Privacy computing node reputation evaluation method and device, storage medium and program product

Abstract

This application provides a method, device, storage medium, and program product for evaluating the reputation of privacy computing nodes, relating to the field of communication technology. The method includes: acquiring attribute reputation scores, interaction reputation scores, historical reputation scores, and predicted reputation scores of the privacy computing node to be evaluated; and determining a comprehensive reputation score of the privacy computing node to be evaluated based on the attribute reputation scores, interaction reputation scores, historical reputation scores, and predicted reputation scores. This method is applicable to the reputation evaluation process of privacy computing nodes and is used to provide a reputation evaluation scheme for privacy computing nodes.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a method, device, storage medium, and program product for evaluating the reputation of privacy computing nodes. Background Technology

[0002] Privacy-preserving computation specifically refers to the analysis and computation of data involving privacy information while protecting the privacy information from being disclosed to the public when processing information such as video, audio, images, graphics, text, numerical values, and ubiquitous network behavioral information streams.

[0003] Privacy-preserving computing nodes inevitably come into contact with private information during privacy-preserving computations. If a privacy-preserving computing node malfunctions, it could lead to the leakage of private information. If a reputation evaluation of privacy-preserving computing nodes could be conducted, the likelihood of such malfunctions could be predicted.

[0004] However, there is currently a lack of reputation evaluation methods for privacy computing nodes. Summary of the Invention

[0005] This application provides a method, device, storage medium, and program product for evaluating the reputation of privacy computing nodes, which can provide a relatively comprehensive method for evaluating the reputation of privacy computing nodes.

[0006] In a first aspect, this application provides a method for evaluating the reputation of a privacy computing node, the method comprising: obtaining the attribute reputation score, interaction reputation score, historical reputation score and predicted reputation score of the privacy computing node to be evaluated;

[0007] Among them, the attribute reputation score is used to evaluate the number and types of provable attribute information provided by the privacy computing node; the provable attribute information is determined based on the attribute information directly provided by the privacy computing node to be evaluated and the attribute information of the privacy computing node to be evaluated provided by a trusted information source; the interaction reputation score is used to evaluate the comprehensive reputation score of interactive privacy computing nodes that have an interactive relationship with the privacy computing node to be evaluated; the historical reputation score is used to evaluate the historical comprehensive reputation score of the privacy computing node to be evaluated; and the predictive reputation score is used to evaluate the behavioral characteristics of the privacy computing node to be evaluated.

[0008] Based on attribute reputation score, interaction reputation score, historical reputation score and predicted reputation score, the comprehensive reputation score of the privacy computing node to be evaluated is determined.

[0009] The privacy computing node reputation evaluation method provided in this application can obtain the attribute reputation score, interaction reputation score, historical reputation score and predicted reputation score of the privacy computing node to be evaluated, and determine the comprehensive reputation score of the privacy computing node to be evaluated based on the attribute reputation score, interaction reputation score, historical reputation score and predicted reputation score. It can comprehensively consider the reputation scores of multiple dimensions, thereby providing a more comprehensive and accurate reputation evaluation method to ensure the credibility of privacy computing nodes in heterogeneous environments.

[0010] Optionally, the trusted information source includes the operation logs of the privacy computing node to be evaluated; obtaining the attribute reputation score of the privacy computing node to be evaluated includes: receiving attribute information sent from the privacy computing node to be evaluated; obtaining the operation logs of the privacy computing node to be evaluated; determining provable attribute information from the operation logs; wherein, provable attribute information is attribute information in the operation logs that is the same as the attribute information sent by the privacy computing node to be evaluated; determining the attribute reputation score based on the number of types of provable attribute information; the attribute reputation score is positively correlated with the number of types of provable attribute information.

[0011] Optionally, the interaction reputation score of the privacy computing node to be evaluated is obtained, including: weighting and summing the comprehensive reputation scores of the privacy computing nodes that have interaction relationships with the privacy computing node to be evaluated to obtain the interaction reputation score.

[0012] Optionally, the weights corresponding to interactive privacy computing nodes are determined based on the reputation weights and connection weights of the interactive privacy computing nodes; the reputation weights are positively correlated with the degree of scoring deviation; the degree of scoring deviation is the degree of deviation between the overall reputation score and the average reputation score of the corresponding interactive privacy computing node; the average reputation score is the average of the overall reputation scores of all privacy computing nodes in the trusted domain where the privacy computing node to be evaluated is located; the trusted domain is a regional network composed of privacy computing nodes that have been issued CA certificates; and the connection weights are positively correlated with the number of privacy computing nodes connected to the corresponding interactive privacy computing node.

[0013] Optionally, the weights corresponding to the interactive privacy computation nodes are calculated using the following formula:

[0014] Weight i = (Reputation Weight) i +Connection weight i ) / 2;

[0015] Among them, weight i Represents the weight corresponding to the i-th interactive privacy computing node that has an interaction relationship with the privacy computing node to be evaluated; reputation weight. i This represents the reputation weight and connection weight corresponding to the i-th interactive privacy computing node that has an interaction relationship with the privacy computing node to be evaluated.i This represents the connection weight corresponding to the i-th interactive privacy computing node that has an interactive relationship with the privacy computing node to be evaluated.

[0016]

[0017] Among them, the comprehensive credit score i The comprehensive reputation score represents the i-th interactive privacy computing node that has an interactive relationship with the privacy computing node to be evaluated.

[0018]

[0019] Among them, the number of connected nodes i The number of privacy computing nodes connected to the i-th interactive privacy computing node that has an interactive relationship with the privacy computing node to be evaluated; the total number of node connections represents the sum of the number of privacy computing nodes connected to each privacy computing node in the trusted domain.

[0020] Optionally, the historical reputation score of the privacy computing node to be evaluated is obtained, including: exponentially decaying the historical comprehensive reputation score of the privacy computing node to be evaluated to obtain the historical reputation score of the privacy computing node to be evaluated.

[0021] Optionally, the historical comprehensive reputation score of the privacy computing node to be evaluated is exponentially decayed to obtain the historical reputation score of the privacy computing node to be evaluated, including:

[0022] Exponential decay is performed according to the following formula:

[0023]

[0024] Here, Adjusted Score represents historical credit score; Score j t represents the historical comprehensive credit score at the j-th evaluation; λ represents the decay factor, which is positively correlated with the decay rate; j This represents the unit time difference between the j-th evaluation and the current time.

[0025] Optionally, obtaining the predicted reputation score of the privacy computing node to be evaluated includes: obtaining behavioral feature data of the privacy computing node to be evaluated; determining the probability of abnormal behavior of the privacy computing node to be evaluated based on the behavioral feature data and the reputation prediction model; the reputation prediction model is used to predict the probability of abnormal behavior based on the behavioral feature data of the privacy computing node; and determining the predicted reputation score of the privacy computing node to be evaluated based on the probability of abnormal behavior.

[0026] Optionally, the predicted reputation score of the privacy computing node to be evaluated is determined based on the probability of its anomalous behavior, including: calculating the predicted reputation score of the privacy computing node to be evaluated according to the following formula:

[0027] Predicted reputation score = (1 - probability of abnormal behavior) × the most recent historical comprehensive reputation score of the privacy computing node to be evaluated.

[0028] Optionally, the method further includes: obtaining an initial training sample set; the initial training sample set includes a first proportion of positive samples and a second proportion of unlabeled samples; wherein each positive sample includes behavioral feature data of a privacy computing node and an abnormal behavior probability label of the privacy computing node; the abnormal behavior probability label of the positive sample is higher than a probability threshold; each unlabeled sample includes behavioral feature data of a privacy computing node; the second proportion is higher than the first proportion; using a Naive Bayes model to label the unlabeled samples with abnormal behavior probability labels, classifying the unlabeled samples into positive samples and negative samples; the abnormal behavior probability label of the negative samples is lower than a probability threshold; training the initial model based on the positive samples in the initial sample set and the positive and negative samples labeled by the Naive Bayes model to obtain a reputation prediction model.

[0029] Optionally, a reputation prediction model is trained based on positive samples in the initial sample set and positive and negative samples labeled by the Naive Bayes model. This includes: training the initial model based on positive samples in the initial sample set and positive and negative samples labeled by the Naive Bayes model until the training stops; validating the initial model based on a validation sample set to obtain the prediction result of the initial model; the validation sample set includes multiple positive samples; calculating the validation score of the initial model based on the prediction result of the initial model; and determining the initial model as a reputation prediction model if the validation score is higher than a score threshold.

[0030] Optionally, based on the prediction results of the initial model, a validation score for the initial model is calculated, including:

[0031] Calculate the validation score of the initial model using the following formula:

[0032]

[0033] Where S represents the validation score; r represents the recall rate; and Pr[f(x)=1] represents the probability that a sample in the validation set is identified as a positive sample based on the probability of abnormal behavior predicted by the initial model.

[0034] Optionally, based on attribute reputation score, interaction reputation score, historical reputation score, and predicted reputation score, a comprehensive reputation score for the privacy computing node to be evaluated is determined, including: weighted summation of attribute reputation score, interaction reputation score, historical reputation score, and predicted reputation score to obtain the comprehensive reputation score for the privacy computing node to be evaluated; wherein the weight corresponding to the attribute reputation score is greater than or equal to the weight corresponding to the predicted reputation score, the weight corresponding to the predicted reputation score is greater than the weight corresponding to the historical reputation score, and the weight corresponding to the historical reputation score is greater than or equal to the weight corresponding to the interaction reputation score.

[0035] Secondly, this application provides a credit evaluation device, which includes various functional modules for the method described in the first aspect above.

[0036] Thirdly, this application provides a computer program product comprising: computer instructions; when the computer instructions are executed in an electronic device, the electronic device causes the electronic device to perform the method described in the first aspect above.

[0037] Fourthly, this application provides an electronic device comprising: a processor and a memory; the memory storing processor-executable instructions; when the processor is configured to execute the instructions, causing the electronic device to perform the method described in the first aspect above.

[0038] Fifthly, this application provides a computer-readable storage medium comprising: software instructions; when the software instructions are executed in an electronic device, they cause the electronic device to perform the method described in the first aspect above.

[0039] The beneficial effects of the second to fifth aspects mentioned above can be referred to the first aspect, and will not be repeated here. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 A schematic diagram illustrating the composition of the privacy computing node reputation evaluation system provided in this application embodiment;

[0042] Figure 2 A flowchart illustrating the privacy computing node reputation evaluation method provided in this application embodiment;

[0043] Figure 3 A schematic diagram of the credit assessment process provided in this application embodiment;

[0044] Figure 4 A schematic diagram illustrating the composition of the credit evaluation device provided in the embodiments of this application;

[0045] Figure 5 This is a schematic diagram illustrating the composition of an electronic device provided in an embodiment of this application. Detailed Implementation

[0046] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0047] It should be noted that in the embodiments of this application, the words "exemplarily" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design scheme described as "exemplarily" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplarily" or "for example" is intended to present the relevant concepts in a specific manner.

[0048] To facilitate a clear description of the technical solutions of the embodiments of this application, the terms "first" and "second" are used in the embodiments of this application to distinguish the same or similar items with essentially the same function and effect. Those skilled in the art can understand that the terms "first" and "second" are not intended to limit the quantity or execution order.

[0049] Privacy-preserving computation specifically refers to the analysis and computation of data involving privacy information while protecting the privacy information from being disclosed to the public when processing information such as video, audio, images, graphics, text, numerical values, and ubiquitous network behavioral information streams.

[0050] Privacy-preserving computing nodes inevitably come into contact with private information during privacy-preserving computations. If a privacy-preserving computing node malfunctions, it could lead to the leakage of private information. If a reputation evaluation of privacy-preserving computing nodes could be conducted, the likelihood of such malfunctions could be predicted.

[0051] However, there is currently a lack of reputation evaluation methods for privacy computing nodes.

[0052] Based on this, embodiments of this application provide a method, device, storage medium, and program product for evaluating the reputation of privacy computing nodes, which can provide a relatively comprehensive method for evaluating the reputation of privacy computing nodes.

[0053] The following description is provided in conjunction with the accompanying drawings.

[0054] Figure 1 This is a schematic diagram illustrating the composition of the privacy computing node reputation evaluation system provided in this application embodiment. Figure 1 As shown, the system includes a privacy computing node 100 and a central node 200, and the privacy computing node 100 and the central node 200 have a communication connection.

[0055] Privacy computing node 100 may include multiple privacy computing nodes ( Figure 1 (The example uses three privacy computing nodes: privacy computing node A, privacy computing node B, and privacy computing node C).

[0056] Privacy computing node 100 can be an electronic device with computing processing capabilities, such as a computer or server.

[0057] The server can be a single server or a server cluster consisting of multiple servers. In some embodiments, the server cluster can also be a distributed cluster. Optionally, the server can also be implemented on a cloud platform, such as a private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, inter-cloud, and multi-cloud, or any combination thereof. This application does not impose any limitations on this.

[0058] Privacy computing node 100 can be used to perform privacy computing.

[0059] The central node 200 can also be an electronic device with computing and processing capabilities, such as a computer or server. For details, please refer to the description of the privacy computing node 100, which will not be repeated here.

[0060] The central node 200 can be used to evaluate the reputation of the privacy computing node 100. The specific process can be referred to in the privacy computing node reputation evaluation method provided in the following method embodiment, which will not be repeated here.

[0061] In some embodiments, the central node 200 can also be used to receive trusted domain joining application information sent by the privacy computing node 100. The trusted domain joining application information may include the identity information of the privacy computing node and attribute information specific to the privacy computing architecture (such as the deployed privacy computing framework, version, deployed services, and the catalog of data services that can be provided externally).

[0062] In this context, the trusted domain can be understood as a regional network composed of privacy computing nodes that have been issued Certificate Authority (CA) certificates by the central node 200 (Certificate Authority).

[0063] In some embodiments, the central node 200 can also, in response to a trusted domain joining application, generate an identity mapping ID based on the identity information of the privacy computing node and the attribute information of the privacy computing architecture using a hash algorithm, and generate a CA certificate, sending the identity mapping ID and CA certificate to the privacy computing node 100. The identity mapping ID and the CA integer issued by the central node are the basis for the privacy computing node 100 to perform interconnection tasks within the trusted domain, such as data interaction and joint modeling.

[0064] The entity executing the privacy computing node reputation evaluation method provided in this application embodiment is a reputation evaluation device. This reputation evaluation device can be the aforementioned central node 200. For example, as described above, the central node 200 can be an electronic device with computing processing capabilities, such as a computer or server. Optionally, the reputation evaluation device can also be a processor (e.g., a central processing unit, CPU) in the aforementioned electronic device; or, the reputation evaluation device can also be an application with reputation evaluation functionality installed in the aforementioned electronic device; or, the reputation evaluation device can also be a software system or platform in the aforementioned electronic device; or, the reputation evaluation device can also be a functional module in the aforementioned electronic device used to execute the privacy computing node reputation evaluation method, etc. This application embodiment does not impose any limitations on these aspects.

[0065] For the sake of simplicity, the following description will take the reputation evaluation device as the execution subject of the privacy computing node reputation evaluation method provided in the embodiments of this application as an example.

[0066] Figure 2 This is a flowchart illustrating the privacy computing node reputation evaluation method provided in an embodiment of this application. Figure 2 As shown, the method includes the following steps:

[0067] S101. Obtain the attribute reputation score, interaction reputation score, historical reputation score, and predicted reputation score of the privacy computing node to be evaluated.

[0068] Among them, the attribute reputation score is used to evaluate the number of types of provable attribute information provided by the privacy computing node. The provable attribute information is determined based on the attribute information directly provided by the privacy computing node to be evaluated and the attribute information of the privacy computing node to be evaluated provided by a trusted information source. The trusted information source may include the operation log of the privacy computing node to be evaluated, a trusted third-party organization, or a digital certificate issued by a trusted third-party organization that includes the attribute information of the privacy computing node to be evaluated. The specific types of trusted information sources are not limited in the embodiments of this application.

[0069] Interactive reputation scoring is used to evaluate the overall reputation score of interactive privacy computing nodes that have interactive relationships with the privacy computing node to be evaluated.

[0070] Historical reputation score is used to assess the overall historical reputation of the privacy computing node to be evaluated. Predicted reputation score is used to assess the behavioral characteristics of the privacy computing node to be evaluated.

[0071] The specific determination process of attribute reputation score, interaction reputation score, historical reputation score and predicted reputation score can be referred to in the following embodiments, and will not be repeated here.

[0072] S102. Based on attribute reputation score, interaction reputation score, historical reputation score and predicted reputation score, determine the comprehensive reputation score of the privacy computing node to be evaluated.

[0073] In one possible implementation, the reputation evaluation device can perform a weighted summation of attribute reputation scores, interaction reputation scores, historical reputation scores, and predicted reputation scores to obtain a comprehensive reputation score for the privacy computing node to be evaluated.

[0074] Among them, the weight corresponding to the attribute reputation score is greater than or equal to the weight corresponding to the predicted reputation score, the weight corresponding to the predicted reputation score is greater than the weight corresponding to the historical reputation score, and the weight corresponding to the historical reputation score is greater than or equal to the weight corresponding to the interaction reputation score.

[0075] For example, a credit rating device can calculate a comprehensive credit score using the following formula:

[0076] Overall Reputation Score = Weight 1 × Attribute Reputation Score + Weight 2 × Predicted Reputation Score + Weight 3 × Historical Reputation Score + Weight 4 × Interaction Reputation Score Formula (1)

[0077] In formula (1), weight1 + weight2 + weight3 + weight4 = 1, and weight1 ≥ weight2 > weight3 ≥ weight4.

[0078] In the privacy computing node reputation evaluation method provided in this application embodiment, the reputation evaluation device can obtain the attribute reputation score, interaction reputation score, historical reputation score and predicted reputation score of the privacy computing node to be evaluated, and determine the comprehensive reputation score of the privacy computing node to be evaluated based on the attribute reputation score, interaction reputation score, historical reputation score and predicted reputation score. It can comprehensively consider the reputation scores of multiple dimensions, thereby providing a more comprehensive and accurate reputation evaluation method to ensure the credibility of privacy computing nodes in heterogeneous environments.

[0079] The following describes the process of obtaining attribute reputation scores in S101 above.

[0080] In some possible embodiments, as described above, the trusted information source may include the runtime logs of the privacy computing node to be evaluated. In this case, the step of obtaining the attribute reputation score in S101 above may specifically include the following steps:

[0081] Step 1a: Receive attribute information sent from the privacy computing node to be evaluated.

[0082] The attribute information may include at least one of the following: type and version of privacy computing framework, catalog of data services provided to external parties, data encryption measures, and encryption algorithms.

[0083] For example, as described above, the reputation evaluation device (central node 200) can receive trusted domain joining application information sent from the privacy computing node 100, which may include the attribute information of the privacy computing node.

[0084] Step 2a: Obtain the running logs of the privacy computing node to be evaluated.

[0085] For example, the privacy computing framework used by the privacy computing node to be evaluated may have built-in log collection and management functions. The reputation evaluation device can communicate and interact with the module corresponding to the log collection and management functions to obtain the operating logs of the privacy computing node to be evaluated.

[0086] Step 3a: Determine provable attribute information from the runtime log.

[0087] Among them, the provable attribute information is the attribute information in the operation log that is the same as the attribute information sent by the privacy computing node to be evaluated.

[0088] For example, a reputation assessment device can identify attribute information that is identical to the running log in the attribute information directly sent by the privacy computing node to be evaluated as provable attribute information.

[0089] Step 4a: Determine the attribute reputation score based on the number of types of verifiable attribute information.

[0090] Among them, the attribute reputation score is positively correlated with the number of types of provable attribute information.

[0091] In one possible implementation, the reputation evaluation device can multiply the number of types of provable attribute information of the privacy computing node to be evaluated by a proportional coefficient to obtain the attribute reputation score of the privacy computing node to be evaluated.

[0092] The scaling factor is greater than 0. For example, the scaling factor can be set to 1, 10, or 20, etc. The embodiments of this application do not limit the specific value of the scaling factor.

[0093] In another possible implementation, the reputation evaluation device may have a pre-set correspondence between the number of types of provable attribute information and the attribute reputation score. The reputation device can determine the attribute reputation score of the privacy computing node to be evaluated based on the number of types of provable attribute information and the aforementioned correspondence.

[0094] For example, a reputation evaluation device can use the number of types of provable attribute information of the privacy computing node to be evaluated as an index to traverse and search the aforementioned correspondence, and use the attribute reputation score corresponding to the number of types of provable attribute information of the privacy computing node to be evaluated in the aforementioned correspondence as the attribute reputation score of the privacy computing node to be evaluated.

[0095] For example, the correspondence between the number of provable attribute information types and attribute reputation scores can be specifically shown in Table 1 below:

[0096] Table 1

[0097]

[0098] As shown in Table 1, this table includes a category for the number of types of provable attribute information and a category for attribute reputation scores. The category for the number of types of provable attribute information includes Quantity 1, Quantity 2, and Quantity 3. The category for attribute reputation scores includes Rating 1, Rating 2, and Rating 3. There is a correspondence between Quantity 1 and Rating 1; a correspondence between Quantity 2 and Rating 2; and a correspondence between Quantity 3 and Rating 3.

[0099] It should be understood that the reputation evaluation of privacy computing nodes in related technologies is mainly concentrated within a single privacy computing framework, lacking cross-framework interoperability. This means that privacy computing nodes between different privacy computing frameworks are difficult to achieve effective interconnection and interoperability.

[0100] In the privacy computing node reputation evaluation method provided in this application embodiment, the attribute information sent by the privacy computing node obtained by the reputation evaluation device may include the privacy computing framework and version deployed by the privacy computing node. In this way, the reputation evaluation device can evaluate the reputation of privacy computing nodes across privacy computing frameworks, which can improve the interoperability between different privacy computing frameworks and realize cross-framework interconnection.

[0101] The following describes the steps in S101 above for obtaining the interactive reputation score of the privacy computing node to be evaluated.

[0102] In some possible embodiments, the step of obtaining the interaction reputation score of the privacy computing node to be evaluated in S101 above may specifically include the following steps:

[0103] Step 1b: Weighted summation of the comprehensive reputation scores of interactive privacy computing nodes that have interactive relationships with the privacy computing node to be evaluated to obtain the interactive reputation score.

[0104] In this context, having an interactive relationship with the privacy computing node to be evaluated means being connected to the privacy computing node to be evaluated, or being connected to the privacy computing node to be evaluated and having an interaction history.

[0105] For example, the reputation evaluation device can have a pre-defined network topology in the trusted domain, and the reputation evaluation device can determine the interactive privacy computing nodes connected to the privacy computing nodes to be evaluated based on the network topology.

[0106] For example, a reputation evaluation device can obtain the operation logs of the privacy computing node to be evaluated, and then determine the interactive privacy computing nodes that have an interaction history with the privacy computing node to be evaluated based on the operation logs.

[0107] Optionally, the reputation evaluation device may specifically perform a weighted summation of the comprehensive reputation scores of interactive privacy computing nodes that have an interactive relationship with the privacy computing node to be evaluated, according to the following formula (2):

[0108]

[0109] In formula (2), k represents the number of all interactive privacy computing nodes that have an interactive relationship with the privacy computing node to be evaluated; comprehensive reputation score i This represents the overall reputation score of the i-th interactive privacy computing node that has an interaction relationship with the privacy computing node being evaluated. Weight i This represents the weight of the i-th interactive privacy computing node that has an interactive relationship with the privacy computing node to be evaluated.

[0110] Optionally, the weights corresponding to the interactive privacy computing nodes (i.e., the weights mentioned above) i The value is determined based on the reputation weight and connection weight of the interactive privacy computing node.

[0111] Among them, the reputation weight is positively correlated with the degree of scoring deviation. The degree of scoring deviation is the degree of deviation between the comprehensive reputation score of the corresponding interactive privacy computing node and the average reputation score. The average reputation score is the average of the comprehensive reputation scores of all privacy computing nodes in the trusted domain where the privacy computing node to be evaluated is located.

[0112] The connection weight is positively correlated with the number of privacy computing nodes connected to the corresponding interactive privacy computing node.

[0113] As an example, weight i It can be calculated using the following formulas (3) to (5):

[0114] Weight i = (Reputation Weight) i +Connection weight i ) / 2 formula (3)

[0115] In formula (3), reputation weight i This represents the reputation weight and connection weight corresponding to the i-th interactive privacy computing node that has an interaction relationship with the privacy computing node to be evaluated. i This represents the connection weight corresponding to the i-th interactive privacy computing node that has an interactive relationship with the privacy computing node to be evaluated.

[0116]

[0117] In formula (4), the comprehensive credit score i This represents the overall reputation score of the i-th interactive privacy computing node that has an interactive relationship with the privacy computing node to be evaluated.

[0118]

[0119] In formula (5), the number of connected nodes i This represents the number of privacy computing nodes connected to the i-th interactive privacy computing node that has an interaction relationship with the privacy computing node to be evaluated. The total number of node connections represents the sum of the number of privacy computing nodes connected to each privacy computing node in the trusted domain.

[0120] For example, taking a trusted domain that includes privacy computing node A, privacy computing node B, and privacy computing node C as an example, assuming that privacy computing node A is connected to 2 other privacy computing nodes, privacy computing node B is connected to 2 other privacy computing nodes, and privacy computing node C is connected to 2 other privacy computing nodes, then the total number of node connections is 2 + 2 + 2 = 6.

[0121] The following describes the steps in S101 above for obtaining the historical reputation score of the privacy computing node to be evaluated.

[0122] In some possible embodiments, the step of obtaining the historical reputation score of the privacy computing node to be evaluated in S101 above may specifically include the following steps:

[0123] Step 1c: Exponentially decay the historical comprehensive reputation score of the privacy computing node to be evaluated to obtain the historical reputation score of the privacy computing node to be evaluated.

[0124] Alternatively, the privacy evaluation device can be exponentially decayed according to the following formula (6):

[0125]

[0126] Here, Adjusted Score represents the historical credit score. Score j denoted as the historical comprehensive credit score at the j-th evaluation. λ represents the decay factor, which is positively correlated with the decay rate. j This represents the unit time difference between the j-th evaluation and the current time. The unit time difference can be in units of 1 day, 1 week, or 1 month; however, this application embodiment does not limit the specific unit of the unit time difference.

[0127] For example, taking the comprehensive reputation score of the privacy computing node to be evaluated as 90 in January, 85 in February, 92 in March, and 88 in April as an example, assuming that the unit time difference is one month, the decay factor is 0.1, and the current time is a certain point in April, then referring to the above formula (6) and specifically substituting the values, we can obtain the following formulas (7) and (8): Adjusted Score=90×e -0.1×3 +85×e -0.1×2 +92×e -0.1×1 +88×e -0.1×0

[0128] Formula (7)

[0129] Adjusted Score = 90 × e -0.3 +85×e -0.2 +92×e -0.1 +88 formula(8)

[0130] The following describes the steps in S101 above for obtaining the predicted reputation score of the privacy computing node to be evaluated.

[0131] In some possible embodiments, the step of obtaining the predicted reputation score of the privacy computing node to be evaluated in S101 above may specifically include the following steps:

[0132] Step 1d: Obtain behavioral characteristic data of the privacy computing node to be evaluated.

[0133] The behavioral characteristic data includes at least one of the following: node activity (or frequency of participation in privacy computing tasks), task contribution, node response speed to requests, number of successfully completed tasks, number of failed tasks, running logs, number of trusted algorithms on the node, and amount of data transmitted by the node.

[0134] Step 2d: Determine the probability of abnormal behavior of the privacy computing node to be evaluated based on behavioral feature data and reputation prediction model.

[0135] The reputation prediction model is used to predict the probability of abnormal behavior based on the behavioral characteristic data of privacy computing nodes. For example, the reputation prediction model can be used to predict the current probability of abnormal behavior based on the current behavioral characteristic data of privacy computing nodes, or it can be used to predict the probability of future abnormal behavior based on the current behavioral characteristics of privacy computing nodes. This application embodiment does not limit this. The specific training process of the reputation prediction model can be referred to in the following embodiments, and will not be repeated here.

[0136] Step 3d: Determine the predicted reputation score of the privacy computing node to be evaluated based on the probability of abnormal behavior of the node to be evaluated.

[0137] Optionally, the reputation evaluation device can specifically calculate the predicted reputation score of the privacy computing node to be evaluated according to the following formula (9):

[0138] Predicted reputation score = (1 - probability of abnormal behavior) × the most recent historical comprehensive reputation score of the privacy computing node to be evaluated (Formula 9)

[0139] In some embodiments, prior to step 1d above, the reputation evaluation device may also acquire the trained reputation prediction model.

[0140] In one possible implementation, the reputation evaluation device can directly obtain the trained reputation prediction model from other devices.

[0141] For example, a reputation assessment device can obtain a trained reputation prediction model from other devices by downloading or transferring it to an intermediate storage medium.

[0142] In another possible implementation, the reputation evaluation device can also train a reputation prediction model using training samples. In this case, prior to step 1d above, the method may further include the following steps:

[0143] Step 1e: Obtain the initial training sample set; the initial training sample set includes a first proportion of positive samples and a second proportion of unlabeled samples.

[0144] Each positive sample includes behavioral feature data of a privacy computing node and an abnormal behavior probability label for that privacy computing node. The abnormal behavior probability label of the positive sample is higher than a probability threshold. The probability threshold can be preset in the reputation evaluation device; for example, the probability threshold can be set to 50%, 60%, or 70%. This application embodiment does not limit the specific value of the probability threshold. Each unlabeled sample includes behavioral feature data of a privacy computing node. The second ratio is higher than the first ratio.

[0145] For example, the initial training sample set may include 20% positive samples and 80% unlabeled samples; or 30% positive samples and 70% unlabeled samples, etc. This application does not limit the specific values ​​of the first and second proportions in its embodiments.

[0146] As an example, each positive or unlabeled sample is associated with an identity mapping ID of a privacy computing node.

[0147] Step 2e: Use the Naive Bayes model to label the unlabeled samples with probability labels of abnormal behavior, and classify the unlabeled samples into positive samples and negative samples.

[0148] Among them, the abnormal behavior probability label of negative samples is lower than the probability threshold.

[0149] Step 3e: Based on the positive samples in the initial sample set and the positive and negative samples labeled by the Naive Bayes model, train the initial model to obtain the reputation prediction model.

[0150] Optionally, the reputation evaluation device can input one or more positive or negative samples into the initial model each time to obtain the predicted probability of abnormal behavior output by the initial model. Then, Pipeline is used to realize the serial pipelined processing of parameter optimization, and GridSearch is used to realize the free combination of multiple parameter indicators in the parameter optimization process to automatically select the optimal parameters.

[0151] Alternatively, the credit rating device can first train the initial model until the training stopping condition is met, and then use a validation sample set to validate the initial model to obtain a credit prediction model. In this case, step 3e above can specifically include the following steps:

[0152] Step 3.1e: Train the initial model based on the positive samples in the initial sample set and the positive and negative samples labeled by the Naive Bayes model until the training stops.

[0153] The training stopping conditions may include: the number of times the reputation evaluation device inputs training samples (positive or negative samples) into the initial model reaches a threshold, and / or the error between the predicted abnormal behavior probability of the initial model and the abnormal behavior probability label of the training samples (positive or negative samples) is less than an error threshold.

[0154] The frequency threshold can be preset in the credit evaluation device. For example, the frequency threshold can be set to 500 times, 1000 times, or 10000 times. This application embodiment does not limit the specific value of the frequency threshold. The error threshold can also be preset in the credit evaluation device. For example, the error threshold can be set to 5%, 10%, or 15%. This application embodiment does not limit the specific value of the error threshold.

[0155] Step 3.2e: Validate the initial model based on the validation sample set to obtain the prediction results of the initial model.

[0156] The validation sample set includes multiple positive samples.

[0157] Step 3.3e: Calculate the validation score of the initial model based on the prediction results of the initial model.

[0158] As an example, a reputation rating device can use precision, recall, accuracy, or F1 score as a verification score.

[0159] As another example, the credit rating device can calculate the validation score of the initial model according to the following formula (10):

[0160]

[0161] In formula (10), S represents the validation score; r represents the recall rate; and Pr[f(x)=1] represents the probability that a sample in the validation set is identified as a positive sample based on the probability of abnormal behavior predicted by the initial model.

[0162] Step 3.4e: If the verification score is higher than the score threshold, the initial model is determined as the reputation prediction model.

[0163] Based on the understanding of the above embodiments, Figure 3 This is a schematic diagram of the credit assessment process provided in an embodiment of this application. Figure 3 As shown, the central node can send data to the privacy computing node (…). Figure 3 (Taking privacy computing nodes A, B, and C as an example, a total of three privacy computing nodes are issued CA certificates.)

[0164] Taking privacy computing node A as an example, the central node can determine the attribute reputation score based on the types and quantities of provable attribute information of privacy computing node A, determine the comprehensive interaction reputation score based on the comprehensive reputation scores of privacy computing nodes B and C, obtain the interaction reputation score of privacy computing node A, and then determine the historical comprehensive reputation score of privacy computing node A. Figure 3 (Using historical reputation score 1, historical reputation score 2, historical reputation score 3, etc. as examples, the historical reputation score is determined, and then the scores are statistically analyzed to determine the historical reputation score.) The reputation evaluation device can obtain a predicted reputation score based on multiple behavioral feature data (of privacy computing node A) using a semi-supervised classification model. Finally, the reputation evaluation device can obtain a comprehensive reputation score for privacy computing node A based on attribute reputation score, interaction reputation score, historical reputation score, and predicted reputation score.

[0165] The foregoing primarily describes the solutions provided by the embodiments of this application from a methodological perspective. To achieve the aforementioned functions, it includes corresponding hardware structures and / or software modules for executing each function. Those skilled in the art should readily recognize that, in conjunction with the units and algorithm steps of the various examples described in the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0166] In an exemplary embodiment, this application also provides a reputation evaluation device. Figure 4 A schematic diagram illustrating the composition of the credit evaluation device provided in the embodiments of this application. Figure 4 As shown, the device includes an acquisition module 401 and a processing module 402.

[0167] The acquisition module 401 is used to acquire the attribute reputation score, interaction reputation score, historical reputation score and predicted reputation score of the privacy computing node to be evaluated;

[0168] Among them, the attribute reputation score is used to evaluate the number and types of provable attribute information provided by the privacy computing node; the provable attribute information is determined based on the attribute information directly provided by the privacy computing node to be evaluated and the attribute information of the privacy computing node to be evaluated provided by a trusted information source; the interaction reputation score is used to evaluate the comprehensive reputation score of interactive privacy computing nodes that have an interactive relationship with the privacy computing node to be evaluated; the historical reputation score is used to evaluate the historical comprehensive reputation score of the privacy computing node to be evaluated; and the predictive reputation score is used to evaluate the behavioral characteristics of the privacy computing node to be evaluated.

[0169] The processing module 402 is used to determine the comprehensive reputation score of the privacy computing node to be evaluated based on attribute reputation score, interaction reputation score, historical reputation score and predicted reputation score.

[0170] In some possible embodiments, the trusted information source includes the operation log of the privacy computing node to be evaluated; the acquisition module 401 is specifically used to receive attribute information sent from the privacy computing node to be evaluated; acquire the operation log of the privacy computing node to be evaluated; determine provable attribute information from the operation log; wherein, provable attribute information is attribute information in the operation log that is the same as the attribute information sent by the privacy computing node to be evaluated; determine an attribute reputation score based on the number of types of provable attribute information; the attribute reputation score is positively correlated with the number of types of provable attribute information.

[0171] In other possible embodiments, the acquisition module 401 is specifically used to perform a weighted summation of the comprehensive reputation scores of interactive privacy computing nodes that have an interactive relationship with the privacy computing node to be evaluated, so as to obtain an interactive reputation score.

[0172] In some other possible embodiments, the weights corresponding to interactive privacy computing nodes are determined based on the reputation weights and connection weights of the interactive privacy computing nodes; the reputation weights are positively correlated with the degree of scoring deviation; the degree of scoring deviation is the degree of deviation between the overall reputation score and the average reputation score of the corresponding interactive privacy computing node; the average reputation score is the average of the overall reputation scores of all privacy computing nodes in the trusted domain where the privacy computing node to be evaluated is located; the trusted domain is a regional network composed of privacy computing nodes that have been issued CA certificates; and the connection weights are positively correlated with the number of privacy computing nodes connected to the corresponding interactive privacy computing node.

[0173] In some other possible embodiments, the weights corresponding to the interactive privacy computing nodes are calculated using the following formula:

[0174] Weight i = (Reputation Weight) i +Connection weight i ) / 2;

[0175] Among them, weight i Represents the weight corresponding to the i-th interactive privacy computing node that has an interaction relationship with the privacy computing node to be evaluated; reputation weight. i This represents the reputation weight and connection weight corresponding to the i-th interactive privacy computing node that has an interaction relationship with the privacy computing node to be evaluated. i This represents the connection weight corresponding to the i-th interactive privacy computing node that has an interactive relationship with the privacy computing node to be evaluated.

[0176]

[0177] Among them, the comprehensive credit score i The comprehensive reputation score represents the i-th interactive privacy computing node that has an interactive relationship with the privacy computing node to be evaluated.

[0178]

[0179] Among them, the number of connected nodes i The number of privacy computing nodes connected to the i-th interactive privacy computing node that has an interactive relationship with the privacy computing node to be evaluated; the total number of node connections represents the sum of the number of privacy computing nodes connected to each privacy computing node in the trusted domain.

[0180] In some other possible embodiments, the acquisition module 401 is specifically used to exponentially decay the historical comprehensive reputation score of the privacy computing node to be evaluated, thereby obtaining the historical reputation score of the privacy computing node to be evaluated.

[0181] In some other possible embodiments, the acquisition module 401 is specifically used to exponentially decay the historical comprehensive reputation score of the privacy computing node to be evaluated, thereby obtaining the historical reputation score of the privacy computing node to be evaluated, including:

[0182] Exponential decay is performed according to the following formula:

[0183]

[0184] Here, Adjusted Score represents historical credit score; Score j t represents the historical comprehensive credit score at the j-th evaluation; λ represents the decay factor, which is positively correlated with the decay rate; j This represents the unit time difference between the j-th evaluation and the current time.

[0185] In some other possible embodiments, the acquisition module 401 is specifically used to acquire behavioral feature data of the privacy computing node to be evaluated; determine the probability of abnormal behavior of the privacy computing node to be evaluated based on the behavioral feature data and the reputation prediction model; the reputation prediction model is used to predict the probability of abnormal behavior based on the behavioral feature data of the privacy computing node; and determine the predicted reputation score of the privacy computing node to be evaluated based on the probability of abnormal behavior.

[0186] In some other possible embodiments, the acquisition module 401 is specifically used to calculate the predicted reputation score of the privacy computing node to be evaluated according to the following formula:

[0187] Predicted reputation score = (1 - probability of abnormal behavior) × the most recent historical comprehensive reputation score of the privacy computing node to be evaluated.

[0188] In some other possible embodiments, the acquisition module 401 is further configured to acquire an initial training sample set; the initial training sample set includes a first proportion of positive samples and a second proportion of unlabeled samples; wherein each positive sample includes behavioral feature data of a privacy computing node and an abnormal behavior probability label of the privacy computing node; the abnormal behavior probability label of the positive sample is higher than a probability threshold; each unlabeled sample includes behavioral feature data of a privacy computing node; the second proportion is higher than the first proportion;

[0189] The processing module 402 is also used to label unlabeled samples with abnormal behavior probability labels using a Naive Bayes model, classifying unlabeled samples into positive samples and negative samples; the abnormal behavior probability label of negative samples is lower than the probability threshold; and the initial model is trained based on the positive samples in the initial sample set and the positive and negative samples labeled by the Naive Bayes model to obtain a reputation prediction model.

[0190] In some other possible embodiments, the processing module 402 is specifically used to train the initial model based on the positive samples in the initial sample set and the positive and negative samples labeled by the Naive Bayes model until the training stops; to validate the initial model based on the validation sample set and obtain the prediction result of the initial model; the validation sample set includes multiple positive samples; to calculate the validation score of the initial model based on the prediction result of the initial model; and to determine the initial model as a reputation prediction model if the validation score is higher than the score threshold.

[0191] In some other possible embodiments, processing module 402 is specifically used to calculate the validation score of the initial model according to the following formula:

[0192]

[0193] Where S represents the validation score; r represents the recall rate; and Pr[f(x)=1] represents the probability that a sample in the validation set is identified as a positive sample based on the probability of abnormal behavior predicted by the initial model.

[0194] In some other possible embodiments, the processing module 402 is specifically used to perform a weighted summation of the attribute reputation score, the interaction reputation score, the historical reputation score, and the predicted reputation score to obtain the comprehensive reputation score of the privacy computing node to be evaluated; wherein, the weight corresponding to the attribute reputation score is greater than or equal to the weight corresponding to the predicted reputation score, the weight corresponding to the predicted reputation score is greater than the weight corresponding to the historical reputation score, and the weight corresponding to the historical reputation score is greater than or equal to the weight corresponding to the interaction reputation score.

[0195] In an exemplary embodiment, this application also provides an electronic device, and the aforementioned credit evaluation device can be applied to the aforementioned electronic device. Figure 5 This is a schematic diagram illustrating the composition of an electronic device provided in an embodiment of this application. For example... Figure 5 As shown, the electronic device includes:

[0196] The processor 10, memory 20, communication line 30, communication interface 40, and input / output interface 50 are included.

[0197] The processor 10, memory 20, communication interface 40, and input / output interface 50 can be connected via communication line 30.

[0198] Processor 10 is used to execute instructions stored in memory 20 to implement the privacy computing node reputation evaluation method provided in the above embodiments of this application. Processor 10 can be a CPU, a network processor (NP), a digital signal processor (DSP), a microprocessor, a microcontroller (MCU) / single-chip microcomputer / microcontroller, a programmable logic device (PLD), or any combination thereof. Processor 10 can also be any other device with processing capabilities, such as a circuit, device, or software module; this application embodiment does not limit this. In one example, processor 10 may include one or more CPUs, for example... Figure 5 CPU0 and CPU1 are mentioned. As an optional implementation, the electronic device may include multiple processors; for example, in addition to processor 10, it may also include processor 60. Figure 5 (The example shown is a dashed line).

[0199] The memory 20 is used to store instructions. For example, the instructions may be computer programs. Optionally, the memory 20 may be a read-only memory (ROM) or other types of static storage devices that can store static information and / or instructions; it may also be a random access memory (RAM) or other types of dynamic storage devices that can store information and / or instructions; it may also be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media, or other magnetic storage devices, etc. The embodiments of this application do not limit this.

[0200] It should be noted that the memory 20 can exist independently of the processor 10 or it can be integrated with the processor 10. The memory 20 can be located inside or outside the electronic device, and this application embodiment does not impose any restrictions on this.

[0201] Communication line 30 is used to transmit information between the components included in the electronic device.

[0202] Communication interface 40 is used to communicate with other devices or other communication networks. These other communication networks can be Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc. Communication interface 40 can be a module, circuit, transceiver, or any device capable of enabling communication.

[0203] Input / output interface 50 is used to enable human-computer interaction between users and electronic devices. For example, it enables action interaction or information exchange between users and electronic devices.

[0204] For example, the input / output interface 50 can be a mouse, keyboard, display screen, or touch screen. Action or information interaction between the user and the electronic device can be achieved through a mouse, keyboard, display screen, or touch screen.

[0205] It should be noted that, Figure 5 The structures shown do not constitute a limitation on electronic devices, except... Figure 5 In addition to the components shown, electronic devices may include more or fewer components than illustrated, or combinations of certain components, or different component arrangements.

[0206] In an exemplary embodiment, this application also provides a computer program product including computer instructions that, when executed in an electronic device, cause the electronic device to implement the methods described in the foregoing method embodiments.

[0207] In an exemplary embodiment, this application also provides a readable storage medium including software instructions that, when executed in an electronic device, cause the electronic device to implement the methods described in the foregoing method embodiments. The readable storage medium may be a non-transitory computer-readable storage medium, such as a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device.

[0208] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented using software programs, implementation can be, in whole or in part, in the form of a computer program product. This computer program product includes one or more computer-executable instructions. When these computer-executable instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer-executable instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer-executable instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means.

[0209] Although this application has been described herein in conjunction with various embodiments, those skilled in the art, by reviewing the accompanying drawings, disclosure, and appended claims, will understand and implement other variations of the disclosed embodiments in carrying out the claimed application. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude multiple components. A single processor or other unit can implement several functions listed in the claims. While different dependent claims may recite certain measures, this does not mean that these measures cannot be combined to produce good results.

[0210] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely exemplary illustrations of this application as defined by the appended claims, and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from the spirit and scope of this application. Thus, if such modifications and modifications of this application fall within the scope of the claims of this application and their equivalents, this application is also intended to include such modifications and modifications.

[0211] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for evaluating the reputation of privacy-preserving computing nodes, characterized in that, The method includes: Obtain the attribute reputation score, interaction reputation score, historical reputation score, and predicted reputation score of the privacy computing node to be evaluated; The attribute reputation score is used to assess the number and types of provable attribute information provided by the privacy computing node to be evaluated. The provable attribute information is determined based on attribute information directly provided by the privacy computing node to be evaluated and attribute information provided by trusted information sources. The interaction reputation score is used to assess the comprehensive reputation score of interactive privacy computing nodes that have an interactive relationship with the privacy computing node to be evaluated. The historical reputation score is used to assess the historical comprehensive reputation score of the privacy computing node to be evaluated. The predicted reputation score is used to assess the behavioral characteristics of the privacy computing node to be evaluated. Based on the attribute reputation score, the interaction reputation score, the historical reputation score, and the predicted reputation score, the comprehensive reputation score of the privacy computing node to be evaluated is determined.

2. The method according to claim 1, characterized in that, The trusted information source includes the operation logs of the privacy computing node to be evaluated; obtaining the attribute reputation score of the privacy computing node to be evaluated includes: Receive attribute information sent from the privacy computing node to be evaluated; Obtain the operation logs of the privacy computing node to be evaluated; The provable attribute information is determined from the operation log; Wherein, the provable attribute information is the same attribute information in the operation log that is sent by the privacy computing node to be evaluated; The attribute reputation score is determined based on the number of types of provable attribute information; the attribute reputation score is positively correlated with the number of types of provable attribute information.

3. The method according to claim 1, characterized in that, The process of obtaining the interaction reputation score of the privacy computing node to be evaluated includes: The interaction reputation score is obtained by weighted summing of the comprehensive reputation scores of the interactive privacy computing nodes that have interaction relationships with the privacy computing node to be evaluated according to the following formula: ； in, This represents the number of all interactive privacy computing nodes that have an interaction relationship with the privacy computing node to be evaluated. The first node that interacts with the privacy-preserving computation node to be evaluated. A comprehensive reputation score for each interactive privacy computing node; The first node that interacts with the privacy-preserving computation node to be evaluated. The weights corresponding to each interactive privacy computing node.

4. The method according to claim 3, characterized in that, The weight corresponding to the interactive privacy computing node is determined based on the reputation weight and the connection weight corresponding to the interactive privacy computing node. The reputation weight is positively correlated with the degree of scoring deviation; the degree of scoring deviation is the degree of deviation between the comprehensive reputation score and the average reputation score of the corresponding interactive privacy computing node. The average reputation score is the average of the comprehensive reputation scores of all privacy computing nodes in the trusted domain where the privacy computing node to be evaluated is located; the trusted domain is a regional network composed of privacy computing nodes that have been issued CA certificates. The connection weight is positively correlated with the number of privacy computing nodes connected to the corresponding interactive privacy computing node.

5. The method according to claim 4, characterized in that, The weights corresponding to the interactive privacy computing nodes are calculated using the following formula: ； in, The first node that interacts with the privacy-preserving computation node to be evaluated. The weights corresponding to each interactive privacy computing node; The first node that interacts with the privacy-preserving computation node to be evaluated. The reputation weight corresponding to each interactive privacy computing node The first node that interacts with the privacy-preserving computation node to be evaluated. The connection weights corresponding to each interactive privacy computing node; ； in, The first node that interacts with the privacy-preserving computation node to be evaluated. A comprehensive reputation score for each interactive privacy computing node; ； in, The first node that interacts with the privacy-preserving computation node to be evaluated. The number of privacy computing nodes connected to each interactive privacy computing node; This represents the total number of privacy computing nodes connected to each privacy computing node in the trusted domain.

6. The method according to claim 1, characterized in that, The process of obtaining the historical reputation score of the privacy computing node to be evaluated includes: The historical comprehensive reputation score of the privacy computing node to be evaluated is obtained by exponentially decaying the historical reputation score of the privacy computing node to be evaluated.

7. The method according to claim 6, characterized in that, The process of exponentially decaying the historical comprehensive reputation score of the privacy computing node to be evaluated to obtain the historical reputation score of the privacy computing node to be evaluated includes: Exponential decay is performed according to the following formula: ; in, This refers to the historical reputation score; Indicates the first Historical comprehensive credit score at the time of this evaluation; Indicates the attenuation factor. Positively correlated with decay rate; Indicates the first The unit time difference between the current evaluation and the present moment.

8. The method according to claim 1, characterized in that, The process of obtaining the predicted reputation score of the privacy computing node to be evaluated includes: Obtain behavioral characteristic data of the privacy computing nodes to be evaluated; Based on the behavioral feature data and the reputation prediction model, the probability of abnormal behavior of the privacy computing node to be evaluated is determined; the reputation prediction model is used to predict the probability of abnormal behavior based on the behavioral feature data of the privacy computing node. The predicted reputation score of the privacy computing node to be evaluated is determined based on the probability of abnormal behavior of the node.

9. The method according to claim 8, characterized in that, The step of determining the predicted reputation score of the privacy computing node to be evaluated based on the probability of abnormal behavior of the node to be evaluated includes: The predicted reputation score of the privacy-preserving computation node to be evaluated is calculated using the following formula: Predicted reputation score = (1 - probability of abnormal behavior) × the most recent historical comprehensive reputation score of the privacy computing node to be evaluated.

10. The method according to claim 8, characterized in that, The reputation prediction model is trained in the following way: Obtain an initial training sample set; the initial training sample set includes a first proportion of positive samples and a second proportion of unlabeled samples; Each positive sample includes behavioral feature data of a privacy computing node and an abnormal behavior probability label of that privacy computing node; the abnormal behavior probability label of the positive sample is higher than a probability threshold; each unlabeled sample includes behavioral feature data of a privacy computing node; the second ratio is higher than the first ratio; The unlabeled samples are labeled with probability labels for abnormal behavior using a Naive Bayes model, and the unlabeled samples are classified into positive samples and negative samples; the probability label for abnormal behavior of the negative samples is lower than the probability threshold. Based on the positive samples in the initial training sample set and the positive and negative samples labeled by the Naive Bayes model, the initial model is trained to obtain the reputation prediction model.

11. The method according to claim 10, characterized in that, The reputation prediction model is obtained by training the initial model with positive samples from the initial sample set and positive and negative samples labeled by the Naive Bayes model, including: Based on the positive samples in the initial sample set and the positive and negative samples labeled by the Naive Bayes model, the initial model is trained until the training stops. The initial model is validated based on the validation sample set to obtain the prediction results of the initial model; the validation sample set includes multiple positive samples. Based on the prediction results of the initial model, the validation score of the initial model is calculated; If the verification score is higher than the score threshold, the initial model is determined as the reputation prediction model.

12. The method according to claim 11, characterized in that, The calculation of the validation score of the initial model based on the prediction results of the initial model includes: The validation score of the initial model is calculated using the following formula: ; in, This represents the verification score; Indicates recall rate; This represents the probability that a sample in the validation set is determined to be a positive sample based on the probability of abnormal behavior predicted by the initial model.

13. The method according to any one of claims 1-12, characterized in that, The determination of the comprehensive reputation score of the privacy computing node to be evaluated based on the attribute reputation score, the interaction reputation score, the historical reputation score, and the predicted reputation score includes: The attribute reputation score, the interaction reputation score, the historical reputation score, and the predicted reputation score are weighted and summed to obtain the comprehensive reputation score of the privacy computing node to be evaluated. Wherein, the weight corresponding to the attribute reputation score is greater than or equal to the weight corresponding to the predicted reputation score, the weight corresponding to the predicted reputation score is greater than the weight corresponding to the historical reputation score, and the weight corresponding to the historical reputation score is greater than or equal to the weight corresponding to the interaction reputation score.

14. An electronic device, characterized in that, The electronic device includes: a processor and a memory; The memory stores instructions that the processor can execute; When the processor is configured to execute the instructions, it causes the electronic device to implement the method as described in any one of claims 1-13.

15. A readable storage medium, characterized in that, The readable storage medium includes: software instructions; When the software instructions are executed in an electronic device, the electronic device causes the electronic device to perform the method as described in any one of claims 1-13.

16. A computer program product, characterized in that, The computer program product includes: computer instructions; When the computer instructions are executed in an electronic device, the electronic device causes the electronic device to perform the method as described in any one of claims 1-13.

Images