Block chain dynamic trust evaluation system and method based on artificial intelligence

Abstract

The invention provides a block chain dynamic trust evaluation system and method based on artificial intelligence, and the method comprises the steps: carrying out the time sequence modeling of the transaction data of a block chain node and the consensus historical record of the node participation consensus, obtaining the behavior credibility of the block chain node, calculating the resource contribution degree of each transaction node, and carrying out the calculation of the resource contribution degree. Fusing the behavior credibility and each resource contribution degree to obtain a consensus score of each transaction node; constructing a node graph neural network of the block chain based on the transaction topological relation and the interaction frequency between the transaction nodes, detecting a plurality of malicious nodes in the block chain by using the node graph neural network, and further evaluating a consensus influence coefficient of the current malicious node aggregation on the block chain node consensus; and performing punitive correction on the consensus score of each malicious node through the consensus influence coefficient to obtain a final trust score of each transaction node in the block chain. Based on the above scheme, punishment correction of malicious node aggregation in blockchain node trust evaluation can be realized.

Description

Technical Field

[0001] The present application relates to the field of blockchain technology, and more specifically, to an artificial intelligence-based blockchain dynamic trust evaluation system and method. Background Art

[0002] Blockchain dynamic trust management utilizes the distributed ledger, immutability and consensus mechanism of blockchain to dynamically evaluate and adjust the trust relationship between participating entities. In the blockchain network, when the transaction node behaves well, the trust value increases; if there is a violation, the trust value decreases, which can effectively incentivize nodes to abide by the rules and ensure the security and stable operation of the blockchain network. It is suitable for various scenarios such as digital currency transactions, supply chain traceability, and IoT device management.

[0003] Traditional blockchain trust assessment methods employ a single reputation scoring mechanism that calculates static reputation values ​​based solely on independent historical node behavior data, ignoring the topological correlations and dynamic interaction patterns between nodes. This results in an inability to identify the threat of group attacks posed by malicious nodes through high-frequency trading pairs, dense interaction networks, or coordinated behavior. Furthermore, malicious nodes can disguise their reputation through periodic compliance behavior, converging during critical periods to launch coordinated attacks. Lacking graph neural network analysis of transaction topology, dynamic modeling of temporal behavior patterns, and quantitative assessment of group influence, traditional methods are unable to effectively detect anomalous correlation patterns between nodes or accurately penalize and suppress coordinated attacks, ultimately exposing blockchain networks to serious security risks. Therefore, how to implement penalty correction for malicious node aggregation in blockchain node trust assessments has become a challenging issue for the industry. Summary of the Invention

[0004] The present application provides an artificial intelligence-based blockchain dynamic trust assessment system and method, which can realize the penalty correction of malicious node aggregation in blockchain node trust assessment.

[0005] In a first aspect, the present application provides a blockchain dynamic trust assessment method based on artificial intelligence, comprising: Obtain the transaction data of the blockchain node and the consensus history of the node's participation in the consensus; Performing time series modeling on the transaction data and the consensus history records to obtain the behavioral credibility of the blockchain nodes, calculating the resource contribution of each transaction node based on the verified block information recorded in the blockchain ledger, and integrating the behavioral credibility and each resource contribution to obtain the consensus score of each transaction node; A blockchain node graph neural network is constructed based on the transaction topology and interaction frequency between each transaction node. The node graph neural network is used to detect multiple malicious nodes in the blockchain, and then the impact coefficient of the current malicious node aggregation on the blockchain node consensus is evaluated; The consensus score of each malicious node is penalized and modified using the consensus influence coefficient to obtain the final trust score of each transaction node in the blockchain.

[0006] In some embodiments, performing time series modeling on the transaction data and the consensus history to obtain the behavioral credibility of the blockchain node specifically includes: Performing multi-time window alignment on the transaction data and the consensus history records to obtain a transaction field set and a consensus field set for each time window; The abnormal transaction ratio of each time window is determined by each transaction field set, and the consensus stability index of each time window is determined by each consensus field set; Construct a time series feature matrix of blockchain nodes based on all abnormal transaction ratios and consensus stability indicators; The time series feature matrix is ​​input into the pre-trained long short-term memory network model to output the behavioral credibility of the blockchain node.

[0007] In some embodiments, calculating the resource contribution of each transaction node based on the verified block information recorded in the blockchain ledger specifically includes: Obtain verification block information recorded in the blockchain ledger; For each transaction node, extract the transaction node's computing power contribution, token pledge contribution, and storage space contribution from the verified block information; The resource contribution of the transaction node is determined by the computing power contribution, the token pledge contribution and the storage space contribution, thereby obtaining the resource contribution of each transaction node.

[0008] In some embodiments, integrating the behavior credibility and the contribution of each resource to obtain the consensus score of each transaction node specifically includes: For each transaction node, normalize the behavior credibility and the resource contribution of the transaction node; Set the fusion weight of the dynamic trust score based on the network status of the transaction node; Based on the fusion weight, the normalized behavior credibility and resource contribution are weighted and fused to obtain the consensus score of the transaction node, and then the consensus score of each transaction node is obtained.

[0009] In some embodiments, constructing a node graph neural network of a blockchain based on the transaction topology relationship and interaction frequency between each transaction node specifically includes: Obtain the transaction topology relationship and interaction frequency between each transaction node; Extracting the node address of each transaction node from the transaction topology relationship; The node addresses of each node are used as graph nodes and the interaction frequencies are used as edges to generate a node graph neural network of the blockchain.

[0010] In some embodiments, detecting multiple malicious nodes in a blockchain using the node graph neural network specifically includes: Aggregating neighbor information of each graph node in the node graph neural network; Get the node embedding value in each neighbor information; The blockchain nodes are screened for anomalies based on the embedded values ​​of each node, and multiple malicious nodes are obtained.

[0011] In some embodiments, the consensus score of each malicious node is penalized and modified using the consensus influence coefficient to obtain the final trust score of each transaction node in the blockchain, specifically including: For each malicious node, the maximum value of all historical consensus influence coefficients of the malicious node in the historical time period is obtained as the correction benchmark value; Determine a penalty coefficient for the malicious node trust score based on the consensus influence coefficient and the revised benchmark value; Determine the trust score of the malicious node by using the penalty coefficient and the consensus score of the malicious node, and then obtain the trust score of each malicious node; The trust score of each malicious node and the consensus score of each transaction node are combined to obtain the final trust score of each transaction node in the blockchain.

[0012] In a second aspect, the present application provides an artificial intelligence-based blockchain dynamic trust evaluation system, comprising an evaluation unit, wherein the evaluation unit comprises: The acquisition module is used to obtain the transaction data of the blockchain node and the consensus history of the node's participation in the consensus; A processing module is configured to perform time series modeling on the transaction data and the consensus history records to obtain the behavioral credibility of the blockchain nodes, calculate the resource contribution of each transaction node based on the verified block information recorded in the blockchain ledger, and integrate the behavioral credibility and each resource contribution to obtain a consensus score for each transaction node; The processing module is further configured to construct a node graph neural network of the blockchain based on the transaction topology relationship and interaction frequency between each transaction node, and use the node graph neural network to detect multiple malicious nodes in the blockchain, thereby evaluating the influence coefficient of the current malicious node aggregation on the blockchain node consensus; The execution module is used to punitively correct the consensus score of each malicious node using the consensus influence coefficient to obtain the final trust score of each transaction node in the blockchain.

[0013] In a third aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device executes the above-mentioned artificial intelligence-based blockchain dynamic trust assessment method.

[0014] In a fourth aspect, the present application provides a computer-readable storage medium, which stores instructions or codes. When the instructions or codes are run on a computer, the computer implements the above-mentioned artificial intelligence-based blockchain dynamic trust assessment method when executed.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: In an artificial intelligence-based blockchain dynamic trust assessment system and method provided by the present application, transaction data of blockchain nodes and consensus history records of nodes participating in consensus are obtained; time series modeling is performed on the transaction data and the consensus history records to obtain the behavioral credibility of the blockchain nodes, and the resource contribution of each transaction node is calculated based on the verification block information recorded in the blockchain ledger. The behavioral credibility and each resource contribution are integrated to obtain the consensus score of each transaction node; a node graph neural network of the blockchain is constructed based on the transaction topology relationship and interaction frequency between each transaction node, and the node graph neural network is used to detect multiple malicious nodes in the blockchain, and then the consensus influence coefficient of the current malicious node aggregation on the blockchain node consensus is evaluated; the consensus score of each malicious node is punitively corrected by the consensus influence coefficient to obtain the final trust score of each transaction node in the blockchain.

[0016] It can be seen that in this application, the consensus score of each malicious node is penalized and corrected through the consensus influence coefficient to obtain the final trust score of each transaction node in the blockchain; first, the consensus score is determined to obtain a comprehensive evaluation index of the fusion behavior credibility and resource contribution, thereby providing an accurate judgment basis for the penalty correction of malicious node aggregation, which can avoid the one-sidedness of single-dimensional evaluation, for example: relying only on resources to be vulnerable to witch attacks, and identifying potential malicious nodes with high resources but low credibility, thereby ensuring that the demotion of malicious nodes takes into account both their individual threats and their actual occupation of network resources, and fundamentally suppresses the risk of malicious nodes gaining trust through resource stacking and then launching attacks; then, By determining the consensus influence coefficient, we can obtain the quantitative threat value of the current malicious node aggregation to the consensus network, thereby achieving a precise strike against coordinated attacks. The consensus influence coefficient not only reflects the degree of individual malice, but also captures the group synergy effect, and can effectively identify hidden threats such as identity networks in malicious attacks and collusion in staking pools in blockchains. Based on the punitive correction of the consensus influence coefficient, the power reduction intensity can be dynamically adjusted, and exponential penalties can be imposed on high-density malicious groups to disintegrate their synergy advantages; linear penalties are used for isolated malicious nodes to avoid excessive suppression, thereby significantly improving the blockchain's defense capabilities against new types of coordinated attacks; in summary, based on the above scheme, the penalty correction of malicious node aggregation in blockchain node trust assessment can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0018] Figure 1 This is an exemplary flow chart of a blockchain dynamic trust assessment method based on artificial intelligence according to some embodiments of the present application; Figure 2 is a schematic diagram of a process for implementing punitive correction according to some embodiments of the present application; Figure 3 is a schematic structural diagram of an evaluation unit according to some embodiments of the present application; Figure 4 This is a structural diagram of a computer device that implements an artificial intelligence-based blockchain dynamic trust assessment method according to some embodiments of the present application. DETAILED DESCRIPTION

[0019] In order to better understand the technical solution of the present application, the technical solution of the present application will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0020] refer to Figure 1 , which is an exemplary flow chart of an artificial intelligence-based blockchain dynamic trust assessment method according to some embodiments of the present application. The artificial intelligence-based blockchain dynamic trust assessment method mainly includes the following steps: In step 101, the transaction data of the blockchain node and the consensus history record of the node participating in the consensus are obtained.

[0021] It should be noted that in this application, transaction data refers to the operational information of transfers and smart contract calls recorded in the blockchain network. The transaction data includes the sender, receiver, amount, and timestamp. Transaction data is the basic unit of blockchain state change; consensus history records refer to the behavior logs of transaction nodes in the consensus mechanism. The consensus history records include block verification, voting signatures, proposal responses and other processes, reflecting the node's contribution to network security; the transaction data of blockchain nodes and the consensus history records of nodes participating in consensus are obtained from the blockchain distributed ledger.

[0022] In specific implementation, a data query request is first initiated through the target node's application programming interface (API) or full-node client. If the node supports lightweight queries, index information such as transaction hashes and block heights can be directly obtained, and the full transaction content can be retrieved based on the index. If historical data archiving is required, block files must be traversed or an off-chain database must be used. For private chains, permission control can be used to directly access the local ledger. For public chains, remote procedure calls (RPCs) or third-party blockchain browser interfaces are required to filter the target data by transaction hash or address, and the query results are used as the blockchain node's transaction data. Next, if the blockchain uses a transparent consensus mechanism, the node's public key can be associated with block production rights by analyzing the validator signature or random number in the block header. For consortium chains, the node's proposal voting records must be extracted from the ordering service log. If the network provides a consensus audit interface, node participation rounds, state synchronization time, and other information can be directly queried. Otherwise, local log files must be parsed or network broadcast messages monitored to reconstruct the consensus participation sequence based on timestamps and block heights. Some chains store evidence of malicious behavior, requiring additional verification of the integrity of the historical records, which are then used as the consensus history of the blockchain node's participation.

[0023] In step 102, the transaction data and the consensus history records are subjected to time series modeling to obtain the behavioral credibility of the blockchain nodes, and the resource contribution of each transaction node is calculated based on the verification block information recorded in the blockchain ledger. The behavioral credibility and each resource contribution are integrated to obtain the consensus score of each transaction node.

[0024] In some embodiments, performing time series modeling on the transaction data and the consensus history to obtain the behavioral credibility of the blockchain node can be achieved by using the following steps: Performing multi-time window alignment on the transaction data and the consensus history records to obtain a transaction field set and a consensus field set for each time window; The abnormal transaction ratio of each time window is determined by each transaction field set, and the consensus stability index of each time window is determined by each consensus field set; Construct a time series feature matrix of blockchain nodes based on all abnormal transaction ratios and consensus stability indicators; The time series feature matrix is ​​input into the pre-trained long short-term memory network model to output the behavioral credibility of the blockchain node.

[0025] It should be noted that in this application, behavior credibility is a scalar value that comprehensively represents the degree of credibility of the node's historical behavior. The higher the behavior credibility, the more reliable it is; the transaction field set refers to the transaction data set divided by a fixed time window; the consensus field set refers to the behavior data set of the node participating in the consensus within the same time window; the abnormal transaction ratio refers to the percentage of transactions marked as abnormal in a single time window to the total number of transactions; the consensus stability index is a weighted comprehensive score used to quantify the reliability of the node consensus behavior.

[0026] In specific implementation, first, the transaction data and the consensus history records are aligned across multiple time windows. The transaction field sets and consensus field sets for each time window can be obtained using the following method: A sliding time window is set (the default is a 30-second time window). For each time window, the records for that time window are extracted from the transaction data and consensus history records. After the transaction data is aggregated by time window, statistical features such as transaction volume, number of unique addresses, and average handling fee are generated as the transaction field set for the time window. Consensus data is extracted from indicators such as the number of blocks produced, number of verification failures, and message broadcast delay within each time window as the consensus field set for the time window. If the data timestamps are discontinuous, linear interpolation or forward filling is used to fill in missing values ​​to ensure that the lengths of each window are consistent. During the alignment process, conflicting data caused by time zone differences and chain forks need to be processed, and finally a structured multi-dimensional time series sequence is output. The transaction field set and consensus field set of each time window can be obtained through the above method; secondly, the abnormal transaction ratio of each time window is determined through each transaction field set, and the consensus stability index of each time window is determined through each consensus field set. This can be achieved in the following way, namely: matching each transaction field set based on a preset matching rule (for example: the deviation of the handling fee from the mean exceeds a specified value) to identify abnormal transactions, and counting the proportion of abnormal transactions in each time window as the abnormal transaction ratio of the corresponding time window, so as to obtain the abnormal transaction ratio of each time window; for each time window, extract the key indicators of the time window (for example: the number of consecutive unproduced blocks, the voting disagreement rate) from the consensus field sequence, and determine the key indicators of each key window through historical experience. The weights of the key indicators are finally calculated using the weighted average method to output a normalized score of 0 to 1 as the consensus stability indicator of the time window. The consensus stability indicator of each time window can be obtained in the above manner. Then, the time series feature matrix of the blockchain node is constructed based on all abnormal transaction ratios and consensus stability indicators. This can be achieved by concatenating the abnormal transaction ratios and consensus stability indicators of each time window into a matrix in chronological order, with each row representing the feature vector of a complete window, to obtain the time series feature matrix of the blockchain node. Finally, the time series feature matrix is ​​input into a pre-trained long short-term memory network (LSTM) model. The behavior credibility of the output blockchain node can be achieved by the following manner: the pre-trained LSTM model needs to include an input layer (adapting the feature matrix dimension), a hidden layer (capturing long-term dependencies), and a fully connected output layer (mapped to credibility). During training, labeled data (for example, a historical matrix of known malicious / normal nodes) is used, and cross-entropy is used as the loss function to optimize parameters. During the inference phase, the time series feature matrix of the target node is input into the model. The LSTM automatically learns the dynamic relationship between transaction anomalies and consensus stability (for example, the consensus failure rate increases after a surge in abnormal transactions), and finally outputs probabilistic credibility as the behavioral credibility of the blockchain node.

[0027] In some embodiments, the calculation of the resource contribution of each transaction node based on the verified block information recorded in the blockchain ledger can be achieved by the following steps: Obtain verification block information recorded in the blockchain ledger; For each transaction node, extract the transaction node's computing power contribution, token pledge contribution, and storage space contribution from the verified block information; The resource contribution of the transaction node is determined by the computing power contribution, the token pledge contribution and the storage space contribution, thereby obtaining the resource contribution of each transaction node.

[0028] It should be noted that in this application, resource contribution is a weighted score that quantifies the level of support each transaction node provides to blockchain network resources; block verification information refers to the recorded data of each transaction node participating in block generation and verification in the blockchain network; computing power contribution refers to the computing resources invested by the transaction node in the consensus process; token pledge contribution refers to the number and duration of tokens locked by the transaction node in the proof-of-stake mechanism; storage space contribution refers to the distributed storage capacity provided by the transaction node for the blockchain network.

[0029] In practice, obtaining verified block information from blockchain ledger records can be accomplished by querying the block headers within the target timeframe from the blockchain node's local ledger database and extracting fields relevant to verification, including the effective computational effort per unit time, the number and duration of tokens locked by the node in proof-of-stake mechanisms, the amount of stored data, and the time of availability. For consortium chains, verification records must be obtained from ordering service logs or smart contract events; for public chains, block data can be pulled in batches using a blockchain browser interface. If the authenticity of the data needs to be verified, it is necessary to compare the ledger copies of multiple nodes to ensure that the information has not been tampered with, and at the same time deal with the conflicting blocks caused by possible chain forks, so that the relevant field results obtained from the query are used as the verification block information recorded in the blockchain ledger; then, for each transaction node, the computing power contribution, token pledge contribution and storage space contribution of the transaction node are extracted from the verification block information. This can be achieved in the following way, namely: for each transaction node, the effective computing amount of the transaction node per unit time in the verification block information is used as the computing power contribution of the transaction node; the product of the number and duration of tokens locked by the transaction node in the proof-of-stake mechanism in the verification block information is used as the transaction node The token pledge contribution of the point; the product of the storage data volume of the transaction node in the verified block information and the availability time is used as the storage space contribution of the transaction node; finally, the resource contribution of the transaction node is determined by the computing power contribution, the token pledge contribution and the storage space contribution, and then the resource contribution of each transaction node is obtained. This can be achieved in the following way, namely: the influence weights of computing power contribution, token pledge contribution and storage space contribution on the transaction node can be preset based on historical experience, so as to calculate the weighted sum of the normalized computing power contribution, token pledge contribution and storage space contribution as the resource contribution of the transaction node. The resource contribution of each transaction node can be obtained in the above way.

[0030] In some embodiments, the following steps may be used to integrate the behavior credibility and the contribution of each resource to obtain a consensus score for each transaction node: For each transaction node, normalize the behavior credibility and the resource contribution of the transaction node; Set the fusion weight of the dynamic trust score based on the network status of the transaction node; Based on the fusion weight, the normalized behavior credibility and resource contribution are weighted and fused to obtain the consensus score of the transaction node, and then the consensus score of each transaction node is obtained.

[0031] It should be noted that in this application, the consensus score is an evaluation value used to measure the priority of a node in participating in the consensus. The higher the consensus score, the more suitable the transaction node is to be selected as a block producer or validator. In specific implementation, for each transaction node, the behavior credibility and the resource contribution of the transaction node are normalized. The following implementation can be adopted, namely: using minimum-maximum normalization, the behavior credibility (such as the 0~1 probability value output by LSTM) is directly retained, and the resource contribution is scaled to the same interval according to the maximum and minimum values ​​of the indicator in all transaction nodes, so as to complete the normalization of the behavior credibility and resource contribution; then, the dynamic trust score fusion is set through the network status of the transaction node. The weight can be implemented in the following way, namely: the weight corresponding to the network status of the transaction node is filtered from the network-weight mapping table of the blockchain as the fusion weight of the dynamic trust score, and the fusion weight is used to adjust the proportion of behavior credibility and resource contribution in the final consensus score; finally, based on the fusion weight, the normalized behavior credibility and resource contribution are weighted and fused to obtain the consensus score of the transaction node, and then the consensus score of each transaction node can be achieved in the following way, namely: the weighted sum of the normalized behavior credibility and resource contribution is calculated using the fusion weight as the consensus score of the transaction node. The consensus score of each transaction node can be obtained in the above way.

[0032] In step 103, a node graph neural network of the blockchain is constructed based on the transaction topology relationship and interaction frequency between each transaction node. The node graph neural network is used to detect multiple malicious nodes in the blockchain, and then the consensus influence coefficient of the current malicious node aggregation on the blockchain node consensus is evaluated.

[0033] In some embodiments, constructing a node graph neural network of a blockchain based on the transaction topology relationship and interaction frequency between each transaction node can be achieved by using the following steps: Obtain the transaction topology relationship and interaction frequency between each transaction node; Extracting the node address of each transaction node from the transaction topology relationship; The node addresses of each node are used as graph nodes and the interaction frequencies are used as edges to generate a node graph neural network of the blockchain.

[0034] It should be noted that in this application, the node graph neural network is a graph structure data model with blockchain addresses as nodes and interaction frequency as edge weights. The node graph neural network can be used to analyze node behavior and network characteristics through deep learning; the transaction topology relationship represents the connection structure formed by transactions between different nodes in the blockchain network; the interaction frequency refers to the frequency of message transmission between each transaction node within a specified time, and the interaction frequency can be used to quantify the strength of the association between transaction nodes; the node address refers to a string that uniquely identifies the participant in the blockchain, and the node address can be used to represent the entity node in the graph neural network.

[0035] In specific implementation, first, the transaction topology relationship and interaction frequency between each transaction node can be obtained by the following method, namely: extracting the original transaction data from the blockchain ledger, including the sender address, receiver address, timestamp and transaction type of each transaction; for large-scale data, a distributed computing framework (such as the Spark framework) can be used for batch processing, and the number of transactions between each transaction node is counted according to the time window (the default is every hour) to form an interaction frequency matrix; if a smart contract is involved, it is necessary to further parse the contract event log and identify the call relationship between the contract and the user node as a transaction topology relationship, where the node represents the address and the edge represents the interaction frequency; then, extracting the node address of each transaction node from the transaction topology relationship can be achieved by the following method, namely: taking the node in the transaction topology relationship as the node address of the corresponding transaction node, the node address of each transaction node can be obtained; finally, taking each node address as a graph node and each interaction frequency as an edge to generate a node graph neural network of the blockchain.

[0036] In some embodiments, detecting multiple malicious nodes in a blockchain using the node graph neural network may be achieved by using the following steps: Aggregating neighbor information of each graph node in the node graph neural network; Get the node embedding value in each neighbor information; The blockchain nodes are screened for anomalies based on the embedded values ​​of each node, and multiple malicious nodes are obtained.

[0037] It should be noted that in this application, malicious nodes refer to nodes with abnormal behavior that threaten the security of the blockchain; neighbor information refers to other nodes directly connected to the target node in the graph neural network and their interaction relationship data; node embedding value is a low-dimensional dense vector used to characterize the node's topological structure, behavior pattern and its semantic information in the global network.

[0038] In specific implementation, first, aggregating the neighbor information of each graph node in the node graph neural network can be implemented in the following manner, namely: for each graph node in the node graph neural network, the nodes directly connected to the graph node are screened out from the graph neural network as neighbor nodes, and the account activity, transaction mode encoding and connection edge weight (i.e., normalized transaction frequency) of each neighbor node are collected, so that the average of all normalized transaction frequencies is used as the neighbor information of the graph node. The neighbor information of each graph node in the node graph neural network can be obtained by the above method; then, obtaining the node embedding value in each neighbor information can be implemented in the following manner, namely: taking the average of the normalized transaction frequency in each neighbor information as the node embedding value in the corresponding neighbor information, so as to obtain the node embedding value in each neighbor information; finally, performing abnormal screening on the blockchain node based on the embedding value of each node, and obtaining multiple malicious nodes can be implemented in the following manner, namely: the abnormal threshold of the blockchain node can be preset through historical experience, and the graph nodes corresponding to the neighbor information with a node embedding value lower than the abnormal threshold are regarded as malicious nodes, so as to obtain multiple malicious nodes.

[0039] In some embodiments, evaluating the consensus influence coefficient of the current malicious node aggregation on the blockchain node consensus can be achieved by using the following steps: Obtain the consensus score of each malicious node; Determining the inner edge density of each malicious node in the node graph neural network; The consensus impact coefficient of the current malicious node aggregation on the blockchain node consensus is determined through all consensus scores and internal edge density.

[0040] It should be noted that, in the present application, the consensus influence coefficient is a threat quantification value that integrates the individual capabilities of malicious nodes and the degree of group collaboration; in specific implementation, first, the consensus score of each malicious node can be obtained in the following way, namely: the consensus score of the transaction node corresponding to each malicious node is used as the consensus score of the malicious node, and the consensus score of each malicious node can be obtained; then, the inner edge density of each malicious node in the node graph neural network can be determined in the following way, namely: for each malicious node, the node graph neural network that only contains the malicious node and the edges between them is used as the induced subgraph of the malicious node, and the actual interaction edges in the induced subgraph are counted. The number of edges is divided by the theoretical maximum number of edges (i.e., the number of edges when the malicious node is fully connected), and the calculated result is used as the inner edge density of the malicious node. The inner edge density of each malicious node in the node graph neural network can be obtained in the above way; the inner edge density is an indicator value used to quantify the closeness between malicious nodes. The higher the inner edge density, the greater the possibility of coordinated malicious behavior; finally, the consensus influence coefficient of the current malicious node cluster on the blockchain node consensus is determined by all consensus scores and inner edge densities. This can be achieved in the following way: the sum of all consensus scores multiplied by the inner edge density is used as the consensus influence coefficient of the current malicious node cluster on the blockchain node consensus.

[0041] In step 104, the consensus score of each malicious node is penalized and modified using the consensus influence coefficient to obtain the final trust score of each transaction node in the blockchain.

[0042] In some embodiments, the consensus score of each malicious node is penalized and modified by the consensus influence coefficient to obtain the final trust score of each transaction node in the blockchain. Figure 2 As described above, this figure is a schematic diagram of the process of implementing punitive correction in some embodiments of the present application. In this embodiment, punitive correction can be implemented by the following steps: In step 1041, for each malicious node, the maximum value of all historical consensus influence coefficients of the malicious node in the historical time period is obtained as a correction reference value; In step 1042, a penalty coefficient for the malicious node trust score is determined based on the consensus influence coefficient and the revised reference value; In step 1043, the trust score of the malicious node is determined by the penalty coefficient and the consensus score of the malicious node, thereby obtaining the trust score of each malicious node; In step 1044, the trust scores of each malicious node and the consensus scores of each transaction node are combined to obtain the final trust score of each transaction node in the blockchain.

[0043] It should be noted that in this application, the final trust score is the full set of credibility of blockchain network nodes after evaluation based on unified standards; the corrected baseline value is a quantitative value of the maximum potential threat posed by a malicious node to the blockchain network in which it is located; the penalty coefficient is an adjustment factor used to quantify the proportion of reduction in the trust score of a malicious node; and the trust score is a comprehensive score reflecting the credibility of the transaction node at the current moment.

[0044] In specific implementation, first, for each malicious node, the maximum value of all historical consensus influence coefficients of the malicious node in the historical time period is obtained as the correction reference value, which can be achieved in the following way: for each malicious node, the maximum value of all historical consensus influence coefficients of the malicious node in the historical time period (the default is the most recent 1 week) is obtained as the correction reference value; secondly, the penalty coefficient of the malicious node trust score is determined according to the consensus influence coefficient and the correction reference value, which can be achieved in the following way: the ratio of the current consensus influence coefficient to the correction reference value is used as the penalty coefficient of the malicious node trust score; then, the trust score of the malicious node is determined by the penalty coefficient and the consensus score of the malicious node, and then the trust score of each malicious node is obtained. This can be achieved in the following way: the product of the penalty coefficient and the consensus score of the malicious node is used as the trust score of the malicious node. The trust score of each malicious node can be obtained by the above method. Finally, the trust score of each malicious node and the consensus score of each transaction node are combined to obtain the final trust score of each transaction node in the blockchain. This can be achieved in the following way: for each transaction node in the blockchain, if the transaction node is not a malicious node, the consensus score of the transaction node is directly used as the final trust score of the transaction node. If the transaction node is a malicious node, the trust score of the transaction node corresponding to the malicious node is used as the final trust score of the transaction node. The final trust score of each transaction node in the blockchain can be obtained by the above method.

[0045] In addition, in another aspect of the present application, in some embodiments, the present application provides a blockchain dynamic trust evaluation system based on artificial intelligence, the system includes an evaluation unit, reference Figure 3 , which is a schematic diagram of the structure of an evaluation unit according to some embodiments of the present application. The evaluation unit includes: an acquisition module 201, a processing module 202 and an execution module 203, which are described as follows: Acquisition module 201, in this application, acquisition module 201 is mainly used to obtain transaction data of blockchain nodes and consensus history records of nodes participating in consensus; Processing module 202, in this application, is used to perform time series modeling on the transaction data and the consensus history records to obtain the behavioral credibility of the blockchain nodes, calculate the resource contribution of each transaction node based on the verified block information recorded in the blockchain ledger, and integrate the behavioral credibility and each resource contribution to obtain the consensus score of each transaction node; It should be noted that the processing module 202 is also used to build a node graph neural network of the blockchain based on the transaction topology relationship and interaction frequency between each transaction node, and use the node graph neural network to detect multiple malicious nodes in the blockchain, and then evaluate the consensus influence coefficient of the current malicious node aggregation on the blockchain node consensus; Execution module 203, in this application, execution module 203 is mainly used to punitively correct the consensus score of each malicious node through the consensus influence coefficient to obtain the final trust score of each transaction node in the blockchain.

[0046] The above describes in detail the examples of the artificial intelligence-based blockchain dynamic trust assessment system and method provided by the embodiments of the present application. It can be understood that in order to realize the above functions, the corresponding device includes hardware structures and / or software modules corresponding to the execution of each function. It should be easily appreciated by those skilled in the art that, in combination with the units and algorithm steps of each example described in the embodiments disclosed herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0047] In some embodiments, the present application also provides a computer device, comprising a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device executes the above-mentioned artificial intelligence-based blockchain dynamic trust assessment method.

[0048] In some embodiments, reference Figure 4 , the dotted line in the figure indicates that the unit or module is optional. The figure is a structural diagram of a computer device for implementing a dynamic trust evaluation method for blockchain based on artificial intelligence according to an embodiment of the present application. The dynamic trust evaluation method for blockchain based on artificial intelligence described in the above embodiment can be Figure 4 The computer device shown in the figure is implemented, and the computer device includes at least one processor 301, a memory 302 and at least one communication unit 305. The computer device can be a terminal device, a server or a chip.

[0049] The processor 301 may be a general-purpose processor or a dedicated processor. For example, the processor 301 may be a central processing unit (CPU), which may be used to control the computer device, execute software programs, and process data from the software programs. The computer device may also include a communication unit 305 for inputting (receiving) and outputting (transmitting) signals.

[0050] For example, the computer device may be a chip, the communication unit 305 may be an input and / or output circuit of the chip, or the communication unit 305 may be a communication interface of the chip, and the chip may be a component of a terminal device, a network device, or other device.

[0051] For another example, the computer device may be a terminal device or a server, and the communication unit 305 may be a transceiver of the terminal device or the server, or the communication unit 305 may be a transceiver circuit of the terminal device or the server.

[0052] The computer device may include one or more memories 302, on which a program 304 is stored. The program 304 can be executed by the processor 301 to generate instructions 303, so that the processor 301 executes the method described in the above method embodiment according to the instructions 303. Optionally, data (such as a target audit model) can also be stored in the memory 302. Optionally, the processor 301 can also read data stored in the memory 302. The data can be stored at the same storage address as the program 304, or at a different storage address from the program 304.

[0053] The processor 301 and the memory 302 may be provided separately or integrated together, for example, integrated on a system on chip (SOC) of a terminal device.

[0054] It should be understood that each step of the above method embodiment can be completed by a hardware-based logic circuit or software-based instructions in the processor 301. The processor 301 can be a CPU, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, such as discrete gates, transistor logic devices, or discrete hardware components.

[0055] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0056] For example, in some embodiments, the present application also provides a computer-readable storage medium, which stores instructions or codes. When the instructions or codes are run on a computer, the computer implements the above-mentioned artificial intelligence-based blockchain dynamic trust assessment method when executing.

[0057] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application.

[0058] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A blockchain dynamic trust assessment method based on artificial intelligence, characterized in that: The steps include: Obtain the transaction data of the blockchain node and the consensus history of the node's participation in the consensus; Performing time series modeling on the transaction data and the consensus history records to obtain the behavioral credibility of the blockchain nodes, calculating the resource contribution of each transaction node based on the verified block information recorded in the blockchain ledger, and integrating the behavioral credibility and each resource contribution to obtain the consensus score of each transaction node; A blockchain node graph neural network is constructed based on the transaction topology and interaction frequency between each transaction node. The node graph neural network is used to detect multiple malicious nodes in the blockchain, and then the impact coefficient of the current malicious node aggregation on the blockchain node consensus is evaluated; The consensus score of each malicious node is penalized and modified using the consensus influence coefficient to obtain the final trust score of each transaction node in the blockchain.

2. The method according to claim 1, wherein Performing time series modeling on the transaction data and the consensus history to obtain the behavioral credibility of the blockchain node specifically includes: Performing multi-time window alignment on the transaction data and the consensus history records to obtain a transaction field set and a consensus field set for each time window; The abnormal transaction ratio of each time window is determined by each transaction field set, and the consensus stability index of each time window is determined by each consensus field set; Construct a time series feature matrix of blockchain nodes based on all abnormal transaction ratios and consensus stability indicators; The time series feature matrix is ​​input into the pre-trained long short-term memory network model to output the behavioral credibility of the blockchain node.

3. The method according to claim 1, wherein The resource contribution of each transaction node is calculated based on the verified block information recorded in the blockchain ledger, including: Obtain verification block information recorded in the blockchain ledger; For each transaction node, extract the transaction node's computing power contribution, token pledge contribution, and storage space contribution from the verified block information; The resource contribution of the transaction node is determined by the computing power contribution, the token pledge contribution and the storage space contribution, thereby obtaining the resource contribution of each transaction node.

4. The method according to claim 1, wherein Combining the credibility of the behavior and the contribution of each resource, the consensus score of each transaction node is obtained, which specifically includes: For each transaction node, normalize the behavior credibility and the resource contribution of the transaction node; Set the fusion weight of the dynamic trust score based on the network status of the transaction node; Based on the fusion weight, the normalized behavior credibility and resource contribution are weighted and fused to obtain the consensus score of the transaction node, and then the consensus score of each transaction node is obtained.

5. The method according to claim 1, wherein The node graph neural network of the blockchain is constructed based on the transaction topology relationship and interaction frequency between each transaction node. Specifically, it includes: Obtain the transaction topology relationship and interaction frequency between each transaction node; Extracting the node address of each transaction node from the transaction topology relationship; The node addresses of each node are used as graph nodes and the interaction frequencies are used as edges to generate a node graph neural network of the blockchain.

6. The method according to claim 1, wherein The node graph neural network is used to detect multiple malicious nodes in the blockchain, specifically including: Aggregating neighbor information of each graph node in the node graph neural network; Get the node embedding value in each neighbor information; The blockchain nodes are screened for anomalies based on the embedded values ​​of each node, and multiple malicious nodes are obtained.

7. The method according to claim 1, wherein The consensus score of each malicious node is penalized and modified using the consensus influence coefficient to obtain the final trust score of each transaction node in the blockchain. Specifically, the following are included: For each malicious node, the maximum value of all historical consensus influence coefficients of the malicious node in the historical time period is obtained as the correction benchmark value; Determine a penalty coefficient for the malicious node trust score based on the consensus influence coefficient and the revised benchmark value; Determine the trust score of the malicious node by using the penalty coefficient and the consensus score of the malicious node, and then obtain the trust score of each malicious node; The trust score of each malicious node and the consensus score of each transaction node are combined to obtain the final trust score of each transaction node in the blockchain.

8. A blockchain dynamic trust evaluation system based on artificial intelligence, the system includes an evaluation unit, characterized in that: The evaluation unit comprises: The acquisition module is used to obtain the transaction data of the blockchain node and the consensus history of the node's participation in the consensus; A processing module is configured to perform time series modeling on the transaction data and the consensus history records to obtain the behavioral credibility of the blockchain nodes, calculate the resource contribution of each transaction node based on the verified block information recorded in the blockchain ledger, and integrate the behavioral credibility and each resource contribution to obtain a consensus score for each transaction node; The processing module is further configured to construct a node graph neural network of the blockchain based on the transaction topology relationship and interaction frequency between each transaction node, and use the node graph neural network to detect multiple malicious nodes in the blockchain, thereby evaluating the influence coefficient of the current malicious node aggregation on the blockchain node consensus; The execution module is used to punitively correct the consensus score of each malicious node using the consensus influence coefficient to obtain the final trust score of each transaction node in the blockchain.

9. A computer device, characterized in that: The computer device includes a memory and a processor, the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device executes the artificial intelligence-based blockchain dynamic trust assessment method described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores instructions or codes, which, when executed on a computer, enable the computer to implement the artificial intelligence-based blockchain dynamic trust assessment method as described in any one of claims 1 to 7.

Images