Trust cascade evolutionary game governance method, system, electronic device and storage medium

Abstract

This invention discloses a trust-cascading evolutionary game governance method, system, electronic device, and storage medium. The method includes: evolving an evolutionary game model based on multi-source situational data provided by participating entities across the entire chain to determine a first governance strategy for each participating entity; determining technological trust levels, organizational trust levels, and ecological trust levels based on multi-source situational data and a trust assessment model; the trust assessment model describes the constraints between technological trust levels, organizational trust levels, and ecological trust levels; obtaining target constraint parameters based on multi-source situational data and a pre-set legal mapping rule base; and optimizing the first governance strategy based on the technological trust level, organizational trust level, ecological trust level, and target constraint parameters to obtain a second governance strategy, which is then used to govern the participating entities. This invention addresses the problems of fragmented trust research, simplified game models, and the disconnect between law and technology.

Description

Technical Field

[0001] This invention relates to the field of data-computing-model governance technology, and in particular to a trust-cascaded evolutionary game governance method, system, electronic device and storage medium. Background Technology

[0002] The digital economy has entered a new stage of deep coupling between data, computing power, and models. The collaboration between data, computing power, and models has evolved from a simple aggregation of resources into a socio-technical system problem involving multiple dimensions such as technology, organization, institutions, and stakeholders. Simultaneously, with the trillion-level leap in the scale of large-scale models, cross-domain collaboration is no longer a purely technical issue, but a complex systems engineering project involving cross-organizational trust transfer, multi-stakeholder interest games, institutional compliance constraints, and agile governance simulations. Existing research often proceeds in segments along the lines of data flow, computing network coordination, and model application, leading to fragmented trust research, simplified game theory models, and a disconnect between law and technology. Summary of the Invention

[0003] This invention provides a trust-cascaded evolutionary game governance method, system, electronic device, and storage medium to address the problems of fragmented trust research, simplified game models, and the disconnect between law and technology.

[0004] According to one aspect of the present invention, a trust-cascading evolutionary game governance method is provided, the method comprising: Acquire multi-source situational data, which is relevant data provided by the participating entities across the entire chain of the data computing and modeling collaborative project; the participating entities include regulators, data computing and modeling platforms, data providers, algorithm model developers, and end users; The evolutionary game model is evolved based on the multi-source situational data to determine the first governance strategy for each participating entity. The strategy learning process of the evolutionary game model is determined based on strategy revenue, compliance cost, expected penalty, and risk loss. The strategy revenue is constituted by the strategy space of each participating entity, which describes the relationship between the governance strategies of the participating entities and their corresponding revenues. The compliance cost is determined by the compliance cost coefficient of each participating entity, which measures the economic cost of the participating entity's compliance investment. The expected penalty is determined by the probability of detection by algorithmic audit and the intensity of institutional penalties. The probability of detection by algorithmic audit reflects the regulatory authority's regulatory technical capabilities and the frequency of spot checks, while the intensity of institutional penalties describes the cost of violations by data, algorithms, and models. The risk loss is used to quantify the degree of risk of model failure. Based on the aforementioned multi-source situational data and trust assessment model, the levels of technological trust, organizational trust, and ecological trust are determined. The trust assessment model describes the constraints among these three levels. The technological trust level is determined based on data credibility, algorithm interpretability, model evaluability, and computing power reliability. The organizational trust level is determined based on the reputation values ​​of all participating entities, contract fulfillment among them, and the matching of rights and responsibilities. The ecological trust level is determined based on social network analysis and structural hole theory. Social network analysis describes the interaction structure among the participating entities. Based on the multi-source situational data and the preset legal mapping rule library, target constraint parameters are obtained; the preset legal mapping rule library is used to store the mapping rules between legal norms and technical constraint parameters in the scenarios corresponding to the multi-source situational data. Based on the technological trust level, the organizational trust level, the ecological trust level, and the target constraint parameters, the first governance strategy is optimized to obtain a second governance strategy, which is then used to govern the participating entities.

[0005] According to another aspect of the present invention, a trust cascade evolutionary game governance system is provided, which executes the trust cascade evolutionary game governance method. The system includes: a multi-source situational data acquisition module, a trust cascade calculation module, an institutional parameterized evolutionary game simulation module, a compliance mapping module, and a strategy optimization module. The multi-source situational data acquisition module is used to acquire multi-source situational data, which is relevant data provided by the participating entities across the entire chain of the data computing and model collaboration project; the participating entities include regulators, data computing and model platforms, data providers, algorithm model developers, and end users; The parameterized evolutionary game simulation module is used to evolve the evolutionary game model based on the multi-source situational data to determine the first governance strategy for each participating entity. The strategy learning process of the evolutionary game model is determined based on strategy revenue, compliance cost, expected penalty, and risk loss. The strategy revenue is composed of the strategy space of each participating entity, which describes the relationship between the governance strategies of the participating entities and their corresponding revenues. The compliance cost is determined by the compliance cost coefficient of each participating entity, which measures the economic cost of the participating entity's compliance investment. The expected penalty is determined by the probability of detection by algorithmic audit and the intensity of institutional penalties. The probability of detection by algorithmic audit reflects the regulatory authority's regulatory technical capabilities and the frequency of spot checks, while the intensity of institutional penalties describes the cost of violations by data, algorithms, and models. The risk loss is used to quantify the degree of risk of model failure. The trust cascade measurement module is used to determine the levels of technological trust, organizational trust, and ecological trust based on the multi-source situational data and the trust assessment model. The trust assessment model describes the constraints between these three levels. The technological trust level is determined based on data credibility, algorithm interpretability, model evaluability, and computing power reliability. The organizational trust level is determined based on the reputation values ​​of all participating entities, contract fulfillment, and the matching of rights and responsibilities among them. The ecological trust level is determined based on social network analysis and structural hole theory. Social network analysis describes the interaction structure among the participating entities. The compliance mapping module is used to obtain target constraint parameters based on the multi-source situational data and the preset legal mapping rule library; the preset legal mapping rule library is used to store the mapping rules between legal norms and technical constraint parameters in the scenarios corresponding to the multi-source situational data. The strategy optimization module is used to optimize the first governance strategy based on the technology trust level, the organization trust level, the ecosystem trust level and the target constraint parameters to obtain a second governance strategy, so as to use the second governance strategy to govern the participating entities.

[0006] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to execute the trust cascading evolution game governance method according to any embodiment of the present invention.

[0007] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions, the computer instructions being configured to cause a processor to execute and implement the trust cascading evolution game governance method described in any embodiment of the present invention.

[0008] The technical solution of this invention involves acquiring multi-source situational data, which is relevant data provided by all participating entities across the entire chain. These entities include regulators, data computing and modeling platforms, data providers, algorithm model developers, and end-users. The introduction of multi-party data ensures a more accurate reflection of the impact of each participating entity in the collaborative data computing and modeling project, as well as the influence between them. Then, based on the multi-source situational data, an evolutionary game model is used to determine the first governance strategy for each participating entity. The strategy learning process of the evolutionary game model is determined based on strategy revenue, compliance costs, expected penalties, and risk losses. Strategy revenue is composed of each participating entity's strategy space, which describes the relationship between the participating entity's governance strategies and corresponding returns. Compliance costs are determined by the compliance cost coefficient of each participating entity, which measures the economic cost of the participating entity's compliance investment. Expected penalties are determined by the probability of detection by algorithmic audits and the intensity of institutional penalties. The probability of detection by algorithmic audits reflects the regulator's regulatory technical capabilities and the frequency of spot checks. The severity of penalties is used to describe the cost of violating data, algorithms, and models; risk loss is used to quantify the degree of risk of model failure; by introducing an evolutionary game model, the strategy probing, marginal game, and evolutionary convergence laws of bounded rationality agents are realistically reproduced, ensuring more accurate output of governance strategies for each participating entity; simultaneously, the levels of technological trust, organizational trust, and ecological trust are determined based on multi-source situational data and a trust assessment model; since the trust assessment model is used to describe the constraints between technological trust, organizational trust, and ecological trust levels; the technological trust level is determined based on data credibility, algorithm interpretability, model evaluability, and computing power reliability; the organizational trust level is determined based on the reputation values ​​of all participating entities, contract fulfillment among participating entities, and the matching of rights and responsibilities; the ecological trust level is determined based on social network analysis and structural hole theory; social network analysis is used to describe the interaction structure among the participating entities; therefore, introducing the trust assessment model into the data-computing-model collaborative project can achieve top-down trust transmission, cross-organizational natural attenuation, and bottom-up reverse constraints. Simultaneously, based on multi-source situational data and the preset legal mapping rule library, target constraint parameters are obtained. The preset legal mapping rule library stores the mapping rules between legal norms and technical constraint parameters in the scenarios corresponding to the multi-source situational data, realizing a two-way adaptation mechanism of code as law. Legal provisions are automatically encoded into technical parameters for the data-computer-model collaborative project, thereby accurately determining the target constraint parameters. Finally, based on the level of technical trust, organizational trust, ecological trust, and target constraint parameters, the first governance strategy is optimized to obtain a second governance strategy. This achieves precise optimization of the governance strategies of each participating entity, enabling precise governance of the participating entities using the second governance strategy. This invention effectively solves the problems of fragmented trust research, simplified game theory models, and the disconnect between law and technology.

[0009] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart of a trust cascading evolution game governance method provided by an embodiment of the present invention; Figure 2 This is an interaction diagram of the five participating entities applicable to the embodiments of the present invention; Figure 3 This is a schematic diagram of the phase convergence of the game strategy evolution applicable to the embodiments of the present invention; Figure 4 This is a topology diagram illustrating the technology-organization-ecology trust cascade transmission applicable to embodiments of the present invention; Figure 5 This is a schematic diagram of the structure of a trust-cascaded evolutionary game governance system provided by an embodiment of the present invention; Figure 6 This is a hierarchical architecture diagram of another trust-cascaded evolutionary game governance system provided by an embodiment of the present invention; Figure 7 This is a schematic diagram of the interactions between various modules within a trust-cascaded evolutionary game governance system applicable to embodiments of the present invention. Figure 8 This is a schematic diagram of the structure of an electronic device that implements the trust cascade evolution game governance method of the present invention, according to an embodiment of the present invention. Detailed Implementation

[0012] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0013] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0014] Figure 1 This is a flowchart illustrating a trust-cascading evolutionary game governance method provided by an embodiment of the present invention. This embodiment is applicable to trust-cascading evolutionary game governance driven by institutional constraints in a data-computer-model collaboration scenario. The method can be executed by a trust-cascading evolutionary game governance system, which can be implemented in hardware and / or software. This trust-cascading evolutionary game governance system can be configured in any electronic device with network communication capabilities. Figure 1 As shown, the trust-cascaded evolutionary game governance method of the present invention may include the following process: S110. Obtain multi-source situational data. Multi-source situational data refers to relevant data provided by all participating entities in the entire chain of the data computing and model collaboration project. Participating entities include regulators, data computing and model platforms, data providers, algorithm model developers, and end users.

[0015] Among them, multi-source situational data aligns with the cross-domain flow characteristics of heterogeneous elements in data computing and modeling. Multi-source situational data includes data dimensions, computing power operation dimensions, model service dimensions, multi-participant behavior dimensions, and compliance system dimensions. The core of the cross-domain flow of heterogeneous elements in data computing and modeling is heterogeneous compatibility, full-domain collaboration, security and controllability, and value closed loop, exhibiting five typical characteristics: strong heterogeneity, task-driven, network awareness, dynamic adaptation, and security and compliance constraints.

[0016] Among them, the data (i.e., data elements) in the heterogeneous elements of the data computing model can refer to cross-domain multi-source heterogeneous data (structured / unstructured, encrypted / plaintext, different ownership), and can include training data, inference data, parameter data or gradient data.

[0017] In the heterogeneous elements of data computing and modeling, "computing" (i.e., computing power elements) can refer to heterogeneous chips or computing power pools. For example, heterogeneous chips can be CPUs, GPUs, FPGAs, ASICs, etc.; computing power pools can be cross-regional, cross-vendor, or cross-architecture computing power pools, such as supercomputing, intelligent computing, and edge computing. "Model" (i.e., model elements) in the heterogeneous elements of data computing and modeling can refer to large models, industry models, small models, heterogeneous frameworks (PyTorch / TensorFlow), or distributed sharding (tensor or pipelined parallelism). "Cross-domain" can be understood as cross-geographical domains (e.g., East Data West Computing), cross-administrative / enterprise domains, cross-resource domains (e.g., cloud-edge-device), and cross-technology domains (e.g., heterogeneous chips / frameworks), etc.

[0018] Specifically, such as Figure 2 The interaction diagram of the five participating entities shown illustrates the interactions between regulators, data computing platforms, data providers, algorithm model developers, and end users. Therefore, it allows for the acquisition of primary and secondary data provided by all participating entities across the entire value chain. Figure 2 The interactive relationships shown extract cross-domain flow characteristics of heterogeneous elements in the first and second data sets that align with the computational model, thereby obtaining multi-source situational data. This multi-source situational data is then preprocessed to obtain processed multi-source situational data, providing standardized input for subsequent calculations. The first data set consists of data directly related to the participating entities, while the second data set is indirectly related. Preprocessing can include cleaning, desensitization, and standardization.

[0019] S120. Based on multi-source situational data, the evolutionary game model is evolved to determine the first governance strategy of each participating entity; the strategy learning process of the evolutionary game model is determined based on strategy income, compliance costs, expected penalties, and risk losses.

[0020] The strategy income consists of the strategy space of each participant. The strategy space describes the relationship between the governance strategies of the participants and the corresponding income, that is, it describes the income of different governance strategies of the participants. As shown in Table 1, the participants and strategy space of the five-party game. The participants in Table 1 are the participating entities.

[0021] Table 1. Participants and Strategy Space in a Five-Party Game ; Compliance costs are comprised of compliance cost coefficients for each participating entity, which measure the economic cost of compliance investment by each entity. Expected penalties are determined by the probability of detection by algorithmic audits and the intensity of institutional penalties. The probability of detection by algorithmic audits reflects the regulatory authority's technical capabilities and the frequency of spot checks, while the intensity of institutional penalties describes the cost of violations by data, algorithms, and models. Risk losses quantify the degree of risk of model failure. Here, the cost of violation can be understood as the cost of violating legal norms in a pre-defined legal mapping rule base. The pre-defined legal mapping rule base stores the mapping rules between legal norms and technical constraint parameters in scenarios corresponding to multi-source situational data.

[0022] Among them, regulatory technical capabilities describe the comprehensive technical strength possessed by regulators when performing algorithm audits in a data-driven collaborative governance scenario. Specifically, regulatory technical capabilities can be understood as the use of technological means by regulators to improve audit effectiveness, including at least the following three aspects: 1. Technical Tools and Infrastructure: This refers to the hardware and software tools and systems used by regulators to perform audits, such as hardware tools relying on computing server clusters, distributed storage, and multi-source interface gateways. The audit logic of software tools may include: using rule engines and matching algorithms for automated audit analysis, enabling a shift from traditional sampling audits to comprehensive audits; 2. Data Acquisition and Analysis Capabilities: This refers to the regulator's ability to effectively acquire, process, and analyze the full volume of multi-source data from the audited entity (such as data modeling platforms, data providers, etc.), and to utilize advanced technologies such as machine learning for in-depth analysis. This capability directly impacts the probability of detecting violations. 3. Rule and Standard Enforcement Capability: This refers to the ability to transform abstract laws and regulations into quantifiable and enforceable technical parameters and audit rules. For example, mapping algorithmic fairness principles to specific parameters such as group equality indicators; In summary, regulatory technical capability refers to the regulator's comprehensive ability to conduct efficient and accurate algorithmic audits of the entire process of data-driven simulation collaborative projects using advanced technological means. The level of regulatory technical capability directly determines the probability of detecting violations, i.e., the probability of detection by algorithmic audits, and is a key force driving the entire evolutionary game model towards a compliant equilibrium.

[0023] Optionally, the strategy learning process of the evolutionary game model can be represented by the following formula: ; in, For the expected return of the participating subjects, i.e. when i∈R, For the regulator's expected return, when i∈P, For the expected return of the numerical modeling platform, when i∈D, For the expected revenue of the data provider, when i∈U, For the expected benefit of the end user, when i∈A, The expected revenue for the algorithm model developer; This provides strategic space for regulator R; The strategy space of the computational model platform P; The strategy space for data provider D; Let A be the strategy space of algorithm model developer A; Let i be the policy space of end user U; i ∈ R, P, D, U, A; C is the compliance cost coefficient. F represents the probability of detection by algorithm audit; L represents the severity of institutional penalties; and L represents the loss from model failure, which can be used to quantify the degree of negative social impact caused by algorithmic accidents. For strategic revenue; For compliance costs; For the purpose of punishment; For risk loss; As a switch indicating whether a failure event has occurred in the flag model, it controls whether risk losses are deducted; for example, It can be an indicator function to mark whether a model failure event has occurred. It can be 1 or 0, where 1 represents failure and 0 represents not failure. E can be expectation, which can be understood as the average (expected) value of the penalty Fi under a specific policy. Calculating an average expected penalty value is used to quantify the uncertain cost of illegality. Among them, the severity of institutional penalties ranges from the severity of penalties for data misuse. The severity of penalties for algorithmic discrimination And the severity of penalties for model safety accidents The components are as follows: the severity of penalties for data misuse reflects the cost of violating data security laws; the severity of penalties for algorithm discrimination reflects the cost of violating algorithm compliance laws; and the severity of penalties for model security incidents reflects the cost of violating artificial intelligence applications.

[0024] Optionally, evolving the evolutionary game model based on multi-source situational data to determine the first governance strategy for each participating entity may include: evolving the evolutionary game model based on multi-source situational data to determine the probability of algorithmic audit detection, the intensity of institutional penalties, and the compliance cost coefficient for the current round of evolution; if the probability of algorithmic audit detection exceeds a first threshold, or the ratio of the intensity of institutional penalties to the compliance cost coefficient exceeds a second threshold, then the governance strategy for each participating entity corresponding to the current round of evolution is determined as the first governance strategy for each participating entity.

[0025] Specifically, replicator dynamic equations and Jacobian matrix stability analysis can be used to evolve the evolutionary game model based on multi-source situational data, thereby clarifying the critical conditions and convergence paths for the system to transition to an evolutionarily stable strategy (ESS). The critical conditions could be that the probability of algorithmic audit findings exceeds a first threshold or the ratio of institutional penalties to compliance cost coefficients exceeds a second threshold. Based on these critical conditions, the corresponding convergence path is determined, and the governance strategy of each participating entity corresponding to the convergence path is then defined as the first governance strategy for each participating entity. Figure 3 The diagram shows the phase convergence of the game strategy evolution.

[0026] Optionally, to address the complexity of the five-party game, this invention can employ a hierarchical dimensionality reduction strategy. Specifically, it first fixes the regulator's strategy and analyzes the equilibrium of the four-party game involving the computational modeling platform, data provider, algorithm model developer, and end-user; then, it introduces the regulator's dynamic response to examine the impact of endogenous adjustments in regulatory intensity. By solving the replicator dynamic equation, the set of equilibrium points is identified, and the stability of the evolutionary game model during its evolution is analyzed using the eigenvalues ​​of the Jacobian matrix. The focus is on the critical condition for the transition from a suboptimal equilibrium of "low quality - low compliance - high risk" to a superior equilibrium of "high quality - high compliance - sustainability": when the probability of algorithm auditing is detected... When the first threshold is exceeded, or the ratio of the institutional penalty intensity F to the compliance cost coefficient C exceeds the second threshold, the second threshold can be the critical value for the transition, and the system may bifurcate, achieving a qualitative change to the equilibrium state.

[0027] S130. Based on multi-source situational data and a trust assessment model, determine the levels of technological trust, organizational trust, and ecological trust. The trust assessment model is used to describe the constraints among these levels. The technological trust level is determined based on data credibility, algorithm interpretability, model evaluability, and computing power reliability. The organizational trust level is determined based on the reputation values ​​of all participating entities, contract fulfillment among participating entities, and the matching of rights and responsibilities. The ecological trust level is determined based on social network analysis and structural hole theory. Social network analysis is used to describe the interaction structure among participating entities.

[0028] Specifically, the trust assessment model includes a first model, a second model, and a third model. These models enable the analysis and processing of multi-source situational data to obtain levels of technological trust, organizational trust, and ecological trust. The first model possesses algorithms for determining data credibility, algorithm interpretability, model evaluability, and computational reliability based on multi-source situational data. The second model has algorithms for determining the reputation values ​​of participating entities, contract fulfillment among participating entities, and the matching of rights and responsibilities based on multi-source situational data. The third model incorporates Social Network Analysis (SNA) and structural hole theory.

[0029] Furthermore, analyzing and processing multi-source situational data based on the first, second, and third models in the trust assessment model to obtain the levels of technological trust, organizational trust, and ecological trust may include steps A1-A3: Step A1: Input multi-source situational data into the first model, outputting data credibility, algorithm interpretability, model evaluability, and computing power reliability. Then, use the Analytic Hierarchy Process (AHP) and entropy weighting method to assign combined weights to these four metrics. Based on the weights assigned to each metric, perform a weighted calculation to obtain a quantifiable and comparable level of technical trust. Specifically, data credibility is determined based on data source compliance certification, quality inspection coverage, version consistency, and anonymization verifiability; algorithm interpretability is determined based on feature visualization, decision traceability, robustness testing, and fairness assessment; model evaluability is determined based on benchmark standardization, drift monitoring real-time performance, failure warning accuracy, and the completeness of the decommissioning mechanism; and computing power reliability is determined based on resource scheduling consistency, training reproducibility, service availability, and energy consumption transparency.

[0030] Step A2: Input the multi-source situational data into the second model and output the reputation value of each participating entity, the contract performance and the matching relationship of rights and responsibilities among the participating entities, so as to determine the quantifiable and comparable organizational trust level based on the reputation value of all participating entities and the contract performance and the matching relationship of rights and responsibilities among the participating entities; Step A3: Input the multi-source situational data into the third model, and then measure the multi-source situational data through social network analysis (SNA) and structural hole theory to output the ecological trust level.

[0031] In an embodiment of the present invention, the trust assessment model may optionally be expressed by the following formula: ; in, The target level of trust; As a measure of technological trust level; As a measure of organizational trust levels; α represents the level of ecological trust; β represents the weighting coefficient of the level of technological trust; θ represents the weighting coefficient of the level of organizational trust; and γ represents the efficiency parameter for the transformation of technological trust into organizational trust. This describes the hindering effect on organizational trust formation when there is insufficient technological trust. Insufficient technological trust is determined as follows: when the level of technological trust is less than a preset threshold, or when at least one of the following—data credibility, algorithm interpretability, model evaluability, and computing power reliability—meets preset conditions, insufficient technological trust is identified.

[0032] Specifically, a target trust level is set based on the actual situation of the data-computing-model collaboration project. Then, the trust assessment model is evolved based on multi-source situational data until the target trust level is reached or the iteration rounds are met, and the corresponding technical trust level, organizational trust level, and ecological trust level are output.

[0033] In addition, such as Figure 4 As shown, the calculation process for determining the levels of technological trust, organizational trust, and ecological trust, based on multi-source situational data and a trust assessment model, is determined by a technology-organization-ecological trust cascade attenuation assessment system. This system is a three-layer cascaded topology, comprising a technological trust layer, an organizational trust layer, and an ecological trust layer. The technological trust layer outputs the technological trust level, the organizational trust layer outputs the organizational trust level, and the ecological trust layer outputs the ecological trust level. The technological trust layer serves as the bottom-level support, the organizational trust layer as the middle-level link, and the ecological trust layer as the top-level consensus. The calculation processes among these three levels exhibit a cascade effect: an increase in technological trust reduces uncertainty, thereby promoting the formation of organizational trust; and the accumulation of organizational trust reduces transaction costs, thus creating conditions for the formation of ecological trust. This invention's three-layer cascaded topology achieves top-down trust transmission, natural cross-organizational attenuation, and bottom-up reverse constraints.

[0034] Optionally, in an embodiment of the present invention, the process of determining the reputation value of a participating entity may include: when the data computing and modeling collaborative project is successfully delivered and its performance meets the standards, the reputation value increases according to a preset growth function; when a preset risk event occurs in the data computing and modeling collaborative project, the reputation value decreases according to a preset exponential decay function.

[0035] Among them, the preset risk events can be at least one of the following: data security events, algorithm compliance risks and system operation failure events, and model reliability risk events; data security events can be data leakage; algorithm compliance risks can be algorithm discrimination; and model reliability risk events can be model failure.

[0036] The preset growth function is expressed by the following formula: ; Where R is the credit score and K is the adjustment coefficient; The actual performance of the numerical simulation collaborative project; This serves as the expected performance baseline for the computational-model collaborative project. The preset exponential decay function is expressed by the following formula: ; Where λ is the attenuation coefficient; D is the severity and scope of the preset risk event.

[0037] The process of determining the reputation value of the participating entities in this invention can realize trust quantification, dynamic iteration of reputation value, and risk node tracing.

[0038] S140. Based on multi-source situational data and a preset legal mapping rule library, obtain target constraint parameters; the preset legal mapping rule library is used to store the mapping rules between legal norms and technical constraint parameters in the scenarios corresponding to the multi-source situational data.

[0039] The pre-defined legal mapping rule base may include mapping rules between legal norms and technical constraint parameters: The “data minimization principle” is mapped to privacy budget limit e (differential privacy parameter), data field access whitelist, data retention period threshold Tretention, and verifiable proof of data deletion; The "principle of algorithmic fairness" is mapped to the group equality index Ademographic (demographic equality difference), the individual fairness constraint eindividuai, and the frequency of algorithmic decision bias detection. The “model interpretability principle” is mapped to the interpretation method coverage (Tcoverage, a gradation standard from black box to fully interpretable), the granularity of feature importance disclosure, and the generativeability of counterfactual examples of decisions. The "human oversight principle" is mapped to the proportion of human-machine collaborative decision-making (phuman), the triggering conditions for manual review of key decisions, and the response time for manual takeover when the model fails (tresponse).

[0040] This invention establishes a pre-defined legal mapping rule base to make legal requirements coded, executable, and verifiable, providing support for the implementation of "code as law" in the integrated data and computing model scenario, and realizing the coding of systems and the automation of compliance.

[0041] S150. Based on the level of technological trust, organizational trust, ecological trust, and target constraint parameters, the first governance strategy is optimized to obtain the second governance strategy, which is then used to govern the participating entities.

[0042] Specifically, using TOIA multidimensional constraints, based on the levels of technological trust, organizational trust, ecological trust, and target constraint parameters, the first governance strategy is optimized to obtain the second governance strategy.

[0043] Furthermore, the data after S110 to S150 of this invention are executed can be fed back and iterated to optimize the evolutionary game model, trust assessment model, and preset legal mapping rule base, thereby forming a business logic driven by theoretical models and with modules operating in tandem.

[0044] The technical solution of this invention involves acquiring multi-source situational data, which is relevant data provided by all participating entities across the entire chain. These entities include regulators, data computing and modeling platforms, data providers, algorithm model developers, and end-users. The introduction of multi-source data ensures a more accurate reflection of the impact of each participating entity in the collaborative data computing and modeling project, as well as the influence between them. Then, based on the multi-source situational data, an evolutionary game model is used to determine the first governance strategy for each participating entity. The strategy learning process of the evolutionary game model is determined based on strategy revenue, compliance costs, expected penalties, and risk losses. Strategy revenue is composed of each participating entity's strategy space, which describes the relationship between the participating entity's governance strategies and corresponding returns. Compliance costs are composed of each participating entity's compliance cost coefficient, which measures the economic cost of the participating entity's compliance investment. Expected penalties are determined by the probability of detection by algorithmic audits and the intensity of institutional penalties. The probability of detection by algorithmic audits reflects the regulator's regulatory technical capabilities and the frequency of spot checks, while the intensity of institutional penalties... The system uses a multi-source approach to describe the cost of violating data, algorithms, and models; risk loss is used to quantify the degree of risk of model failure; by introducing an evolutionary game model, it realistically recreates the strategy probing, marginal game, and evolutionary convergence patterns of bounded rationality agents, ensuring more accurate output of governance strategies for each participating entity; simultaneously, it determines the levels of technological trust, organizational trust, and ecological trust based on multi-source situational data and a trust assessment model; since the trust assessment model describes the constraints between these levels; the technological trust level is determined based on data credibility, algorithm interpretability, model evaluability, and computing power reliability; the organizational trust level is determined based on the reputation values ​​of all participating entities, contract fulfillment, and the matching of rights and responsibilities among them; and the ecological trust level is determined based on social network analysis (SNA) and structural hole theory; social network analysis describes the interaction structure among the participating entities; therefore, introducing the trust assessment model into the data-computing-model collaborative project can achieve top-down trust transmission, cross-organizational natural attenuation, and bottom-up reverse constraints. Simultaneously, based on multi-source situational data and the preset legal mapping rule library, target constraint parameters are obtained. The preset legal mapping rule library stores the mapping rules between legal norms and technical constraint parameters in the scenarios corresponding to the multi-source situational data, realizing a two-way adaptation mechanism of code as law. Legal provisions are automatically encoded into technical parameters for the data-computer-model collaborative project, thereby accurately determining the target constraint parameters. Finally, based on the level of technical trust, organizational trust, ecological trust, and target constraint parameters, the first governance strategy is optimized to obtain a second governance strategy. This achieves precise optimization of the governance strategies of each participating entity, enabling precise governance of the participating entities using the second governance strategy. This invention effectively solves the problems of fragmented trust research, simplified game theory models, and the disconnect between law and technology.

[0045] Figure 5 This is a schematic diagram of a trust-cascading evolutionary game governance system provided in an embodiment of the present invention. This embodiment is applicable to trust-cascading evolutionary game governance driven by institutional constraints in a data-computer-model collaboration scenario. This trust-cascading evolutionary game governance system can be implemented in hardware and / or software, and can be configured in any electronic device with network communication capabilities. Figure 5 As shown, the trust cascade evolutionary game governance system of the present invention includes: a multi-source situational data acquisition module 210, a trust cascade calculation module 220, an institutional parameterized evolutionary game simulation module 230, a compliance mapping module 240, and a strategy optimization module 250. The multi-source situational data acquisition module 210 is used to acquire multi-source situational data, which is relevant data provided by the participating entities across the entire chain; the participating entities include regulators, data computing platforms, data providers, algorithm model developers, and end users; The institutional parameterized evolutionary game simulation module 220 is used to evolve the evolutionary game model based on the multi-source situational data to determine the first governance strategy for each participating entity. The strategy learning process of the evolutionary game model is determined based on strategy revenue, compliance cost, expected penalty, and risk loss. The strategy revenue is composed of the strategy space of each participating entity, which describes the relationship between the governance strategies of the participating entities and their corresponding revenues. The compliance cost is determined by the compliance cost coefficient of each participating entity, which measures the economic cost of the participating entity's compliance investment. The expected penalty is determined by the probability of detection by algorithmic audit and the intensity of institutional penalties. The probability of detection by algorithmic audit reflects the regulatory authority's regulatory technical capabilities and sampling frequency, while the intensity of institutional penalties describes the cost of violations by data, algorithms, and models. The risk loss is used to quantify the degree of risk of model failure. The trust cascade measurement module 230 is used to determine the levels of technological trust, organizational trust, and ecological trust based on the multi-source situational data and the trust assessment model. The trust assessment model describes the constraints between these three levels. The technological trust level is determined based on data credibility, algorithm interpretability, model evaluability, and computing power reliability. The organizational trust level is determined based on the reputation values ​​of all participating entities, contract fulfillment, and the matching of rights and responsibilities among them. The ecological trust level is determined based on social network analysis (SNA) and structural hole theory. Social network analysis describes the interaction structure among the participating entities. The compliance mapping module 240 is used to obtain target constraint parameters based on the multi-source situational data and the preset legal mapping rule library; the preset legal mapping rule library is used to store the mapping rules between legal norms and technical constraint parameters in the scenarios corresponding to the multi-source situational data. The strategy optimization module 250 is used to optimize the first governance strategy based on the technology trust level, the organization trust level, the ecological trust level and the target constraint parameters to obtain a second governance strategy, so as to use the second governance strategy to govern the participating entities.

[0046] Based on the above embodiments, optionally, the strategy learning process of the evolutionary game model is represented by the following formula: ; in, The expected return for the participating entities, This provides strategic space for regulator R; The strategy space of the computational model platform P; The strategy space for data provider D; Let A be the strategy space of algorithm model developer A; Let i be the policy space of end user U; i ∈ R, P, D, U, A; C is the compliance cost coefficient. F represents the probability of detection by algorithm audit; F represents the severity of institutional penalties; L represents the loss due to model failure. For strategic revenue; For compliance costs; For the purpose of punishment; For risk loss; It acts as a switch to indicate whether a failure event has occurred in the flag model, and is used to control whether risk losses are deducted.

[0047] Based on the above embodiments, optionally, the severity of the system's penalties consists of penalties for data misuse, penalties for algorithm discrimination, and penalties for model security incidents; the penalties for data misuse are used to reflect the cost of data security violations; the penalties for algorithm discrimination are used to reflect the cost of algorithm compliance violations; and the penalties for model security incidents are used to reflect the cost of illegal applications of artificial intelligence.

[0048] Based on the above embodiments, optionally, the institutional parameterized evolutionary game simulation module includes an evolution unit, which is used to: evolve the evolutionary game model based on the multi-source situational data, and determine the algorithm audit detection probability, institutional penalty intensity, and compliance cost coefficient of the current round of evolution; if the algorithm audit detection probability exceeds a first threshold, or the ratio of the institutional penalty intensity to the compliance cost coefficient exceeds a second threshold, then the governance strategy of each participating entity corresponding to the current round of evolution is determined as the first governance strategy of each participating entity.

[0049] Based on the above embodiments, optionally, the trust assessment model is expressed by the following formula: ; in, The target level of trust; As for the level of trust in technology; As a measure of organizational trust levels; α represents the level of ecological trust; β represents the weighting coefficient of the level of technological trust; θ represents the weighting coefficient of the level of organizational trust; and γ represents the efficiency parameter for the transformation of technological trust into organizational trust. This describes the hindering effect on organizational trust formation when there is insufficient trust in technology.

[0050] Based on the above embodiments, optionally, the process of determining the reputation value of the participating entity includes: When the digital computing and simulation collaborative project is successfully delivered and its performance meets the standards, the reputation value increases according to a preset growth function; When a preset risk event occurs in the digital computing and model collaboration project, the reputation value decreases according to a preset exponential decay function.

[0051] Based on the above embodiments, optionally, the preset growth function is expressed by the following formula: ; Where R is the credit score and K is the adjustment coefficient; These are the actual performance parameters of the numerical simulation collaborative project; The expected performance baseline parameters for the numerical simulation collaborative project; The preset exponential decay function is expressed by the following formula: ; Wherein, λ is the attenuation coefficient; D is the quantitative parameter of the severity and impact range of the preset risk event.

[0052] Optionally, such as Figure 6 As shown, the trust-cascaded evolutionary game governance system of the present invention may include an application operation and maintenance layer, a core business layer, and a hardware support layer.

[0053] The core business layer can serve as the carrier of the trust-cascade evolutionary game governance method of this invention. The core business layer may include a multi-source situational data acquisition module, a trust-cascade calculation module, a system parameterized evolutionary game simulation module, a compliance mapping module, a strategy optimization module, and a closed-loop execution and feedback control module; wherein, for example... Figure 7 As shown, the closed-loop execution and feedback control module is used to feed back the governance strategy output by the strategy optimization module to the multi-source situational data acquisition module, the trust cascade calculation module, the institutional parameterized evolutionary game simulation module, and the compliance mapping module for iterative optimization, thereby ensuring the accuracy of the trust cascade evolutionary game governance system.

[0054] The hardware support layer consists of a computing server cluster, distributed storage, multi-source interface gateways, and edge operation and maintenance terminals, providing hardware computing power and access support for the trust cascading calculation module, the institutional parameterized evolutionary game simulation module, and the compliance mapping module in the core business layer. This invention adopts standard universal interfaces, divided into four categories: data access, internal modules, control and management distribution, and feedback transmission. It can seamlessly interface with data computing model platforms, computing nodes, large model interfaces, and regulatory systems. The interface logic is customized around the business processes of this invention, exhibiting strong adaptability and specialization.

[0055] The application operations layer is used to support functions such as governance visualization, situation monitoring, parameter configuration, policy distribution, risk warning, and evolution phase display. It provides a human-computer interaction interface for regulatory and operations personnel, thereby intuitively presenting the trust chain and game evolution (e.g. Figure 3 The diagram shows the phase convergence of the game strategy evolution, and the distribution of compliance risks, among other things.

[0056] The trust-cascaded evolutionary game governance system provided in this embodiment of the invention can execute the trust-cascaded evolutionary game governance method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0057] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0058] Figure 8A schematic diagram of an electronic device is shown that can be used to implement the trust cascading evolutionary game governance method of embodiments of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0059] like Figure 8 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 may also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0060] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0061] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, central processing unit (CPU), graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as the trust cascade evolutionary game governance method.

[0062] In some embodiments, the trust cascading evolutionary game governance method can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via read-only memory (ROM) 12 and / or communication unit 19. When the computer program is loaded into random access memory (RAM) 13 and executed by processor 11, one or more steps of the trust cascading evolutionary game governance method described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to execute the trust cascading evolutionary game governance method by any other suitable means (e.g., by means of firmware).

[0063] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0064] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0065] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0066] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0067] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0068] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0069] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0070] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A trust-cascading evolutionary game governance method, characterized in that, The method includes: Acquire multi-source situational data, which is relevant data provided by the participating entities across the entire chain of the data computing and modeling collaborative project; the participating entities include regulators, data computing and modeling platforms, data providers, algorithm model developers, and end users; The evolutionary game model is evolved based on the multi-source situational data to determine the first governance strategy for each participating entity. The strategy learning process of the evolutionary game model is determined based on strategy revenue, compliance cost, expected penalty, and risk loss. The strategy revenue is constituted by the strategy space of each participating entity, which describes the relationship between the governance strategies of the participating entities and their corresponding revenues. The compliance cost is determined by the compliance cost coefficient of each participating entity, which measures the economic cost of the participating entity's compliance investment. The expected penalty is determined by the probability of detection by algorithmic audit and the intensity of institutional penalties. The probability of detection by algorithmic audit reflects the regulatory authority's regulatory technical capabilities and the frequency of spot checks, while the intensity of institutional penalties describes the cost of violations by data, algorithms, and models. The risk loss is used to quantify the degree of risk of model failure. Based on the aforementioned multi-source situational data and trust assessment model, the levels of technological trust, organizational trust, and ecological trust are determined. The trust assessment model describes the constraints among these three levels. The technological trust level is determined based on data credibility, algorithm interpretability, model evaluability, and computing power reliability. The organizational trust level is determined based on the reputation values ​​of all participating entities, contract fulfillment among them, and the matching of rights and responsibilities. The ecological trust level is determined based on social network analysis and structural hole theory. Social network analysis describes the interaction structure among the participating entities. Based on the multi-source situational data and the preset legal mapping rule library, target constraint parameters are obtained; the preset legal mapping rule library is used to store the mapping rules between legal norms and technical constraint parameters in the scenarios corresponding to the multi-source situational data. Based on the technological trust level, the organizational trust level, the ecological trust level, and the target constraint parameters, the first governance strategy is optimized to obtain a second governance strategy, which is then used to govern the participating entities.

2. The method according to claim 1, characterized in that, The strategy learning process of the evolutionary game model is represented by the following formula: ； in, The expected return for the participating entities, This provides strategic space for regulator R; The strategy space of the computational model platform P; The strategy space for data provider D; Let A be the strategy space of algorithm model developer A; Let i be the policy space of end user U; i ∈ R, P, D, U, A; C is the compliance cost coefficient. F represents the probability of detection by algorithm audit; F represents the severity of institutional penalties; L represents the loss due to model failure. For strategic revenue; For compliance costs; For the purpose of punishment; For risk loss; It acts as a switch to indicate whether a failure event has occurred in the flag model, and is used to control whether risk losses are deducted.

3. The method according to claim 2, characterized in that, The severity of penalties under the system consists of penalties for data misuse, penalties for algorithm discrimination, and penalties for model security incidents. The penalties for data misuse are used to reflect the cost of violating data security laws; the penalties for algorithm discrimination are used to reflect the cost of violating algorithm compliance laws; and the penalties for model security incidents are used to reflect the cost of violating artificial intelligence applications.

4. The method according to claim 3, characterized in that, Based on the multi-source situational data, the evolutionary game model is evolved to determine the first governance strategy for each participating entity, including: The evolutionary game model is evolved based on the multi-source situational data to determine the algorithm audit detection probability, institutional penalty intensity, and compliance cost coefficient of the current round of evolution; If the probability of the algorithm audit exceeds the first threshold, or the ratio of the system penalty intensity to the compliance cost coefficient exceeds the second threshold, then the governance strategy of each participating entity corresponding to the current round of evolution will be determined as the first governance strategy of each participating entity.

5. The method according to claim 1, characterized in that, The trust assessment model is expressed by the following formula: ； in, The target level of trust; As a measure of technological trust level; As a measure of organizational trust levels; α represents the level of ecological trust; β represents the weighting coefficient of the level of technological trust; θ represents the weighting coefficient of the level of organizational trust; and γ represents the efficiency parameter for the transformation of technological trust into organizational trust. This describes the hindering effect on organizational trust formation when there is insufficient trust in technology.

6. The method according to claim 1 or 5, characterized in that, The process of determining the reputation value of the participating entities includes: When the digital computing and simulation collaborative project is successfully delivered and its performance meets the standards, the reputation value increases according to a preset growth function; When a preset risk event occurs in the digital computing model collaborative project, the reputation value decreases according to a preset exponential decay function.

7. The method according to claim 6, characterized in that, The preset growth function is expressed by the following formula: ； Where R is the credit score and K is the adjustment coefficient; These are the actual performance parameters of the numerical simulation collaborative project; The expected performance baseline parameters for the numerical simulation collaborative project; The preset exponential decay function is expressed by the following formula: ； Wherein, λ is the attenuation coefficient; D is a quantitative parameter representing the severity and scope of the preset risk event.

8. A trust-cascading evolutionary game governance system, characterized in that, The system implementing the trust cascade evolutionary game governance method according to any one of claims 1-7 includes: a multi-source situational data acquisition module, a trust cascade calculation module, an institutional parameterized evolutionary game simulation module, a compliance mapping module, and a strategy optimization module; The multi-source situational data acquisition module is used to acquire multi-source situational data, which is relevant data provided by the participating entities across the entire chain of the data computing and model collaboration project; the participating entities include regulators, data computing and model platforms, data providers, algorithm model developers, and end users; The parameterized evolutionary game simulation module is used to evolve the evolutionary game model based on the multi-source situational data to determine the first governance strategy for each participating entity. The strategy learning process of the evolutionary game model is determined based on strategy revenue, compliance cost, expected penalty, and risk loss. The strategy revenue is composed of the strategy space of each participating entity, which describes the relationship between the governance strategies of the participating entities and their corresponding revenues. The compliance cost is determined by the compliance cost coefficient of each participating entity, which measures the economic cost of the participating entity's compliance investment. The expected penalty is determined by the probability of detection by algorithmic audit and the intensity of institutional penalties. The probability of detection by algorithmic audit reflects the regulatory authority's regulatory technical capabilities and the frequency of spot checks, while the intensity of institutional penalties describes the cost of violations by data, algorithms, and models. The risk loss is used to quantify the degree of risk of model failure. The trust cascade measurement module is used to determine the levels of technological trust, organizational trust, and ecological trust based on the multi-source situational data and the trust assessment model. The trust assessment model describes the constraints between these three levels. The technological trust level is determined based on data credibility, algorithm interpretability, model evaluability, and computing power reliability. The organizational trust level is determined based on the reputation values ​​of all participating entities, contract fulfillment, and the matching of rights and responsibilities among them. The ecological trust level is determined based on social network analysis and structural hole theory. Social network analysis describes the interaction structure among the participating entities. The compliance mapping module is used to obtain target constraint parameters based on the multi-source situational data and the preset legal mapping rule library; the preset legal mapping rule library is used to store the mapping rules between legal norms and technical constraint parameters in the scenarios corresponding to the multi-source situational data. The strategy optimization module is used to optimize the first governance strategy based on the technology trust level, the organization trust level, the ecosystem trust level and the target constraint parameters to obtain a second governance strategy, so as to use the second governance strategy to govern the participating entities.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the trust cascading evolution game governance method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that are used to cause a processor to execute the trust cascading evolution game governance method according to any one of claims 1-7.

Images