CN-122027264-A - Cross-station identity authentication method and system for trust continuity and behavior evolution

CN122027264ACN 122027264 ACN122027264 ACN 122027264ACN-122027264-A

Abstract

The invention discloses a cross-site identity authentication method and a system for trust continuity and behavior evolution, which belong to the technical field of platform safety identity authentication and access control, and specifically comprise the steps of S1, collecting anonymous cross-site user information and uniformly standardizing, S2, constructing a heterogeneous structure with cooperative modeling of attribute features and behavior features to realize joint characterization of identity features, S3, constructing an identity structure consistency alignment ISCAM module, evaluating consistency deviation degree of the identity features and a historical identity structure and extracting cross-site identity feature representation under structural constraint, S4, constructing a multi-task constraint association hypothesis generation AHGM module, and adaptively generating multipath cross-site identity association hypothesis when consistency conflict exists, S5, constructing a trust continuity backtracking verification TCC-BVM module to realize evaluation and backtracking correction of time continuity and behavior evolution consistency of candidate paths, and effectively inhibiting cross-site identity drift and error association, so as to realize reliable authentication of cross-site identity under an anonymous network environment.

Inventors

JIN LIRONG
LIU YI
CAO YANG
SHI XUAN
Wang Nongyi
TANG JIAXUAN
JIANG QIAN

Assignees

广东工业大学

Dates

Publication Date: 20260512
Application Date: 20260203

Claims (7)

1. A cross-station identity authentication method for trust continuity and behavior evolution is characterized by comprising the following steps: S1, collecting user related data from a plurality of anonymous network stations, carrying out unified standardization processing on user attribute information and user generated content, and eliminating differences of different stations in data format, expression mode and time scale; s2, constructing a static identity attribute feature modeling branch and a behavior feature evolution modeling branch, and carrying out static-dynamic heterogeneous double-branch modeling on the identity features of users at different sites; s3, constructing an identity structure consistency alignment ISCAM module, realizing consistency deviation degree evaluation between the cross-site identity characteristics and user history identity structure reference distribution, and generating cross-site identity characteristic representation; s4, constructing a multi-task constraint association hypothesis generation AHGM module to obtain a plurality of candidate cross-station identity association hypotheses after identity structure conflicts occur; s5, under the structural deviation constraint feedback of the ISCAM module, constructing a trust continuity backtracking verification TCC-BVM module, and realizing backtracking verification and correction of candidate cross-station identity association assumptions; and S6, outputting a cross-station user identity association result which simultaneously meets trust continuity constraint and identity structure stability, and realizing cross-station identity authentication in an anonymous network environment.
2. The cross-site identity authentication method for trust continuity and behavior evolution according to claim 1, wherein the specific steps of S2 are as follows: S21, inputting the user name and the personal profile into a static identity attribute feature modeling branch to obtain an identity attribute representation reflecting the stability of the user identity ; S22, inputting user generated content, behavior time information and context associated information into a behavior feature evolution modeling branch to obtain identity behavior representation reflecting user behavior mode and content style evolution feature 。
3. The cross-site identity authentication method for trust continuity and behavior evolution according to claim 2, wherein the specific steps of S21 are as follows: S211, inputting a text set formed by user names and personal profiles of users into a cross-language pre-training semantic coding model to unify semantic mapping to obtain high-dimensional static attribute semantic vectors ; S212, under the constraint of information bottleneck, for the said Information decoupling modeling of variation inference is carried out to obtain identity core components representing essential personality preferences of users And an ambient noise component characterizing site expression style and context bias ; S213 for the said Discretized mapping, in global codebook Selecting the cluster center vector with the smallest distance to the cluster center vector Obtaining quantitative static identity characterization Simultaneously applying the above-mentioned materials And (3) with Input countermeasure learning network, suppression Interference to identity discrimination features, static identity fingerprint representation with unchanged output environment ; S214, static identity fingerprints of the same user on different sites Modeling as identity latent variable distribution And introducing a commitment loss constraint to minimize cross-station identity latent variable distribution as a main optimization target Realizing cross-station alignment of static identity attribute at distribution level, and finally outputting steady-state identity attribute representation ; The specific steps of the step S22 are as follows: s221, performing content semantic coding, time period coding and context association coding on discrete behaviors of users in a plurality of anonymous websites to obtain users Behavior state vector sequence at discrete time points ; S222, combining the above Inputting continuous time behavior evolution model of structure consistency constraint, and controlling user cross-station behavior in continuous time interval Internal modeling to obtain behavior evolution track representation ; S223, the method Mapping to public behavior feature space to obtain identity behavior feature representations of the same user on different anonymous sites And suppressing correlated noise in the common behavior feature space by cross-site consistency contrast constraint.
4. A cross-site identity authentication method for trust continuity and behavioral evolution according to claim 3, wherein the specific steps of S3 are as follows: s31, constructing an identity structure consistency alignment ISCAM module; S32, distributing historical identity structure references And a dual branch feature representation 、 Inputting the structural deviation constraint information into ISCAM module to obtain the consistent deviation degree of the structural deviation constraint information And structural departure direction constraint And based on the Alignment of And Obtaining a cross-station identity characteristic representation modulated by identity structure consistency constraint ; The step S32 is to use an identity structure consistency alignment ISCAM module to carry out consistency correction on the current cross-station identity characteristics, and the method comprises the following specific steps: s321, representing the double-branch characteristic And (3) with Co-mapping to a reference space defined by a user's historical identity structure, in combination with the user In the first place Fused cross-site identity feature representation of individual historical observation moments Construction of historic identity structure reference distribution Achieving consistency alignment; s322 for the said Generated state jump delta Constructing a jump-induced uncertainty metric matrix Is obtained at Under the constraint of (1) to the current observation point Obtained identity state support point And said at least one of Support point of (2) Degree of deviation of consistency between And then pass through Obtaining the structure deviation direction of the current state pointing to the history reference center by the optimal transmission solution of (2) ; S323 is the following And (3) with And Performing direction consistency alignment and weighted fusion, and inhibiting feature components inconsistent with the historical identity structure to obtain cross-station identity feature representation modulated by identity structure consistency constraint 。
5. The cross-site identity authentication method for trust continuity and behavior evolution according to claim 4, wherein the specific steps of S4 are as follows: s41, designing a correlation hypothesis generation AHGM module of the multi-task constraint; S42, the consistency deviation degree of S32 Input to AHGM module when Obtaining a plurality of candidate cross-station identity association hypotheses when a preset threshold is exceeded; The step S42 is to use a multi-task constraint association hypothesis generation AHGM module to realize cross-station identity association hypothesis under the common constraint of identity structure consistency and behavior evolution, and the method comprises the following specific steps: S421, the method The cross-station identity feature representation exceeding a preset threshold is constructed as a conflict feature set, the conflict feature set is mapped into a dynamic evolution manifold space defined by a behavior evolution model, and evolution constraints of the identity features in a continuous time dimension are uniformly described; S422, associating paths for any candidate cross-station identity Respectively calculating static identity attribute consistency cost Cost of consistency of behavior content and style evolution Interest evolution distribution consistency cost ; S423, each searching individual in the optimized searching algorithm is encoded into a candidate cross-station identity association path And making a random walk step Degree of deviation from consistency Adaptive association by random walk term when user experiences interruption of behavior sequence due to long offline Probability jump crossing the non-observation time interval is realized, and potential identity association paths between discontinuous time nodes are established; S424 for the said Inputting the multi-path cross-station identity association hypothesis set into the multi-task consistency cost to calculate a consistency score, and combining the evolution span screening and sorting of the candidate paths in the time dimension to obtain the multi-path cross-station identity association hypothesis set 。
6. The cross-site identity authentication method for trust continuity and behavior evolution according to claim 5, wherein the specific steps of S5 are as follows: s51, designing a trust continuity backtracking verification TCC-BVM module; S52, under the constraint of path-level identity feature evolution and time continuity, carrying out trust continuity assessment, self-adaptive threshold judgment and backtracking correction on the candidate identity associated paths; The step S52 is to utilize a trust continuity backtracking verification TCC-BVM module to carry out path level verification on candidate cross-station identity association paths, and the method comprises the following specific steps: S521, associating paths with any candidate cross-station identity Reconstruction of path-level identity feature sequences in chronological order At the said Is used for constructing corresponding covariance matrix under the induction of (1) And to the said Constructing a direction projection operator Calculating local trust continuity deviation of adjacent site identity feature distribution in path And combining the time span sensing weight between adjacent stations in the path For the said Weighted aggregation, while introducing path level decision threshold adaptively varying with identity structure consistency bias Obtaining a path-level trust continuity scoring function ; S522, according to the For candidate paths Performing trace-back verification of path level structure consistency to obtain a path set meeting trust continuity constraint 。
7. A cross-site identity authentication system for trust continuity and behavioral evolution, comprising: The identity feature double-branch module is used for carrying out heterogeneous double-branch modeling on user identity information from different anonymous network stations so as to realize collaborative characterization of user identity stability features and behavior evolution features; ISCAM module for evaluating cross-site identity feature and user history identity structure reference distribution The degree of consistent deviation between the two, and the degree of consistent deviation of the structural deviation constraint information is generated And structural departure direction constraint And being structurally constrained against deviation Down pair identity attribute feature representation Identity behavior feature representation Alignment is carried out to obtain a cross-station identity characteristic representation modulated by identity structure consistency constraint ; AHGM module for, upon detection of identity structure identity deviation When the preset threshold value is exceeded, generating a plurality of candidate cross-station identity association hypotheses under the multi-task constraint; The TCC-BVM module is used for carrying out trust continuity assessment on candidate cross-station identity association assumptions by combining path-level identity characteristic evolution and time continuity constraint under the feedback of the consistency deviation degree output by the identity structure consistency alignment module, calculating path-level trust continuity scores and combining follow-up Adaptively varying decision thresholds And backtracking verification and correction are carried out on the candidate identity association hypothesis, so that a cross-station user identity association result which simultaneously meets the trust continuity constraint and the identity structure stability is output.

Description

Cross-station identity authentication method and system for trust continuity and behavior evolution Technical Field The invention relates to the technical field of identity authentication and access control of platform security, in particular to a cross-station identity authentication method and system for trust continuity and behavior evolution. Background With the rapid development of anonymous networks, dark network forums and decentralized communities, users have become normal to act in multiple accounts and multiple identities in different anonymous sites. Such networking environments often lack a unified identity management and access control mechanism, allowing users the freedom to change user names, modify personal profiles, or intermittently participate in discussions, thereby creating highly fragmented, heterogeneous identity characterizations across multiple sites. The characteristics enhance the privacy protection capability of users to a certain extent, but also bring significant challenges to network space security management, illegal behavior traceability and cross-platform access control. Aiming at the problems, the existing cross-station identity authentication method has achieved a certain research result. Related technologies are generally based on static identity attribute features such as user names, personal profiles and the like, and combine text semantic similarity calculation or cluster analysis to achieve preliminary association of cross-site user identities. Part of researches further introduce user generated content, posting time and context information, and the robustness of identity authentication is improved through a behavior feature modeling or multi-mode embedding mode. In addition, due to the introduction of the multi-task learning and contrast learning technology, a plurality of identity-related subtasks can be cooperatively optimized in a shared feature space, and the overall accuracy and generalization capability of cross-station identity recognition are improved to a certain extent. However, most of the above methods focus on local feature similarity or static discrimination results, and still have certain limitations. Firstly, the existing behavior modeling method is mostly based on discrete time sequences, and is difficult to effectively describe continuous evolution characteristics of users under the influence of cross-site migration, intermittent activity or emergency, secondly, explicit modeling of a user history identity structure is lacked, consistency deviation degree between current identity features and history identities is difficult to evaluate from an overall structure level, thirdly, when structural conflict occurs to the identity features of different sites, the existing method generally only outputs a single matching result, and a multi-hypothesis generation and backtracking verification mechanism is lacked, so that requirements of network security scenes on stability and credibility of identity authentication results are difficult to meet. Disclosure of Invention In order to solve the technical problems, the invention provides the following technical scheme: A cross-site identity authentication method for trust continuity and behavior evolution, comprising the steps of: S1, collecting user related data from a plurality of anonymous network stations, carrying out unified standardization processing on user attribute information and user generated content, and eliminating differences of different stations in data format, expression mode and time scale; s2, constructing a static identity attribute feature modeling branch and a behavior feature evolution modeling branch, and carrying out static-dynamic heterogeneous double-branch modeling on the identity features of users at different sites; s3, constructing an identity structure consistency alignment ISCAM module, realizing consistency deviation degree evaluation between the cross-site identity characteristics and user history identity structure reference distribution, and generating cross-site identity characteristic representation; s4, constructing a multi-task constraint association hypothesis generation AHGM module to obtain a plurality of candidate cross-station identity association hypotheses after identity structure conflicts occur; s5, under the structural deviation constraint feedback of the ISCAM module, constructing a trust continuity backtracking verification TCC-BVM module, and realizing backtracking verification and correction of candidate cross-station identity association assumptions; and S6, outputting a cross-station user identity association result which simultaneously meets trust continuity constraint and identity structure stability, and realizing cross-station identity authentication in an anonymous network environment. As a preferable scheme of the cross-station identity authentication method for trust continuity and behavior evolution, the method comprises the following specific ste