virtual environment. This active deception mechanism generates false data and operation processes, causing the attacker to mistakenly believe that he has obtained valuable information, thereby preventing him from further trying to access real data. By dynamically adjusting the complexity of the virtual environment, the system can keep the attacker confused in the false data, and by recording his behavior in the virtual environment, it provides an important reference for the subsequent optimization of protection strategies. This mechanism not only enhances the security of the data, but also avoids the problem of alerting the attacker in traditional protection measures.

[0099] Furthermore, the present invention's proximal policy optimization algorithm, combined with a meta-policy gradient algorithm, dynamically generates and optimizes decoy strategies within a deceptive environment, enabling the system to make real-time adjustments based on the attacker's behavioral patterns, enhancing defensive flexibility. The system continuously learns and optimizes through feedback from the attacker's operational behavior, ensuring the continued effectiveness of the decoy strategy. This feature makes the system highly adaptable to unknown or complex attacks, enabling it to adopt the most appropriate protective measures based on the specific attack behavior.

[0100] Finally, the multi-level security policy adjustment mechanism of this invention effectively addresses security threats at different levels. Based on the administrator's operation log and abnormal behavior detection results, the system triggers corresponding security policies. From simple operation logging to enhanced verification strength, all the way to account locking or data export restriction, it responds to different levels of threats step by step, ensuring that the system effectively protects data security under all circumstances. This multi-level response mechanism is relatively rare in the existing technology and effectively fills the gap in security management systems in dynamically responding to complex attack behaviors. BRIEF DESCRIPTION OF THE DRAWINGS

[0101] 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:

[0102] Figure 1 This is a flowchart of a deep learning-based administrator face recognition data export protection method proposed by the present invention;

[0103] Figure 2 This is a schematic diagram of the structure of a multimodal biometric verification method for administrator face recognition data export protection based on deep learning proposed by the present invention;

[0104] Figure 3 This is a schematic diagram of the generation of false data and false operation interface and active induction strategy for the deep learning-based administrator face recognition data export protection method proposed by the present invention. DETAILED DESCRIPTION

[0105] The present invention will now be described in further detail with reference to the accompanying drawings, which are simplified schematic diagrams that illustrate the basic structure of the present invention in a schematic manner.

[0106] refer to Figure 1-3 , a method for protecting administrator face recognition data export based on deep learning, comprising the following steps:

[0107] S1. When the administrator requests data export, the system initiates multimodal biometric verification and uses a multi-task learning neural network to verify the administrator's identity;

[0108] S2. After identity verification is passed, the system dynamically adjusts the verification strength and complexity based on the administrator's permissions and the sensitivity of the requested data;

[0109] S3. The system monitors the administrator's operational behavior and uses the Transformer model to analyze the operation sequence to detect abnormal behavior or potential attacks.

[0110] S4. Generate potential attack behavior samples through generative adversarial networks, and combine imitation learning technology to learn historical attacker behavior patterns to train and optimize the anomaly detection module;

[0111] S5. When the system detects abnormal behavior or potential attacks, it activates the enhanced active deception module, generates multi-layered false data and false operation interfaces, and uses inductive data manipulation and hierarchical virtualization technology to confuse attackers into entering the virtual environment.

[0112] S6. In a virtual environment, the system generates and optimizes the induction strategy through a proximal policy optimization algorithm combined with a meta-policy gradient algorithm. It dynamically adjusts the induction method based on the attacker's behavior pattern and records all the attacker's operations.

[0113] S7. The backend system records all operation logs and triggers multi-level security policy adjustments based on abnormal situations, including increasing verification strength, locking accounts, or restricting data export permissions, and retaining complete operation records.

[0114] In this embodiment, S1 specifically includes:

[0115] S11. When the system starts, it receives an authentication request from the administrator and collects the administrator's multimodal biometric data in real time, including facial images, fingerprints, irises, and voice data;

[0116] S12, inputting the collected multimodal biometric data into a multi-task learning neural network model, wherein the multi-task learning neural network includes a feature extraction layer and a task branching layer;

[0117] S13, the feature extraction layer uses a convolutional neural network to extract features of each biological feature and generate a feature vector;

[0118] S14, the task branch layer performs classification tasks based on the feature vectors of the biological features, the face feature vectors are classified and judged using the fully connected layer, the fingerprint feature vectors are matched using the support vector machine model, the iris feature vectors are compared after adjusting the weights using the attention mechanism, and the voice feature vectors are processed by the long short-term memory network to generate a matching result;

[0119] S15. The biometric verification results are comprehensively judged using a weighted fusion method. The weights are dynamically adjusted based on current environmental factors, including light intensity, sound background, and fingerprint clarity.

[0120] R=w1·r1+w2·r2+…+w n ·r n ;

[0121] Among them, R represents the final authentication result, w i represents the weight of the i-th biometric feature, r i represents the classification result of the i-th biometric feature;

[0122] S16. When the final authentication result R is greater than the threshold T set by the system r When the administrator identity verification is passed; if R≤T r , the verification fails, the administrator's operation request is rejected, and the security alert mechanism is triggered.

[0123] In this embodiment, S2 specifically includes:

[0124] S21, the system receives the authority information after the administrator passes the identity authentication, and the authority information includes the administrator's role, authority level, historical operation records and risk assessment of the current operation. The administrator authority matrix is ​​P a ={p ij}, where p ij Indicates that the administrator is in the i-th system operation module at time t j The permission level at the time;

[0125] S22, the system reads the data sensitivity information requested by the administrator to be exported. The data sensitivity changes dynamically according to the data category, encryption level and business importance. The requested data sensitivity matrix is ​​D s ={d kl}, where d kl Indicates that the k-th data at time t l sensitivity;

[0126] S23, the system according to the administrator authority matrix P a and the request data sensitivity matrix D s The multi-dimensional mapping relationship between them is used to calculate the comprehensive adjustment coefficient C:

[0127]

[0128] Among them, γ ik (t j ,t l ) represents the dynamic weight adjustment coefficient between administrator privileges and data sensitivity, p ij Indicates the administrator's authority level in the i-th system module at time t j The permission value when d kl Indicates the sensitivity of the kth category of data requested by the administrator at time t l α represents the nonlinear adjustment coefficient of the permission value, β represents the nonlinear adjustment coefficient of the sensitivity value, λ represents the time attenuation coefficient, n represents the number of system permission modules, and m represents the number of data types.

[0129] S24. Verify the strength based on the comprehensive adjustment coefficient C:

[0130]

[0131] Where V(C) represents the strength verification function, V1 represents low strength verification, V2 represents medium strength verification, V3 represents high strength verification, V4 represents the highest strength verification, and T1, T2, and T3 represent system safety thresholds.

[0132] S25. The system dynamically adjusts the complexity of the verification process based on the dynamic changes in the operating environment over time, and ultimately determines the administrator's data export range;

[0133] S26. When the comprehensive adjustment coefficient C is lower than the set minimum threshold T3, the system rejects the administrator's data export request, triggers an alarm, and records the abnormal operation behavior.

[0134] In this embodiment, S3 specifically includes:

[0135] S31. The system receives and processes the operation sequence log generated by the administrator during the data export process. The operation sequence log includes operation time, data access frequency, data type and operation device information. The operation sequence is O={o1, o2, ..., o N}, where o i represents the administrator’s i-th operation record, and N represents the total number of operations;

[0136] S32. Preprocess the operation sequence and extract the time interval of each operation:

[0137] Δt i =t i+1 -t i ;

[0138] Where Δt i Indicates the time interval between the i-th operation and the next operation, t i Indicates the timestamp of the administrator’s i-th operation, t i+1 Indicates the timestamp of the administrator's i+1th operation;

[0139] S33. The system inputs the preprocessed operation sequence into the Transformer-based anomaly detection model. The Transformer model uses a multi-head self-attention mechanism to capture the temporal dependencies in the operation sequence and calculate the attention weights between operations:

[0140]

[0141] in, represents the attention weight between the i-th operation and the j-th operation under the g-th attention head, represents the query vector of the i-th operation under the g-th attention head, represents the key vector of the jth operation under the gth attention head, d q represents the dimension of the vector, η ij Indicates dynamic adjustment of bias term, softmax indicates normalization function;

[0142] S34. Aggregate the operation dependencies under all attention heads into an operation feature vector. The model merges the outputs of multiple heads through a weighted multi-head attention mechanism to generate the final operation feature vector:

[0143]

[0144] Among them, Y i represents the final aggregated feature vector of the i-th operation, G represents the number of attention heads, and V g represents the learnable weight of each attention head, Represents the value vector under the g-th attention head;

[0145] S35, according to Y i and Δt i , the system generates an abnormal behavior score for the operation:

[0146]

[0147] Among them, S op represents the abnormal behavior score of the entire operation sequence, B ij represents the attention weight between operations, κ i represents the adaptive adjustment factor of the i-th operation, ζ represents the time attenuation coefficient, t i and tj Indicates the timestamp of the operation;

[0148] S36, when abnormal behavior score S op Exceeds the set threshold T op When the score is lower than the threshold, the system determines that the operation is abnormal, triggers the protection mechanism and records the abnormal operation log; if the score is lower than the threshold, the operation is considered normal and the system continues to allow the data export operation.

[0149] In this embodiment, the S4 specifically includes:

[0150] S41, the system extracts normal operation and attack behavior features based on the administrator's operation log and historical attack behavior data. The normal behavior feature set is N = {n1, n2, ..., n p}, where n i represents the normal behavior feature of the i-th type, and the attack behavior feature set is A={a1,a2,…,a q}, where a j represents the characteristics of the j-th attack behavior;

[0151] S42. The system inputs the attack behavior features into the Generative Adversarial Network. The Generative Adversarial Network consists of a generator G and a discriminator D. The generator generates potential attack behavior samples G(z) based on random noise z0:

[0152] G(z)=G(θ g ,z0);

[0153] Among them, θ g Represents the learnable parameters of the generator, z0 represents the random noise vector input to the generator;

[0154] S43, the imitation learning part uses the historical attack behavior dataset P mimic The characteristic operation mode M={m1,m2,…,m r} to train the generator so that its output G(z) is close to the real attack pattern:

[0155]

[0156] Among them, L G represents the loss function of the generator, E represents the mathematical expectation, P z represents the random noise distribution, D(G(z)) represents the output probability of the discriminator for the generated sample, λ g Represents the adjustment coefficient of imitation learning, ∥G(z)-m∥ 2 represents the difference between the generated sample and the imitated attack behavior, and m represents the historical attack behavior pattern;

[0157] S44, the discriminator D is used to distinguish between real attack behaviors and fake attack behaviors. The optimization goal is to maximize the recognition probability of real attack samples:

[0158]

[0159] Among them, L D represents the loss function of the discriminator, P data represents the distribution of real attack behaviors, and x represents real attack samples;

[0160] S45. By optimizing the generation of potential attack behavior samples through a generative adversarial network, the system continuously trains the anomaly detection module. The anomaly detection module improves its ability to detect complex and covert attack behaviors by learning from the generated attack samples. When a potential attack behavior is detected, the protection mechanism is triggered, the attack operation is recorded, and the model is updated.

[0161] S46. Based on the continuously updated operation log, the system continuously optimizes the generator's generation capabilities by combining generative adversarial networks with imitation learning technology.

[0162] Optionally, the S5 specifically includes:

[0163] S51. When the system detects abnormal behavior or potential attack, it starts the enhanced active deception module to generate multi-level false data and false operation interface. The false data feature set is F = {f1, f2, ..., f m}, where f i Represents the characteristics of the i-th type of false data:

[0164] f i =g(d i ,θ f );

[0165] Among them, d i represents the true data feature of the i-th category, θ f represents the transformation parameters for generating false data, g(·) represents the data transformation function;

[0166] S52, the system generates a fake operation interface through virtualization technology to induce the attacker to operate. The set of virtual operation interfaces is V = {v1, v2, ..., v k}, where v j Represents the jth virtual operation interface:

[0167] v j =h(u j ,θ v );

[0168] Among them, u j represents the operation behavior of the jth attacker, θ vrepresents the transformation parameter for generating the fake interface, h(·) represents the mapping function between the attacker's operation and the fake interface;

[0169] S53, the system dynamically generates induced operations through the induced data operation mechanism, and the induced operation set is I={i1,i2,…,i e}, where i l Indicates the first induction operation and calculates the ineffectiveness score of the induction operation:

[0170]

[0171] Among them, S i represents the invalidity score of the induced operation, e represents the total number of induced operations, α l represents the weight of the first induction operation, Indicates the result of the attacker's first fake operation in a fake environment. Indicates the first real operation result, μ l Represents the adaptive adjustment coefficient, Δt l Indicates the time interval between the first operation and the last operation, γ l represents the time decay coefficient;

[0172] S54, the system scores S based on the ineffectiveness of the induced operation i Dynamically adjust the complexity of fake data and fake interfaces, and dynamically optimize the fake environment as the score increases:

[0173]

[0174] Among them, C f represents the complexity of the false environment, m represents the total number of false data categories, ρ i represents the weight of the false data of type i, δ i represents the adaptive complexity adjustment coefficient, and T represents the scoring threshold set by the system;

[0175] S55, the system dynamically adjusts the characteristics of the fake environment according to the attacker's behavior and operation records, and records the operation logs, and compares the attacker's fake operation records. and corresponding real operation records Calculate the matching score M r :

[0176]

[0177] Among them, θ q represents the weight of the qth operation.

[0178] In this embodiment, S6 specifically includes:

[0179] S61. The system monitors the attacker's operational behavior in a virtual environment and collects behavioral data in real time. The system initializes the induction strategy based on the operation sequence and uses a proximal strategy optimization algorithm to generate the initial strategy.

[0180] S62. The system dynamically optimizes the initial strategy through the meta-policy gradient algorithm. The meta-policy gradient algorithm adjusts the induced strategy in real time according to the attacker's operation mode:

[0181]

[0182] Among them, L Meta (θ) represents the loss function of the meta-policy gradient algorithm, θ k represents the parameters of the k-th strategy, represents the kth strategy, Represents the parameter θ k The gradient, E represents the mathematical expectation, R t represents the reward function at time step t, and T represents the total length of the time step;

[0183] S63. The system continuously optimizes the inducement strategy based on the feedback from the attacker's operation behavior:

[0184]

[0185] Among them, F t represents the feedback value at time step t, r i represents the reward value of the attacker's ith operation in the fake environment, λ3 represents the time interval sensitivity coefficient, γ f represents the discount factor, t i -t i-1 Indicates the time difference between two operations;

[0186] S64. Based on the feedback function and the results of meta-policy gradient optimization, the system dynamically adjusts the induction strategy:

[0187]

[0188] Among them, θ′ represents the updated policy parameters, θ represents the policy parameters before the update, and τ represents the learning rate;

[0189] S65. The system dynamically adjusts the inducement operation in the fake environment based on the optimized policy parameters. According to the attacker's operation mode, the system generates a new inducement interface and data.

[0190] Example 1:

[0191] To verify the feasibility of the present invention, it was applied to a large financial institution. This financial institution has repeatedly faced the risk of internal administrators exporting highly sensitive data without sufficient verification or with incorrect permission control. In particular, during the permission management and data export process, existing security protection systems are often unable to effectively prevent potential unauthorized operations by internal personnel or external malicious attacks.

[0192] At this financial institution, administrators are responsible for exporting sensitive data from various company departments and regularly uploading it to the headquarters server for analysis. This operation involves multiple key steps, including identity authentication, data permission control, and the export of sensitive data. However, traditional single-authentication mechanisms and static permission management methods are unable to adapt to the dynamically changing business needs and security threats in the financial environment. Therefore, the financial institution decided to introduce the deep learning-based administrator facial recognition data export protection method of the present invention to address security vulnerabilities in current identity authentication, dynamic permission management, and abnormal operation detection.

[0193] After applying the invention, the system first performs multimodal biometric verification on the administrator's identity. Whenever the administrator requests to export data, the system collects the administrator's facial image, fingerprint, iris and voice features through the camera, and inputs these features into the multi-task learning neural network for verification in real time. In a typical business operation scenario, an administrator needs to export the transaction data of a VIP customer. During the operation, the system extracts features from the facial image, fingerprint and iris, and classifies them through a convolutional neural network. At the same time, based on environmental changes (such as insufficient light or sound interference), the system dynamically adjusts the weights of different modal biometrics. For example, when the light is poor, fingerprint and iris features will have a larger weight, and when the facial features are interfered with, the weight of the voice features will increase. This multimodal fusion verification method effectively prevents attacks by forged biometrics.

[0194] By analyzing the test data of the financial institution, the system processed a total of 632 administrator data export requests within a week. Among them, 45 verification requests would result in verification failure under single-modal verification conditions due to fuzzy or unrecognizable biometric features. However, after using the multimodal weighted fusion verification technology of the present invention, the system successfully passed 38 verifications in complex environments, with a verification success rate of 84.4%. Before the introduction of this invention, the pass rate of traditional single biometric verification methods in similar scenarios was only 65.8%. This data fully demonstrates the effectiveness of multimodal verification, especially in harsh environmental conditions, which significantly improves the robustness and accuracy of verification.

[0195] Once the administrator has passed the verification, the system will dynamically adjust the verification strength based on their permissions and the sensitivity of the data being requested to be exported. In the authority management system of the financial institution, the administrator's permission level and the sensitivity of the exported data will be mapped into a permission matrix and a data sensitivity matrix. For example, when exporting low-sensitivity data (such as ordinary customer data), the system only requires double verification of face and fingerprint. However, when it comes to exporting highly sensitive data (such as VIP customer transaction data or financial statements), the system will require the administrator to pass multiple verifications, including comprehensive verification of face, fingerprint, iris and voice. Through the reinforcement learning mechanism, the system can also automatically adjust the verification strength according to the risk level of the operation. For example, during a VIP customer transaction data export operation, the system detected a deviation between the administrator's operation log and the historical record. The system proactively increased the verification strength, upgrading from double verification to multiple verification to ensure the security of the data export process.

[0196] During a month-long data test, 165 requests were made to export highly sensitive data. The system dynamically adjusted verification strength based on data sensitivity, automatically upgrading verification in 37 cases and preventing potential unauthorized access. Traditional static permission management systems are nearly incapable of dynamically adjusting verification complexity in similar environments, which can easily lead to overly broad or narrow permissions, increasing security risks. This system's dynamic permission adjustment mechanism makes the export of highly sensitive data more secure, demonstrating its superior performance in sensitive data management within financial institutions.

[0197] In addition, the system generates potential attack samples in the background by combining generative adversarial network (GAN) technology with imitation learning technology, and trains the anomaly detection module with data from historical attack behaviors. When the system detects potential abnormal behavior, it automatically generates false data and false operation interfaces to confuse the attacker. In a simulated network attack, the attacker attempted to disguise himself as an internal administrator to export highly sensitive data. The system analyzed the attacker's operational behavior, identified abnormal behavior, and automatically generated false transaction data and a false export interface to lure the attacker into a virtual environment. The attacker mistakenly believed that he had successfully obtained real data, but all operations in the virtual environment were recorded and used as a basis for subsequent optimization of protection strategies. After testing, the system successfully induced the attacker into a false environment in 15 simulated attack scenarios and recorded all operational behaviors. The introduction of this proactive defense strategy effectively prevents the risk of external malicious attackers obtaining real data.

[0198] Data analysis following the implementation of this invention has significantly improved data export protection for financial institutions. During a one-month application period, the system processed 1,021 administrator data export requests, successfully averting 11 potential data breaches. Through multiple measures, including proactive induction, anomaly detection, and dynamic permission adjustments, the system minimizes security threats during data export, significantly improving the security and reliability of financial institutions' data management.

[0199] Table 1 Comparison of data export security between the traditional system and the system of the present invention

[0200] Comparison Item Traditional systems System of the present invention Total verification requests 632 632 Single modality verification success rate 65.8％ - Multimodal verification success rate - 84.4％ Number of privilege escalations 0 37 Number of potential data breaches 11 0 Number of simulated attacks successfully blocked 5 15

[0201] Table 1 above clearly demonstrates the significant improvement of the proposed system over the traditional system in several key metrics. First, the total number of verification requests in the table remains consistent; both systems processed 632 administrator data export requests under the same verification environment. However, in terms of verification success rate, the proposed system significantly outperformed the traditional system.

[0202] For traditional systems, the verification process relies on single-modal verification technology, with a success rate of only 65.8%. This verification method is prone to verification failure when encountering unfavorable environments (such as insufficient light, unclear fingerprints, etc.). In the system of the present invention, multimodal biometric verification is adopted, and multiple biometric features such as face, fingerprint, iris, and voice are weighted and fused. The weight of each feature is dynamically adjusted through a deep learning algorithm, successfully improving the verification success rate to 84.4%. This shows that the identity authentication of the present invention in complex environments is more reliable, and the verification robustness has been significantly enhanced.

[0203] Regarding the number of privilege escalations, the traditional system lacked the ability to dynamically adjust verification strength based on data sensitivity. Consequently, verification strength remained unchanged for all 632 operations. In contrast, the proposed system, leveraging reinforcement learning to assess data sensitivity and operational risk in real time, automatically upgraded verification strength for 37 high-risk data export requests. This mechanism effectively prevented administrators from exporting highly sensitive data through unauthorized operations, ensuring data security.

[0204] Another key difference is the number of potential data breaches. The traditional system detected 11 potential data breaches in a month, while the system we introduced detected zero. This result is due to the dynamic permissions management and abnormal operation detection mechanisms we introduced. The system not only adjusts authentication strength in real time but also detects abnormal behavior and proactively prevents potential data breaches.

[0205] The number of successful blocks of simulated attacks is also a significantly improved indicator. The traditional system only successfully blocked 5 attacks when facing simulated attacks, while the system of the present invention successfully blocked 15 simulated attacks by generating adversarial networks (GANs) and actively inducing defense strategies. The present invention generates false data and false operation interfaces to induce attackers to enter a virtual environment and record their operation behaviors, thereby effectively preventing attackers from obtaining real data. This mechanism shows that the present invention can more proactively defend and confuse attackers when facing complex external attacks, greatly improving the security of the system.

[0206] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A method for protecting administrator face recognition data export based on deep learning. It is characterized in that The steps include: S1. When the administrator requests data export, the system starts multimodal biometric verification and uses a multi-task learning neural network to verify the administrator's identity; S2. After identity verification is passed, the system dynamically adjusts the verification strength and complexity based on the administrator's permissions and the sensitivity of the requested data; S3, the system monitors the administrator's operation behavior, uses the Transformer model to analyze the operation sequence, and detects whether there is abnormal behavior or potential attack; S4. Generate potential attack behavior samples through generative adversarial networks, and combine imitation learning technology to learn historical attacker behavior patterns, train and optimize the anomaly detection module; S5. When the system detects abnormal behavior or potential attacks, it starts the enhanced active deception module to generate multi-level false data and false operation interfaces, and confuses attackers into entering the virtual environment through inductive data operations and hierarchical virtualization technology; S6. In a virtual environment, the system generates and optimizes the induction strategy through the proximal policy optimization algorithm combined with the meta-policy gradient algorithm, dynamically adjusts the induction method according to the attacker's behavior pattern, and records all the attacker's operations; S7. The backend system records all operation logs and triggers multi-level security policy adjustments based on abnormal situations, including increasing verification strength, locking accounts or restricting data export permissions, and retaining complete operation records.

2. According to the deep learning-based administrator face recognition data export protection method of claim 1, It is characterized in that The S1 specifically includes: S11. When the system starts, it receives the administrator's identity authentication request and collects the administrator's multimodal biometric data in real time, including facial images, fingerprints, irises, and voice data; S12, inputting the collected multimodal biometric data into a multi-task learning neural network model, wherein the multi-task learning neural network includes a feature extraction layer and a task branching layer; S13, the feature extraction layer uses a convolutional neural network to extract features from each biological feature and generate a feature vector; S14, the task branch layer performs classification tasks according to the feature vectors of biological features, the face feature vectors are classified and judged using the fully connected layer, the fingerprint feature vectors are matched using the support vector machine model, the iris feature vectors are compared after adjusting the weights through the attention mechanism, and the voice feature vectors are processed through the long short-term memory network to generate matching results; S15. The verification results of the biometric features are comprehensively judged by a weighted fusion method. The weights are dynamically adjusted according to the current environmental factors. The environmental factors include light intensity, sound background and fingerprint clarity: R=w 1 ·r 1 +w 2 ·r 2 +…+w n ·r n ; Among them, R represents the final authentication result, w i represents the weight of the i-th biometric feature, r i represents the classification result of the i-th biometric feature; S16. When the final identity verification result R is greater than the threshold T set by the system r When the administrator identity verification is passed; if R≤T r , the verification fails, the administrator's operation request is rejected, and the security alert mechanism is triggered.

3. According to the deep learning-based administrator face recognition data export protection method of claim 1, It is characterized in that The S2 specifically includes: S21, the system receives the permission information of the administrator after the identity authentication is passed, and the permission information includes the administrator's role, permission level, historical operation records and risk assessment of the current operation. The administrator permission matrix is ​​P a ={p ij }, where p ij Indicates that the administrator is in the i-th system operation module at time t j The permission level at the time; S22, the system reads the data sensitivity information requested by the administrator to be exported. The data sensitivity changes dynamically according to the data category, encryption level and business importance. The requested data sensitivity matrix is ​​D s ={d kl }, where d kl Indicates that the k-th data at time t l sensitivity; S23, the system according to the administrator authority matrix P a and the request data sensitivity matrix D s The multi-dimensional mapping relationship between them is used to calculate the comprehensive adjustment coefficient C: Among them, γ ik (t j ,t l ) represents the dynamic weight adjustment coefficient between administrator privileges and data sensitivity, p ij represents the administrator's permission level in the i-th system module at time t j The permission value when d kl represents the sensitivity of the kth category of data requested by the administrator at time t l The data sensitivity value at that time, α represents the nonlinear adjustment coefficient of the permission value, β represents the nonlinear adjustment coefficient of the sensitivity value, λ represents the time attenuation coefficient, n represents the number of system permission modules, and m represents the number of data types; S24. Verify the strength according to the comprehensive adjustment coefficient C: Among them, V(C) represents the strength verification function, V 1 Indicates low-intensity verification, V 2 Indicates medium strength verification, V 3 Indicates high-intensity verification, V 4 Indicates the highest strength verification, T 1 , T 2 and T 3 Indicates the system safety threshold; S25. The system dynamically adjusts the complexity of the verification process according to the dynamic changes of the operating environment over time, and finally determines the data export scope of the administrator; S26, when the comprehensive adjustment coefficient C is lower than the set minimum threshold T 3 When the system rejects the administrator's request to export data, it triggers an alarm and records the abnormal operation behavior.

4. According to the deep learning-based administrator face recognition data export protection method described in claim 1, It is characterized in that The S3 specifically includes: S31, the system receives and processes the operation sequence log generated by the administrator during the data export process, the operation sequence log includes the operation time, data access frequency, data type and operation device information, the operation sequence is O={o 1 ,o 2 ,…,o N }, where o i represents the administrator’s i-th operation record, and N represents the total number of operations; S32, pre-process the operation sequence and extract the time interval of each operation: Δt i =t i+1 -t i ; Among them, Δt i represents the time interval between the i-th operation and the next operation, t i represents the timestamp of the administrator’s i-th operation, t i+1 Indicates the timestamp of the administrator's i+1th operation; S33. The system inputs the preprocessed operation sequence into the Transformer-based anomaly detection model. The Transformer model uses a multi-head self-attention mechanism to capture the temporal dependencies in the operation sequence and calculate the attention weights between operations: in, represents the attention weight between the i-th operation and the j-th operation under the g-th attention head, represents the query vector of the i-th operation under the g-th attention head, represents the key vector of the jth operation under the gth attention head, d q represents the dimension of the vector, η ij Indicates dynamic adjustment of bias term, and softmax indicates normalization function; S34. Aggregate the operation dependencies under all attention heads into an operation feature vector. The model merges the outputs of multiple heads through a weighted multi-head attention mechanism to generate the final operation feature vector: Among them, Y i represents the final aggregated feature vector of the ith operation, G represents the number of attention heads, and V g represents the learnable weight of each attention head, Represents the value vector under the g-th attention head; S35. According to Y i and Δt i , the system generates an operational abnormal behavior score: Among them, S op represents the abnormal behavior score of the entire operation sequence, B ij represents the attention weight between operations, κ i represents the adaptive adjustment factor of the ith operation, ζ represents the time decay coefficient, t i and t j Indicates the timestamp of the operation; S36, when abnormal behavior score S op Exceeding the set threshold T op When the score is lower than the threshold, the system determines that the operation is abnormal, triggers the protection mechanism and records the abnormal operation log; if the score is lower than the threshold, the operation is considered normal and the system continues to allow data export operations.

5. According to the deep learning-based administrator face recognition data export protection method of claim 1, It is characterized in that The S4 specifically includes: S41. The system extracts normal operation and attack behavior features based on the administrator's operation log and historical attack behavior data. The normal behavior feature set is N = {n 1 ,n 2 ,…,n p }, where n i represents the i-th normal behavior feature, and the attack behavior feature set is A = {a 1 ,a 2 ,…,a q }, where a j represents the characteristics of the j-th attack behavior; S42, the system inputs the attack behavior features into the generative adversarial network, which consists of a generator G and a discriminator D. The generator generates a random noise z 0 Generate potential attack behavior samples G(z): G(z))=G(θ g ,With 0 ); Among them, θ g represents the learnable parameters of the generator, z 0 A random noise vector representing the input to the generator; S43, the imitation learning part uses the historical attack behavior dataset P mimic The characteristic operation mode M in 1 ,m 2 ,…,m r } to train the generator’s output G(z) to be close to the real attack pattern: Among them, L G represents the loss function of the generator, E represents the mathematical expectation, P z represents the random noise distribution, D(G(z)) represents the output probability of the discriminator for the generated sample, λ g represents the adjustment coefficient of imitation learning, ||G(z)-m|| 2 represents the difference between the generated sample and the imitated attack behavior, and m represents the historical attack behavior pattern; S44, discriminator D is used to distinguish between real attack behaviors and fake attack behaviors, and the optimization goal is to maximize the recognition probability of real attack samples: Among them, L D represents the loss function of the discriminator, P data represents the distribution of real attack behaviors, and x represents real attack samples; S45. Generate and optimize potential attack behavior samples through generative adversarial networks. The system continuously trains the anomaly detection module. The anomaly detection module improves its ability to detect complex and hidden attack behaviors by learning the generated attack samples. When a potential attack behavior is detected, the protection mechanism is triggered to record the attack operation and update the model. S46. Based on the continuously updated operation log, the system continuously optimizes the generation capability of the generator by combining generative adversarial networks with imitation learning technology.

6. According to the deep learning-based administrator face recognition data export protection method of claim 1, It is characterized in that The S5 specifically includes: S51. When the system detects abnormal behavior or potential attack, the enhanced active deception module is activated to generate multi-level false data and false operation interfaces. The false data feature set is F = {f 1 ,f 2 ,…,f m }, where f i Represents the characteristics of the i-th type of false data: f i =g(d i ,i f ); Among them, d i represents the real data feature of the i-th category, θ f represents the transformation parameters for generating false data, g(·) represents the data transformation function; S52, the system generates a false operation interface through virtualization technology to induce the attacker to operate. The set of virtual operation interfaces is V = {v 1 ,v 2 ,…,v k }, where v j Represents the jth virtual operation interface: v j =h(u j ,i v ); Among them, u j represents the operation behavior of the jth attacker, θ v represents the transformation parameter for generating the fake interface, h(·) represents the mapping function between the attacker's operation and the fake interface; S53, the system dynamically generates induced operations through the induced data operation mechanism, and the induced operation set is I = {i 1 ,i 2 ,…,i e }, where i l Indicates the first induced operation, and calculates the ineffectiveness score of the induced operation: Among them, S i represents the ineffectiveness score of the induced operation, e represents the total number of induced operations, α l represents the weight of the first induction operation, It indicates the result of the first fake operation performed by the attacker in the fake environment. Indicates the first real operation result, μ l Represents the adaptive adjustment coefficient, Δt l Indicates the time interval between the first operation and the last operation, γ l represents the time attenuation coefficient; S54, the system scores S according to the ineffectiveness of the induced operation i Dynamically adjust the complexity of fake data and fake interface, and dynamically optimize the fake environment as the score increases: Among them, C f represents the complexity of the false environment, m represents the total number of false data types, ρ i represents the weight of the i-th type of false data, δ i represents the adaptive complexity adjustment coefficient, and T represents the scoring threshold set by the system; S55, the system dynamically adjusts the characteristics of the fake environment according to the attacker's behavior and operation records. By recording the operation log, the system compares the attacker's fake operation records and the corresponding real operation records Calculate the matching score M r : Among them, θ q Represents the weight of the qth operation.

7. According to the deep learning-based administrator face recognition data export protection method described in claim 1, It is characterized in that The S6 specifically includes: S61. The system monitors the attacker's operation behavior in a virtual environment and collects behavior data in real time. The system initializes the induction strategy according to the operation sequence and generates the initial strategy using the proximal strategy optimization algorithm. S62. The system dynamically optimizes the initial strategy through the meta-policy gradient algorithm. The meta-policy gradient algorithm adjusts the induced strategy in real time according to the attacker's operation mode: Among them, L Meta (θ) represents the loss function of the meta-policy gradient algorithm, θ k represents the parameters of the k-th strategy, represents the kth strategy, Represents the parameter θ k The gradient of, E represents the mathematical expectation, R t represents the reward function at time step t, and T represents the total length of the time step; S63. The system continuously optimizes the inducement strategy based on the feedback from the attacker’s operation behavior: Among them, F t represents the feedback value at time step t, r i represents the reward value of the attacker's ith operation in the fake environment, λ 3 represents the time interval sensitivity coefficient, γ f represents the discount factor, t i -t i-1 Indicates the time difference between two operations; S64. Based on the feedback function and the results of meta-strategy gradient optimization, the system dynamically adjusts the induction strategy: Among them, θ ′ represents the updated policy parameters, θ represents the policy parameters before updating, and τ represents the learning rate; S65. The system dynamically adjusts the induced operation in the fake environment according to the optimized strategy parameters, and generates new induced interface and data according to the attacker's operation mode.

Images