A blockchain consensus method and system for professional title review

Abstract

The application provides a blockchain consensus method and system for professional title review, which comprises the following steps: dividing the behaviors of all nodes into positive behaviors and negative behaviors, and calculating the credit scores of all review agency nodes according to the first and second integral rules; screening all the nodes to be elected that meet the condition that the credit scores are greater than a first threshold value, and selecting the node to be elected corresponding to the maximum election probability from all the election probabilities as the master node of the current round of consensus; controlling the master node to generate a pre-preparation message containing a unique task serial number, and sending the pre-preparation message to the consensus nodes for verification; and the master node broadcasts a submission message to the whole network, and each consensus node verifies and executes an expert extraction algorithm to obtain an extraction result. The application can fundamentally solve the fairness, security and efficiency problems in the professional title review expert extraction process.

Description

Technical Field

[0001] This invention relates to the field of expert review technology, and in particular to a blockchain consensus method and system for professional title review. Background Technology

[0002] Professional title evaluation is a core component of the talent evaluation system, and expert selection, as the starting point of the evaluation process, directly determines the credibility of the evaluation results in terms of fairness, security, and efficiency. Traditional expert selection systems for professional title evaluation often employ centralized architectures or simple distributed protocols, which have several key shortcomings in practical applications. First, the selection process lacks transparency. Traditional systems rely on manual operation or centralized servers to manage the expert database and selection rules, resulting in a lack of transparent and traceable oversight mechanisms for aspects such as expert screening, task allocation, and result generation, which easily raises questions about the fairness of the evaluation.

[0003] Secondly, data security and integrity face significant risks. Experts' sensitive personal information, including their professional titles, research fields, and review history, as well as extraction rule parameters and final review results, is typically stored in centralized databases, making it susceptible to tampering or leakage. Traditional systems often employ simple encryption methods, such as single symmetric encryption, which are insufficient to defend against attacks from malicious nodes, such as data forgery or denial-of-service attacks. Once the database is compromised, it will directly lead to the interruption of the review process or the distortion of results.

[0004] Poor node fault tolerance and scalability are also major shortcomings of traditional systems. Traditional distributed review systems often use consensus algorithms such as RAFT or classic PBFT. RAFT cannot tolerate malicious node behavior and only supports fault recovery; while classic PBFT has Byzantine fault tolerance, it suffers from fixed nodes, high communication complexity, and random master node election. When the scale of review increases, such as in cross-provincial or cross-industry joint reviews, the increase in the number of participating nodes leads to a significant decrease in the throughput of classic PBFT and a substantial increase in transaction latency, making it unable to meet the requirements of high-concurrency extraction. At the same time, the continued presence of malicious nodes consumes network resources, further reducing system stability.

[0005] Furthermore, traditional systems lack a mechanism to constrain node behavior, failing to quantitatively assess and constrain the actions of participating nodes such as review bodies and supervisory units. Malicious nodes may submit invalid expert information or refuse to participate in consensus. While this may not directly undermine the consensus outcome, it can lead to repeated retries in the consensus process, increasing extraction latency. If a low-performance node is randomly selected as a core node, such as a master node, it may also cause process bottlenecks, affecting review efficiency. Summary of the Invention

[0006] The purpose of this invention is to provide a blockchain consensus method and system for professional title evaluation, aiming to solve at least one of the problems in the background technology.

[0007] In a first aspect, the present invention provides a blockchain consensus method for professional title evaluation, the method comprising: Construct a blockchain network for professional title evaluation that includes management nodes and review agency nodes; The behavior of all nodes is divided into positive behavior and negative behavior, and a first integration rule and a second integration rule are constructed for the positive behavior and the negative behavior respectively. The reputation score of all review agency nodes is calculated according to the first integration rule and the second integration rule. All candidate nodes that meet the requirement of having a reputation score greater than a first threshold are selected, and the election probability of all candidate nodes is calculated based on the reputation score. The candidate node corresponding to the highest election probability is selected from all election probabilities as the master node for this round of consensus. When the management node initiates an expert extraction request to the master node, it controls the master node to generate a pre-preparation message containing a unique task sequence number and sends the pre-preparation message to the consensus node for verification. If the verification is successful, a preparation message is broadcast to the entire network. When the number of preparation messages collected by the master node is greater than the second threshold, the master node broadcasts a submission message to the entire network. After verification by each consensus node, an expert extraction algorithm is executed to obtain the extraction result. The consensus nodes are all review agency nodes except the master node.

[0008] Secondly, the present invention provides a blockchain consensus system for professional title evaluation, the system comprising: The blockchain network construction module is used to build a blockchain network for professional title evaluation that includes management nodes and review agency nodes. The reputation score calculation module is used to divide the behavior of all nodes into positive behavior and negative behavior, and to construct a first score rule and a second score rule for the positive behavior and the negative behavior respectively, and to calculate the reputation score of all review agency nodes according to the first score rule and the second score rule. The master node selection module is used to filter out all candidate nodes that meet the requirement that the reputation score is greater than the first threshold, calculate the election probability of all candidate nodes based on the reputation score, and select the candidate node corresponding to the maximum election probability from all election probabilities as the master node of this round of consensus. The verification module is used to control the master node to generate a pre-preparation message containing a unique task sequence number when the management node initiates an expert extraction request to the master node, and to send the pre-preparation message to the consensus node for verification. The extraction execution module is used to broadcast a preparation message to the entire network if the verification is successful. When the number of preparation messages collected by the master node is greater than the second threshold, the master node broadcasts a submission message to the entire network. After verification by each consensus node, the expert extraction algorithm is executed to obtain the extraction result. The consensus nodes are all review agency nodes except the master node.

[0009] Thirdly, the present invention provides a storage medium that stores one or more programs that, when executed by a processor, implement the aforementioned blockchain consensus method for professional title evaluation.

[0010] Fourthly, the present invention provides an electronic device, the electronic device comprising a memory and a processor, wherein: The memory is used to store computer programs; When the processor executes the computer program stored in the memory, it implements the aforementioned blockchain consensus method for professional title evaluation.

[0011] Compared with the prior art, the present invention has the following advantages: This invention improves upon traditional PBFT consensus by introducing a reputation mechanism and dynamic master node election, and fundamentally solves the challenges of fairness, security, and efficiency in the expert selection process for professional title evaluation by combining implicit association modeling and fairness constraints. Specifically, firstly, by constructing refined node behavior scoring rules and calculating reputation scores in real time, quantitative evaluation and dynamic constraints on participating review institution nodes are achieved, effectively curbing malicious behavior. Based on this, high-reputation, high-performance nodes are dynamically elected as master nodes based on reputation scores and network performance, significantly improving the reliability and efficiency of the consensus process and eliminating process delays or result biases caused by poor master node performance or malicious intent. Secondly, in the expert selection algorithm, an implicit association evaluation model based on semantics and graph neural networks is creatively introduced, which can intelligently identify and avoid potential association risks between experts and applicants. Simultaneously, by establishing a fairness constraint extraction model aimed at minimizing distribution bias, the problem of excessive expert concentration is systematically avoided, thus ensuring broad representativeness and fairness in expert selection at the algorithm level. Finally, the entire consensus process and extraction results are stored on the blockchain and distributed storage technology to ensure transparency, immutability and traceability throughout the entire chain from demand initiation and consensus verification to result generation, providing a solid technical foundation for building a modern professional title evaluation system with strong credibility, scalability and high availability. Attached Figure Description

[0012] Figure 1 This is a flowchart of a blockchain consensus method for professional title evaluation proposed in an embodiment of the present invention; Figure 2This is a schematic diagram of the structure of a blockchain consensus system for professional title evaluation proposed in an embodiment of the present invention.

[0013] The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Unless otherwise defined, the technical or scientific terms used herein should have the ordinary meaning understood by those skilled in the art. The terms "comprising" and similar expressions used herein mean that the element or object preceding the word covers the element or object listed after the word and its equivalents, but does not exclude other elements or objects.

[0015] like Figure 1 As shown, an embodiment of the present invention proposes a blockchain consensus method for professional title evaluation, the method comprising steps S101 to S105, wherein: Step S101: Construct a blockchain network for professional title evaluation that includes management nodes and review agency nodes; It should be noted that the management node is the competent department for professional title evaluation, which is responsible for configuring evaluation rules, managing node permissions, and handling anomalies; the evaluation institution node is the unit with evaluation qualifications, which is responsible for providing expert database resources and participating in the consensus process; and the supervision node is the third-party auditing institution, which is responsible for monitoring the extraction process and verifying the integrity of the results. It has no consensus voting rights but has the right to audit the results.

[0016] In addition, in some embodiments, when a new node submits identity authentication materials to the management node, the identity authentication materials include an institution qualification certificate, a node public key, and contact information. After verifying the authenticity of the materials, the management node initiates a network-wide node consensus verification to broadcast the new node's authentication request to all admitted nodes. When the management node collects more than a first preset proportion of consent signatures from approved nodes, it determines that the new node has passed authentication. At this time, a third-party CA (Certificate Authority) generates a digital certificate for the newly authenticated node, the management node assigns it an initial reputation score (100 points by default), adds the new node to the consensus group, and synchronizes the entire network's node list.

[0017] By requiring new nodes to submit qualification materials, and through a consensus mechanism initiated by the management node and passed by more than a first preset proportion of existing nodes, a strict but decentralized admission committee mechanism has been established. This not only ensures the legitimacy and credibility of newly joined nodes and prevents the influx of malicious nodes, but also maintains the autonomy and consensus foundation of the network community by granting voting rights to existing nodes, enabling the system to flexibly adapt to the collaborative needs of cross-regional and large-scale reviews.

[0018] Step S102: Divide the behavior of all nodes into positive behavior and negative behavior, and construct a first integration rule and a second integration rule for the positive behavior and the negative behavior respectively, and calculate the reputation score of all review agency nodes according to the first integration rule and the second integration rule. In this step, the first scoring rule is: if the review agency node successfully participates in an expert consensus drawing once, the first bonus item is executed; If the review agency's qualification verification is valid, the second bonus item will be applied; The time taken from receiving a message to responding is recorded as the message response delay. If the message response delay is less than or equal to the first time threshold, the third bonus item is executed once. The cumulative score is calculated using the following formula: ; in, To accumulate points, , , These are the first bonus item, the second bonus item, and the third bonus item, respectively. n is the cumulative number of consensuses reached by the review agency node in the expert consensus drawing process, and v is the number of valid qualification verifications of the review agency node. The second scoring rule is as follows: if the cumulative number of consensus-building attempts refused to participate in expert consensus-building within a unit of time exceeds the first preset threshold, then the first deduction item will be executed. , where m is the cumulative number of consensus-building attempts by those who refused to participate in the expert consensus-building process; If the message response delay exceeds the first duration threshold for three consecutive times, the second deduction item will be executed. Where d is the delay in milliseconds; If the expert qualification information verification is invalid, the third deduction item will be implemented according to the level of invalidity. For example, invalid ratings are divided into three levels: minor (0.1 points deducted), moderate (0.2 points deducted), and malicious (0.3 points deducted). The cumulative deduction is calculated using the following formula. : .

[0019] In summary, by designing mathematical formulas with different weights for positive behaviors such as "successful participation in consensus," "providing valid verification," and "responding with delay," as well as negative behaviors such as "submitting invalid information," "refusing to participate," and "continuous delay," a refined and differentiated measurement of node contribution and fault degree can be achieved. For example, using a logarithmic function to reward long-term participation encourages continuity while preventing unlimited expansion of points, providing accurate data for subsequent fair and credible reputation ranking and node status classification.

[0020] Step S103: Filter out all candidate nodes that meet the condition that the reputation score is greater than the first threshold, calculate the election probability of all candidate nodes based on the reputation score, and select the candidate node corresponding to the maximum election probability from all election probabilities as the master node of this round of consensus. It should be noted that review agency nodes with a credit score less than or equal to the first threshold are considered low-credit nodes. Low-credit nodes that are removed must undergo a 30-day observation period. During this period, if they have no negative behavior and have completed 10 data synchronization tasks, they can submit a recovery application to the management node. After the management node approves the application, its credit score will be reset to 50 points, and it will be rejoined to the consensus group (initially as a normal node).

[0021] Furthermore, in some embodiments, the reputation score is calculated according to the following formula: ; in, , These are the credit scores at time t and time t-1, respectively.

[0022] Before each round of expert selection, the management node triggers the master node election process. The election algorithm dynamically calculates weights based on node reputation scores and network performance. The election probability formula is as follows: ; in, Let be the election probability of the i-th node to be elected at time t. , Let be the reputation scores of the i-th and j-th candidate nodes at time t, respectively. , Let be the network delays of the i-th and j-th nodes to be elected at time t, respectively. , These are reputation weight and performance weight, respectively. This represents the maximum network latency of all nodes to be elected at time t.

[0023] To address the critical issue of dynamically linking the reputation mechanism with the consensus process, the reputation score must reflect the node's latest behavior in real time; otherwise, its evaluation will become outdated and invalid. More importantly, the election probability formula is designed to scientifically combine a node's historical reputation with its real-time network performance as the basis for election. This avoids the possibility of electing an inefficient master node with high network latency based solely on reputation, or an unreliable node with poor reputation based solely on performance. This comprehensively optimizes the reliability and efficiency of the master node, ensuring the quality of each round of consensus initiation.

[0024] Step S104: When the management node initiates an expert extraction request to the master node, it controls the master node to generate a pre-preparation message containing a unique task sequence number and sends the pre-preparation message to the consensus node for verification. Step S105: If the verification is successful, the preparation message is broadcast to the entire network. When the number of preparation messages collected by the master node is greater than the second threshold, the master node broadcasts the submission message to the entire network. After verification by each consensus node, the expert extraction algorithm is executed to obtain the extraction result. The consensus nodes are all review agency nodes except the master node.

[0025] It should be noted that the management node initiates an expert extraction request, including extraction parameters: review area, expert professional title level, and expert avoidance rules; after receiving the extraction request, the master node generates a unique extraction task sequence number and loads the encrypted expert database (corresponding to CID) from IPFS. , The CID is the identifier for the expert database on IPFS (CID is the IPFS content-unique identifier), packaged into a pre-prepared message (including task sequence number, ...). The master node signature is broadcast to all consensus nodes; each consensus node (reviewing node) verifies the message's legitimacy upon receiving it: verifying the master node signature (based on the master node's public key), verifying... Validity (checking the existence of the CID via the IPFS gateway), verification of whether the extracted parameters conform to the preset review rules, which refer to the review rules set by the review committee based on professional and technical titles; nodes that pass verification broadcast preparation messages (including node signatures and message digests) to the entire network and trigger positive adjustments to reputation scores. The master node collects preparation messages, and when the number of valid preparation messages exceeds a second threshold, for example, this second threshold equals... If f is the maximum number of malicious nodes, then the submission phase begins; if not enough messages are collected (e.g., a malicious node refuses to send), then a view switch is triggered, and a new master node is elected.

[0026] The master node then broadcasts a submission message (including a digest of the preparation message set and the master node's signature) to the entire network. After each consensus node verifies the integrity of the preparation message set, it executes an expert extraction algorithm: Based on extraction parameters, candidate experts are selected from the encrypted expert database to meet the review field and expert title level, while excluding experts who trigger avoidance rules; a secure random number is generated, and several review experts are extracted from the candidate experts based on the secure random number, generating the extraction result; the extraction result is symmetrically encrypted and stored in a distributed storage system, the result identifier is obtained, and a submission response is sent to the master node; the master node collects the submission responses, and if the number of submission responses is greater than a third threshold, the extraction result is considered valid; otherwise, it is invalid.

[0027] In summary, the pre-processing message generation and verification steps ensure that expert extraction tasks are securely and reliably initialized and distributed within the consensus system. Requiring the management node to initiate a request containing specific review parameters, and for the master node to generate a unique sequence number and load the designated expert database, guarantees the uniqueness, integrity, and traceability of the task. Furthermore, requiring consensus nodes to verify the validity of the master node signature, the expert database identifier (CID), and the compliance of the parameters establishes a rigorous pre-verification checkpoint. This design effectively prevents forged task requests, tampered expert data, or non-compliant review parameters from entering the subsequent consensus process, thus establishing a robust defense against tampering and ensuring compliance from the very first step.

[0028] By employing parametric screening to ensure experts meet basic requirements, utilizing secure random numbers incorporating blockchain factors to guarantee unpredictability in the selection process, and finally encrypting and storing the results in a distributed system (IPFS), with only the fingerprint (hash) and address (CID) stored on-chain, this design leverages the tamper-proof nature of blockchain to ensure impartiality while addressing the bottleneck of limited blockchain data capacity through off-chain storage, achieving a balance between security and practicality.

[0029] Furthermore, in some embodiments, the extraction results obtained from the above steps constitute an initial candidate expert set. Traditional random extraction still lacks substantive fairness. For example, there may be explicit teacher-student or colleague relationships between review experts and review subjects, as well as implicit academic collaborations and community connections. Based on this, in order to further improve the fairness of extraction, it is necessary to construct an implicit association evaluation model. The implicit association evaluation model takes the semantic feature vectors and structured representations of any two entities as input and outputs an implicit association score characterizing the association strength between the two. The entities include experts in the expert database and the applicants and review subjects in this review. The semantic feature vectors are obtained by analyzing the entity information. The information is semantically represented, and the structured representation is obtained by constructing a relationship graph based on the cooperation relationship between entities and processing it using a graph neural network. For each candidate expert in the initial candidate expert set, the implicit association score between the candidate expert and the applicant or review object is calculated according to the implicit association evaluation model. If the implicit association score exceeds the fourth threshold, the candidate expert is removed, resulting in a candidate expert set after implicit avoidance screening. Finally, at least one fairness constraint is introduced, and under the premise of satisfying all fairness constraints, the optimization solution is performed with the goal of minimizing the distribution deviation of the final expert set to obtain the final candidate expert set, and the final candidate expert set is uploaded to the blockchain.

[0030] By constructing an evaluation model that integrates semantics and graph structure, these implicit connections can be deeply mined and quantified, enabling intelligent avoidance. Furthermore, introducing fairness constraints such as historical sampling frequency and community concentration for optimization aims to proactively optimize the allocation of expert resources from a statistical perspective, avoiding uneven expert workload and thereby enhancing the fairness and credibility of the review process at a deeper level.

[0031] Specifically, in some embodiments, the implicit association score is calculated according to the following formula: ; in, The implicit association between entity a and entity b is classified. These are the weighting coefficients. Let be the semantic feature vector of entity a. Let be the semantic feature vector of entity b. This is a structured representation of entity a output by a graph neural network. This is a structured representation of entity b output by the graph neural network. This is a similarity calculation function; Construct an objective function based on the following formula, which aims to minimize the distribution bias of the final expert set: ; in, For the final set of candidate experts, This is the set of candidate experts selected through implicit avoidance screening. , for The distribution of P is the baseline distribution of the expert database corresponding to the subject of the review object, which is a pre-defined distribution. The total variation distance, For fairness constraints, such as the final expert candidate set The number of experts from the same unit shall not exceed u, and the proportion shall be... , This is the preset threshold. Representing constraints It is true.

[0032] Specifically, in some embodiments, the master node will extract the result corresponding to... , This represents the address or fingerprint of the extracted result file in IPFS, used to locate the complete encrypted extracted result data off-chain. The result hash (SHA-256 calculation), the list of participating consensus nodes, and the record of reputation score changes are packaged into an "extraction result block" and submitted to the blockchain. The encrypted extraction result (AES encrypted) is stored in IPFS; the blockchain only stores... The result hash not only ensures that the data is immutable, but also solves the problem of limited storage capacity in blockchain.

[0033] In addition, the supervisory nodes obtain information from the blockchain. The encrypted extraction results are downloaded through the IPFS gateway, and the AES key is decrypted using the management node's public key (pre-distributed) to decrypt the results. The hash value of the result is then recalculated and compared with the hash value stored on the blockchain. If they match, the verification is successful, a verification report is generated and uploaded to the blockchain. If they do not match, an abnormal alarm is triggered, the management node suspends the review process, and troubleshooting is initiated (such as checking the integrity of IPFS data and the behavior records of consensus nodes).

[0034] The management node will publish the verified extraction results on the government affairs platform, and objections will be accepted during the publicity period. Anyone can query the extraction task ID, the list of participating nodes, the consensus process log, and the result hash through the blockchain explorer, achieving full traceability of the process.

[0035] When the master node fails, any consensus node can initiate a view switch request to collect data. After a signature is supported, f represents the maximum number of malicious nodes allowed, and a new master node is elected. Then, malicious nodes are identified. If a node submits invalid preparations or messages three times consecutively, the management node marks it as a malicious node, removes it from the consensus group, resets its reputation score to zero, and prohibits it from reapplying for rejoining for six months. When data is corrupted, if the expert database or result data stored in IPFS is corrupted, it is restored through the IPFS multi-node backup mechanism (IPFS automatically obtains complete data from other nodes).

[0036] The management node can adjust the consensus parameters according to the scale of the review; adjustments require approval. The agreement takes effect after the consensus of the management nodes is reached; the reputation score rules (such as positive / negative behavior scores) can be optimized according to the actual operation.

[0037] In summary, this invention improves the traditional PBFT consensus by introducing a reputation mechanism and dynamic master node election, and combines implicit association modeling and fairness constraints to fundamentally solve the challenges of fairness, security, and efficiency in the expert selection process for professional title evaluation. Specifically, firstly, by constructing refined node behavior scoring rules and calculating reputation scores in real time, quantitative evaluation and dynamic constraints on participating review institution nodes are achieved, effectively curbing malicious behavior. Based on this, high-reputation, high-performance nodes are dynamically elected as master nodes based on reputation scores and network performance, significantly improving the reliability and efficiency of the consensus process and eliminating process delays or result biases caused by poor master node performance or malicious intent. Secondly, in the expert selection algorithm, an implicit association evaluation model based on semantics and graph neural networks is creatively introduced, which can intelligently identify and avoid potential association risks between experts and applicants. Simultaneously, by establishing a fairness constraint extraction model aimed at minimizing distribution bias, the problem of excessive expert concentration is systematically avoided, thus ensuring broad representativeness and fairness in expert selection at the algorithm level. Finally, the entire consensus process and extraction results are stored on the blockchain and distributed storage technology to ensure transparency, immutability and traceability throughout the entire chain from demand initiation and consensus verification to result generation, providing a solid technical foundation for building a modern professional title evaluation system with strong credibility, scalability and high availability.

[0038] like Figure 2 As shown, one embodiment of the present invention also proposes a blockchain consensus system for professional title evaluation, the system comprising: Blockchain network construction module 10 is used to build a professional title evaluation blockchain network that includes management nodes and review agency nodes; The reputation score calculation module 20 is used to divide the behavior of all nodes into positive behavior and negative behavior, and to construct a first score rule and a second score rule for the positive behavior and the negative behavior respectively, and to calculate the reputation score of all review agency nodes according to the first score rule and the second score rule. The master node selection module 30 is used to filter out all candidate nodes that meet the requirement that the reputation score is greater than the first threshold, calculate the election probability of all candidate nodes based on the reputation score, and select the candidate node corresponding to the maximum election probability from all election probabilities as the master node of this round of consensus. The verification module 40 is used to control the master node to generate a pre-preparation message containing a unique task sequence number when the management node initiates an expert extraction request to the master node, and to send the pre-preparation message to the consensus node for verification. The extraction execution module 50 is used to broadcast a preparation message to the entire network if the verification is successful. When the number of preparation messages collected by the master node is greater than the second threshold, the master node broadcasts a submission message to the entire network. After verification by each consensus node, the expert extraction algorithm is executed to obtain the extraction result. The consensus nodes are all review agency nodes except the master node.

[0039] In another aspect, the present invention also proposes a storage medium on which one or more programs are stored, which, when executed by a processor, implement the aforementioned blockchain consensus method for professional title evaluation.

[0040] In another aspect, the present invention also proposes an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory to implement the aforementioned blockchain consensus method for professional title evaluation.

[0041] Those skilled in the art will understand that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can mean any means that can contain stored, communicated, propagated, or transmitted programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0042] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0043] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0044] While embodiments of the present invention have been described in detail above, it will be apparent to those skilled in the art that various modifications and variations can be made to these embodiments. However, it should be understood that such modifications and variations fall within the scope and spirit of the invention as set forth in the claims. Furthermore, the invention described herein may have other embodiments and can be implemented or carried out in various ways.

Claims

1. A blockchain consensus method for professional title evaluation, characterized in that, The method includes: Construct a blockchain network for professional title evaluation that includes management nodes and review agency nodes; The behavior of all nodes is divided into positive behavior and negative behavior, and a first integration rule and a second integration rule are constructed for the positive behavior and the negative behavior respectively. The reputation score of all review agency nodes is calculated according to the first integration rule and the second integration rule. All candidate nodes that meet the requirement of having a reputation score greater than a first threshold are selected, and the election probability of all candidate nodes is calculated based on the reputation score. The candidate node corresponding to the highest election probability is selected from all election probabilities as the master node for this round of consensus. When the management node initiates an expert extraction request to the master node, it controls the master node to generate a pre-preparation message containing a unique task sequence number and sends the pre-preparation message to the consensus node for verification. If the verification is successful, a preparation message is broadcast to the entire network. When the number of preparation messages collected by the master node is greater than the second threshold, the master node broadcasts a submission message to the entire network. After verification by each consensus node, an expert extraction algorithm is executed to obtain the extraction result. The consensus nodes are all review agency nodes except the master node.

2. The blockchain consensus method for professional title evaluation according to claim 1, characterized in that, The steps of dividing the behavior of all nodes into positive and negative behaviors, and constructing a first integration rule and a second integration rule for the positive and negative behaviors respectively, include: The first scoring rule is: if the review agency node successfully participates in an expert consensus drawing once, the first bonus item will be applied; If the review agency's qualification verification is valid, the second bonus item will be applied; The time taken from receiving a message to responding is recorded as the message response delay. If the message response delay is less than or equal to the first time threshold, the third bonus item is executed once. The cumulative score is calculated using the following formula: ； in, To accumulate points, , , These are the first bonus item, the second bonus item, and the third bonus item, respectively. n is the cumulative number of consensuses reached by the review agency node in the expert consensus drawing process, and v is the number of valid qualification verifications of the review agency node. The second scoring rule is as follows: if the cumulative number of consensus-building attempts refused to participate in expert consensus-building within a unit of time exceeds the first preset threshold, then the first deduction item will be executed. , where m is the cumulative number of consensus-building attempts by those who refused to participate in the expert consensus-building process; If the message response delay exceeds the first duration threshold for three consecutive times, the second deduction item will be executed. Where d is the delay in milliseconds; If the expert's qualification information verification is invalid, the third deduction item will be applied. The cumulative deduction points are calculated according to the following formula. : 。 3. The blockchain consensus method for professional title evaluation according to claim 2, characterized in that, The step of calculating the reputation score of all review agency nodes according to the first scoring rule and the second scoring rule includes: The credit score is calculated using the following formula: ； in, , These are the credit scores at time t and time t-1, respectively. The step of selecting all candidate nodes that meet the requirement of having a reputation score greater than a first threshold, and calculating the election probability of all candidate nodes based on the reputation score, includes: The election probability is calculated using the following formula: ； in, Let be the election probability of the i-th node to be elected at time t. , Let be the reputation scores of the i-th and j-th candidate nodes at time t, respectively. , Let be the network delays of the i-th and j-th nodes to be elected at time t, respectively. , These are reputation weight and performance weight, respectively. This represents the maximum network latency of all nodes to be elected at time t.

4. The blockchain consensus method for professional title evaluation according to claim 3, characterized in that, When the management node initiates an expert extraction request to the master node, the step of controlling the master node to generate a pre-preparation message containing a unique task sequence number and sending the pre-preparation message to the consensus node for verification includes: When the management node initiates an extraction request containing extraction parameters, including the review area, professional title level, and expert avoidance rules, the master node generates a unique extraction task sequence number, loads the encrypted expert library from the distributed storage system, packages it into a pre-prepared message, and broadcasts it. After receiving the pre-preparation message, each consensus node verifies the legality of the pre-preparation message. The verification process includes verifying the master node signature, verifying the validity of the unique extraction task sequence number, and verifying whether the extraction parameters meet the preset review rules.

5. The blockchain consensus method for professional title evaluation according to claim 4, characterized in that, The expert extraction algorithm includes: Candidate experts are selected based on extraction parameters to select experts from the encrypted expert database who meet the review field and expert title level, and to exclude experts who trigger the avoidance rules. Generate a secure random number, and select a number of review experts from the candidate experts based on the secure random number, and generate the selection results; The extracted results are symmetrically encrypted and stored in a distributed storage system. The result identifier is obtained, and a submission response is sent to the master node. The master node collects submission responses, and if the number of submission responses exceeds the third threshold, the extraction result is considered valid.

6. The blockchain consensus method for professional title evaluation according to claim 1, characterized in that, The method further includes: When a new node submits identity authentication materials to the management node, the identity authentication materials include the organization's qualification certificate, the node's public key, and contact information. After verifying the authenticity of the materials, the management node initiates a network-wide node consensus verification to broadcast the new node's authentication request to all approved nodes. When the management node collects more than the first preset proportion of consent signatures from admitted nodes, it determines that the new node has passed authentication.

7. The blockchain consensus method for professional title evaluation according to claim 1, characterized in that, The method includes: An implicit association evaluation model is constructed. The implicit association evaluation model takes the semantic feature vectors and structured representations of any two entities as inputs and outputs an implicit association score that represents the strength of the association between the two entities. The entities include experts in the expert database and the applicants and review subjects of this review. The semantic feature vectors are obtained by semantically representing the entity information. The structured representations are obtained by constructing a relationship graph based on the cooperation relationship between entities and processing it using a graph neural network. For each candidate expert in the initial candidate expert set, the implicit association score between the candidate expert and the applicant or review subject is calculated according to the implicit association evaluation model. If the implicit association score exceeds the fourth threshold, the candidate expert is removed, resulting in a set of candidate experts after implicit avoidance screening. Introduce at least one fairness constraint, and under the premise of satisfying all fairness constraints, optimize the solution with the objective of minimizing the distribution deviation of the final expert set to obtain the final candidate expert set, and upload the final candidate expert set to the blockchain.

8. The blockchain consensus method for professional title evaluation according to claim 7, characterized in that, Calculate the implicit association score using the following formula: ； in, The implicit association between entity a and entity b is classified. These are the weighting coefficients. Let be the semantic feature vector of entity a. Let be the semantic feature vector of entity b. This is a structured representation of entity a output by a graph neural network. This is a structured representation of entity b output by the graph neural network. This is a similarity calculation function; Construct an objective function based on the following formula, which aims to minimize the distribution bias of the final expert set: ； in, For the final set of candidate experts, This is the set of candidate experts selected through implicit avoidance screening. , for The distribution of P, where P is the baseline distribution of the expert database corresponding to the subject being reviewed. The total variation distance, For the sake of fairness, Representing constraints It is true.

9. A blockchain consensus system for professional title evaluation, characterized in that, The system includes: The blockchain network construction module is used to build a blockchain network for professional title evaluation that includes management nodes and review agency nodes. The reputation score calculation module is used to divide the behavior of all nodes into positive behavior and negative behavior, and to construct a first score rule and a second score rule for the positive behavior and the negative behavior respectively, and to calculate the reputation score of all review agency nodes according to the first score rule and the second score rule. The master node selection module is used to filter out all candidate nodes that meet the requirement that the reputation score is greater than the first threshold, calculate the election probability of all candidate nodes based on the reputation score, and select the candidate node corresponding to the maximum election probability from all election probabilities as the master node of this round of consensus. The verification module is used to control the master node to generate a pre-preparation message containing a unique task sequence number when the management node initiates an expert extraction request to the master node, and to send the pre-preparation message to the consensus node for verification. The extraction execution module is used to broadcast a preparation message to the entire network if the verification is successful. When the number of preparation messages collected by the master node is greater than the second threshold, the master node broadcasts a submission message to the entire network. After verification by each consensus node, the expert extraction algorithm is executed to obtain the extraction result. The consensus nodes are all review agency nodes except the master node.

10. A storage medium, characterized in that, The storage medium stores one or more programs that, when executed by a processor, implement the blockchain consensus method for professional title evaluation as described in any one of claims 1-8.

Images