Bank virtual space transaction risk control method and device, storage medium and equipment

Abstract

The invention provides a bank virtual space transaction risk control method and device, a storage medium and equipment, and is applied to the technical field of artificial intelligence. According to the method, when a user enters a meta universe bank space, multi-modal biological characteristics and interactive behavior characteristics are collected, and a dynamic behavior key is generated and subjected to cross-chain storage; when a user triggers a transaction, acquiring real-time transaction behavior data, and generating a temporary key in combination with the multi-modal biological characteristics; performing similarity matching on the temporary key and the on-chain dynamic behavior key through a dynamic time warping method, and calculating a deviation value; when the deviation value reaches the preset risk threshold value, a corresponding risk defense strategy is started, continuous dynamic verification and active risk defense of the user identity are achieved, and the safety and reliability of identity authentication in the complex virtual environment are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to methods, apparatus, storage media and equipment for risk control of virtual space transactions in banks. Background Technology

[0002] Traditional identity authentication systems often employ static passwords, biometric recognition (such as fingerprints and facial recognition), or two-factor authentication, relying on preset keys and single-modal recognition, and using centralized servers for data processing and storage. Regarding data traceability, some platforms utilize single-chain structures to record user behavior or transaction information, but lack cross-chain collaboration mechanisms, making it difficult to uniformly verify identity information and creating "data silos." Furthermore, traditional systems largely rely on post-event accountability, lacking proactive defense capabilities and struggling to effectively counter simulated attacks and spoofing.

[0003] Existing methods suffer from three major shortcomings: first, static keys and single-modal features are easily forged, and dynamic behavior consistency judgment is lacking; second, cross-chain collaboration is lacking, hindering identity sharing and verification in multi-chain environments; and third, the security model is passive, with limited defense capabilities. These deficiencies are particularly prominent in high-risk scenarios such as metaverse finance.

[0004] Therefore, how to improve the security and reliability of identity authentication in complex virtual environments has become a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] In view of the above problems, the present invention provides a method, apparatus, storage medium and device for controlling risks in virtual space transactions in banks that overcomes or at least partially solves the above problems. The technical solution is as follows:

[0006] A method for controlling risks in virtual space transactions for banks includes:

[0007] Obtain multimodal biometric information of users when they enter the Metaverse Bank space;

[0008] Obtain the user's interactive behavior characteristics information in the metaverse banking space;

[0009] Using the multimodal biometric information and the interaction behavior information, a dynamic behavior key is generated;

[0010] The dynamic behavior key is stored across chains;

[0011] When the user triggers a transaction in the metaverse bank space, the user's real-time transaction behavior data is obtained;

[0012] A temporary key is generated using the multimodal biometric information and the real-time transaction behavior data;

[0013] The temporary key is matched with the dynamic behavior key on the chain using dynamic time warping similarity to obtain a similarity deviation value;

[0014] Determine whether the similarity deviation value is equal to or greater than a preset risk threshold. If so, activate the risk defense strategy corresponding to the similarity deviation value.

[0015] Optionally, the cross-chain storage of the dynamic behavior key includes:

[0016] The dynamic behavior key is decomposed into a private component and a public component;

[0017] The private component is homomorphically encrypted and stored on a private consortium blockchain. The public component is bound to the user's financial account and written into a public blockchain. A verifiable mapping relationship is then constructed between the encrypted private component on the private consortium blockchain and the public component on the public blockchain.

[0018] Optionally, obtaining the user's interactive behavior characteristics information in the metaverse bank space includes:

[0019] Record the sequence of user interaction behaviors in the metaverse banking space;

[0020] The interaction behavior sequence is subjected to time series analysis to extract the user's interaction behavior feature information in the metaverse bank space.

[0021] Optionally, recording the user's interaction sequence in the metaverse bank space includes:

[0022] The user's operation events within the Metaverse Bank Space are mapped to a sequence of coordinate points in a preset coordinate system, wherein the preset coordinate system is a three-dimensional spatiotemporal coordinate system with the virtual counter coordinate position within the Metaverse Bank Space as the reference origin;

[0023] Record the user's behavioral entropy value within the metaverse bank space;

[0024] Using the coordinate point sequence and the behavior entropy value, the user's interaction behavior sequence in the metaverse bank space is generated.

[0025] Optionally, the step of performing time-series analysis on the interaction behavior sequence to extract the user's interaction behavior feature information in the metaverse bank space includes:

[0026] The interaction sequence is spatiotemporally embedded using a Transformer encoder to obtain a spatiotemporal embedding vector.

[0027] Based on the spatiotemporal embedding vector, an adversarial generative network is used to create a behavioral feature enhancement dataset corresponding to the user.

[0028] The dataset is augmented using the aforementioned behavioral features, and the feature extraction model is optimized through contrastive learning.

[0029] The optimized feature extraction model is used to extract the user's interactive behavior feature information in the metaverse bank space from the spatiotemporal embedding vector.

[0030] Optionally, before determining whether the similarity deviation value is equal to or greater than a preset risk threshold, the method further includes:

[0031] Random number seeds are generated using a quantum entropy source;

[0032] The preset risk threshold is dynamically adjusted based on the random number seed.

[0033] Optionally, activating the risk defense strategy corresponding to the similarity deviation value includes:

[0034] When the similarity deviation value corresponds to low risk, an invisible decoy element containing a fake API interface and fake data packets is injected into the user's current transaction interface;

[0035] In cases where the similarity deviation value corresponds to medium risk, the current transaction process is forcibly interrupted, and multimodal review and verification are enabled.

[0036] If the similarity deviation value corresponds to a high risk, the user's financial account will be frozen, and a cross-chain audit will be initiated to trace the attack path.

[0037] A bank virtual space transaction risk control device includes: a multimodal biometric information acquisition unit, an interactive behavior feature information acquisition unit, a dynamic behavior key generation unit, a cross-chain storage unit, a real-time transaction behavior data acquisition unit, a temporary key generation unit, a similarity deviation value acquisition unit, a risk judgment unit, and a risk defense activation unit.

[0038] The multimodal biometric information acquisition unit is used to acquire multimodal biometric information of the user when entering the Metaverse Bank space;

[0039] The interactive behavior feature information acquisition unit is used to acquire the interactive behavior feature information of the user in the metaverse bank space;

[0040] The dynamic behavior key generation unit is used to generate a dynamic behavior key using the multimodal biometric information and the interaction behavior feature information;

[0041] The cross-chain storage unit is used to store the dynamic behavior key across chains;

[0042] The real-time transaction behavior data acquisition unit is used to acquire the user's real-time transaction behavior data when it detects that the user has triggered a transaction in the metaverse bank space.

[0043] The temporary key generation unit is used to generate a temporary key using the multimodal biometric information and the real-time transaction behavior data;

[0044] The similarity deviation value acquisition unit is used to perform dynamic time-normalized similarity matching between the temporary key and the dynamic behavior key on the chain to obtain a similarity deviation value.

[0045] The risk judgment unit is used to determine whether the similarity deviation value is equal to or greater than a preset risk threshold. If so, the risk defense activation unit is triggered.

[0046] The risk defense activation unit is used to activate the risk defense strategy corresponding to the similarity deviation value.

[0047] A computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the aforementioned bank virtual space transaction risk control method.

[0048] An electronic device includes at least one processor, at least one memory connected to the processor, and a bus; wherein the processor and the memory communicate with each other via the bus; the processor is used to call program instructions in the memory to execute the aforementioned bank virtual space transaction risk control method.

[0049] By employing the above technical solutions, the present invention provides a method, apparatus, storage medium, and device for controlling transaction risks in a virtual banking space. These methods acquire multimodal biometric information of a user entering the metaverse banking space; acquire user interaction behavior characteristics within the metaverse banking space; generate a dynamic behavior key using the multimodal biometric information and interaction behavior characteristics; store the dynamic behavior key across blockchains; acquire real-time transaction behavior data of the user when a transaction is detected in the metaverse banking space; generate a temporary key using the multimodal biometric information and real-time transaction behavior data; perform dynamic time-warping similarity matching between the temporary key and the on-chain dynamic behavior key to obtain a similarity deviation value; and determine whether the similarity deviation value is equal to or greater than a preset risk threshold. If so, the risk defense strategy corresponding to the similarity deviation value is activated. The present invention, by collecting multimodal biometric and interaction behavior characteristics of users in the metaverse banking space, generating and cross-chain storing dynamic behavior keys, dynamically generating temporary keys in conjunction with real-time transaction behavior data, and judging risks through dynamic time-warping similarity matching, achieves continuous dynamic verification of identity and proactive risk defense, effectively improving the security and reliability of identity authentication in complex virtual environments.

[0050] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0051] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0052] Figure 1 The diagram illustrates a flowchart of one embodiment of the bank virtual space transaction risk control method provided by the present invention.

[0053] Figure 2 This is a flowchart illustrating a specific implementation of step S130 in the bank virtual space transaction risk control method provided in an embodiment of the present invention;

[0054] Figure 3 This is a flowchart illustrating a specific implementation of step S110 in the bank virtual space transaction risk control method provided in an embodiment of the present invention;

[0055] Figure 4The diagram shows a specific implementation of step S300 in the bank virtual space transaction risk control method provided in this embodiment of the invention.

[0056] Figure 5 This is a flowchart illustrating a specific implementation of step S310 in the bank virtual space transaction risk control method provided in an embodiment of the present invention;

[0057] Figure 6 This diagram illustrates the structure of a bank virtual space transaction risk control device provided in an embodiment of the present invention.

[0058] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present invention is shown. Detailed Implementation

[0059] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

[0060] With the rapid development of information technology, identity authentication systems play a central role in various application scenarios. Traditional identity authentication systems generally use static passwords, biometric recognition, or token-based two-factor authentication to verify user identity. These traditional methods typically rely on preset keys or single-modality biometric recognition technology, complete the identity authentication process through simple encryption algorithms, and mostly depend on centralized servers for data processing and storage.

[0061] In terms of data traceability and security management, some platforms have attempted to use the single-chain structure of blockchain to record user behavior or transaction information, aiming to achieve immutability and secure traceability of identity information. However, due to the lack of an effective cross-chain collaboration mechanism, user identity information and behavioral data are difficult to manage and share uniformly in a multi-chain environment, resulting in "data silos" between systems and limiting the efficiency and reliability of identity verification. Furthermore, traditional identity authentication systems primarily rely on post-event accountability, lacking proactive defense capabilities and making it difficult to promptly identify and block potential risky operations.

[0062] Traditional identity authentication methods suffer from the following main shortcomings: Firstly, static keys and single-modal biometrics are easily forged or stolen, lacking dynamic judgment of user real-time behavioral consistency. Secondly, the lack of cross-chain collaboration mechanisms makes it difficult to uniformly verify user identities in a multi-chain environment, resulting in data isolation. Furthermore, existing security models tend to rely on passive detection, lacking sufficient defense capabilities against complex security threats such as simulated attacks and spoofing. These limitations are particularly evident in application environments like Metaverse Finance, which have extremely high requirements for interactivity and security. There is an urgent need for an intelligent identity security solution that integrates multimodal biometrics and dynamic behavioral analysis, supports cross-chain collaboration, and possesses real-time proactive defense capabilities to improve the security and reliability of user identity authentication.

[0063] Based on this, this invention provides a method for risk control in virtual space transactions within a bank. Specifically for the metaverse financial environment, it constructs a multimodal dynamic identity authentication and secure transaction method, addressing issues such as frequent identity forgery, lack of behavioral consistency verification, inability to unify cross-chain identities, and lagging risk defense mechanisms. This invention integrates user multimodal biometrics with three-dimensional behavioral trajectories to generate encrypted dynamic keys. Combined with proactive risk mitigation technology and a dual-chain collaborative cross-chain identity anchoring mechanism, it achieves high-strength anti-counterfeiting verification, real-time risk detection, and fully traceable trusted interaction, significantly improving identity security and reliability in complex virtual environments.

[0064] like Figure 1 The diagram shows a flowchart of one embodiment of the bank virtual space transaction risk control method provided by this invention. The method may include:

[0065] S100: Obtain multimodal biometric information of the user when entering the Metaverse Bank space.

[0066] Among them, the Metaverse Bank Space refers to a banking service environment built on a virtual reality or augmented reality platform. Users enter this space through digital identities to conduct financial interactions and transactions, reflecting a financial service scenario that integrates virtual and reality.

[0067] Among them, multimodal biometric information refers to the collection of various biometric data collected by users in the metaverse bank space, which may include visible light face images, infrared thermal imaging images and voiceprints, etc., and form a biometric vector with high robustness and anti-counterfeiting capabilities through fusion algorithms.

[0068] Specifically, when a user enters the Metaverse Bank space through a wearable device, their visible light facial image, infrared thermal image, and voiceprint features are simultaneously collected. A visible light-infrared dual-mode fusion neural network is used to extract features from the images, and an initial biometric vector is generated by splicing and fusing the images. This vector is then combined with facial blood flow distribution heatmaps obtained from infrared thermal imaging to filter abnormal textures, providing real-time defense against 3D mask spoofing attacks and ensuring highly robust identity feature collection.

[0069] As some examples, when a user logs into the bank's metaverse via a VR (Virtual Reality) / AR (Augmented Reality) device, the system synchronously collects their biometric information using integrated multimodal sensors: 1) using a visible light camera to collect facial images with a resolution of at least 1920x1080. ;2) Use an infrared thermal imaging camera to collect facial infrared thermal images. The image can display a real-time thermal map of subcutaneous blood flow distribution on the face, which can be used to detect abnormal body temperature or 3D mask attacks; 3) Collect at least 3 seconds of voice commands through a microphone array and extract voiceprint feature vectors. The three types of raw data were collected and processed using a visible light-infrared dual-mode fusion algorithm. The infrared image was used to filter out anomalous texture disturbances (such as the static texture of a silicone mask) in the visible light image. Finally, the data were processed by a feature extraction network. Features are extracted separately, and then a high-dimensional, robust initial biofeedback vector is generated through tensor concatenation. This serves as the static biological basis for subsequently generating dynamic behavior keys.

[0070] S110. Obtain user interaction behavior characteristics information in the metaverse bank space.

[0071] Interactive behavior feature information refers to data that reflects users' operating habits and behavior patterns in virtual space, such as operation delay, interface click path, voice command interval, and three-dimensional behavior trajectory, which are used to capture users' behavioral habits.

[0072] Specifically, embodiments of the present invention can record multi-dimensional behavioral sequences generated during user interaction with the virtual environment, and use advanced time series models to perform in-depth analysis and feature extraction on the behavioral sequences, thereby extracting interactive behavioral feature information that can characterize the user's unique behavioral patterns.

[0073] S120. Utilize multimodal biometric information and interactive behavior information to generate dynamic behavior keys.

[0074] Among them, the dynamic behavior key refers to the encryption key generated based on the fusion of multimodal biometric information and interaction behavior information. The dynamic behavior key possesses dynamic and high-entropy properties, effectively preventing forgery and duplication, and serves as a credential for identity authentication and secure transactions.

[0075] Specifically, embodiments of the present invention can utilize interactive behavior feature information. With behavioral entropy splicing into behavioral tensors The improved Transformer encoder is used for spatiotemporal embedding, and the position encoding employs a Fibonacci spiral-based function to enhance spatial correlation, thereby obtaining the encoded behavioral feature sequence. Generative adversarial networks are used to generate pseudo-samples of user-specific behaviors, and comparative learning is performed to optimize the behavior feature extractor. Extracting behavioral key representation . Action key With the initial biological feature vector splicing, input key fusion network Generate dynamic behavior keys This enables dynamic, high-entropy, and non-replicable identity authentication keys.

[0076] S130, Perform cross-chain storage of dynamic behavior keys.

[0077] Specifically, embodiments of the present invention can parse the dynamic behavior key into two components, one for privacy protection and the other for public verification. Using homomorphic encryption and distributed storage technology, the privacy component is securely stored in a controlled private blockchain network. Simultaneously, the public component is reliably bound to the user's financial identity and written to a public blockchain with public auditing capabilities. Through zero-knowledge proof technology, a verifiable association between the corresponding components on the private and public blockchains is constructed and maintained without disclosing privacy information, thereby achieving secure storage and collaborative management of the key in a cross-chain environment.

[0078] S140. When a user triggers a transaction in the Metaverse Bank space, obtain the user's real-time transaction behavior data.

[0079] Real-time transaction behavior data refers to the instant interactive behavior data of users when triggering financial transactions, including the current operation sequence, transaction parameters and behavioral characteristics, which are used to assist in generating temporary authentication keys and risk assessment.

[0080] Specifically, embodiments of the present invention can collect real-time interaction behavior data when a user initiates a financial transaction, including the operation sequence, transaction parameters, and related behavioral characteristics, such as click paths and voice command durations. Through sensor devices and a virtual space monitoring module, this data is transformed into behavioral input for identity verification, providing real-time dynamic information for temporary key generation.

[0081] As examples, embodiments of the present invention can immediately initiate real-time data acquisition upon detecting a user's transaction request, such as a transfer or payment: capturing all interaction events within 10 seconds before and after the transaction is triggered, and regenerating the coordinate point sequence. Simultaneously, it collects the user's real-time voice command intervals, operation delays, and interface click paths during the current session. It also utilizes an accelerometer and eye tracker to acquire micro-vibration frequencies and visual focus persistence data during the transaction. This real-time data constitutes the behavioral sequence of the current transaction, and its acquisition dimensions and feature types are completely consistent with the interaction behavior feature information acquisition process in step S110, ensuring subsequent comparability.

[0082] S150. Generate a temporary key using multimodal biometric information and real-time transaction behavior data.

[0083] Among them, the temporary key, which is a short-term key dynamically generated by combining multimodal biometrics and real-time transaction behavior data, is used to verify the continuity and consistency of user identity during the transaction process, thereby enhancing security.

[0084] Specifically, embodiments of the present invention can combine real-time transaction behavior data with collected multimodal biometrics to quickly extract behavioral features and fuse them with biometrics, generating a temporary authentication key that is only used at the current transaction moment. This temporary key has timeliness and dynamism, ensuring the continuity and security of identity during the transaction process.

[0085] As examples, embodiments of the present invention can input multimodal biometric information and real-time transaction behavior data into a pre-trained user-specific behavior feature extraction model to extract real-time behavior feature vectors. A key fusion network identical to that used in step S120 is employed. The system integrates multimodal biometric information and real-time transaction behavior data to generate a temporary key for verifying this transaction.

[0086] S160. Perform dynamic time-warped similarity matching between the temporary key and the dynamic behavior key on the chain to obtain the similarity deviation value.

[0087] The similarity deviation value refers to the measurement of the difference between the temporary key calculated by the Dynamic Time Warping (DTW) algorithm and the on-chain dynamic behavior key. The value reflects the degree of similarity between the current behavior and the historical behavior and is a key indicator for risk assessment.

[0088] Specifically, embodiments of the present invention may employ a dynamic time warping algorithm to perform temporal similarity matching between temporary keys and dynamic behavior keys stored on the blockchain, calculate the optimal alignment path and distance between the two on the time axis, and obtain a similarity deviation value.

[0089] As examples, embodiments of the present invention can retrieve the historical sequence of dynamic behavior keys stored after a user's most recent N successful authentications from the blockchain. A dynamic time warping (DTW) operation is performed on the temporary key and each historical key sequence to calculate the minimum cumulative distance between them. By taking the minimum DTW distance value, or by taking a weighted average of multiple distance values, a similarity deviation value between 0 and 1 is obtained. The smaller this value, the more closely the current transaction behavior matches the historical normal pattern.

[0090] S170. Determine whether the similarity deviation value is equal to or greater than the preset risk threshold. If so, proceed to step S180.

[0091] Among them, the preset risk threshold refers to the critical point of similarity deviation value set in advance according to the security policy. When the calculated deviation value reaches or exceeds the threshold, it is determined that there is a potential risk and the corresponding defense mechanism is triggered.

[0092] Specifically, when the similarity deviation value calculated in this embodiment of the invention reaches or exceeds the threshold, it is determined that there is an abnormality or potential fraud risk, triggering a risk response mechanism and entering a risk defense process.

[0093] Understandably, if the similarity deviation is less than the preset risk threshold, the normal transaction process will continue.

[0094] S180, Initiate the risk defense strategy corresponding to the similarity deviation value.

[0095] Among them, the risk defense strategy refers to the layered security response measures taken according to the risk level. These measures may include displaying fake transaction interfaces to lure attackers, multimodal verification, and freezing accounts and initiating cross-chain auditing and tracing mechanisms when the risk is high, so as to ensure transaction security and identity authenticity.

[0096] Specifically, in this embodiment of the invention, the system can dynamically adapt and execute graded proactive defense measures based on the risk level identified by the similarity deviation value: for low-risk levels, a monitoring strategy of implanting hidden decoys is adopted; for medium-risk levels, a blocking strategy including process interruption and enhanced verification is initiated; and for high-risk levels, a containment strategy of account freezing and path tracing is implemented.

[0097] This invention provides a method for risk control of virtual space transactions in banks. The method includes: obtaining multimodal biometric information of a user entering a metaverse banking space; obtaining interactive behavior feature information of the user in the metaverse banking space; generating a dynamic behavior key using the multimodal biometric information and interactive behavior feature information; storing the dynamic behavior key across chains; obtaining real-time transaction behavior data of the user when a transaction is detected in the metaverse banking space; generating a temporary key using the multimodal biometric information and real-time transaction behavior data; performing dynamic time-warping similarity matching between the temporary key and the on-chain dynamic behavior key to obtain a similarity deviation value; determining whether the similarity deviation value is equal to or greater than a preset risk threshold; if so, activating a risk defense strategy corresponding to the similarity deviation value. This invention, by collecting multimodal biometric and interactive behavior features of users in the metaverse banking space, generating and storing dynamic behavior keys across chains, dynamically generating temporary keys in conjunction with real-time transaction behavior data, and judging risks through dynamic time-warping similarity matching, achieves continuous dynamic verification of identity and proactive risk defense, effectively improving the security and reliability of identity authentication in complex virtual environments.

[0098] Optional, based on Figure 1 The method shown is as follows: Figure 2 The diagram shows a specific implementation of step S130 in the bank virtual space transaction risk control method provided in this embodiment of the invention. Step S130 may specifically include:

[0099] S200: Decompose the dynamic behavior key into a private component and a public component.

[0100] The private component refers to the sensitive part extracted from the dynamic behavior key, which contains the user's privacy information or key identity authentication data.

[0101] The public component refers to the publicly available portion of the dynamic behavior key, used to bind to the user's financial account information and recorded on a public blockchain. The public component does not contain sensitive privacy information, but serves as a public credential for identity authentication and transaction verification, supporting cross-platform identity verification and audit traceability.

[0102] Specifically, this invention, based on a secret-sharing key segmentation algorithm, mathematically decomposes a dynamic behavior key that integrates user biometrics and behavioral patterns into two components with different functions and security attributes. The private component contains core biometric data and high-precision behavioral patterns that directly identify sensitive user information, designed to ensure absolute privacy. The public component primarily contains behavioral sequence hash digests or non-sensitive attribute tags used for on-chain verification and not directly revealing privacy. This decomposition process uses a cryptographically secure random number generator to ensure the unpredictability of the segmentation, making it impossible to reconstruct the complete original key from a single component, thus laying the foundation for subsequent dual-chain storage and cross-chain verification.

[0103] S210. Store the private component in the private consortium blockchain after homomorphic encryption, write the public component into the public blockchain after binding it with the user's financial account, and construct a verifiable mapping relationship between the encrypted private component on the private consortium blockchain and the public component on the public blockchain.

[0104] Private consortium blockchains refer to a controlled blockchain network that is jointly maintained by multiple parties, including banks and other financial institutions, and has an access control mechanism.

[0105] Public blockchains refer to blockchain networks that are open to the public and accessible and verifiable by any participant. They are used to store public tokens and identity credentials linked to users' financial accounts, ensuring the transparency and immutability of identity information, and supporting cross-chain identity authentication and auditing functions.

[0106] Verifiable mappings refer to the correspondence between private and public components established through cryptographic techniques. This relationship can be verified by smart contracts or verification nodes on the public blockchain without disclosing private information. Verifiable mappings ensure the consistency and trustworthiness of identity information between the two chains, enabling cross-chain identity authentication and security auditing.

[0107] Specifically, the private component is first encrypted using a homomorphic encryption algorithm (such as Paillier or elliptic curve homomorphic encryption) to generate an encryption key, ensuring that some operations are still supported in the encrypted state, thus improving privacy protection and computational efficiency. The encrypted private component is distributed and stored on controlled private consortium blockchain nodes, enabling multi-party collaborative management and tamper-proof protection. Simultaneously, the public component is bound to the user's financial account information and written to the public blockchain via a smart contract, ensuring the transparency and immutability of the identity credential. This embodiment of the invention uses zero-knowledge proof technology to construct a mapping verification protocol between the encrypted private component on the private consortium blockchain and the public component on the public blockchain. Verification nodes can confirm the validity of the proof without disclosing private key information, ensuring the consistency and trustworthiness of identity information between the two chains, and achieving cross-chain identity anchoring and security auditing.

[0108] It should be noted that the embodiments of this invention can construct an attribute-based encryption scheme based on elliptic curve bilinear pairing, subdividing user financial account permissions into multiple attribute fragments. Each attribute fragment, after homomorphic encryption, is distributed and stored across multiple nodes of the consortium blockchain, ensuring that a single node cannot recover the complete permission information, thereby achieving secure partitioning and protection of permissions. Secondly, smart contracts are deployed on the public blockchain to uniformly manage the key mapping relationship between private and public components. The execution condition of the smart contract is that at least two-thirds of the nodes in the consortium blockchain jointly verify the validity of the zero-knowledge proof of the user's private component. When the verification by more than two-thirds of the nodes meets this threshold, the identity mapping is confirmed as valid, ensuring the consistency and trustworthiness of the dual-chain identity information. Furthermore, a lightweight cross-chain relay is introduced, utilizing a Merkel-Patricia tree structure to structurally compress the data returned by the consortium blockchain nodes, ensuring the controllability of the verification path and data integrity, while controlling the cross-chain audit latency to within 50 milliseconds, meeting the requirements for real-time identity invocation and audit traceability.

[0109] This invention generates dynamic behavior keys based on multimodal biometric and interactive behavioral information collected from users within the Metaverse Bank space. These keys are then combined with real-time transaction data for identity verification and risk assessment. A cross-chain storage mechanism is employed to split the dynamic behavior key into private and public components, stored separately on a private consortium blockchain and a public blockchain. By homomorphically encrypting the private component and establishing a verifiable mapping between the private consortium blockchain and the public blockchain, the privacy of sensitive identity information is protected while ensuring the transparency and immutability of identity credentials. This achieves trusted binding and efficient auditing of cross-chain identities. This layered storage and verification mechanism effectively enhances the security of dynamic behavior keys and the overall system protection capabilities, significantly reducing the risks of identity theft and data leakage, and strengthening the credibility and real-time risk response capabilities of Metaverse financial transactions.

[0110] Optional, based on Figure 1 The method shown is as follows: Figure 3 The diagram shows a specific implementation of step S110 in the bank virtual space transaction risk control method provided in this embodiment of the invention. Step S110 may specifically include:

[0111] S300 records the sequence of user interaction behaviors in the Metaverse Bank space.

[0112] Specifically, embodiments of the present invention can construct a three-dimensional spatiotemporal coordinate system based on key virtual objects (such as virtual counters), and map user operation events to coordinate points in this coordinate system in real time. Simultaneously, the uncertainty or regularity of user operations is quantified and recorded, generating behavioral entropy values. These coordinate point sequences and behavioral entropy values ​​are then integrated to form a complete sequence of interactive behaviors.

[0113] S310. Perform time-series analysis on the interaction behavior sequence to extract the user's interaction behavior feature information in the metaverse bank space.

[0114] Specifically, embodiments of the present invention can employ a deep sequence model to embed behavioral sequences in a high dimension, obtaining a vector representation containing spatiotemporal contextual information. An augmented dataset that fully reflects user behavior patterns is generated through data augmentation techniques, and this dataset is used to optimize the feature extraction model. The optimized model is then applied to extract interactive behavioral feature information representing unique user behavior patterns from the embedded representation.

[0115] This invention, based on multimodal biometrics and interactive behavior information collected when a user enters the Metaverse Bank space, records the user's interactive behavior sequence within the Metaverse Bank space and performs time-series analysis on this sequence to accurately extract the user's interactive behavior features. This allows for a more comprehensive and dynamic reflection of the user's behavioral habits and operating patterns, thereby not only improving the accuracy and personalization of dynamic behavior key generation but also enhancing the reliability of subsequent risk assessment through dynamic time-warping similarity matching. This effectively improves the security of identity verification and the accuracy of risk defense strategies, ensuring the credibility and real-time response capability of transactions within the Metaverse Bank space.

[0116] Optional, based on Figure 1 The method shown is as follows: Figure 4 The diagram shows a specific implementation of step S300 in the bank virtual space transaction risk control method provided in this embodiment of the invention. Step S300 may specifically include:

[0117] S400. Map the user's operation events in the metaverse bank space to a sequence of coordinate points in a preset coordinate system, wherein the preset coordinate system is a three-dimensional spatiotemporal coordinate system with the virtual counter coordinate position in the metaverse bank space as the reference origin.

[0118] Specifically, in this embodiment of the invention, a virtual counter coordinate position within the metaverse bank space can be established as the origin. A three-dimensional spatiotemporal coordinate system is used to capture user operation events in real time using a spatial tracking system; then, the user's actions at each time point are... The operation actions are converted into three-dimensional coordinate points by sensors or position tracking devices. This forms a continuous sequence of coordinate points. It fully maps the user's behavioral trajectory in the virtual space, providing spatial foundation data for subsequent behavioral feature extraction.

[0119] S410, Record the entropy value of the user's behavior within the Metaverse Bank space.

[0120] Among them, behavioral entropy refers to an indicator that reflects the complexity and uncertainty of a user's interactive behavior by quantifying the diversity of user operations and the distribution characteristics of attention in the metaverse banking space.

[0121] Specifically, embodiments of the present invention can acquire the micro-vibration frequency signal of the user's handheld controller through an accelerometer. and the duration of visual focus recorded by the eye tracker By combining the two types of behavioral signals, a joint probability distribution can be constructed. Based on this probability distribution, calculate the user behavior entropy value. This value quantifies the complexity of user operations and the degree of distraction, providing a behavioral diversity indicator for dynamic behavior key generation.

[0122] S420. Using the coordinate point sequence and behavior entropy value, generate the user's interaction behavior sequence in the metaverse bank space.

[0123] Specifically, in this embodiment of the invention, the three-dimensional coordinate point sequence obtained in step S400 can be used... The behavioral entropy value calculated in step S410 By splicing the sequences, a behavioral tensor sequence containing information about spatial location and behavioral complexity is constructed. Each of them Subsequently, the behavior tensor sequence is input into an improved spatiotemporal feature encoding model (such as a Transformer-based encoder) to achieve spatiotemporal feature extraction and fusion of user interaction behavior, generating an interaction behavior sequence that reflects the user's personalized operation mode, providing rich feature support for the accurate generation of dynamic behavior keys.

[0124] This invention maps user actions within the metaverse banking space into a sequence of three-dimensional spatiotemporal coordinates centered on the virtual counter. Combined with recorded behavioral entropy values, this highly integrates the user's spatial behavioral trajectory and complexity, comprehensively and dynamically reflecting the user's personalized interaction patterns in the virtual space. Generating interactive behavior sequences using coordinate point sequences and behavioral entropy values ​​not only enhances the refined expression of behavioral features but also significantly strengthens the uniqueness and non-replicability of dynamic behavioral keys, effectively preventing behavioral features from being forged or replayed. Simultaneously, this approach provides higher-dimensional and more accurate data support for subsequent risk detection, identity verification, and abnormal behavior identification, improving the system's immediate protection against security risks such as virtual identity impersonation and abnormal transactions. This significantly enhances the reliability and robustness of user identity authentication and transaction security within the metaverse banking space.

[0125] Optional, based on Figure 1 The method shown is as follows: Figure 5 The diagram shows a specific implementation of step S310 in the bank virtual space transaction risk control method provided in this embodiment of the invention. Step S310 may specifically include:

[0126] S500: Use the Transformer encoder to perform spatiotemporal embedding processing on the interaction behavior sequence to obtain the spatiotemporal embedding vector.

[0127] Among them, the spatiotemporal embedding vector refers to the high-dimensional vector representation obtained by fusing spatiotemporal information of the user's interactive behavior sequence in the metaverse bank space through the Transformer encoder. The spatiotemporal embedding vector integrates the spatial location information and temporal sequence features of the user's operation, and can effectively capture the temporal dynamics and spatial correlations in the behavior trajectory, so as to achieve a deep expression and encoding of the user's behavior pattern.

[0128] Specifically, in this embodiment of the invention, the three-dimensional operation coordinate sequence collected by the user in the metaverse bank space through a spatial tracking system is concatenated with the calculated behavior entropy value to form a behavior tensor sequence containing spatiotemporal information. An improved Transformer encoder based on Fibonacci spiral function position encoding is designed and applied. This behavior tensor sequence is input into the model, and through multi-layer self-attention mechanism and feedforward network encoding, deep embedding processing of spatiotemporal dynamic features is achieved, ultimately outputting a high-dimensional spatiotemporal embedding vector containing both temporal and spatial information.

[0129] As some examples, embodiments of the present invention use an interaction behavior sequence tensor that integrates spatial coordinates and behavioral entropy values. As input, firstly, to effectively model the spatiotemporal location information of each element in the sequence, a position encoding function based on an improved Fibonacci spiral sequence is introduced, which is used for each time step. Generate position encoding vector For example, through a function Calculation, where For the first The Fibonacci numbers are used to enhance the mathematical representation of spatial relationships. Subsequently, the positions are encoded. With the original input Element-wise addition is performed to form an enhanced input containing spatiotemporal location information. This enhanced sequence is then fed into a multi-layer Transformer encoder for deep feature extraction. The encoder, through its self-attention mechanism, can dynamically capture the long-range dependencies between any two operation events in a sequence. For example, it can analyze the intrinsic relationship between a user's operation pattern of moving from point A to point B and their subsequent behavior of staying at point C. After being processed by multiple layers of encoder stacking, the original sequence X is transformed into a new feature sequence containing rich spatiotemporal context information. Each of them These are the spatiotemporal embedding vectors at the corresponding time points. These vectors not only encode "where the operation occurred," but also "when the operation occurred" and "the relationship with the operations before and after," forming the deep representation basis of behavioral features.

[0130] S510. Based on spatiotemporal embedding vectors, use adversarial generative networks to create user-corresponding enhanced datasets of behavioral features.

[0131] The behavior feature enhancement dataset is an expanded sample set generated by the adversarial generative network based on existing spatiotemporal embedding vectors. It contains diverse pseudo-samples of real user behavior features to enrich the training sample space, improve the model's ability to recognize and its robustness against different behavioral variations, and assist in optimizing the generalization performance of the behavior feature extraction model.

[0132] Specifically, in this embodiment of the invention, spatiotemporal embedding vectors can be used as input to real samples. The generator of the adversarial generative network generates pseudo samples that are similar to user behavior features based on random noise distribution, while the discriminator evaluates the authenticity of the samples. Through the alternating training of the generator and the discriminator, the authenticity and diversity of pseudo samples are continuously improved, and finally an enhanced dataset covering the diversity of user behavior features is formed to expand the training sample space and improve the robustness and generalization ability of feature learning.

[0133] As some examples, embodiments of the present invention can use the user's real spatiotemporal embedded vector sequence As a real data distribution, construct a generator. and a discriminator Generator By learning from real data The latent distribution of Z* = G(z) can generate highly realistic pseudo-samples Z* = G(z) that simulate user behavior characteristics from random noise. These generated pseudo-samples... Compared with real samples Together, they form a richer and more diverse "behavioral feature enhancement dataset," providing ample training data for subsequent model optimization.

[0134] S520. Enhance the dataset using behavioral features and optimize the feature extraction model through comparative learning.

[0135] The feature extraction model refers to a model trained on a deep learning architecture, used to automatically learn and extract key representational features of user interaction behavior from spatiotemporal embedding vectors and their augmented datasets. This model is optimized through contrastive learning, aiming to maximize the similarity between samples of the same user behavior and minimize the similarity between different user samples, thereby improving the accuracy and security of dynamic behavior key generation.

[0136] Specifically, embodiments of the present invention can use a contrastive learning framework to train a behavioral feature extraction model, construct positive and negative sample pairs using an augmented dataset, and the model can effectively distinguish individual feature differences by maximizing the feature similarity between different samples of the same user and minimizing the similarity between samples of different users. During the training process, a temperature coefficient is introduced to adjust the similarity distribution, optimize the model's discriminative ability and feature expression accuracy, thereby forming an accurate and stable behavioral feature extractor.

[0137] As examples, in order to train a feature extractor that can extract more discriminative and robust behavioral features. A contrastive learning framework is employed for optimization. Within this framework, "positive sample pairs" and "negative sample pairs" are constructed from the behavioral feature enhancement dataset. For the same user, their real samples... Compared to corresponding pseudo-samples generated through data augmentation (such as slight perturbations) or adversarial generative networks Samples from different users are considered a pair of positive samples because they essentially represent the same behavioral pattern. Samples from different users are considered negative samples. Feature extractor The goal is to map positive sample pairs from the same user to very close locations in the feature space, while mapping negative sample pairs from different users to more distant locations. This objective is achieved by minimizing the contrastive loss function. To achieve this, in which, The temperature coefficient is used. The feature extractor is trained on a large scale using an augmented dataset, guided by this loss function. The parameters are continuously optimized, thereby learning to ignore irrelevant noise and natural fluctuations within individuals in behavioral data, while capturing stable and unique core behavioral pattern features that can distinguish different users.

[0138] S530. Use the optimized feature extraction model to extract user interaction behavior feature information in the metaverse bank space from the spatiotemporal embedding vector.

[0139] Specifically, in this embodiment of the invention, the spatiotemporal embedding vector can be input into the optimized feature extraction model. The model automatically extracts the core interactive behavior features of the user in the metaverse banking space and outputs a high-dimensional, highly discriminative behavior representation vector, i.e., interactive behavior feature information. The interactive behavior feature information, combined with multimodal biometric vectors, is further processed through a key fusion network and serves as a key input for dynamic behavior key generation, ensuring high reliability of identity authentication and transaction security.

[0140] As some examples, in feature extraction models After optimization through comparative learning, it is put into practical use. For any user session, the system first generates a spatiotemporal embedding vector sequence of its interaction behavior sequence according to the aforementioned steps. Then, the entire sequence (Or, depending on the task requirements, take a specific moment or the aggregated representation) and input it into the optimized feature extractor. The model, based on its knowledge learned from massive amounts of augmented data, performs deep analysis and feature transformation on the input spatiotemporal embedding vector. Its final output is denoted as... This refers to the extracted, highly condensed "interaction behavior feature information." This feature... It is a low-dimensional, dense vector representation that discards redundant details and temporal fluctuations in the original behavior sequence, but retains to the greatest extent possible the abstract pattern information that can uniquely represent the user's specific operating habits and interaction style. This feature It is then fused with biometric feature vectors to generate dynamic behavior keys. Its high-quality extraction is a core component and is key to ensuring the high accuracy and strong security of the entire identity authentication system.

[0141] This invention constructs an interactive behavior sequence based on the coordinate point sequence and behavioral entropy value of user operation events mapped to a three-dimensional spatiotemporal coordinate system within the Metaverse Bank space. By introducing a Transformer encoder for spatiotemporal embedding processing, it can deeply capture the temporal dynamics and spatial correlation features of user behavior, obtaining a high-dimensional spatiotemporal embedding vector. Subsequently, a generative adversarial network is used to generate a rich behavioral feature enhancement dataset based on this embedding vector, expanding sample diversity and improving the model's adaptability to behavioral variations. Through comparative learning to optimize the feature extraction model, the discriminative power and robustness of behavioral features are significantly enhanced. Finally, using this optimized model, unique interactive behavior feature information of users is accurately extracted from the spatiotemporal embedding vector, greatly improving the non-replicability and security of dynamic behavior keys, effectively preventing identity impersonation and spoofing attacks, and ensuring the transaction security and user identity authenticity in the Metaverse Bank virtual space.

[0142] Optionally, in the above Figure 1 Based on one or more corresponding embodiments, in another optional embodiment provided by the present invention, before step S170, the method may further include:

[0143] A random number seed is generated using a quantum entropy source. Based on the random number seed, a preset risk threshold is dynamically adjusted.

[0144] Specifically, embodiments of the present invention can integrate a quantum entropy source hardware module, which utilizes the inherent randomness of quantum physics as the entropy source. The quantum entropy source operates continuously, periodically collecting these microscopic quantum events and converting them into a discrete binary bit stream. This bit stream is inherently truly random, unpredictable, and unbiased. To generate a random number seed for security applications, the system sets a fixed sampling period (e.g., every millisecond or when each transaction is triggered), and extracts a bit sequence of a specified length from the real-time bit stream output by the quantum entropy source. This bit sequence This is the generated quantum random number seed, followed by adjusting the preset risk threshold. ,in, It is a normalized random factor that is uniformly distributed within a predetermined range.

[0145] This invention introduces a quantum entropy source to generate a high-quality random number seed. Based on this random number seed, a preset risk threshold is dynamically adjusted, which significantly improves the randomness and unpredictability of risk assessment. This prevents attackers from using statistical or machine learning methods to infer fixed thresholds and bypass attacks, thereby enhancing the system's sensitivity to potential risks and its defense capabilities. This further ensures the security of transactions in the Metaverse Bank space and the authenticity of user identities.

[0146] Optionally, in the above Figure 1Based on one or more corresponding embodiments, in another optional embodiment provided by the present invention, step S180 may specifically include:

[0147] If the similarity deviation value corresponds to low risk, inject a hidden decoy element containing a fake API interface and fake data packets into the user's current transaction interface; if the similarity deviation value corresponds to medium risk, forcibly interrupt the current transaction process and enable multimodal review and verification; if the similarity deviation value corresponds to high risk, freeze the user's financial account and initiate cross-chain audit to trace the attack path.

[0148] Specifically, when the similarity deviation value corresponds to low risk, an invisible decoy element is automatically injected into the user's current virtual transaction interface. This decoy includes a fake API interface and a fake data packet that is encrypted but has abnormal logic. These elements are seamlessly embedded in the interface and do not interfere with normal user operations. However, once triggered by an attacker, the operation node is immediately marked as a potential source of malicious activity in the background, and the information is written to the local security log, forming an early attack identification and trapping mechanism.

[0149] As examples, embodiments of the present invention can dynamically generate a visually completely invisible (e.g., 0% transparency or perfect blending with the background) and logically highly realistic interactive element based on the current transaction scenario (e.g., transfer, query). This could be a fake API interface disguised as a "confirm" button, or an invisible data packet with compliant encrypted format but containing false transaction information. These decoy elements are seamlessly injected into the user's current VR / AR transaction interface rendering stream, completely invisible and uninteractive to normal users. The system backend synchronously establishes a "decoy-request" mapping table and starts listening. If the invisible decoy is triggered by any form of interactive request (e.g., click, data call), the system will immediately determine that the request source node is a malicious attacker, automatically record its device ID, IP, behavioral fingerprint, etc., and mark it as a malicious node, dynamically updating it to a local risk list for subsequent tracking and escalation processing. This proactively captures and marks potential attack probes without interfering with normal users.

[0150] Specifically, in cases where the similarity deviation value corresponds to medium risk, the current transaction process is immediately and forcibly interrupted to prevent its continued execution, and a multimodal verification process is initiated. This includes activating the voiceprint recognition module to perform voiceprint confirmation based on an anti-synthetic attack phase perturbation algorithm, while simultaneously activating the gesture recognition module to accurately identify the user's hand joint movement trajectory using a convolutional attitude machine. The user is required to complete a specific gesture sequence corresponding to the private key mapping space vector. Only after this dual verification ensures the authenticity of the user's identity can the transaction be resumed.

[0151] As examples, embodiments of the present invention can immediately send an interrupt command to the transaction processing engine, forcibly suspending all subsequent operations of the current transaction process, and displaying a prominent security warning and review prompt on the user interface. Subsequently, based on a preset "risk-response matrix," a set of multimodal review and verification processes is dynamically selected and initiated. Typical combinations include: 1) Voiceprint review: requiring the user to read aloud a randomly generated dynamic phrase. The system injects a random frequency offset during collection to resist synthetic attacks and matches it with the voiceprint features stored on the chain; 2) Gesture review: calling a convolutional gesture machine model through a VR controller or camera, requiring the user to execute a set of gesture sequences according to a specific spatial vector trajectory mapped by their private key (such as drawing a specific shape in the air). Only after all review modalities have passed verification will the transaction process be resumed and reauthorized; failure in any step will raise the risk level to high risk.

[0152] Specifically, when a similarity deviation value corresponds to a high risk, the system automatically freezes the user's financial account, prohibits all trading activities, and prevents further asset losses. At the same time, it triggers a cross-chain audit mechanism, calling data from both the consortium blockchain and the public blockchain, verifying identity mapping through zero-knowledge proofs, and using a lightweight cross-chain relay to collect and compress transaction behavior evidence, quickly reconstructing attack paths. This supports the security team in tracing and investigating abnormal behavior and handling emergencies, comprehensively ensuring the safety of user funds and the compliance of platform transactions.

[0153] As examples, embodiments of the present invention can send instructions to the core account management system to immediately freeze all financial accounts (including deposits, wealth management, payments, etc.) associated with a user, blocking any further asset transfer operations, and informing the user through notification channels. Simultaneously, cross-chain auditing and tracing are initiated: the cross-chain identity anchoring module is activated, calling the smart contract on the public chain, initiating a joint verification request to the consortium chain based on zero-knowledge proofs, retrieving and decrypting all on-chain private component operation logs, behavioral key generation records, and network node information related to this abnormal transaction. A lightweight cross-chain relay efficiently synchronizes data, and the audit engine quickly reconstructs the complete attack chain from attack probing, triggering decoys, abnormal behavior to the final transaction request, locating the attack entry point and potentially leaked links, generating a detailed audit report, and providing an immutable chain of evidence for subsequent judicial evidence collection, system hardening, and loss recovery.

[0154] This invention, in its embodiments, injects stealth decoy elements to lure potential attackers under low-risk conditions, enhancing early threat identification capabilities; under medium-risk conditions, it forcibly interrupts transactions and enables multimodal verification, strengthening identity authentication; and under high-risk conditions, it freezes accounts and initiates cross-chain auditing and tracing, ensuring rapid disruption of the attack chain and comprehensive source tracing. This tiered response mechanism effectively improves the security of Metaverse Bank transactions and the level of user fund protection, realizing a dynamic, intelligent, and layered risk control system.

[0155] Although the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. In certain environments, multitasking and parallel processing may be advantageous.

[0156] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0157] Corresponding to the above method embodiments, this invention also provides a bank virtual space transaction risk control device, the structure of which is as follows: Figure 6 As shown, it may include: a multimodal biometric information acquisition unit 10, an interaction behavior feature information acquisition unit 20, a dynamic behavior key generation unit 30, a cross-chain storage unit 40, a real-time transaction behavior data acquisition unit 50, a temporary key generation unit 60, a similarity deviation value acquisition unit 70, a risk judgment unit 80, and a risk defense activation unit 90.

[0158] The multimodal biometric information acquisition unit 10 is used to acquire multimodal biometric information of the user when entering the Metaverse Bank space.

[0159] The interactive behavior feature information acquisition unit 20 is used to acquire the interactive behavior feature information of users in the metaverse bank space.

[0160] The dynamic behavior key generation unit 30 is used to generate dynamic behavior keys by utilizing multimodal biometric information and interaction behavior information.

[0161] Cross-chain storage unit 40 is used for cross-chain storage of dynamic behavior keys.

[0162] The real-time transaction behavior data acquisition unit 50 is used to acquire the user's real-time transaction behavior data when it detects that the user has triggered a transaction in the metaverse bank space.

[0163] The temporary key generation unit 60 is used to generate temporary keys using multimodal biometric information and real-time transaction behavior data.

[0164] The similarity deviation value acquisition unit 70 is used to perform dynamic time-normalized similarity matching between the temporary key and the dynamic behavior key on the chain to obtain the similarity deviation value.

[0165] Risk assessment unit 80 is used to determine whether the similarity deviation value is equal to or greater than the preset risk threshold. If so, risk defense activation unit 90 is triggered.

[0166] Risk defense activation unit 90 is used to activate the risk defense strategy corresponding to the similarity deviation value.

[0167] Optionally, the cross-chain storage unit 40 can be used to decompose the dynamic behavior key into a private component and a public component; store the private component in a private consortium chain after homomorphic encryption; write the public component into the public blockchain after binding it with the user's financial account; and construct a verifiable mapping relationship between the encrypted private component on the private consortium chain and the public component on the public blockchain.

[0168] Optionally, the interaction behavior feature information acquisition unit 20 may specifically include: an interaction behavior sequence recording subunit and a feature information extraction subunit.

[0169] The interaction behavior sequence recording sub-unit is used to record the user's interaction behavior sequence in the metaverse bank space.

[0170] The feature information extraction subunit is used to perform time-series analysis on the interaction behavior sequence and extract the user's interaction behavior feature information in the metaverse bank space.

[0171] Optionally, the interactive behavior sequence recording sub-unit can be used to map user operation events within the metaverse bank space into a sequence of coordinate points in a preset coordinate system, wherein the preset coordinate system is a three-dimensional spatiotemporal coordinate system with the virtual counter coordinate position within the metaverse bank space as the reference origin; record the user's behavioral entropy value within the metaverse bank space; and generate the user's interactive behavior sequence within the metaverse bank space using the coordinate point sequence and behavioral entropy value.

[0172] Optionally, the feature information extraction subunit can be used to perform spatiotemporal embedding processing on the interaction behavior sequence using a Transformer encoder to obtain a spatiotemporal embedding vector; based on the spatiotemporal embedding vector, use an adversarial generative network to create a behavior feature enhancement dataset corresponding to the user; use the behavior feature enhancement dataset to optimize the feature extraction model through contrastive learning; and use the optimized feature extraction model to extract the user's interaction behavior feature information in the metaverse bank space from the spatiotemporal embedding vector.

[0173] Optionally, the bank's virtual space transaction risk control device may also include: a risk threshold dynamic adjustment unit.

[0174] The risk threshold dynamic adjustment unit is used to generate a random number seed using a quantum entropy source before the risk judgment unit 80 judges whether the similarity deviation value is equal to or greater than the preset risk threshold; based on the random number seed, the preset risk threshold is dynamically adjusted.

[0175] Optionally, the risk defense initiation unit 90 can be used to inject a hidden decoy element containing a fake API interface and fake data packets into the user's current transaction interface when the similarity deviation value corresponds to low risk; to forcibly interrupt the current transaction process and enable multimodal review and verification when the similarity deviation value corresponds to medium risk; and to freeze the user's financial account and initiate cross-chain audit to trace the attack path when the similarity deviation value corresponds to high risk.

[0176] This invention provides a risk control device for virtual space transactions in banks. The device is used to: obtain multimodal biometric information of a user entering a metaverse banking space; obtain user interaction behavior characteristics within the metaverse banking space; generate a dynamic behavior key using the multimodal biometric information and interaction behavior characteristics; store the dynamic behavior key across blockchains; obtain real-time transaction behavior data of the user when a transaction is detected in the metaverse banking space; generate a temporary key using the multimodal biometric information and real-time transaction behavior data; perform dynamic time-warping similarity matching between the temporary key and the on-chain dynamic behavior key to obtain a similarity deviation value; determine whether the similarity deviation value is equal to or greater than a preset risk threshold; if so, activate the risk defense strategy corresponding to the similarity deviation value. This invention, by collecting multimodal biometric and interaction behavior characteristics of users in the metaverse banking space, generating and storing dynamic behavior keys across blockchains, dynamically generating temporary keys based on real-time transaction behavior data, and judging risks through dynamic time-warping similarity matching, achieves continuous dynamic verification of identity and proactive risk defense, effectively improving the security and reliability of identity authentication in complex virtual environments.

[0177] Regarding the apparatus in the above embodiments, the specific manner in which each unit performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0178] The bank's virtual space transaction risk control device includes a processor and a memory. The aforementioned multimodal biometric information acquisition unit 10, interactive behavior feature information acquisition unit 20, dynamic behavior key generation unit 30, cross-chain storage unit 40, real-time transaction behavior data acquisition unit 50, temporary key generation unit 60, similarity deviation value acquisition unit 70, risk judgment unit 80, and risk defense activation unit 90 are all stored as program units in the memory. The processor executes the aforementioned program units stored in the memory to achieve the corresponding functions.

[0179] The processor contains a kernel that retrieves the corresponding program units from memory. One or more kernels can be configured. By adjusting kernel parameters, when a user enters the Metaverse Bank space, multimodal biometrics and interaction behavior characteristics are collected, generating a dynamic behavior key and storing it across the chain. When a user triggers a transaction, real-time transaction behavior data is acquired and combined with multimodal biometrics to generate a temporary key. A dynamic time warping method is used to match the temporary key with the on-chain dynamic behavior key to calculate the deviation value. When the deviation value reaches a preset risk threshold, an appropriate risk defense strategy is activated. This achieves continuous dynamic verification of user identity and proactive risk defense, significantly improving the security and reliability of identity authentication in complex virtual environments.

[0180] This invention provides a computer-readable storage medium storing a program that, when executed by a processor, implements a method for controlling risks in virtual space transactions within a bank.

[0181] This invention provides a processor for running a program, wherein the program executes a bank virtual space transaction risk control method during runtime.

[0182] like Figure 7 As shown, this embodiment of the invention provides an electronic device 1000, which includes at least one processor 1001, at least one memory 1002 connected to the processor 1001, and a bus 1003. The processor 1001 and the memory 1002 communicate with each other via the bus 1003. The processor 1001 is used to call program instructions in the memory 1002 to execute the aforementioned bank virtual space transaction risk control method. The electronic device in this document can be a server, PC, PAD, mobile phone, etc.

[0183] The present invention also provides a computer program product that, when executed on an electronic device, is suitable for executing the steps of an initialization method for risk control of virtual space transactions in banking.

[0184] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatuses, electronic devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0185] In a typical configuration, an electronic device includes one or more processors (CPUs), memory, and a bus. The electronic device may also include input / output interfaces, network interfaces, etc.

[0186] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, like read-only memory (ROM) or flash RAM, and memory includes at least one memory chip. Memory is an example of computer-readable media.

[0187] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0188] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this invention are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0189] It is understood that before using the technical solutions disclosed in the various embodiments of this disclosure, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this disclosure in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0190] In the description of this invention, it should be understood that if the terms "upper", "lower", "front", "rear", "left" and "right" are used to indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the position or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention.

[0191] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0192] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0193] The above are merely embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the present invention.

Claims

1. A method for controlling risks in virtual space transactions for banks, characterized in that, include: Obtain multimodal biometric information of users when they enter the Metaverse Bank space; Obtain the user's interactive behavior characteristics information in the metaverse banking space; Using the multimodal biometric information and the interaction behavior information, a dynamic behavior key is generated; The dynamic behavior key is stored across chains; When the user triggers a transaction in the metaverse bank space, the user's real-time transaction behavior data is obtained; A temporary key is generated using the multimodal biometric information and the real-time transaction behavior data; The temporary key is matched with the dynamic behavior key on the chain using dynamic time warping similarity to obtain a similarity deviation value; Determine whether the similarity deviation value is equal to or greater than a preset risk threshold. If so, activate the risk defense strategy corresponding to the similarity deviation value.

2. The method according to claim 1, characterized in that, The cross-chain storage of the dynamic behavior key includes: The dynamic behavior key is decomposed into a private component and a public component; The private component is homomorphically encrypted and stored on a private consortium blockchain. The public component is bound to the user's financial account and written into a public blockchain. A verifiable mapping relationship is then constructed between the encrypted private component on the private consortium blockchain and the public component on the public blockchain.

3. The method according to claim 1, characterized in that, The process of obtaining the user's interactive behavior characteristics information in the metaverse banking space includes: Record the sequence of user interaction behaviors in the metaverse banking space; The interaction behavior sequence is subjected to time series analysis to extract the user's interaction behavior feature information in the metaverse bank space.

4. The method according to claim 3, characterized in that, The recording of the user's interaction sequence in the metaverse bank space includes: The user's operation events within the Metaverse Bank Space are mapped to a sequence of coordinate points in a preset coordinate system, wherein the preset coordinate system is a three-dimensional spatiotemporal coordinate system with the virtual counter coordinate position within the Metaverse Bank Space as the reference origin; Record the user's behavioral entropy value within the metaverse bank space; Using the coordinate point sequence and the behavior entropy value, the user's interaction behavior sequence in the metaverse bank space is generated.

5. The method according to claim 4, characterized in that, The step of performing time-series analysis on the interaction behavior sequence to extract the user's interaction behavior feature information in the metaverse bank space includes: The interaction sequence is spatiotemporally embedded using a Transformer encoder to obtain a spatiotemporal embedding vector. Based on the spatiotemporal embedding vector, an adversarial generative network is used to create a behavioral feature enhancement dataset corresponding to the user. The dataset is augmented using the aforementioned behavioral features, and the feature extraction model is optimized through contrastive learning. The optimized feature extraction model is used to extract the user's interactive behavior feature information in the metaverse bank space from the spatiotemporal embedding vector.

6. The method according to claim 1, characterized in that, Before determining whether the similarity deviation value is equal to or greater than a preset risk threshold, the method further includes: Random number seeds are generated using a quantum entropy source; The preset risk threshold is dynamically adjusted based on the random number seed.

7. The method according to claim 1, characterized in that, The activation of the risk defense strategy corresponding to the similarity deviation value includes: When the similarity deviation value corresponds to low risk, an invisible decoy element containing a fake API interface and fake data packets is injected into the user's current transaction interface; In cases where the similarity deviation value corresponds to medium risk, the current transaction process is forcibly interrupted, and multimodal review and verification are enabled. If the similarity deviation value corresponds to a high risk, the user's financial account will be frozen, and a cross-chain audit will be initiated to trace the attack path.

8. A bank virtual space transaction risk control device, characterized in that, include: The system includes a multimodal biometric information acquisition unit, an interactive behavior feature information acquisition unit, a dynamic behavior key generation unit, a cross-chain storage unit, a real-time transaction behavior data acquisition unit, a temporary key generation unit, a similarity deviation value acquisition unit, a risk judgment unit, and a risk defense activation unit. The multimodal biometric information acquisition unit is used to acquire multimodal biometric information of the user when entering the Metaverse Bank space; The interactive behavior feature information acquisition unit is used to acquire the interactive behavior feature information of the user in the metaverse bank space; The dynamic behavior key generation unit is used to generate a dynamic behavior key using the multimodal biometric information and the interaction behavior feature information; The cross-chain storage unit is used to store the dynamic behavior key across chains; The real-time transaction behavior data acquisition unit is used to acquire the user's real-time transaction behavior data when it detects that the user has triggered a transaction in the metaverse bank space. The temporary key generation unit is used to generate a temporary key using the multimodal biometric information and the real-time transaction behavior data; The similarity deviation value acquisition unit is used to perform dynamic time-normalized similarity matching between the temporary key and the dynamic behavior key on the chain to obtain a similarity deviation value. The risk judgment unit is used to determine whether the similarity deviation value is equal to or greater than a preset risk threshold. If so, the risk defense activation unit is triggered. The risk defense activation unit is used to activate the risk defense strategy corresponding to the similarity deviation value.

9. A computer-readable storage medium having a program stored thereon, characterized in that, When the program is executed by the processor, it implements the bank virtual space transaction risk control method as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, The electronic device includes at least one processor, at least one memory connected to the processor, and a bus; wherein the processor and the memory communicate with each other through the bus; the processor is used to call program instructions in the memory to execute the bank virtual space transaction risk control method as described in any one of claims 1 to 7.

Images