Multi-factor fraud link tracking method and system based on artificial intelligence

Abstract

The invention discloses a multi-factor fraud link tracking method and system based on artificial intelligence, and relates to the field of data security tracking. The method comprises the following steps: constructing a multi-modal behavior disturbance structure model, extracting disturbance response through a multi-dimensional behavior source signal, and calculating disturbance absorption intensity; constructing an inter-entity energy flow coupling relation graph based on the disturbance absorption intensity; performing structure propagation and embedding representation updating according to the energy flow coupling relation graph, and iteratively generating entity structure embedding representation by adopting an edge weight driven nonlinear propagation mechanism; calculating a link credibility score for the path sequence between any entities, and constructing a fraud link set; and in combination with the link credibility score and the entity structure embedded representation, calculating the reverse sensitivity of the entity in the fraud link, and generating a response control strategy to realize traceability control. And through multi-modal disturbance modeling, energy flow coupling structure construction and a multi-factor path evaluation mechanism, the fraudulent behavior identification precision is remarkably improved, and the chain modeling capability and the link screening robustness are coordinated.

Description

Technical Field

[0001] The present invention relates to the field of data security tracking, and specifically to an artificial intelligence-based multi-factor fraud link tracking method and system. Background Art

[0002] With the development of information technology and the continuous expansion of digital transaction scenarios, multi-source, heterogeneous, high-frequency, and dynamic fraud has emerged in online environments. Traditional fraud detection methods based on rule matching, feature clustering, or static graph mining have shown problems such as insufficient recognition, delayed response, and difficulty in tracing the source of complex collaborative fraud patterns. In applications such as financial transactions, e-commerce activities, social networks, and smart IoT, fraudsters often evade the discriminative mechanisms of conventional detection algorithms by constructing hidden collaborative links, manipulating behavioral perturbations, and avoiding feature homogeneity, creating systemic risk control blind spots.

[0003] Existing approaches attempt to incorporate algorithms such as graph neural networks and graph embedding learning for structurally enhanced modeling, aiming to improve the accuracy of fraudulent entity identification. However, these approaches often rely on static adjacency relationships and fixed propagation rules, making it difficult to dynamically characterize the evolution of behavioral perturbations and the structure of cross-modal collaborative behaviors. Furthermore, they lack systematic modeling mechanisms for trustworthy path link modeling, perturbation consistency constraints, and behavioral complexity characterization.

[0004] Furthermore, current fraud tracking solutions generally overlook the differences in entities' ability to absorb external disturbances and the dynamic evolution of coupled response strength in multimodal behavior scenarios, resulting in inaccurate identification of potential collaborative relationships between entities and the establishment of link credibility. Furthermore, the lack of a unified modeling mechanism for structural collaboration and disturbance response within a path makes it difficult to develop an interpretable and controllable fraud link assessment method, limiting the ability to achieve global perception and reverse tracing of fraud chains.

[0005] Therefore, it is urgent to propose a new fraud link tracking technology that integrates multimodal disturbance analysis, energy flow coupling modeling, structural propagation mechanism and path credibility assessment to achieve highly robust identification of complex fraud behaviors, linkage relationship modeling and refined traceability control. Summary of the Invention

[0006] Based on the above-mentioned shortcomings of the prior art, the purpose of the present invention is to provide a multi-factor fraud link tracking method and system based on artificial intelligence to solve the above-mentioned technical problems.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a multi-factor fraud link tracing method based on artificial intelligence, comprising: Construct a multimodal behavioral disturbance structure model, extract the disturbance response through the temporal evolution of multidimensional behavioral source signals in multiple modal domains, and calculate the disturbance absorption intensity; Based on the disturbance absorption intensity and the nonlinear interaction between behaviors, an energy flow coupling relationship diagram between entities is constructed, and entities with potential fraudulent collaborative behaviors are marked as high-intensity connection edges. According to the energy flow coupling relationship graph, the structure propagation and embedding representation update are carried out, and the entity structure embedding representation is iteratively generated using the edge weight driven nonlinear propagation mechanism; Evaluate the credibility of path sequences between any entities, calculate link credibility scores, and construct a set of fraudulent links with perturbation consistency, path coordination, and low behavioral complexity differences; Combining the link credibility score and entity structure embedding representation, a path reverse contribution function is constructed to calculate the reverse sensitivity of the entity in the fraud link, and a response control strategy is generated to achieve traceability control.

[0008] The present invention is further configured such that the step of constructing the multimodal behavior disturbance structure model comprises: Model the disturbance response of the time evolution of multi-dimensional behavioral source signals in different modal domains and extract the behavioral dynamic disturbance trajectory with nonlinear change characteristics; Construct disturbance energy consumption based on behavioral dynamic disturbance trajectory and obtain disturbance absorption energy under modal conditions; The absorption energy under all modes is integrated to calculate the inter-modal disturbance absorption intensity, which measures the response ability and absorption characteristics of the behavioral entity to external disturbances under different modes.

[0009] The present invention is further configured such that the process of constructing the energy flow coupling relationship diagram between entities includes: Generate the behavioral coupling responsiveness between entity pairs according to the temporal evolution results of the multimodal behavioral interaction function; Generate energy flow induction weights between entity pairs according to disturbance absorption intensity and behavioral coupling responsiveness; According to the preset coupling strength threshold, the entity pairs with high-weight connections in the graph structure are determined as the connection nodes of the edges to form an energy flow coupling relationship diagram between the entities.

[0010] The present invention is further configured such that the generation of the entity structure embedding representation includes: Assigning an initial entity structure embedding representation based on the perturbation absorption strength of each entity; According to the constructed energy flow coupling relationship graph, based on the adjacency relationship of each entity, the diffusion contribution of adjacent entities to the current round of entity structure embedding representation is calculated through the edge weight driven nonlinear propagation mechanism; The entity structure embedding representation of the previous round is nonlinearly combined with the diffusion contribution of the current round to form the entity structure embedding representation of the next round; Multiple rounds of propagation and updating are performed until the entity structure embedding representation converges or reaches a preset iteration limit, and finally a stable entity structure embedding representation of each entity is obtained.

[0011] The present invention is further configured such that the constructing of the fraud link set includes: Extract the path sequence between any entity pairs and construct a set of path factors based on the disturbance absorption strength of each node in the path, the entity structure embedding representation and the complex coding characteristics of the behavior; Constructing a disturbance consistency factor based on the disturbance absorption intensity in the path; Constructing path synergy factors based on the embedded representation of entity structures in the path; Constructing behavioral complexity difference factors based on the behavioral complexity coding in the path; Calculate the link credibility score corresponding to the path according to the preset factor fusion rules; Combined with the link credibility score threshold, the path sequences with high credibility features in all paths are screened to construct a fraudulent link set.

[0012] The present invention is further configured such that the construction logic of the disturbance consistency factor is: Extract the disturbance absorption strength of each entity node based on the path factor set; The consistency of disturbance response within a path is measured based on the nonlinear variation characteristics of the disturbance absorption strength between adjacent nodes. The disturbance absorption offset and power control factor are set, and the disturbance difference pairs within the path are multiplied and accumulated to construct the overall disturbance consistency factor.

[0013] The present invention is further configured such that the construction logic of the path synergy factor is: According to any pair of adjacent entity nodes in any path, the entity structure embedding representation is obtained, and the power control operation is applied to each pair of adjacent entity nodes, and then element-by-element mapping and fusion are performed. The full path embedding coupling sequence is compressed and represented by a chain nested product method. A nonlinear feature mapping function is constructed to extract the stable kernel response of the multi-level embedding product features and generate the path structure synergy factor.

[0014] The present invention is further configured such that the calculation of the behavioral complex coding includes: Obtain the original behavioral event sequence of the target entity in multiple behavioral modal domains, including interactive behavior, triggering behavior, response behavior, and modal jump behavior; Perform structured discretization processing on the original behavioral event sequence to extract a multi-dimensional behavioral feature set; Based on the multidimensional behavioral feature set, the original behavioral complexity factor is constructed; The original behavior complexity factor is perturbation-enhanced and nonlinearly interactively fused to generate behavior complexity coding.

[0015] The present invention is further configured such that the generating response control strategy comprises: Based on the link credibility score and entity structure embedding representation, a path reverse contribution function is constructed to model the degree of structural deviation of the terminal node in the fraud link relative to the path source node; Calculate the reverse sensitivity for each entity in the fraud path in which it participates; All reverse sensitivities are modeled inductively, and a set of response control strategies is constructed by combining the structural embedding distance between entities and the interference modulation factor, indicating the control level distribution of different entities in the traceability chain.

[0016] The present invention also provides an artificial intelligence-based multi-factor fraud link tracking system, the system comprising: Multimodal disturbance extraction module: Constructs a multimodal behavioral disturbance structure model, extracts disturbance responses through the temporal evolution of multidimensional behavioral source signals in multiple modal domains, and calculates disturbance absorption intensity; Energy flow coupling relationship construction module: Based on the disturbance absorption intensity and the nonlinear interaction relationship between behaviors, an energy flow coupling relationship diagram between entities is constructed, and entities with potential fraudulent collaborative behaviors are marked as high-intensity connection edges; Structural Propagation and Embedding Update Module: This module performs structural propagation and embedding representation updates based on the energy flow coupling relationship graph, and uses an edge-weight-driven nonlinear propagation mechanism to iteratively generate entity structure embedding representations. Fraud Link Credibility Assessment Module: This module evaluates the credibility of path sequences between any entities, calculates link credibility scores, and constructs a set of fraud links that exhibit perturbation consistency, path coordination, and low behavioral complexity. Reverse traceability control decision module: Combines the link credibility score and entity structure embedding representation to construct a path reverse contribution function, calculates the reverse sensitivity of the entity in the fraud link, and generates a response control strategy to achieve traceability control.

[0017] The present invention provides a multi-factor fraud link tracing method and system based on artificial intelligence. The method constructs a multimodal behavior disturbance structure model, extracts disturbance responses through the temporal evolution of multidimensional behavior source signals in multiple modal domains, and calculates disturbance absorption strength. Based on the disturbance absorption strength and the nonlinear interaction relationship between behaviors, an energy flow coupling relationship diagram between entities is constructed, and entities with potential fraud collaborative behaviors are marked as high-intensity connection edges. Structural propagation and embedding representation update are performed according to the energy flow coupling relationship diagram, and an edge-weight-driven nonlinear propagation mechanism is used to iteratively generate entity structure embedding representations. The method performs credibility evaluation on the path sequence between any entities, calculates the link credibility score, and constructs a fraud link set with disturbance consistency, path synergy, and low behavioral complexity difference. Combined with the link credibility score and the entity structure embedding representation, a path reverse contribution function is constructed, the reverse sensitivity of the entity in the fraud link is calculated, and a response control strategy is generated to achieve traceability control. The beneficial effects produced include: 1. Improve the accuracy of disturbance response modeling: By constructing a multimodal behavior disturbance structure model, we systematically extract the nonlinear disturbance absorption characteristics of multi-source behavior signals in each modal domain, characterize the entity's dynamic response capability to external disturbances, and enhance the fine-grained expression capability of fraud behavior modeling. 2. Enhanced entity collaboration identification capabilities: By constructing an energy flow coupling relationship diagram, a coupling mechanism between disturbance absorption intensity and behavioral interaction responsiveness is introduced to achieve structural modeling of potential collaborative behavior chains between entities, improving the accuracy of entity association identification in complex fraud behavior networks. 3. Construct a multi-factor path credibility scoring mechanism: By introducing multi-dimensional path factors such as disturbance consistency factor, path synergy factor, and behavioral complexity difference, a fine-grained link credibility scoring model is constructed to significantly improve the recognition accuracy and screening robustness of fraudulent paths.

[0018] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without inventive efforts. In the drawings: Figure 1 This is a flow chart of a multi-factor fraud link tracing method based on artificial intelligence, according to an exemplary embodiment of the present invention; Figure 2 The figure is a schematic structural diagram of a multi-factor fraud link tracking system based on artificial intelligence, which is an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0020] The following describes the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art will readily appreciate the other advantages and benefits of the present invention from the disclosure herein. The present invention may also be implemented or applied through various other specific embodiments, and the various details in this specification may be modified or altered based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are intended only to illustrate the present invention and are not intended to limit the scope of protection of the present invention.

[0021] It should be noted that the illustrations provided in the following embodiments are merely schematic illustrations of the basic concept of the present invention. Therefore, the illustrations only show components related to the present invention and are not drawn according to the number, shape, and size of components in actual implementation. In actual implementation, the type, quantity, and proportion of each component may be changed arbitrarily, and the component layout may also be more complex.

[0022] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it will be apparent to those skilled in the art that the embodiments of the present invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring the embodiments of the present invention.

[0023] Example 1: A multi-factor fraud link tracing method based on artificial intelligence, such as Figure 1 Shown, including: Construct a multimodal behavioral disturbance structure model, extract the disturbance response through the temporal evolution of multidimensional behavioral source signals in multiple modal domains, and calculate the disturbance absorption intensity; Based on the disturbance absorption intensity and the nonlinear interaction between behaviors, an energy flow coupling relationship diagram between entities is constructed, and entities with potential fraudulent collaborative behaviors are marked as high-intensity connection edges. According to the energy flow coupling relationship graph, the structure propagation and embedding representation update are carried out, and the entity structure embedding representation is iteratively generated using the edge weight driven nonlinear propagation mechanism; Evaluate the credibility of path sequences between any entities, calculate link credibility scores, and construct a set of fraudulent links with perturbation consistency, path coordination, and low behavioral complexity differences; Combining the link credibility score and entity structure embedding representation, a path reverse contribution function is constructed to calculate the reverse sensitivity of the entity in the fraud link, and a response control strategy is generated to achieve traceability control.

[0024] The present invention is further configured such that the step of constructing the multimodal behavior disturbance structure model comprises: The disturbance response modeling is performed on the time evolution process of the multi-dimensional behavior source signal in different modal domains, and the behavioral dynamic disturbance trajectory with nonlinear change characteristics is extracted. Specifically, the intra-modal disturbance response structure is constructed. Based on the nonlinear change rate and mutation sensitivity factor of the behavior signal on the time axis, the disturbance response basis is formed. The formula is: , in, Behavioral Entity exist Moment Mode The following behavioral timing signals, is the behavioral acceleration component, used to detect mutation trends. is the modal nonlinear acceleration response coefficient, is the power amplification of the first-order growth rate of the mode, is the perturbation activation bias, is the time growth tension function, where , is the disturbance response mapping function, ; Construct the disturbance energy consumption according to the behavioral dynamic disturbance trajectory, obtain the disturbance absorption energy under the mode, define the disturbance absorption power density function, and construct the disturbance energy consumption by combining the response change rate and the modal amplification function. The formula is: , in, For the The disturbance under the mode absorbs energy, is the intensity of disturbance response change, For modal absorbing converters, trade-off between behavioral response and time variation factors; The absorption energy under all modes is integrated to calculate the inter-modal disturbance absorption intensity, which measures the response ability and absorption characteristics of the behavioral entity to external disturbances under different modes. The disturbance absorption intensity calculation formula is: , in, is the disturbance absorption strength, which reflects the instability of the entity in multimodal behavior and is used for fraud potential modeling. is the modal non-equilibrium amplification index, is the number of modes, is the full modal power reduction coefficient, which controls the convergence rate of modal fusion.

[0025] The present invention is further configured such that the process of constructing the energy flow coupling relationship diagram between entities includes: Generate the behavioral coupling responsiveness between entity pairs based on the temporal evolution of the multimodal behavioral interaction function. Specifically, define the entity set , the interaction function family between behavioral modes , representing an entity and In modal 、 The calculation logic of the cross-behavior response curve on the behavior coupling response is: , in, For entity pairs , The behavioral coupling responsiveness, is the first-order time derivative of the interaction function between behavioral modes, which is used to characterize the interaction rate. is the time-sensitive modulation function, is the regulating factor; The energy flow induction weight between the entity pairs is generated according to the disturbance absorption intensity and the behavioral coupling responsiveness. The disturbance absorption intensity and the behavioral coupling responsiveness of the two entities are jointly nonlinearly reconstructed to construct the energy flow induction weight. ,in, , For Entity 、 The disturbance absorption intensity, , is the absorption weight index, is the behavioral coupling response amplification index, is the adjustment factor, harmonic convergence scale control, is the disturbance heterogeneity enhancement factor, is the energy flow induction weight between the entity pairs; According to the preset coupling strength threshold, the entity pairs with high weight connections in the graph structure are determined as the connection nodes of the edge, forming an energy flow coupling relationship diagram between entities, and defining the graph. , where the edge weight function is given by Determine and set the coupling strength threshold , determine high-risk connection edges, edge set definition formula: ,like , then and Connected edges indicate potential fraudulent coordination.

[0026] The present invention is further configured such that the generation of the entity structure embedding representation includes: Based on the perturbation absorption strength of each entity, the initial entity structure embedding representation is assigned. Specifically, for each entity Allocate initial entity structure representation , and its dimensional elements and perturbation absorption intensity are nonlinearly perturbed and scaled: , ,in, Dimension The initial entity structure embedding representation of is the embedding dimension, is the nonlinear power index, Initialize the noise perturbation for the perturbation; According to the constructed energy flow coupling relationship graph, based on the adjacency relationship of each entity, the diffusion contribution of the adjacent entity to the current round of entity structure embedding representation is calculated through the edge weight driven nonlinear propagation mechanism. The propagation operation with edge weight activation function is defined above, and all entity structure representations of each entity are coupled and diffused through its adjacent edges. The propagation formula is: , in, Contribute to diffusion, is the edge weight activation index, For Entity The set of adjacent entities of is the entity structure embedding representation of the adjacent nodes in the previous round, is the edge-weight driven adaptive perturbation weight vector, Element-wise multiplication operation, is the nonlinear mapping kernel function; The entity structure embedding representation of the previous round is nonlinearly combined with the diffusion contribution of the current round to form the entity structure embedding representation of the next round. The formula is: , , To update the scale factor, Represents a splicing operation, Embedding representation for the entity structure of the previous round, is the perturbation gating function, represents the bitwise power product; Perform multiple propagation and update rounds until the entity structure embedding representation converges or reaches the preset iteration limit, and finally obtain the stable entity structure embedding representation of each entity , is the number of transmissions.

[0027] The present invention is further configured such that the constructing of the fraud link set includes: Extract the path sequence between any entity pair, and construct the path factor set based on the disturbance absorption strength of each node in the path, the entity structure embedding representation, and the complex coding characteristics of the behavior. All the reachable path sequence sets are: , each path Represented as an ordered sequence of entities ; Constructing a disturbance consistency factor based on the disturbance absorption intensity in the path ; Constructing path synergy factor based on entity structure embedding representation in path ; Complex encoding based on behavior in the path Constructing behavioral complexity difference factors , specifically, ,in, is the difference regularization term, The smaller the value, the higher the stability of the path behavior; According to the preset factor fusion rules, the link credibility score corresponding to the path is calculated. The calculation formula is: , in, , , To balance the regulatory factors, is the overall compression modulation parameter, Score the link credibility; Combined with the link credibility score threshold, the path sequence with high credibility features in all paths is screened, a fraud link set is constructed, and a preset link credibility score threshold is set. , then the fraud link set .

[0028] The present invention is further configured such that the construction logic of the disturbance consistency factor is: Extract the disturbance absorption strength of each entity node based on the path factor set; The consistency of disturbance response within a path is measured based on the nonlinear variation characteristics of the disturbance absorption strength between adjacent nodes. Set the disturbance absorption offset and power control factor, multiply and accumulate the disturbance difference pairs in the path, and construct the overall disturbance consistency factor. Specifically, for the path , extract the disturbance absorption code value sequence of all path nodes , construct the perturbation consistency factor , the calculation formula is: , in, For the The node disturbance absorption strength, is the perturbation power factor, Absorbs the offset for disturbance; is the disturbance consistency factor. The closer the value is to 1, the more consistent the disturbance response within the path is. is the number of path nodes.

[0029] The present invention is further configured such that the construction logic of the path synergy factor is: According to any pair of adjacent entity nodes in any path, the entity structure embedding representation is obtained, and the power control operation is applied to each pair of adjacent entity nodes, and then element-by-element mapping and fusion are performed. The chain nested product method is used to compress the full path embedding coupling sequence, construct a nonlinear feature mapping function, extract the stable kernel response of the multi-level embedding product feature, and generate the path synergy factor. Specifically, the path Each pair of consecutive entities , The structural embedding representations of , , defining the path synergy factor The chained kernel embedding compression result of the product of all embedding mapping vectors on the path is: , in, is the structural embedding power operation coefficient, is the element-wise product, is a chained nested product operator, is the mapping compression function, defined as: , , A larger value indicates a stronger path structure synergy.

[0030] The present invention is further configured such that the calculation of the behavioral complex coding includes: Obtain the original behavioral event sequence of the target entity in multiple behavioral modal domains, including interactive behavior, triggering behavior, response behavior, and modal jump behavior; Perform structured discretization processing on the original behavioral event sequence to extract a multi-dimensional behavioral feature set; Based on the multidimensional behavioral feature set, the original behavioral complexity factor is constructed; The original behavior complexity factor is perturbed, enhanced, and nonlinearly interactively fused to generate a behavior complexity code. Specifically, each entity is obtained Observation behavior feature set: : Interaction trigger sequence collection (such as login, payment, transfer, etc.); Time density feature sequence (frequency of behavior per unit time); : Behavior response type mapping (e.g. multi-round interaction, request chain distribution); : Heteromodal jump feature sequence (number and path of behavioral mode jumps); Construct the original behavioral complexity factor: , in It is a structural compression encoding function that outputs a behavioral complexity factor; The original behavior complexity factor is perturbed, enhanced, and nonlinearly interactively fused to generate a behavior complexity code: , in, Complex coding for behavior, For the Behavioral dimension characteristics, is the behavioral feature dimension, is the local inhibition modulation factor, is the singular response correction constant, is the perturbation complexity amplification parameter, is the nonlinear aggregation power index.

[0031] The present invention is further configured such that the generating response control strategy comprises: The path reverse contribution function is constructed based on the entity structure embedding representation to model the structural deviation degree of the terminal node in the fraud link relative to the path source node. The formula for constructing the path reverse contribution function is: , in, For nodes In the dimensional structure embedding component, is the source node suppression modulation factor, is the embedding dimension, is the coordinated offset damping coefficient, is the nonlinear embedding perturbation index, It is a trade-off indicator between the path structure deviation degree and reverse contribution. The starting node of each path , The end node ; Based on the link credibility score and the path reverse contribution function, the reverse sensitivity of each entity in the fraud path in which it participates is calculated as follows: , in, To include nodes The subset of fraud paths, Score the path credibility, is the path reverse contribution function, Nodes in the path The normalized position influence factor to the terminal node is defined as ,in is the position of the node in the path, Multiple trade-off factors, For nodes Reverse sensitivity to terminal fraud; All reverse sensitivities are summarized and modeled, and the response control strategy set is constructed by combining the structural embedding distance between entities and the interference modulation factor, indicating the control level distribution of different entities in the traceability chain, and defining the control strategy set. , generate control levels for each sensitive node , of the following form: , in, is the basic control factor, To control the activation weight, is the sensitivity index compression factor, For Entity The local interference factor, is the structural embedding distance metric between entities, defined as: , is the distance suppression modulation parameter, For the entity The final response control strategy strength.

[0032] Example 2: See also Figure 2 The exemplary artificial intelligence-based multi-factor fraud link tracking system includes: Multimodal disturbance extraction module: Constructs a multimodal behavioral disturbance structure model, extracts disturbance responses through the temporal evolution of multidimensional behavioral source signals in multiple modal domains, and calculates disturbance absorption intensity; Energy flow coupling relationship construction module: Based on the disturbance absorption intensity and the nonlinear interaction relationship between behaviors, an energy flow coupling relationship diagram between entities is constructed, and entities with potential fraudulent collaborative behaviors are marked as high-intensity connection edges; Structural Propagation and Embedding Update Module: This module performs structural propagation and embedding representation updates based on the energy flow coupling relationship graph, and uses an edge-weight-driven nonlinear propagation mechanism to iteratively generate entity structure embedding representations. Fraud Link Credibility Assessment Module: This module evaluates the credibility of path sequences between any entities, calculates link credibility scores, and constructs a set of fraud links that exhibit perturbation consistency, path coordination, and low behavioral complexity. Reverse traceability control decision module: Combines the link credibility score and entity structure embedding representation to construct a path reverse contribution function, calculates the reverse sensitivity of the entity in the fraud link, and generates a response control strategy to achieve traceability control.

[0033] It should be noted that the artificial intelligence-based multi-factor fraud link tracing system provided in the above embodiment and the artificial intelligence-based multi-factor fraud link tracing method provided in the above embodiment are based on the same concept. The specific manner in which each module and unit performs operations has been described in detail in the method embodiment and will not be repeated here. In actual applications, the artificial intelligence-based multi-factor fraud link tracing system provided in the above embodiment can allocate the above functions to different functional modules as needed, that is, divide the internal structure of the system into different functional modules to complete all or part of the functions described above, and this is not limited here.

[0034] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via a wired (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0035] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the associated objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0036] In this application, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or plural.

[0037] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0038] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0039] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0040] In the several embodiments provided in this application, it should be understood that the disclosed system can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0041] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0042] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0043] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0044] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A multi-factor fraud link tracing method based on artificial intelligence, characterized in that: include: Construct a multimodal behavioral disturbance structure model, extract the disturbance response through the temporal evolution of multidimensional behavioral source signals in multiple modal domains, and calculate the disturbance absorption intensity; Based on the disturbance absorption intensity and the nonlinear interaction between behaviors, an energy flow coupling relationship diagram between entities is constructed, and entities with potential fraudulent collaborative behaviors are marked as high-intensity connection edges. According to the energy flow coupling relationship graph, the structure propagation and embedding representation update are carried out, and the entity structure embedding representation is iteratively generated using the edge weight driven nonlinear propagation mechanism; Evaluate the credibility of path sequences between any entities, calculate link credibility scores, and construct a set of fraudulent links with perturbation consistency, path coordination, and low behavioral complexity differences; Combining the link credibility score and entity structure embedding representation, a path reverse contribution function is constructed to calculate the reverse sensitivity of the entity in the fraud link, and a response control strategy is generated to achieve traceability control.

2. The multi-factor fraud link tracing method based on artificial intelligence according to claim 1 is characterized in that: The steps to construct a multimodal behavioral perturbation structure model include: Model the disturbance response of the time evolution of multi-dimensional behavioral source signals in different modal domains and extract the behavioral dynamic disturbance trajectory with nonlinear change characteristics; Construct disturbance energy consumption based on behavioral dynamic disturbance trajectory and obtain disturbance absorption energy under modal conditions; The absorption energy under all modes is integrated to calculate the inter-modal disturbance absorption intensity, which measures the response ability and absorption characteristics of the behavioral entity to external disturbances under different modes.

3. The multi-factor fraud link tracing method based on artificial intelligence according to claim 2 is characterized in that: The process of constructing the energy flow coupling relationship diagram between entities includes: Generate the behavioral coupling responsiveness between entity pairs according to the temporal evolution results of the multimodal behavioral interaction function; Generate energy flow induction weights between entity pairs according to disturbance absorption intensity and behavioral coupling responsiveness; According to the preset coupling strength threshold, the entity pairs with high-weight connections in the graph structure are determined as the connection nodes of the edges to form an energy flow coupling relationship diagram between the entities.

4. The artificial intelligence-based multi-factor fraud link tracing method according to claim 2 or 3, characterized in that: The generation of entity structure embedding representation includes: Assigning an initial entity structure embedding representation based on the perturbation absorption strength of each entity; According to the constructed energy flow coupling relationship graph, based on the adjacency relationship of each entity, the diffusion contribution of adjacent entities to the current round of entity structure embedding representation is calculated through the edge weight driven nonlinear propagation mechanism; The entity structure embedding representation of the previous round is nonlinearly combined with the diffusion contribution of the current round to form the entity structure embedding representation of the next round; Multiple rounds of propagation and updating are performed until the entity structure embedding representation converges or reaches a preset iteration limit, and finally a stable entity structure embedding representation of each entity is obtained.

5. The multi-factor fraud link tracing method based on artificial intelligence according to claim 4 is characterized in that: Constructing a fraud link set includes: Extract the path sequence between any entity pairs and construct a set of path factors based on the disturbance absorption strength of each node in the path, the entity structure embedding representation and the complex coding characteristics of the behavior; Constructing a disturbance consistency factor based on the disturbance absorption intensity in the path; Constructing path synergy factors based on the embedded representation of entity structures in the path; Constructing behavioral complexity difference factors based on the behavioral complexity coding in the path; Calculate the link credibility score corresponding to the path according to the preset factor fusion rules; Combined with the link credibility score threshold, the path sequences with high credibility features in all paths are screened to construct a fraudulent link set.

6. The multi-factor fraud link tracing method based on artificial intelligence according to claim 5 is characterized in that: The construction logic of the perturbation consistency factor is: Extract the disturbance absorption strength of each entity node based on the path factor set; The consistency of disturbance response within a path is measured based on the nonlinear variation characteristics of the disturbance absorption strength between adjacent nodes. The disturbance absorption offset and power control factor are set, and the disturbance difference pairs within the path are multiplied and accumulated to construct the overall disturbance consistency factor.

7. The multi-factor fraud link tracing method based on artificial intelligence according to claim 5 is characterized in that: The construction logic of the path synergy factor is: According to any pair of adjacent entity nodes in any path, the entity structure embedding representation is obtained, and the power control operation is applied to each pair of adjacent entity nodes, and then element-by-element mapping and fusion are performed. The full path embedding coupling sequence is compressed and represented by a chain nested product method. A nonlinear feature mapping function is constructed to extract the stable kernel response of the multi-level embedding product features and generate the path structure synergy factor.

8. The multi-factor fraud link tracing method based on artificial intelligence according to claim 5 is characterized in that: The computations for behavioral complex coding include: Obtain the original behavioral event sequence of the target entity in multiple behavioral modal domains, including interactive behavior, triggering behavior, response behavior, and modal jump behavior; Perform structured discretization processing on the original behavioral event sequence to extract a multi-dimensional behavioral feature set; Based on the multidimensional behavioral feature set, the original behavioral complexity factor is constructed; The original behavior complexity factor is perturbation-enhanced and nonlinearly interactively fused to generate behavior complexity coding.

9. The multi-factor fraud link tracing method based on artificial intelligence according to claim 5, characterized in that: Generating a responsive control strategy includes: Based on the link credibility score and entity structure embedding representation, a path reverse contribution function is constructed to model the degree of structural deviation of the terminal node in the fraud link relative to the path source node; Calculate the reverse sensitivity for each entity in the fraud path in which it participates; All reverse sensitivities are inductively modeled, and a set of response control strategies is constructed by combining the structural embedding distance between entities and the interference modulation factor, indicating the control level distribution of different entities in the traceability chain.

10. An artificial intelligence-based multi-factor fraud link tracing system, used to implement the artificial intelligence-based multi-factor fraud link tracing method according to any one of claims 1 to 9, characterized in that: include: Multimodal disturbance extraction module: Constructs a multimodal behavioral disturbance structure model, extracts disturbance responses through the temporal evolution of multidimensional behavioral source signals in multiple modal domains, and calculates disturbance absorption intensity; Energy flow coupling relationship construction module: Based on the disturbance absorption intensity and the nonlinear interaction relationship between behaviors, an energy flow coupling relationship diagram between entities is constructed, and entities with potential fraudulent collaborative behaviors are marked as high-intensity connection edges; Structural Propagation and Embedding Update Module: This module performs structural propagation and embedding representation updates based on the energy flow coupling relationship graph, and uses an edge-weight-driven nonlinear propagation mechanism to iteratively generate entity structure embedding representations. Fraud Link Credibility Assessment Module: This module evaluates the credibility of path sequences between any entities, calculates link credibility scores, and constructs a set of fraud links that exhibit perturbation consistency, path coordination, and low behavioral complexity. Reverse traceability control decision module: Combines the link credibility score and entity structure embedding representation to construct a path reverse contribution function, calculates the reverse sensitivity of the entity in the fraud link, and generates a response control strategy to achieve traceability control.

Images