A method for constructing a vocational education credible ability authentication system based on multi-chain cooperation

Abstract

The present application relates to a kind of professional education reliable ability authentication system construction method based on multi-chain cooperation, comprising S1, node access AI pre-check and chain weight dynamic constraint configuration;S2, cross-chain protocol post-quantum encryption and synchronization mechanism optimization;S3, cross-chain synchronization period introduces AI prediction adjustment and chain weight association mechanism;S4, establish AI deviation detection and cross-chain data checking rule of quantum security check.This application standardizes multi-chain cooperation to build framework, which can quickly adapt to different levels of vocational colleges and skill evaluation agencies, reduce the cost of platform secondary development and deployment cycle;Multi-chain can be flexibly expanded to access new educational entities and regulatory nodes, continuously maintain the reliable ability of the whole system data in the process of horizontal expansion of the system, and optimize the performance of platform scale landing.

Description

Technical Field

[0001] This invention belongs to the field of trusted capability certification technology, specifically relating to a method for constructing a vocational education trusted capability certification system based on multi-chain collaboration. Background Technology

[0002] Vocational education is the core carrier for cultivating skilled talent. In the current context of digital transformation, the credibility of competency certification and the efficiency of industry-education integration have become pain points in the industry. Driven by policy, schools and enterprises urgently need to break down data barriers, but the traditional system has significant shortcomings: First, data storage relies on centralized platforms, making it susceptible to tampering and loss, and core data such as practical training and enterprise evaluations lack credibility; second, competency certification standards are rigid, with an imbalance between theoretical and practical weighting, making it difficult to dynamically adjust to industry needs; third, cross-school credit transfer lacks a unified basis, skill labels have low matching degrees with enterprise job requirements, and there is a disconnect between industry and education. Furthermore, traditional certification certificates are easily forged, and cross-entity verification is costly, hindering talent mobility. Therefore, it is urgent to build a multi-chain collaborative and trustworthy system to solve the problems of data credibility, dynamic certification, and industry-education collaboration. Summary of the Invention

[0003] The purpose of this invention is to address the shortcomings and deficiencies of existing technologies by providing a method for constructing a vocational education credible competency certification system based on multi-chain collaboration.

[0004] The technical problem solved by this invention is achieved through the following technical solution: A method for constructing a trustworthy vocational education competency certification system based on multi-chain collaboration, the method comprising the following steps: S1. Node admission AI pre-verification and dynamic chain weight constraint configuration; S2, optimization of quantum encryption and synchronization mechanisms after cross-chain protocol; S3, cross-chain synchronization cycle introduces AI prediction adjustment and chain rights association mechanism; S4. Establish cross-chain data verification rules for AI bias detection and quantum security verification.

[0005] Moreover, S1 specifically refers to: S101. Establish a competency certification system The system adopts a multi-chain topology, with the node layer corresponding to a multi-party collaborative governance module, and university nodes. Responsible for academic accreditation, including institutional nodes. Enterprise nodes Education management department node Third-party auditing agency nodes University / College Nodes Responsible for academic education certification, enterprise nodes The education management department is responsible for assessing practical skills. Coordinated standard setting and supervision, third-party auditing institutions as nodes Provides data auditing; the chain structure layer corresponds to the data storage and transfer module, including the credit main chain. University Sub-chain Enterprise subchain Credit main chain It is the core of credit transfer, and the sub-chain of institutions. Enterprise subchain Skills data from both the school and the enterprise are stored separately, and after standardization processing, they are synchronized to the main chain to achieve reliable and traceable data. S102. Establish a node qualification AI pre-review mechanism; S103. The initial trust value calculation model integrates AI scoring and new parameters, and designs differentiated formulas for different node types. S104. Introducing dynamic penalty factor into chain weight calculation. With node activity coefficient The final chain weight formula is obtained by re-optimizing: ; ; in: This refers to the number of times the node has experienced AI pre-screening anomalies in the past 30 days. ≤3, more than 3 times = 0.85, The node's online time over the past 30 days. The duration is normalized to a range of 0 to 1; S105, Deploy a modular multi-chain topology.

[0006] Furthermore, the specific AI pre-screening mechanism for node qualifications in S102 is as follows: 10201) Establish a node qualification AI pre-review mechanism. This mechanism obtains the node subject's digital identity certificate by calling the unified identity authentication interface of the 2025 version of the National Vocational Education Digital Identity System, and synchronizes the institution's unified social credit code digital certificate and vocational education exclusive identity identifier with the node information within the system. 10202) By extracting text and preprocessing it against standard text in the official database, a standard dictionary is constructed and converted into vectors. The cosine similarity algorithm is used to calculate the correlation between the two, and the pre-screening score is calculated. ; in: To extract the cosine similarity between the text and the text in the official database: ; in: , For text vector dimensions, it must satisfy... A value of ≥0.95 allows entry into the core qualification parameter collection stage; otherwise, the extracted text data will be invalid. The scores represent the certificate qualification levels, with Level 1 receiving 1.0, Level 2 receiving 0.8, and Level 3 receiving 0.6. Data integrity is scored equally, with 100% field completeness earning a score of 1.0 and a score decreasing by 0.1 for each missing non-critical field. 10203) The core qualification parameter adds the proportion of dual-qualified teachers in the institution. Its value ranges from 0 to 1, and the calculation method is as follows: ; in: This refers to the number of dual-qualified teachers. This refers to the total number of teachers; Enterprise new apprenticeship system coverage The value range is 0~1, and the calculation method is as follows: ; in: This refers to the number of employees participating in the apprenticeship program. This refers to the total number of practical training positions; Number of digital twin training bases in colleges and universities Enterprise Industrial Internet Platform Access Rate The value ranges from 0 to 1, and the calculation method is as follows: ; in: This refers to the number of training devices connected to IIoT. This refers to the total number of training devices.

[0007] Furthermore, the initial trust value calculation model in S103 integrates AI scoring and parameters as follows: 10301) Determine the node type and perform the corresponding trust value calculation: If it is a school node Then proceed to step 10302); if it is an enterprise node. If it is an education management department node, then execute step 10303); If it is a third-party auditing node, then proceed to step 10304). Then proceed to step 10305). 10302) Targeting university nodes The formula for calculating the trust value is as follows: ; Where: Y represents the number of years of operation (years), A represents the number of provincial-level or above teaching achievements (items), V represents the asset value of practical training equipment (millions of yuan), and C represents the credibility score (0-100 points). This formula guarantees... ∈[35,100], if the number of digital twin training bases When the number is ≥3, the number of additional digital twin training bases will be increased. Trust value bonus; 10303) Targeting enterprise nodes The formula for calculating the trust value is as follows: ; Where: G represents the industry qualification level, from level 1 to 5; P represents the number of practical training cooperation projects in the past 3 years; Z represents the number of core process patents; and the enterprise's industrial internet platform access rate... Additional addition when ≥0.8 Trust value bonus; 10304) Targeting the education management department node The formula for calculating the trust value is as follows: ; Where: F represents the number of policy documents issued, This refers to the amount of policy funding support, of which Map the amount to 0 to 10 cents; 10305) Targeting third-party auditing nodes The formula for calculating the trust value is as follows: ; Where: L represents the audit qualification level score, with scores ranging from A to F, where A is 100, B is 80, ..., F is 20; Acc represents the past audit accuracy rate. The number of security audits per year; among them Map the number of times to a score of 0 to 5.

[0008] Furthermore, the process of re-optimizing the final chain weight formula after introducing a dynamic penalty factor and a node activity coefficient in S104 is as follows: 10401) Determine the node type and re-optimize to obtain the final chain weight formula: If it is a school node Then proceed to step 10402); if it is an enterprise node. If it is an education management department node, then proceed to step 10403). If it is a third-party auditing node, then proceed to step 10404). Then proceed to step 10405). 10402) The final chain weight formula obtained by re-optimizing the university nodes is as follows: ; in: To ensure the teaching resource input rate, a constraint of 55 ≤ ≤85, when The time marker is designated as a core university node; 10403) The final chain weight formula obtained by re-optimizing for enterprise nodes is as follows: ; in: The value represents the coverage rate of training equipment, ranging from 0 to 1, with a constraint of 55 ≤ ≤85, when A node with a value ≥80 is marked as a core enterprise node; 10404) The final chain weight formula obtained by re-optimizing the education management department node is as follows: ; in: The value represents the level of policy support, ranging from 0 to 1, with a constraint of 70 ≤ ≤95; 10405) The final chain weight formula obtained by re-optimizing the third-party auditing agency node is as follows: ; in: The audit coverage ratio is defined as a value between 0 and 1, with a constraint of 60 ≤ ≤90.

[0009] Furthermore, the process of deploying the modular multi-chain topology in S105 is as follows: 10501) Deploys the credit main chain, which supports dynamic sharding, where the number of shards... = , To predict daily cross-chain transaction volume, 500,000 is the daily average capacity limit for a single shard. [1,8]), block generation time optimized to: ; (10502) The university subchain adopts a main chain and side chain architecture. The main chain stores core data, and the side chains store ordinary data. It uses a layered consensus mechanism, with core nodes W≥80 participating in the main chain consensus and ordinary nodes 55≤W<80 participating in the side chain consensus. The block size is... The calculation formula is: ; 10503) The enterprise subchain introduces a parallel transaction pool, the number of transaction pools... , The higher the number of transaction pools, the more transactions are allowed, up to a maximum of 5, which corresponds to the number of Raft consensus nodes. ; in: The average chain weight of the enterprise sub-chain, with a maximum of 7 nodes and a number of channels. ≤80, number of concurrent transactions per channel ≥300; 10504) The cross-chain gateway is deployed in the local data centers of vocational schools and enterprises, and the network latency is high. It adopts a primary / standby dual-machine hot standby mode, with a primary / standby switchover time of [missing information]. Seconds, configuration parameters: data transmission bandwidth B≥500Mbps, peak value Signature verification delay ≤8ms and employs hardware acceleration modules, maximum concurrent connections Conn≥2000, and a distributed memory pool for data buffering, with a memory pool capacity of [missing information]. The more shards, the larger the capacity, up to 70GB, and it supports hot and cold data tiering, including storing hot data in memory and cold data in SSD; the hot / cold data switching time is also supported. .

[0010] Moreover, S2 specifically refers to: 201) The data standardization specification adds 18 key fields to the digital twin training data. These key fields include equipment posture parameters (Pose) and operational force feedback values. Number of collisions in the virtual environment ; 202) Introduce a generative AI field completion mechanism to automatically complete missing non-critical fields; 203) Establish a field mapping matrix The dynamic calibration formula, when the field uses the rate of change of frequency ΔFreq ≥ 0.2, recalculates the mapping coefficients using the analytic hierarchy process: .

[0011] Furthermore, the generative AI field completion mechanism introduced in 202) is as follows: 20201) Constructing a multimodal large model Built with 70B parameters and based on the Transformer-XL architecture, it first completes pre-training for general multimodal feature learning, and then fine-tunes it using digital twin training data; 20202) Import the dataset Perform data completion and calculate the completion accuracy. With completion confidence : ; ; in: The total number of samples to be supplemented. To complete the model with values, The true value is I(), which is an indicator function; it is 1 if they are equal, and 0 otherwise. 20203) Determine the accuracy of completion. With completion confidence The range, if the accuracy is completed ≥98%, complete confidence level If ≥0.95, proceed to step 20204); if 0.9 ≤ When <0.95, execute step 20205). When <0.9, proceed to step 20206). 20204) Directly use data completion; 20205) Manually review and complete the data; 20206) Re-execute step 2020).

[0012] Moreover, S3 specifically refers to: 301) Based on model Predicting data increments in the next 10 minutes The model input features include the incremental sequence of historical data from the past hour. Number of active nodes , distribution ratio of sensitivity levels Prediction error ; 302) After the student skills assessment results are uploaded, calculate the synchronization interval formula to complete the synchronization of sub-chain nodes, ensure data integrity, and the average number of synchronizations per node per day is ≤100. 303) When the amount of data in a single chain increases instantaneously Or fluctuations in data credibility Data synchronization is triggered immediately upon arrival; The specific formula for the synchronization interval in step 302) is as follows: 30201) If it is a sub-chain of an institution, proceed to step 30202); if it is a sub-chain of an enterprise, proceed to step 30203). 30202) Formula for synchronization interval of school subchain: ; in: The standard deviation of chain weights over the past 7 days; 30203) Enterprise Subchain Synchronization Interval Formula: ; 30204) The synchronization interval is corrected by combining the data increment prediction value to obtain the final formula; In step 30204), the final formula obtained by combining the data increment prediction value to correct the synchronization interval is as follows: 3020301) Determine the subchain type and adjust the synchronization interval based on the data increment prediction value to obtain the final formula: If it is a college subchain, execute step 3020302); if it is an enterprise subchain, execute step 3020303). 3020302) School Subchain Correction Synchronization Interval: ; 3020303) Enterprise subchain correction synchronization interval: .

[0013] Moreover, S4 specifically refers to: 401) Format validation uses regular expressions and semantic judgment of the large model to mainly determine whether the semantics of the fields conform to professional specifications; 40101) Access blockchain cross-chain node data and decompose it into structured fields according to the preset data dictionary; 40102) Pre-set vocational education / industry-specific regular expression rules for each field, traverse all fields to perform regular expression matching, and count the matching results; 40103) Based on the "Vocational Education Professional Catalog", "National Vocational Skill Standards" and "Industry Job Specifications", a dedicated corpus is built. The general model is fine-tuned. Fields that have passed the format check are input into the fine-tuned model for semantic analysis to determine whether there are any semantic deviations. This refers to the quantitative results of professional semantic compliance; 4010301) If ≥0.95 and If the value is 0.9, the field verification passes and proceeds to the subsequent quantum security verification stage; 4010302) If 0.95, output "Format error", and mark the specific error fields. The error data is fed back to the cross-chain data source node, and can be re-validated after secondary correction; 402) Range validation sets dynamic thresholds for numeric fields, including practical operation time thresholds. ,in This refers to the number of training steps, with a maximum of 10 minutes and a dimensional accuracy threshold. ,in The skill level is specified, with a maximum thickness of 0.01mm. The pass / fail status is determined based on the threshold of practical operation time. 40201) When the actual operation time is less than the left threshold of the actual operation time If the duration is too short, it will be marked as abnormal and manually reviewed. If the review is successful, it is considered acceptable; otherwise, it is considered unacceptable. 40202) When the practical operation time is at the left threshold ≤Actual operation time≤Practice time right threshold If the time is not met, the test result is deemed satisfactory, and the actual duration is recorded. 40203) When the actual operation time is greater than the right threshold of the actual operation time If the timeout occurs, it will be marked as abnormal (timeout) and manually reviewed. If the review is successful, it is considered qualified; otherwise, it is considered unqualified. 403) Signature verification is a mixed signature The verification formula is: ; The process passes only if both verifications are true (proceed to step 404). If the final result is false, it is marked as "signature verification failed" and execution stops. 404) AI bias detection uses the Isolation Forest algorithm Abnormal data was detected, and the abnormality score was: ; Where: d is the average distance from the data point to the trees in the forest. The average distance is the mean. Standard deviation If the value is ≥0.8, it is considered abnormal, and its execution is intercepted and stopped; if <0.8, considered normal, certification passed.

[0014] The advantages and beneficial effects of this invention are as follows: 1. This invention adopts a multi-chain sharding collaborative evidence storage mechanism, which puts teaching assessment, practical training experience, and skills certification data on separate chains and anchors them across chains. Tampering with a single chain cannot destroy the overall data consensus, thereby improving the trustworthiness of vocational education ability certification data to be tamper-proof and traceable throughout the process.

[0015] 2. This invention uses a multi-chain approach to process business requests from multiple parties, including universities and enterprises, in parallel. Compared to a single-chain architecture, this significantly improves the concurrent throughput of cross-institutional data verification. At the same time, it relies on a cross-chain trusted interaction protocol to ensure the trustworthiness and compliance of cross-departmental data calls, thus solving the performance shortcomings of lack of trust and query lag in multi-entity connections.

[0016] 3. The standardized multi-chain collaborative architecture of this invention can be quickly adapted to vocational schools and skills assessment institutions at different levels, reducing the cost of secondary development and deployment cycle of the platform; the multi-chain can be flexibly expanded to connect new school operators and regulatory nodes, continuously maintaining the credibility of the entire system's data during the horizontal expansion of the system, and optimizing the platform's performance for large-scale deployment. Detailed Implementation

[0017] The present invention will be further described in detail below through specific embodiments. The following embodiments are merely descriptive and not limiting, and should not be used to limit the scope of protection of the present invention.

[0018] An innovative method for constructing a trustworthy vocational education competency certification system based on multi-chain collaboration is described below, with the following steps: S1. Node admission AI pre-verification and dynamic chain weight constraint configuration; S101. Establish a competency certification system The system adopts a multi-chain topology, with the node layer corresponding to a multi-party collaborative governance module, and university nodes. Responsible for academic accreditation, including institutional nodes. Enterprise nodes Education management department node Third-party auditing agency nodes University / College Nodes Responsible for academic education certification, enterprise nodes The education management department is responsible for assessing practical skills. Coordinated standard setting and supervision, third-party auditing institutions as nodes Provides data auditing; the chain structure layer corresponds to the data storage and transfer module, including the credit main chain. University Sub-chain Enterprise subchain Credit main chain It is the core of credit transfer, and the sub-chain of institutions. Enterprise subchain Skills data from both the school and the enterprise are stored separately, and after standardization processing, they are synchronized to the main chain to achieve reliable and traceable data. S102. Establish a node qualification AI pre-review mechanism; S103. The initial trust value calculation model integrates AI scoring and new parameters, and designs differentiated formulas for different node types. S104. Introducing dynamic penalty factor into chain weight calculation. With node activity coefficient The final chain weight formula is obtained by re-optimizing: ; ; in: This refers to the number of times the node has experienced AI pre-screening anomalies in the past 30 days. ≤3, more than 3 times = 0.85, The node's online time over the past 30 days. The duration is normalized to a range of 0 to 1; S105, Deploy a modular multi-chain topology.

[0019] Furthermore, the specific AI pre-screening mechanism for node qualifications in S102 is as follows: 10201) Establish a node qualification AI pre-review mechanism. This mechanism obtains the node subject's digital identity certificate by calling the unified identity authentication interface of the 2025 version of the National Vocational Education Digital Identity System, and synchronizes the institution's unified social credit code digital certificate and vocational education exclusive identity identifier with the node information within the system. 10202) By extracting text and preprocessing it against standard text in the official database, a standard dictionary is constructed and converted into vectors. The cosine similarity algorithm is used to calculate the correlation between the two, and the pre-screening score is calculated. ; in: To extract the cosine similarity between the text and the text in the official database: ; in: , For text vector dimensions, it must satisfy... A value of ≥0.95 allows entry into the core qualification parameter collection stage; otherwise, the extracted text data will be invalid. The scores represent the certificate qualification levels, with Level 1 receiving 1.0, Level 2 receiving 0.8, and Level 3 receiving 0.6. Data integrity is scored equally, with 100% field completeness earning a score of 1.0 and a score decreasing by 0.1 for each missing non-critical field. 10203) The core qualification parameter adds the proportion of dual-qualified teachers in the institution. Its value ranges from 0 to 1, and the calculation method is as follows: ; in: This refers to the number of dual-qualified teachers. This refers to the total number of teachers; Enterprise new apprenticeship system coverage The value range is 0~1, and the calculation method is as follows: ; in: This refers to the number of employees participating in the apprenticeship program. This refers to the total number of practical training positions; Number of digital twin training bases in colleges and universities Enterprise Industrial Internet Platform Access Rate The value ranges from 0 to 1, and the calculation method is as follows: ; in: This refers to the number of training devices connected to IIoT. This refers to the total number of training devices.

[0020] Furthermore, the initial trust value calculation model in S103 integrates AI scoring and parameters as follows: 10301) Determine the node type and perform the corresponding trust value calculation: If it is a school node Then proceed to step 10302); if it is an enterprise node. If it is an education management department node, then execute step 10303); If it is a third-party auditing node, then proceed to step 10304). Then proceed to step 10305). 10302) Targeting university nodes The formula for calculating the trust value is as follows: ; Where: Y represents the number of years of operation (years), A represents the number of provincial-level or above teaching achievements (items), V represents the asset value of practical training equipment (millions of yuan), and C represents the credibility score (0-100 points). This formula guarantees... ∈[35,100], if the number of digital twin training bases When the number is ≥3, the number of additional digital twin training bases will be increased. Trust value bonus; 10303) Targeting enterprise nodes The formula for calculating the trust value is as follows: ; Where: G represents the industry qualification level, from level 1 to 5; P represents the number of practical training cooperation projects in the past 3 years; Z represents the number of core process patents; and the enterprise's industrial internet platform access rate... Additional addition when ≥0.8 Trust value bonus; 10304) Targeting the education management department node The formula for calculating the trust value is as follows: ; Where: F represents the number of policy documents issued, This refers to the amount of policy funding support, of which Map the amount to 0 to 10 cents; 10305) Targeting third-party auditing nodes The formula for calculating the trust value is as follows: ; Where: L represents the audit qualification level score, with scores ranging from A to F, where A is 100, B is 80, ..., F is 20; Acc represents the past audit accuracy rate. The number of security audits per year; among them Map the number of times to a score of 0 to 5.

[0021] Furthermore, the process of re-optimizing the final chain weight formula after introducing a dynamic penalty factor and a node activity coefficient in S104 is as follows: 10401) Determine the node type and re-optimize to obtain the final chain weight formula: If it is a school node Then proceed to step 10402); if it is an enterprise node. If it is an education management department node, then proceed to step 10403). If it is a third-party auditing node, then proceed to step 10404). Then proceed to step 10405). 10402) The final chain weight formula obtained by re-optimizing the university nodes is as follows: ; in: To ensure the teaching resource input rate, a constraint of 55 ≤ ≤85, when The time marker is designated as a core university node; 10403) The final chain weight formula obtained by re-optimizing for enterprise nodes is as follows: ; in: The value represents the coverage rate of training equipment, ranging from 0 to 1, with a constraint of 55 ≤ ≤85, when A node with a value ≥80 is marked as a core enterprise node; 10404) The final chain weight formula obtained by re-optimizing the education management department node is as follows: ; in: The value represents the level of policy support, ranging from 0 to 1, with a constraint of 70 ≤ ≤95; 10405) The final chain weight formula obtained by re-optimizing the third-party auditing agency node is as follows: ; in: The audit coverage ratio is defined as a value between 0 and 1, with a constraint of 60 ≤ ≤90.

[0022] Furthermore, the process of deploying the modular multi-chain topology in S105 is as follows: 10501) Deploys the credit main chain, which supports dynamic sharding, where the number of shards... = , To predict daily cross-chain transaction volume, 500,000 is the daily average capacity limit for a single shard. [1,8]), block generation time optimized to: ; (10502) The university subchain adopts a main chain and side chain architecture. The main chain stores core data, and the side chains store ordinary data. It uses a layered consensus mechanism, with core nodes W≥80 participating in the main chain consensus and ordinary nodes 55≤W<80 participating in the side chain consensus. The block size is... The calculation formula is: ; 10503) The enterprise subchain introduces a parallel transaction pool, the number of transaction pools... , The higher the number of transaction pools, the more transactions are allowed, up to a maximum of 5, which corresponds to the number of Raft consensus nodes. ; in: The average chain weight of the enterprise sub-chain, with a maximum of 7 nodes and a number of channels. ≤80, number of concurrent transactions per channel ≥300; 10504) The cross-chain gateway is deployed in the local data centers of vocational schools and enterprises, and the network latency is high. It adopts a primary / standby dual-machine hot standby mode, with a primary / standby switchover time of [missing information]. Seconds, configuration parameters: data transmission bandwidth B≥500Mbps, peak value Signature verification delay ≤8ms and employs hardware acceleration modules, maximum concurrent connections Conn≥2000, and a distributed memory pool for data buffering, with a memory pool capacity of [missing information]. The more shards, the larger the capacity, up to 70GB, and it supports hot and cold data tiering, including storing hot data in memory and cold data in SSD; the hot / cold data switching time is also supported. .

[0023] S2, optimization of quantum encryption and synchronization mechanisms after cross-chain protocol; 201) The data standardization specification adds 18 key fields to the digital twin training data. These key fields include equipment posture parameters (Pose) and operational force feedback values. Number of collisions in the virtual environment ; 202) Introduce a generative AI field completion mechanism to automatically complete missing non-critical fields; 203) Establish a field mapping matrix The dynamic calibration formula, when the field uses the rate of change of frequency ΔFreq ≥ 0.2, recalculates the mapping coefficients using the analytic hierarchy process: .

[0024] Furthermore, the generative AI field completion mechanism introduced in 202) is as follows: 20201) Constructing a multimodal large model Built with 70B parameters and based on the Transformer-XL architecture, it first completes pre-training for general multimodal feature learning, and then fine-tunes it using digital twin training data; 20202) Import the dataset Perform data completion and calculate the completion accuracy. With completion confidence : ; ; in: The total number of samples to be supplemented. To complete the model with values, The true value is I(), which is an indicator function; it is 1 if they are equal, and 0 otherwise. 20203) Determine the accuracy of completion. With completion confidence The range, if the accuracy is completed ≥98%, complete confidence level If ≥0.95, proceed to step 20204); if 0.9 ≤ When <0.95, execute step 20205). When <0.9, proceed to step 20206). 20204) Directly use data completion; 20205) Manually review and complete the data; 20206) Re-execute step 2020).

[0025] S3, cross-chain synchronization cycle introduces AI prediction adjustment and chain rights association mechanism; 301) Based on model Predicting data increments in the next 10 minutes The model input features include the incremental sequence of historical data from the past hour. Number of active nodes , distribution ratio of sensitivity levels Prediction error ; 302) After the student skills assessment results are uploaded, calculate the synchronization interval formula to complete the synchronization of sub-chain nodes, ensure data integrity, and the average number of synchronizations per node per day is ≤100. 303) When the amount of data in a single chain increases instantaneously Or fluctuations in data credibility Data synchronization is triggered immediately upon arrival; The specific formula for the synchronization interval in step 302) is as follows: 30201) If it is a sub-chain of an institution, proceed to step 30202); if it is a sub-chain of an enterprise, proceed to step 30203). 30202) Formula for synchronization interval of school subchain: ; in: The standard deviation of chain weights over the past 7 days; 30203) Enterprise Subchain Synchronization Interval Formula: ; 30204) The synchronization interval is corrected by combining the data increment prediction value to obtain the final formula; In step 30204), the final formula obtained by combining the data increment prediction value to correct the synchronization interval is as follows: 3020301) Determine the subchain type and adjust the synchronization interval based on the data increment prediction value to obtain the final formula: If it is a college subchain, execute step 3020302); if it is an enterprise subchain, execute step 3020303). 3020302) School Subchain Correction Synchronization Interval: ; 3020303) Enterprise subchain correction synchronization interval: .

[0026] S4. Establish cross-chain data verification rules for AI bias detection and quantum security verification.

[0027] 401) Format validation uses regular expressions and semantic judgment of the large model to mainly determine whether the semantics of the fields conform to professional specifications; 40101) Access blockchain cross-chain node data and decompose it into structured fields according to the preset data dictionary; 40102) Pre-set vocational education / industry-specific regular expression rules for each field, traverse all fields to perform regular expression matching, and count the matching results; 40103) Based on the "Vocational Education Professional Catalog", "National Vocational Skill Standards" and "Industry Job Specifications", a dedicated corpus is built. The general model is fine-tuned. Fields that have passed the format check are input into the fine-tuned model for semantic analysis to determine whether there are any semantic deviations. This refers to the quantitative results of professional semantic compliance; 4010301) If ≥0.95 and If the value is 0.9, the field verification passes and proceeds to the subsequent quantum security verification stage; 4010302) If 0.95, output "Format error", and mark the specific error fields. The error data is fed back to the cross-chain data source node, and can be re-validated after secondary correction; 402) Range validation sets dynamic thresholds for numeric fields, including practical operation time thresholds. ,in This refers to the number of training steps, with a maximum of 10 minutes and a dimensional accuracy threshold. ,in The skill level is specified, with a maximum thickness of 0.01mm. The pass / fail status is determined based on the threshold of practical operation time. 40201) When the actual operation time is less than the left threshold of the actual operation time If the duration is too short, it will be marked as abnormal and manually reviewed. If the review is successful, it is considered acceptable; otherwise, it is considered unacceptable. 40202) When the practical operation time is at the left threshold ≤Actual operation time≤Practice time right threshold If the time is not met, the test result is deemed satisfactory, and the actual duration is recorded. 40203) When the actual operation time is greater than the right threshold of the actual operation time If the timeout occurs, it will be marked as abnormal (timeout) and manually reviewed. If the review is successful, it is considered qualified; otherwise, it is considered unqualified. 403) Signature verification is a mixed signature The verification formula is: ; The process passes only if both verifications are true (proceed to step 404). If the final result is false, it is marked as "signature verification failed" and execution stops. 404) AI bias detection uses the Isolation Forest algorithm Abnormal data was detected, and the abnormality score was: ; Where: d is the average distance from the data point to the trees in the forest. The average distance is the mean. Standard deviation If the value is ≥0.8, it is considered abnormal, and its execution is intercepted and stopped; if <0.8, considered normal, certification passed.

[0028] Based on the present invention, "A Method for Constructing a Vocational Education Trustworthy Competency Certification System Based on Multi-Chain Collaboration", the following comparative analysis experimental scheme and simulation experimental results were designed to verify the system's ability to address pain points in traditional vocational education certification, such as data tampering, rigid certification standards, and a disconnect between industry and education.

[0029] I. Design of the Comparison Method To comprehensively evaluate the performance of this patented system, three different architectures were established as comparative baselines in the experiment: 1. Baseline Method A (Traditional Centralized System): Corresponds to the traditional system mentioned in the patent background technology. Data storage relies on a single centralized platform database, without blockchain structure and without AI-assisted verification.

[0030] 2. Baseline Method B (Traditional Single-Chain System): This method uses a single blockchain architecture to process data from all nodes, employs static permission consensus for fixed nodes, adopts a fixed-period data synchronization strategy, and lacks cross-chain mechanisms and AI data completion capabilities.

[0031] 3. This patented method (multi-chain collaboration and AI enhancement system): Utilizing the architecture proposed in this invention, it comprises a multi-chain topology including a credit main chain, university sub-chains, and enterprise sub-chains; it introduces a dynamic penalty factor. With node activity coefficient The dynamic chain weight mechanism; adopting a multimodal large model Generative AI field completion; application based on Model predicts data increment The dynamic cross-chain synchronization mechanism; and the security rules for hybrid signatures and AI bias detection.

[0032] II. Comparison of Experimental Indicators Based on the technological innovations of the patents, the following four core indicators were extracted for quantitative comparison: 1. Data integrity and repair rate (%): This measures the system's ability to ensure the integrity of the final on-chain data when faced with practical data containing missing fields, such as digital twin training data, and verifies the effectiveness of the generative AI field completion mechanism.

[0033] 2. Cross-chain synchronization latency and throughput (TPS): This measures the system's data processing efficiency in high-concurrency scenarios such as vocational school final exams or enterprise intensive training, and verifies dynamic sharding and... The role of dynamically adjusting the synchronization period in the model.

[0034] 3. Malicious Node Penalty Response Time (hours): Measures the speed at which the system identifies and revokes consensus permissions of abnormal / malicious nodes, verifying the AI ​​pre-verification and dynamic chain power penalty mechanism. Sensitivity of (anomaly detection).

[0035] 4. Abnormal Data Interception Rate (%): Measures the system's ability to identify abnormal data formats, tampering attacks, and non-compliant data (such as abnormal operation time), and verifies the comprehensive protection capabilities of regular expression / semantic verification, quantum security verification, and the isolated forest algorithm.

[0036] III. Comparison of experimental results: The experimental results are shown in Table 1.

[0037] Table 1

[0038] IV. Analysis of Experimental Results 1. Significantly Improved High-Concurrency Adaptability: In simulated high-concurrency training data upload scenarios, traditional single-chain system method B, due to its fixed block generation and synchronization time, suffers from transaction pool backlog and soaring latency. This patented method, through... The model predicts data increments in real time. The synchronization interval formula for dynamically adjusting the sub-chains of enterprises and universities is as follows: and This ensures low latency and high TPS for the system.

[0039] 2. High degree of consistency between data quality and industry-education standards: Faced with noisy or missing field underlying data, traditional methods A and B can only choose to reject or report errors. This patent addresses this through 70B parameters. The pre-trained model not only automatically completes the missing data, but also ensures compliance with professional semantics through a fine-tuned larger model. Verification greatly reduces the cost of manual intervention and solves the problem of "disconnect between industry and education".

[0040] 3. System resilience against quantum computing and malicious behavior: The experiment simulated malicious node injection behavior. This patented system uses text cosine similarity in the node admission process. Calculate the pre-screening score and directly block fake nodes; during operation, trigger immediately upon detecting abnormal behavior. The penalty factor reduces its consensus weight. Meanwhile, hybrid signatures... The introduction of this technology effectively mitigates the risk of future quantum hacking and verifies the system's extremely high security ceiling.

[0041] Although embodiments of the invention have been disclosed for illustrative purposes, those skilled in the art will understand that various substitutions, variations, and modifications are possible without departing from the spirit and scope of the invention and the appended claims. Therefore, the scope of the invention is not limited to the contents disclosed in the embodiments.

Claims

1. A method for constructing a vocational education credible competency certification system based on multi-chain collaboration, characterized in that: The steps of the method are as follows: S1. Node admission AI pre-verification and dynamic chain weight constraint configuration; S2, optimization of quantum encryption and synchronization mechanisms after cross-chain protocol; S3, cross-chain synchronization cycle introduces AI prediction adjustment and chain rights association mechanism; S4. Establish cross-chain data verification rules for AI bias detection and quantum security verification.

2. The method for constructing a vocational education trustworthy competency certification system based on multi-chain collaboration as described in claim 1, characterized in that: Specifically, S1 is: S101. Establish a competency certification system The system adopts a multi-chain topology, with the node layer corresponding to a multi-party collaborative governance module, and university nodes. Responsible for academic accreditation, including institutional nodes. Enterprise nodes Education management department node Third-party auditing agency nodes University / College Nodes Responsible for academic education certification, enterprise nodes The education management department is responsible for assessing practical skills. Coordinated standard setting and supervision, third-party auditing institutions as nodes Provides data auditing; the chain structure layer corresponds to the data storage and transfer module, including the credit main chain. University Sub-chain Enterprise subchain Credit main chain It is the core of credit transfer, and the sub-chain of institutions. Enterprise subchain Skills data from both the school and the enterprise are stored separately, and after standardization processing, they are synchronized to the main chain to achieve reliable and traceable data. S102. Establish a node qualification AI pre-review mechanism; S103. The initial trust value calculation model integrates AI scoring and new parameters, and designs differentiated formulas for different node types. S104. Introducing dynamic penalty factor into chain weight calculation. With node activity coefficient The final chain weight formula is obtained by re-optimizing: ； ； in: This refers to the number of times the node has experienced AI pre-screening anomalies in the past 30 days. ≤3, more than 3 times = 0.85, The node's online time over the past 30 days. The duration is normalized to a range of 0 to 1; S105, Deploy a modular multi-chain topology.

3. The method for constructing a vocational education trustworthy competency certification system based on multi-chain collaboration as described in claim 2, characterized in that: The specific AI pre-screening mechanism for node qualifications in S102 is as follows: 10201) Establish a node qualification AI pre-review mechanism. This mechanism obtains the node subject's digital identity certificate by calling the unified identity authentication interface of the 2025 version of the National Vocational Education Digital Identity System, and synchronizes the institution's unified social credit code digital certificate and vocational education exclusive identity identifier with the node information within the system. 10202) By extracting text and preprocessing it against standard text in the official database, a standard dictionary is constructed and converted into vectors. The cosine similarity algorithm is used to calculate the correlation between the two, and the pre-screening score is calculated. ； in: To extract the cosine similarity between the text and the text in the official database: ； in: , For text vector dimensions, it must satisfy... A value of ≥0.95 allows entry into the core qualification parameter collection stage; otherwise, the extracted text data will be invalid. The scores represent the certificate qualification levels, with Level 1 receiving 1.0, Level 2 receiving 0.8, and Level 3 receiving 0.

6. Data integrity is scored equally, with 100% field completeness earning a score of 1.0 and a score decreasing by 0.1 for each missing non-critical field. 10203) The core qualification parameter adds the proportion of dual-qualified teachers in the institution. Its value ranges from 0 to 1, and the calculation method is as follows: ； in: This refers to the number of dual-qualified teachers. This refers to the total number of teachers; Enterprise new apprenticeship system coverage The value range is 0~1, and the calculation method is as follows: ； in: This refers to the number of employees participating in the apprenticeship program. This refers to the total number of practical training positions; Number of digital twin training bases in colleges and universities Enterprise Industrial Internet Platform Access Rate The value ranges from 0 to 1, and the calculation method is as follows: ； in: This refers to the number of training devices connected to IIoT. This refers to the total number of training devices.

4. The method for constructing a vocational education trustworthy competency certification system based on multi-chain collaboration as described in claim 2, characterized in that: The initial trust value calculation model in S103 integrates AI scoring and parameters as follows: 10301) Determine the node type and perform the corresponding trust value calculation: If it is a school node Then proceed to step 10302); if it is an enterprise node. If it is an education management department node, then execute step 10303); If it is a third-party auditing node, then proceed to step 10304). Then proceed to step 10305). 10302) Targeting university nodes The formula for calculating the trust value is as follows: ； Where: Y represents the number of years of operation (years), A represents the number of provincial-level or above teaching achievements (items), V represents the asset value of practical training equipment (millions of yuan), and C represents the credibility score (0-100 points). This formula guarantees... ∈[35,100], if the number of digital twin training bases When the number is ≥3, the number of additional digital twin training bases will be increased. Trust value bonus; 10303) Targeting enterprise nodes The formula for calculating the trust value is as follows: ； Where: G represents the industry qualification level, from level 1 to 5; P represents the number of practical training cooperation projects in the past 3 years; Z represents the number of core process patents; and the enterprise's industrial internet platform access rate... Additional addition when ≥0.8 Trust value bonus; 10304) Targeting the education management department node The formula for calculating the trust value is as follows: ； Where: F represents the number of policy documents issued, This refers to the amount of policy funding support, of which Map the amount to 0 to 10 cents; 10305) Targeting third-party auditing nodes The formula for calculating the trust value is as follows: ； Where: L represents the audit qualification level score, with scores ranging from A to F, where A is 100, B is 80, ..., F is 20; Acc represents the past audit accuracy rate. The number of security audits per year; among them Map the number of times to a score of 0 to 5.

5. The method for constructing a vocational education trustworthy competency certification system based on multi-chain collaboration as described in claim 2, characterized in that: The specific process of re-optimizing the final chain weight formula after introducing a dynamic penalty factor and a node activity coefficient in S104 is as follows: 10401) Determine the node type and re-optimize to obtain the final chain weight formula: If it is a school node Then proceed to step 10402); if it is an enterprise node. If it is an education management department node, then proceed to step 10403). Then proceed to step 10404). If it is a third-party auditing agency node Then proceed to step 10405). 10402) The final chain weight formula obtained by re-optimizing the university nodes is as follows: ； in: To ensure the teaching resource input rate, a constraint of 55 ≤ ≤85, when The time marker is designated as a core university node; 10403) The final chain weight formula obtained by re-optimizing for enterprise nodes is as follows: ； in: The value represents the coverage rate of training equipment, ranging from 0 to 1, with a constraint of 55 ≤ ≤85, when A node with a value ≥80 is marked as a core enterprise node; 10404) The final chain weight formula obtained by re-optimizing the education management department node is as follows: ； in: The value represents the level of policy support, ranging from 0 to 1, with a constraint of 70 ≤ ≤95; 10405) The final chain weight formula obtained by re-optimizing the third-party auditing agency node is as follows: ； in: The audit coverage ratio is defined as a value between 0 and 1, with a constraint of 60 ≤ ≤90.

6. The method for constructing a vocational education trustworthy competency certification system based on multi-chain collaboration as described in claim 2, characterized in that: The specific process of deploying the modular multi-chain topology in S105 is as follows: 10501) Deploys the credit main chain, which supports dynamic sharding, where the number of shards... = , To predict daily cross-chain transaction volume, 500,000 is the daily average capacity limit for a single shard. [1,8]), block generation time optimized to: ； (10502) The university subchain adopts a main chain and side chain architecture. The main chain stores core data, and the side chains store ordinary data. It uses a layered consensus mechanism, with core nodes W≥80 participating in the main chain consensus and ordinary nodes 55≤W<80 participating in the side chain consensus. The block size is... The calculation formula is: ； 10503) The enterprise subchain introduces a parallel transaction pool, the number of transaction pools... , The higher the number of transaction pools, the more transactions are allowed, up to a maximum of 5, which corresponds to the number of Raft consensus nodes. ； in: The average chain weight of the enterprise sub-chain, with a maximum of 7 nodes and a number of channels. ≤80, number of concurrent transactions per channel ≥300; 10504) The cross-chain gateway is deployed in the local data centers of vocational schools and enterprises, and the network latency is high. It adopts a primary / standby dual-machine hot standby mode, with a primary / standby switchover time of [missing information]. Seconds, configuration parameters: data transmission bandwidth B≥500Mbps, peak value Signature verification delay ≤8ms and employs hardware acceleration modules, maximum concurrent connections Conn≥2000, and a distributed memory pool for data buffering, with a memory pool capacity of [missing information]. The more shards, the larger the capacity, up to 70GB, and it supports hot and cold data tiering, including storing hot data in memory and cold data in SSD; the hot / cold data switching time is also supported. .

7. The method for constructing a vocational education trustworthy competency certification system based on multi-chain collaboration as described in claim 1, characterized in that: Specifically, S2 is: 201) The data standardization specification adds 18 key fields to the digital twin training data. These key fields include equipment posture parameters (Pose) and operational force feedback values. Number of collisions in the virtual environment ; 202) Introduce a generative AI field completion mechanism to automatically complete missing non-critical fields; 203) Establish a field mapping matrix The dynamic calibration formula, when the field uses the rate of change of frequency ΔFreq ≥ 0.2, recalculates the mapping coefficients using the analytic hierarchy process: 。 8. The method for constructing a vocational education trustworthy competency certification system based on multi-chain collaboration as described in claim 7, characterized in that: The generative AI field completion mechanism introduced in section 202) is as follows: 20201) Constructing a multimodal large model Built with 70B parameters and based on the Transformer-XL architecture, it first completes pre-training for general multimodal feature learning, and then fine-tunes it using digital twin training data; 20202) Import the dataset Perform data completion and calculate the completion accuracy. With completion confidence : ； ； in: The total number of samples to be supplemented. To complete the model with values, The true value is I(), which is an indicator function; it is 1 if they are equal, and 0 otherwise. 20203) Determine the accuracy of completion. With completion confidence The range, if the accuracy is completed ≥98%, complete confidence level If ≥0.95, proceed to step 20204); if 0.9 ≤ When <0.95, execute step 20205). When <0.9, proceed to step 20206). 20204) Directly use data completion; 20205) Manually review and complete the data; 20206) Re-execute step 2020).

9. The method for constructing a vocational education trustworthy competency certification system based on multi-chain collaboration as described in claim 1, characterized in that: Specifically, S3 is: 301) Based on model Predicting data increments in the next 10 minutes The model input features include the incremental sequence of historical data from the past hour. Number of active nodes , distribution ratio of sensitivity levels Prediction error ; 302) After the student skills assessment results are uploaded, calculate the synchronization interval formula to complete the synchronization of sub-chain nodes, ensure data integrity, and the average number of synchronizations per node per day is ≤100. 303) When the amount of data in a single chain increases instantaneously Or fluctuations in data credibility Data synchronization is triggered immediately upon arrival; The specific formula for the synchronization interval in step 302) is as follows: 30201) If it is a sub-chain of an institution, proceed to step 30202); if it is a sub-chain of an enterprise, proceed to step 30203). 30202) Formula for synchronization interval of school subchain: ； in: The standard deviation of chain weights over the past 7 days; 30203) Enterprise Subchain Synchronization Interval Formula: ； 30204) The synchronization interval is corrected by combining the data increment prediction value to obtain the final formula; In step 30204), the final formula obtained by combining the data increment prediction value to correct the synchronization interval is as follows: 3020301) Determine the subchain type and adjust the synchronization interval based on the data increment prediction value to obtain the final formula: If it is a college subchain, execute step 3020302); if it is an enterprise subchain, execute step 3020303). 3020302) School Subchain Correction Synchronization Interval: ； 3020303) Enterprise subchain correction synchronization interval: 。 10. The method for constructing a vocational education trustworthy competency certification system based on multi-chain collaboration according to claim 1, characterized in that: Specifically, S4 is: 401) Format validation uses regular expressions and semantic judgment of the large model to mainly determine whether the semantics of the fields conform to professional specifications; 40101) Access blockchain cross-chain node data and decompose it into structured fields according to the preset data dictionary; 40102) Pre-set vocational education / industry-specific regular expression rules for each field, traverse all fields to perform regular expression matching, and count the matching results; 40103) Based on the "Vocational Education Professional Catalog", "National Vocational Skill Standards" and "Industry Job Specifications", a dedicated corpus is built. The general model is fine-tuned. Fields that have passed the format check are input into the fine-tuned model for semantic analysis to determine whether there are any semantic deviations. This refers to the quantitative results of professional semantic compliance; 4010301) If ≥0.95 and If the value is 0.9, the field verification passes and proceeds to the subsequent quantum security verification stage; 4010302) If 0.95, output "Format error", and mark the specific error fields. The error data is fed back to the cross-chain data source node, and can be re-validated after secondary correction; 402) Range validation sets dynamic thresholds for numeric fields, including practical operation time thresholds. ,in This refers to the number of training steps, with a maximum of 10 minutes and a dimensional accuracy threshold. ,in The skill level is specified, with a maximum thickness of 0.01mm. The pass / fail status is determined based on the threshold of practical operation time. 40201) When the actual operation time is less than the left threshold of the actual operation time If the duration is too short, it will be marked as abnormal and manually reviewed. If the review is successful, it is considered acceptable; otherwise, it is considered unacceptable. 40202) When the practical operation time is at the left threshold ≤Actual operation time≤Practice time right threshold If the time is not met, the test result is deemed satisfactory, and the actual duration is recorded. 40203) When the actual operation time is greater than the right threshold of the actual operation time If the timeout occurs, it will be marked as abnormal (timeout) and manually reviewed. If the review is successful, it is considered qualified; otherwise, it is considered unqualified. 403) Signature verification is a mixed signature The verification formula is: ； The process passes only if both verifications are true (proceed to step 404). If the final result is false, it is marked as "signature verification failed" and execution stops. 404) AI bias detection uses the Isolation Forest algorithm Abnormal data was detected, and the abnormality score was: ； Where: d is the average distance from the data point to the trees in the forest. The average distance is the mean. Standard deviation If the value is ≥0.8, it is considered abnormal, and its execution is intercepted and stopped; if <0.8, considered normal, certification passed.

Images