Federal model contribution evaluation method and system for government-enterprise cooperation

Abstract

The embodiment of the invention provides a federal model contribution evaluation method and system for government-enterprise cooperation. The method is applied to the technical field of privacy calculation and federated learning, and comprises the steps that a government affair subject and an enterprise subject train a federated model based on respective local data to generate model update data, and a verifiable encryption method is adopted to generate first encrypted differential data and second encrypted differential data. And in a preset encryption domain, performing secret state feature fusion on the encrypted differential data to obtain a secret state feature result of each party, and generating a first encryption contribution result and a second encryption contribution result based on encryption performance evaluation and a fair distribution rule. The first and second contribution proof are then generated in conjunction with the respective public parameters. According to the scheme, on the premise that local data privacy of all parties is protected, the contribution proof capable of being independently verified is generated through differential data encryption and secret state feature processing, and a credible basis is provided for automatic excitation and fair distribution.

Description

Technical Field

[0001] This disclosure relates to the fields of privacy computing and federated learning technologies, and in particular to a method and system for evaluating the contribution of federated models for government-enterprise collaboration. Background Technology

[0002] With the marketization of artificial intelligence and data elements, the demand for collaborative modeling using federated learning is growing among government and enterprises in areas such as smart cities and public safety. After local training by all parties, a fair and verifiable contribution evaluation mechanism is urgently needed to support incentive allocation and trust building, and promote the secure flow of data value.

[0003] The existing technology involves each participating party uploading local model updates (such as gradients or weight differences) to the central server; the server calculates indicators such as feature importance, gradient similarity, or Shapley value in plaintext or semi-encrypted state to preliminarily assess the contributions of each party; finally, the central node generates a contribution score and notifies the participating parties.

[0004] However, existing technologies require decryption of model update data during aggregation or evaluation, posing a risk of sensitive information leakage; furthermore, the evaluation process is opaque, and participants cannot independently verify the fairness of contribution scores, which can easily lead to trust disputes; at the same time, the lack of verifiable credentials means that the evaluation results are only internal scores and cannot serve as a credible basis for on-chain incentives, audits, or compliance. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this disclosure provides a method and system for evaluating contributions to a federal model used in government-enterprise collaboration. This disclosure resolves the problems of existing technologies, such as the need to decrypt model update data leading to the exposure of sensitive information, the opaque evaluation process making it impossible for all parties to verify impartiality, the lack of verifiable on-chain credentials, and the difficulty in using evaluation results as a credible basis for incentives, audits, or compliance.

[0006] According to a first aspect of this disclosure, a method for evaluating the contribution of a federated model for government-enterprise collaboration is provided, comprising: obtaining first model update data generated by a government entity when training a federated model based on first local data, and second model update data generated by an enterprise entity when training a federated model based on second local data; encrypting the first model update data and the second model update data using a verifiable encryption method to obtain first encrypted differential data and second encrypted differential data. Within a preset encryption domain, the first encrypted differential data and the second encrypted differential data are subjected to encrypted feature fusion processing to obtain the first encrypted feature result of the government entity and the second encrypted feature result of the enterprise entity. Within a preset encryption domain, the encryption performance impact of the first and second encrypted feature results is evaluated and the fair allocation rules are quantified to generate the first encryption contribution result of the government entity and the second encryption contribution result of the enterprise entity. Obtain the first public parameter corresponding to the first encrypted contribution result and the second public parameter corresponding to the second encrypted contribution result; generate a first contribution certificate for the government entity based on the first encrypted contribution result and the first public parameter; and generate a second contribution certificate for the enterprise entity based on the second encrypted contribution result and the second public parameter.

[0007] According to a second aspect of this disclosure, a federated model contribution evaluation system for government-enterprise collaboration is provided, for performing the method as described in the first aspect, comprising: a data encryption module, configured to acquire first model update data generated by the government entity when training a federated model based on first local data, and second model update data generated by the enterprise entity when training a federated model based on second local data, and to encrypt the first model update data and the second model update data using a verifiable encryption method to obtain first encrypted differential data and second encrypted differential data; The encrypted feature fusion module is used to perform encrypted feature fusion processing on the first encrypted differential data and the second encrypted differential data within a preset encrypted domain to obtain the first encrypted feature result of the government entity and the second encrypted feature result of the enterprise entity. The encryption contribution result generation module is used to evaluate the encryption performance impact of the first encryption state feature result and the second encryption state feature result and quantify the fair allocation rules within a preset encryption domain, and generate the first encryption contribution result of the government entity and the second encryption contribution result of the enterprise entity. The contribution proof generation module is used to obtain the first public parameter corresponding to the first encrypted contribution result and the second public parameter corresponding to the second encrypted contribution result, generate the first contribution proof of the government entity based on the first encrypted contribution result and the first public parameter, and generate the second contribution proof of the enterprise entity based on the second encrypted contribution result and the second public parameter.

[0008] According to a third aspect of this disclosure, an electronic device is provided, comprising: a memory and a processor, wherein the memory stores a computer program, and the processor executes the program to implement the method described above.

[0009] In the above-described method and system for evaluating contributions in a federal model for government-enterprise collaboration, this disclosure, while protecting the privacy of local data of all parties, generates independently verifiable contribution proofs through encrypted differential data and encrypted feature processing, providing a credible basis for automated incentives and fair allocation. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this disclosure 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 some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 A schematic flowchart of a federal model contribution assessment method for government-enterprise collaboration according to an embodiment of this disclosure is shown. Figure 2 A schematic flowchart of a federal model contribution assessment method for government-enterprise collaboration according to an embodiment of this disclosure is shown. Figure 3 A schematic block diagram of a federal model contribution assessment system for government-enterprise collaboration according to an embodiment of the present disclosure is shown. Figure 4 A block diagram of an exemplary electronic device according to an embodiment of the present disclosure is shown. Detailed Implementation

[0012] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present disclosure.

[0013] Those skilled in the art will understand that the terms "first," "second," etc., in the embodiments of this disclosure are only used to distinguish different steps, devices, or modules, and do not represent any specific technical meaning, nor do they indicate a necessary logical order between them. It should also be understood that in the embodiments of this disclosure, "multiple" can refer to two or more, and "at least one" can refer to one, two, or more. It should also be understood that any component, data, or structure mentioned in the embodiments of this disclosure can generally be understood as one or more unless explicitly limited or given a contrary indication in the context. Furthermore, the term "and / or" in this disclosure is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this disclosure generally indicates that the related objects before and after are in an "or" relationship. It should also be understood that the descriptions of the various embodiments in this disclosure emphasize the differences between the various embodiments; their similarities or commonalities can be referred to mutually, and for the sake of brevity, they will not be elaborated upon one by one.

[0014] Furthermore, it should be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn to actual scale. The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit this disclosure or its application or use. Techniques, methods, and apparatus known to those skilled in the art will not be discussed in detail, but where appropriate, such techniques, methods, and apparatus should be considered part of the specification. It should be noted that similar reference numerals and letters in the following drawings denote similar items; therefore, once an item is defined in one drawing, it need not be further discussed in subsequent drawings.

[0015] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0016] Figure 1 This is a schematic flowchart illustrating a federal model contribution evaluation method for government-enterprise collaboration, provided as an embodiment of this disclosure. The method of this disclosure aims to achieve accurate detection of both large and small targets in images.

[0017] S101, obtain the first model update data generated by the government entity when training the federated model based on the first local data, and the second model update data generated by the enterprise entity when training the federated model based on the second local data, and encrypt the first model update data and the second model update data using a verifiable encryption method to obtain the first encrypted differential data and the second encrypted differential data.

[0018] In the context of government-enterprise collaboration, a government agency or department with data resources and business application scenarios can be considered as one of the participants in federated learning. It holds training data related to government business locally and participates in the training process of the federated model.

[0019] The enterprise entity can refer to an enterprise or institution with industry data or computing resources in a government-enterprise collaboration scenario. As another participant in federated learning, it holds training data related to the enterprise's business locally and participates in the training process of the federated model.

[0020] First-local data refers to the set of raw data related to government operations held by a government entity within its local data environment. This data may include government transaction records, administrative approval data, public service interaction logs, etc. First-local data is not uploaded or shared; it is used only in the government entity's local computing environment to train the federated model.

[0021] Secondary local data refers to the set of raw data related to the business operations held by the enterprise entity within its local data environment. This data includes enterprise operational data, customer interaction data, industry production data, etc. Secondary local data is used solely within the enterprise entity's local computing environment for training the federated model.

[0022] A federated model can refer to a machine learning model jointly trained by multiple data stakeholders through a federated learning framework without sharing the original data. In this application, the federated model can integrate the data features of government entities and enterprise entities, achieving cross-domain collaborative modeling capabilities.

[0023] The first model update data can refer to the parameter gradient or weight difference information generated by the government entity during the local training process, after iteratively calculating the parameters of the federated model based on local data, and used to update the global model.

[0024] The second model update data can refer to the parameter gradient or weight difference information generated by the enterprise entity during local training, after iteratively calculating the parameters of the federated model based on local data, and used to update the global model.

[0025] Verifiable encryption methods refer to cryptographic methods that, while encrypting data, also enable third parties to verify the correctness of the encryption result. For example, zero-knowledge proofs, homomorphic encryption combined with commitment mechanisms, and other methods can all be used to achieve verifiable encryption of model update data.

[0026] The first encrypted differential data can refer to the ciphertext differential information obtained after the first model update data generated by the government entity is processed by a verifiable encryption method.

[0027] The second encrypted differential data can refer to the ciphertext differential information obtained after the second model update data generated by the enterprise entity is processed by a verifiable encryption method.

[0028] The government entity trains the shared federated model locally based on its own local data. During training, the government entity uses a gradient descent algorithm to iteratively update the model parameters, obtaining first model update data corresponding to the distribution characteristics of the local data. This first model update data is typically represented as a parameter gradient vector or a weight difference matrix, reflecting the contribution of the government entity's data to the global model.

[0029] Simultaneously, the enterprise entity trains the same federated model locally based on its second set of local data. During training, the enterprise entity also uses optimization algorithms to update the model parameters, obtaining corresponding updated second model data. This updated second model data maintains the same data feature dimensions as the updated first model data, but the numerical results reflect the differential features of the enterprise entity's data.

[0030] Next, verifiable encryption methods are applied to the first and second model update data, respectively. The verifiable encryption method can employ homomorphic encryption combined with zero-knowledge proofs: First, a homomorphic encryption algorithm (such as Paillier encryption) is used to perform ciphertext mapping on the first and second model update data, obtaining a first ciphertext update vector and a second ciphertext update vector. This ensures that the encrypted data can still support subsequent addition and aggregation operations without decryption. Second, a corresponding zero-knowledge proof circuit is constructed to generate a correctness proof for each party's submitted ciphertext update vector, enabling the verifier to confirm, without knowing the plaintext, that the encrypted data indeed originates from legitimate model training updates, rather than being forged or tampered with. Through the above processing, the first encrypted differential data and the second encrypted differential data are finally obtained. The first encrypted differential data corresponds to the ciphertext form of the government entity's first model update data, combined with its zero-knowledge correctness proof; the second encrypted differential data corresponds to the ciphertext form of the enterprise entity's second model update data, combined with its zero-knowledge correctness proof.

[0031] S102, perform cryptographic feature fusion processing on the first encrypted differential data and the second encrypted differential data within the preset encryption domain to obtain the first cryptographic feature result of the government entity and the second cryptographic feature result of the enterprise entity.

[0032] A predefined encryption field can refer to a mathematical operation space predefined based on a specific cryptographic algorithm, used to perform calculations on data in ciphertext. This encryption field is usually defined by homomorphic encryption algorithms or multi-party secure computation protocols, such as the modular arithmetic field based on Paillier homomorphic encryption, or the finite field based on elliptic curve cryptography.

[0033] The first encrypted feature result can refer to the result obtained after performing feature fusion operations on the first encrypted differential data within a preset encrypted domain. This result is still in a encrypted state and does not expose any plaintext information, but its value represents the feature expression of the government entity's first model update data under encrypted conditions.

[0034] The second encrypted feature result can refer to the result obtained after performing feature fusion operation on the second encrypted differential data within a preset encryption domain.

[0035] First, a preset encryption domain is determined. This domain can be constructed using a homomorphic encryption algorithm, such as the Paillier homomorphic encryption scheme based on a large integer modulus space. Within this encryption domain, all computational operations are performed in ciphertext form, ensuring that no plaintext information is leaked during the computation process. Subsequently, the first encrypted differential data generated by the government entity and the second encrypted differential data generated by the enterprise entity are input into the encrypted feature fusion process. The encrypted differential data is essentially a ciphertext vector obtained by encrypting the differential results of model parameter updates during local training.

[0036] In the encrypted feature fusion process, the first and second encrypted differential data are first uniformly formatted to place them in the same encrypted computation space, ensuring the feasibility of subsequent fusion. Next, without decryption, a weighted combination operation is performed on each component of the two sets of encrypted differential data. For example, corresponding weight factors can be preset for different data sources and embedded in the fusion operation in an encrypted state, thereby adjusting the weight of contribution differences in the ciphertext state. Through the ciphertext scalar multiplication operation supported by the homomorphic encryption algorithm, the encrypted differential data is weighted without decryption. Specifically, the system utilizes the exponentiation property of encrypted data to directly apply the preset weight factors to each component of the ciphertext vector, thereby differentiating the contribution of different data sources in the ciphertext space. After weighted combination, the fusion result is further normalized within the encrypted domain. The purpose of this normalization is to keep the fused numerical result within a comparable range so that it can be directly used in subsequent on-chain verification or global optimization. The entire process is based on homomorphic operations in the encrypted domain, without any plaintext exposure. Based on an approximate calculation method supported by homomorphic encryption, the fused ciphertext vector is securely normalized. The system adjusts the numerical range of the ciphertext result using preset reference values ​​and a linear scaling mechanism under encryption, ensuring it meets the standardization requirements for subsequent on-chain verification or global optimization, all without requiring decryption. During the ciphertext feature fusion process, the system retains the independent processing paths for each participant's data, ensuring that while completing collaborative computation, the dominant contribution of each entity can still be traced and separated. This design allows the fusion result to be mapped separately to ciphertext outputs reflecting the respective characteristics of the government and enterprise entities. Ultimately, the system outputs two types of ciphertext feature results: one generated with the first encrypted differential data from the government entity as the dominant contribution, and the other generated with the second encrypted differential data from the enterprise entity as the dominant contribution.

[0037] Based on the above technical solution, optionally, the first encrypted differential data and the second encrypted differential data are subjected to encrypted feature fusion processing within a preset encryption domain to obtain the first encrypted feature result of the government entity and the second encrypted feature result of the enterprise entity, including: A shallow nonlinear transformation is performed on the first encrypted differential data within a preset encryption domain to obtain the first preliminary encrypted feature representation of the government entity; and a shallow nonlinear transformation is performed on the second encrypted differential data within a preset encryption domain to obtain the second preliminary encrypted feature representation of the enterprise entity. Within a preset encryption domain, an affine transformation is performed on the first preliminary encrypted feature representation and the second preliminary encrypted feature representation based on preset global statistical parameters to obtain the first normalized encrypted feature representation of the government entity and the second normalized encrypted feature representation of the enterprise entity. Within a preset encryption domain, the first normalized encrypted feature representation and the second normalized encrypted feature representation are projected onto a unified encrypted semantic space through a preset linear mapping matrix, respectively, to obtain the first mapped encrypted feature representation of the government entity and the second mapped encrypted feature representation of the enterprise entity. Within a preset encryption domain, the first mapping cryptographic feature representation and the second mapping cryptographic feature representation are subjected to dimensional coordination processing to obtain the first aligned cryptographic feature representation of the government entity and the second aligned cryptographic feature representation of the enterprise entity. Within a preset encrypted domain, the first aligned encrypted feature representation and the second aligned encrypted feature representation are weighted and combined according to a preset fusion weighting rule to obtain the first encrypted feature result of the government entity and the second encrypted feature result of the enterprise entity.

[0038] In this scheme, the first preliminary encrypted feature representation can refer to the encrypted feature vector or matrix representation obtained by shallow nonlinear transformation of the first encrypted differential data of the government entity within a preset encryption domain. It retains the core feature information of the original encrypted data, but enhances the feature expression capability through nonlinear mapping, while still maintaining encryption security.

[0039] The second preliminary encrypted feature representation can refer to the encrypted feature vector or matrix representation obtained by shallow nonlinear transformation of the second encrypted differential data of the enterprise entity within a preset encryption domain.

[0040] Preset global statistical parameters refer to the set of parameters obtained by the system based on historical or global data during federated modeling or dense feature processing. These parameters are used to standardize or normalize the preliminary feature representations of each party, including statistical quantities such as mean, variance, minimum, and maximum.

[0041] The first normalized dense state feature representation can refer to the dense state feature representation obtained by performing an affine transformation (standardization or normalization) on the first preliminary dense state feature representation of the government subject based on preset global statistical parameters. It is used to eliminate the dimensional differences and offset effects between different feature dimensions.

[0042] The second normalized dense state feature representation can refer to the dense state feature representation obtained by performing an affine transformation on the second preliminary dense state feature representation of the enterprise entity based on preset global statistical parameters.

[0043] The pre-defined linear mapping matrix can refer to the invertible or reducible matrix used in the mapping process of dense semantic space. It is used to map the normalized dense features to a unified dense semantic space, so that multiple features can be compared and fused in the same vector space or feature subspace.

[0044] A unified dense semantic space can refer to the common representation space of normalized dense features from multiple parties after linear mapping. It is used to ensure that dense features of different subjects can interoperate and align at the same semantic scale, so as to carry out dimensional coordination, weighted fusion and contribution evaluation.

[0045] The first mapping dense state feature representation can refer to the result of projecting the first normalized dense state feature representation of the government entity onto the unified dense state semantic space through a preset linear mapping matrix. This is used to preserve the original feature information while making it comparable to the features of the enterprise entity in the same space.

[0046] The second mapping dense state feature representation can refer to the result of projecting the second normalized dense state feature representation of the enterprise entity onto a unified dense state semantic space through a preset linear mapping matrix.

[0047] The first aligned dense state feature representation can refer to the dense state feature representation obtained after dimensional coordination processing of the first mapped dense state feature representation in a unified dense state semantic space. The imbalance between different feature dimensions is eliminated by adjusting the feature scale and variance, which is conducive to weighted fusion.

[0048] The second aligned dense state feature representation can refer to the dense state feature representation after dimensional coordination processing of the second mapped dense state feature representation within a unified dense state semantic space.

[0049] The preset fusion weight rules can refer to the set of rules used to calculate the contribution ratio of each party to the final dense feature result during the dense feature fusion process. These rules include methods such as weight coefficient allocation, weighted summation, and regularization, which are used to ensure that the features of each party retain their independent contributions and generate a globally unified feature representation after fusion.

[0050] The first dense state feature result can refer to the dense state feature result obtained by weighting and combining the first aligned dense state feature representation of the government entity under the influence of the fusion weighting rule.

[0051] The second dense state feature result can refer to the dense state feature result obtained by weighting and combining the second aligned dense state feature representation of the enterprise entity under the influence of the fusion weight rule.

[0052] Within the predefined encryption domain, a shallow nonlinear transformation is first performed on the first encrypted differential data of the government entity. This process includes: applying a dense activation function, such as a dense Sigmoid, a dense ReLU, or a dense Tanh, to the encrypted values ​​of each feature dimension to maintain the homomorphic computability of the data in the encrypted state, while mapping the linear relationship of the original differential data to a nonlinear space to enhance the feature expressiveness, generating the first preliminary dense feature representation of the government entity. Similarly, the same nonlinear transformation is performed on the second encrypted differential data of the enterprise entity to generate the second preliminary dense feature representation of the enterprise entity, ensuring the consistency of data processing across multiple parties.

[0053] Subsequently, an affine transformation is performed on the preliminary encrypted feature representation based on preset global statistical parameters. These global statistical parameters include the mean, variance, and maximum and minimum values ​​for each feature dimension, used to achieve standardization within the encrypted domain. Specifically, the preliminary encrypted feature values ​​are subtracted from their corresponding mean values, then divided by the standard deviation or scaled proportionally to a fixed interval, forming a first normalized encrypted feature representation for government entities and a second normalized encrypted feature representation for enterprise entities, thus eliminating differences in the dimensions of different features.

[0054] After normalization, the normalized encrypted feature representations are projected onto a unified encrypted semantic space using a predefined linear mapping matrix. This linear mapping matrix can be rotated, scaled, and orthogonally projected using encrypted matrix multiplication, ensuring semantic consistency of features from different participants within the same vector space and guaranteeing the effectiveness of homomorphic operations under encrypted conditions. After mapping, a first mapped encrypted feature representation for government entities and a second mapped encrypted feature representation for enterprise entities are generated.

[0055] Next, dimensional harmonization processing is performed on the mapped dense feature representation. This processing includes scaling adjustments, variance matching, offset correction, and feature value range unification for each dimension of the feature values, generating a first-aligned dense feature representation for government entities and a second-aligned dense feature representation for enterprise entities. The purpose of dimensional harmonization is to ensure that the numerical distribution of features from multiple parties is consistent within a unified semantic space.

[0056] Finally, based on preset fusion weight rules, the aligned dense-state feature representations are weighted and combined. The fusion weight rules dynamically allocate weights according to the contribution of each party, the historical performance of the participants, or preset proportions. The final feature values ​​can be calculated using weighted summation, exponential weighting, or regularized weighting methods, yielding the first dense-state feature result for the government entity and the second dense-state feature result for the enterprise entity. These dense-state feature results retain the independent feature information of each party while generating a fused feature representation.

[0057] This solution not only protects data privacy but also ensures the verifiability and fairness of subsequent contribution assessments and incentive allocations, while improving overall system security and multi-party collaboration efficiency.

[0058] S103, within a preset encryption domain, the encryption performance impact assessment and fair allocation rule quantification of the first and second encrypted feature results are performed to generate the first encryption contribution result of the government entity and the second encryption contribution result of the enterprise entity.

[0059] The first encrypted contribution result can refer to a quantitative result generated within a preset encrypted domain, based on the first encrypted characteristic result of the government entity, after quantification through encryption performance impact assessment and fair allocation rules. This result exists in encrypted form and is used to characterize the effective contribution made by the government entity to the global model update or task completion in multi-party joint computation. This result includes not only the size of the government entity's computational contribution but also a weighted evaluation of performance overhead and fairness constraints in an encrypted environment.

[0060] The second encryption contribution result can refer to the quantitative result generated within a preset encryption domain, based on the second encryption state characteristic result of the enterprise entity, after being quantified by the same encryption performance impact assessment and fair allocation rules. This result also exists in encrypted form.

[0061] The system statistically analyzes the number of encryption operations performed on the first and second encrypted feature results during homomorphic addition and multiplication, and estimates the time consumption of each operation based on the key length of the encryption algorithm, thereby obtaining the corresponding computational complexity index. If Paillier encryption is used, the bitwise operation overhead of a single addition or multiplication is determined based on the modulus length. If polynomial homomorphic encryption is used, the computational complexity is estimated by the polynomial expansion degree and coefficient size. Based on the ciphertext length of the encrypted feature results (e.g., 2048 bits, 4096 bits) and the required number of interactions, the system estimates the bandwidth usage and latency required during network transmission, thereby obtaining the communication overhead index. The system statistically analyzes the file size of the encrypted feature results in the encrypted state, calculates the space cost required for storage in the distributed ledger, and obtains the storage overhead index. After the above steps, encryption performance evaluation vectors for the first and second encrypted feature results are obtained respectively. To ensure comparability among different indicators, the system normalizes the computational complexity, communication overhead, and storage overhead indicators, mapping them to standardized values ​​within a unified range. For example, each indicator is scaled to the [0,1] interval. This normalization is implemented within the encrypted domain using a pre-defined linear transformation mechanism, ensuring that decryption is not required throughout the process. In this way, the performance evaluation results from both government entities and enterprise entities can serve as input for subsequent fair allocation calculations. Furthermore, the encrypted performance evaluation vector is also encrypted using a homomorphic encryption algorithm after generation, ensuring that it resides within the same encrypted domain as the first and second encrypted feature results.

[0062] The system employs pre-defined fair allocation strategies, such as: based on proportional allocation rules, weighting contribution and performance overhead to ensure that entities with high contributions but excessive performance consumption do not receive excessive rewards; or based on Shapley value allocation rules, calculating the marginal contribution of each participant to the overall performance improvement during collaborative training. Within a pre-defined encrypted domain, the system homomorphically sums the first encrypted feature result with its corresponding performance evaluation value to form the first encrypted contribution result; and homomorphically sums the second encrypted feature result with its corresponding performance evaluation value to form the second encrypted contribution result. For government entities, the first encrypted contribution result reflects the weighted balance between their feature contribution value and performance overhead; for enterprises, the second encrypted contribution result reflects the weighted balance between their feature contribution value and performance overhead. Ultimately, the system obtains, in encrypted form: the first encrypted contribution result for government entities, whose value includes a comprehensive quantification of their encrypted feature contribution and performance consumption; and the second encrypted contribution result for enterprises, whose value is also represented in encrypted form.

[0063] Based on the above technical solution, optionally, within a preset encryption domain, the encryption performance impact assessment and fair allocation rule quantification of the first and second encrypted feature results are performed to generate the first encryption contribution result for the government entity and the second encryption contribution result for the enterprise entity, including: Within a preset encryption domain, the encryption performance impact of the first and second encrypted feature results is evaluated based on preset performance evaluation indicators, resulting in a first performance impact score for the government entity and a second performance impact score for the enterprise entity. Within a preset encrypted domain, the first performance impact score, the second performance impact score, and the preset fair allocation rules are input into a preset game theory allocation model to generate the first encrypted contribution result for the government entity and the second encrypted contribution result for the enterprise entity.

[0064] In this scheme, the preset performance evaluation indicators can refer to the metrics used to measure the performance impact that may occur when calculating, analyzing or transmitting encrypted feature results in the encrypted domain, including but not limited to computation latency, encryption computation complexity, ciphertext size, data transmission load and resource consumption.

[0065] The first performance impact score can refer to the quantitative score calculated based on the first encrypted feature result of the government entity and according to the preset performance evaluation index. It is used to reflect the degree of impact of the encrypted feature result on the system performance during encryption processing and calculation.

[0066] The second performance impact score can refer to the quantitative score calculated based on the second dense state characteristic results of the enterprise entity and according to the preset performance evaluation indicators.

[0067] The preset fair allocation rules can refer to a set of rules used to determine the allocation ratio or weight of the encrypted contribution results of each participant in a multi-party collaborative environment, including but not limited to allocation based on contribution percentage, allocation based on performance score, and weight adjustment based on historical contribution.

[0068] The pre-defined game-theoretic allocation model can refer to a mathematical model based on game theory principles. This model is used to calculate encrypted contribution results based on the encrypted state characteristics, performance impact scores, and fair allocation rules of each party in a multi-party context. This includes Nash equilibrium analysis, optimal strategy calculation, and revenue allocation strategy generation. Through this model, a first encrypted contribution result for the government entity and a second encrypted contribution result for the enterprise entity can be generated, achieving a secure, fair, and verifiable multi-party incentive allocation.

[0069] Within a pre-defined encrypted domain, the first encrypted feature result for the government entity and the second encrypted feature result for the enterprise entity are first obtained. For each encrypted feature result, an encryption performance impact assessment is performed based on pre-defined performance evaluation indicators. Specific steps include: performing encrypted computation and transmission simulations on the encrypted feature results within the encrypted domain, quantifying the computational latency, ciphertext expansion size, resource consumption, and encryption computational complexity generated during processing, and standardizing these indicators; subsequently, generating a first performance impact score for the government entity and a second performance impact score for the enterprise entity based on the standardized results, reflecting the computational efficiency and resource consumption levels of each party under encrypted processing.

[0070] Subsequently, the first performance impact score, the second performance impact score, and the preset fair allocation rule are input into a preset game theory allocation model. This model calculates the allocation strategy for the encrypted contributions of multiple parties based on the principles of payoff distribution and strategy optimization in game theory. Specifically, the game theory allocation model performs the following processing steps: (1) Constructing the participant payoff function: Based on the input performance impact score and fair allocation rules, the model defines payoff functions for the government entity and the enterprise entity respectively. The payoff function is used to characterize the incentive value or payoff quantification index obtained by each participant under different encryption contribution allocation schemes; (2) Solving for the optimal allocation weight: Under the premise of satisfying fairness constraints, encryption performance efficiency requirements and system security, the allocation weight that makes the overall incentive allocation Pareto optimal or near optimal is calculated by solving Nash equilibrium or other optimization algorithms, thereby determining the weight ratio of the government entity and the enterprise entity in the final contribution result; (3) Generating encryption contribution results: According to the obtained optimal allocation weight, combined with the basic value of the encryption state characteristic contribution of each party, the encryption contribution of the government entity and the enterprise entity is quantified in the encryption domain, and the first encryption contribution result of the government entity and the second encryption contribution result of the enterprise entity are generated respectively. The first and second encrypted contribution results not only reflect the magnitude of each participant's contribution but also incorporate performance overhead and fairness adjustment factors during the encryption process. They can be directly used as trusted inputs for subsequent contribution proof generation and incentive allocation smart contracts. Throughout the process, all calculations are performed within a pre-defined encryption domain, ensuring that the original data, intermediate parameters, and final results are all encrypted.

[0071] The training process for the pre-defined game theory allocation model is as follows: First, the model's internal parameters are initialized, including the payoff function structure for each participant, allocation weight parameters, and fairness constraint coefficients. The payoff function quantifies the expected payoffs for government and enterprise entities under different encryption contribution allocation schemes, comprehensively considering indicators such as encrypted feature contribution values, encryption performance overhead, resource consumption, and computational efficiency. Subsequently, historical encryption contribution data, performance impact scores, and fairness allocation rules are acquired to form an encryption training sample set. Within the encryption domain, the base value of encrypted feature contribution, computational latency, ciphertext expansion size, and resource consumption of each sample are standardized for subsequent payoff calculations. In each training iteration, the model simulates multi-party allocation schemes within the encryption domain, using different allocation strategy combinations as candidate strategies. The payoff function values ​​for each strategy combination are calculated, and the Nash equilibrium condition and Pareto improvement constraint are evaluated to ensure incentive balance and fairness for participants under different allocation weights. Based on the iteration results, the model uses an optimization algorithm to adjust the allocation weight parameters, making the overall payoff approach optimal while satisfying fairness constraints, encryption performance efficiency requirements, and system security. During training, the performance overhead of each allocation scheme is dynamically evaluated through encrypted computation and transmission simulations, and this performance metric is fed back to update the revenue function to enhance the model's adaptability to encryption efficiency and resource consumption. After training, the model can calculate the optimal allocation weights within the encrypted domain based on the real-time input performance impact score and fair allocation rules. It also generates the first and second encrypted contribution results by combining the basic values ​​of each party's encrypted state characteristics contribution, thereby quantifying the contributions of government entities and enterprise entities, integrating encryption performance and fairness adjustment factors, and providing trusted, end-to-end encrypted input for subsequent contribution proof generation and incentive allocation smart contracts.

[0072] This scheme effectively preserves the independent feature contributions of each participant, while achieving secure alignment and unified representation of features from multiple parties, providing reliable and privacy-protected basic data for subsequent contribution quantification and incentive allocation.

[0073] S104, obtain the first public parameter corresponding to the first encrypted contribution result and the second public parameter corresponding to the second encrypted contribution result, generate the first contribution certificate of the government entity based on the first encrypted contribution result and the first public parameter, and generate the second contribution certificate of the enterprise entity based on the second encrypted contribution result and the second public parameter.

[0074] The first public parameter can refer to the verification auxiliary information publicly disclosed within the encryption domain, corresponding to the first encrypted contribution result of the government entity. This public parameter is usually generated by a key generation center or a zero-knowledge proof protocol, and includes the public key, generator, homomorphic operation benchmark value, randomness proof auxiliary quantity, etc. in the homomorphic encryption algorithm. It is used to ensure that the correctness and source of the first encrypted contribution result can be publicly verified without disclosing the original contribution data.

[0075] The second public parameter can refer to the verification auxiliary information disclosed in the encryption domain that corresponds to the second encryption contribution result of the enterprise entity. Its composition is similar to that of the first public parameter, mainly including public key, group parameter, randomness commitment value, etc., which is used to support the verification of the verifiability of the second encryption contribution result.

[0076] The first contribution proof can refer to an encrypted proof generated based on the first encrypted contribution result and its corresponding first public parameters, which can prove the true contribution of the government entity in the training or collaborative computation of the federated model. This proof is generated using zero-knowledge proof, verifiable encryption, or commitment mechanisms, ensuring that any verifier can confirm the validity of the government entity's contribution without decrypting the original contribution data.

[0077] The second contribution proof can refer to the encrypted proof generated based on the second encrypted contribution result and its corresponding second public parameter, which can prove the true contribution of the enterprise entity in federated model training or collaborative computing.

[0078] The system invokes a pre-defined encryption parameter generation mechanism to obtain the first public parameter corresponding to the first encrypted contribution result, and similarly obtains the second public parameter corresponding to the second encrypted contribution result. For the first encrypted contribution result, the system uses the first public parameter to construct a commitment value. For example, in the homomorphic encryption domain, the first encrypted contribution result is combined with a randomness auxiliary quantity to generate an immutable commitment. This commitment ensures that external verifiers can confirm that the contribution result of the government entity is consistent with the pre-defined calculation logic. For the second encrypted contribution result, the system performs the same commitment construction operation using the second public parameter, thereby ensuring the verifiability of the enterprise entity's encrypted contribution result.

[0079] The system constructs a first contribution proof based on the first encrypted contribution result and the first public parameter. Specifically, the system constructs a zero-knowledge proof statement logic within the encrypted domain, uses the encrypted contribution result as the input statement for the proof, and applies a zero-knowledge proof protocol (such as Schnorr proof or Bulletproofs) to prove that "the contribution result of the government entity is indeed generated by a preset encrypted calculation process and satisfies the fair allocation rules," without disclosing the plaintext value of the contribution. The system constructs a second contribution proof based on the second encrypted contribution result and the second public parameter to ensure the authenticity and credibility of the enterprise entity's contribution result. This proof also relies on a zero-knowledge protocol, supporting any verifier to verify the result on-chain. The contribution proof mechanism is introduced to solve the "trust deficiency" problem in multi-party collaboration, ensuring that participating parties cannot forge or exaggerate contributions, while avoiding the privacy leakage risks brought by centralized auditing. Ultimately, the system obtains: the first contribution proof of the government entity; and the second contribution proof of the enterprise entity. These contribution proofs are written into the distributed ledger as immutable on-chain evidence.

[0080] In this embodiment of the application, while protecting the privacy of local data of all parties, a contribution proof that can be independently verified is generated through encrypted differential data and encrypted feature processing, providing a credible basis for automated incentives and fair allocation.

[0081] Based on the above technical solution, optionally, before generating the first contribution certificate of the government entity based on the first encrypted contribution result and the first public parameter, and before generating the second contribution certificate of the enterprise entity based on the second encrypted contribution result and the second public parameter, the method further includes: The system obtains the first identity identifier, the first participation data scale, and the first task context information of the government entity, performs a consistency check on the first identity identifier, the first participation data scale, and the first task context information, and obtains a first check result. The system also obtains the second identity identifier, the second participation data scale, and the second task context information of the enterprise entity, performs a consistency check on the second identity identifier, the second participation data scale, and the second task context information, and obtains a second check result. Accordingly, a first contribution certificate for the government entity is generated based on the first encrypted contribution result and the first public parameter, and a second contribution certificate for the enterprise entity is generated based on the second encrypted contribution result and the second public parameter, including: If the first verification result and the second verification result meet the preset verification conditions, a first contribution commitment value bound to the government entity is generated based on the first encrypted contribution result and the first public parameter, and a second contribution commitment value bound to the enterprise entity is generated based on the second encrypted contribution result and the second public parameter. Based on a preset contribution calculation model, a first contribution commitment value, and a first public parameter, a first verifiable proof is generated by calling a preset zero-knowledge proof protocol; and based on a preset contribution calculation model, a second contribution commitment value, and a second public parameter, a second verifiable proof is generated by calling a preset zero-knowledge proof protocol. Perform integrity encapsulation on the first verifiable proof and the second verifiable proof to obtain the first contribution proof of the government entity and the second contribution proof of the enterprise entity.

[0082] In this scheme, the first identity identifier can be the identity information used to uniquely identify the government entity, which may include the user ID assigned by the government system, public key information, or other unique authentication credentials.

[0083] The first measure of data contribution can be local data metrics provided by government entities in federated learning or multi-party collaborative tasks, such as the number of samples, data feature dimensions, or the total number of data entries, used to assess the magnitude of their contribution.

[0084] The first task context information can describe the background information of the government entity's participation in the task, including task type, training round, task start time and end time, etc., which is used to assist in consistency verification and contribution calculation.

[0085] The first verification result can be the result obtained after performing a consistency check on the first identity identifier, the first participating data scale, and the first task context information. It is used to determine whether the information submitted by the government entity is complete, authentic, and conforms to preset rules.

[0086] The second identity identifier can be identity information used to uniquely identify the enterprise entity, and may include user IDs assigned by the enterprise system, public key information, or other unique authentication credentials.

[0087] The second measure of data contribution can be local data metrics provided by the enterprise entity in federated learning or multi-party collaborative tasks, such as the number of samples, data feature dimensions, or the total number of data entries, used to assess the magnitude of their contribution.

[0088] The second task context information can describe the background information of the enterprise's participation in the task, including task type, training round, task start time and end time, etc., which is used to assist in consistency verification and contribution calculation.

[0089] The second verification result can be the result obtained after performing a consistency check on the second identity identifier, the second participating data scale, and the second task context information. It is used to determine whether the information submitted by the enterprise entity is complete, authentic, and conforms to preset rules.

[0090] The preset verification conditions can be a set of rules or thresholds used to determine the legality of the participant's information, including the uniqueness of identity information, the integrity of data scale and task context, format standardization, and consistency with system records.

[0091] The first contribution commitment value can be an encrypted commitment bound to the government entity, generated based on the first encrypted contribution result and the first public parameter of the government entity, used to prove that its contribution in the task is quantified and cannot be tampered with.

[0092] The second contribution commitment value can be an encrypted commitment bound to the enterprise entity, generated based on the enterprise entity's second encrypted contribution result and second public parameters.

[0093] The preset contribution calculation model can be a mathematical or computational model used to map contribution commitment values ​​and public parameters to verifiable proofs. It supports quantitative calculation, proof generation, and consistency verification, and can include weighting functions, mapping rules, and hash constraints.

[0094] A pre-defined zero-knowledge proof protocol can be a cryptographic protocol that enables participants to generate verifiable proofs of the correctness of their contributions or operations without exposing the actual data content, while maintaining privacy and security.

[0095] The first verifiable proof can be a verifiable proof generated by the government entity using a zero-knowledge proof protocol based on the first contribution commitment value and a preset contribution calculation model, proving the authenticity and consistency of its contribution quantification results.

[0096] The second verifiable proof can be a verifiable proof generated by the enterprise entity using a zero-knowledge proof protocol based on the second contribution commitment value and a preset contribution calculation model.

[0097] Within the pre-defined encrypted domain, the first identity identifier, the first participating data scale, and the first task context information of the government entity are first obtained. The first identity identifier consists of a unique user ID, digital certificate, or public key generated by the government system, used to uniquely identify the entity within the system. The first participating data scale includes the number of data samples provided by the government entity in the federated model training, the dimension of each data item, and the total number of data entries. The first task context information includes the task type, training rounds, participation time window, and task stage information. After obtaining the above information, consistency verification is performed within the encrypted domain. The specific steps are as follows: First, the first identity identifier is compared with the user ID and public key of the registration center to verify its uniqueness and legality. Then, the first participating data scale is compared with the minimum data volume, data format, and dimensional constraints in the task requirements to confirm that the data volume and structure conform to the specifications. Next, the first task context information is matched for time, task type, and stage to ensure consistency with the task plan and historical records. Finally, the comparison result is recorded as the first verification result, the value of which indicates whether the pre-defined verification conditions are met (e.g., "valid" if all verification items pass, otherwise "invalid"). Accordingly, the second identity identifier of the enterprise entity, the second participating data scale, and the second task context information are obtained, and the same comparison and verification steps are performed in the encrypted domain to obtain the second verification result.

[0098] Once both the first and second verification results meet the preset verification conditions, a first contribution commitment value bound to the government entity is generated within the encrypted domain based on the first encrypted contribution result and the first public parameter. Specifically, an encrypted commitment algorithm (such as Pedersen commitment) is used to combine the first encrypted contribution result with the first public parameter to calculate an immutable commitment value. A timestamp is recorded, and a random blind factor is added to ensure zero-knowledge characteristics. Similarly, a second contribution commitment value bound to the enterprise entity is generated based on the second encrypted contribution result and the second public parameter, using the same algorithm to ensure security and immutability. Subsequently, within the encrypted domain, a preset contribution calculation model is used to quantify each contribution commitment value. This model includes: a contribution mapping function that maps the commitment value to a standardized contribution range; a weighted quantization function that generates quantified values ​​based on the contribution weights of each party and encryption performance indicators; and hash constraints used to generate verifiable digests to prevent tampering.

[0099] After the calculation is completed, a preset zero-knowledge proof protocol is invoked to execute the commitment-challenge-response process: the system generates a random challenge, the participants use their contribution commitment value to generate a response, and the zero-knowledge proof algorithm is combined to generate a first verifiable proof and a second verifiable proof, ensuring that the authenticity of the contribution can be verified without disclosing the original contribution data. Finally, the generated verifiable proofs are completely encapsulated, including calculating a hash digest of the proof content, generating a digital signature, attaching a timestamp, and packaging them into the final output: the first contribution proof for the government entity and the second contribution proof for the enterprise entity.

[0100] The training process for the pre-defined contribution calculation model is as follows: First, multi-party encrypted contribution data from historical federated training tasks are collected, including the encrypted contribution results of each party, corresponding public parameters, actual data scale, task context information, and historical contribution results verified by zero-knowledge proof. Then, these data are feature-encoded within the encrypted domain, mapping the contribution results to numerical feature vectors and the public parameters, data scale, and task context information to standardized auxiliary vectors. Next, initial parameters for the quantization mapping function and weighted quantization function are constructed and embedded into the model structure within the encrypted domain. Subsequently, the model parameters are iteratively updated within the encrypted domain using gradient descent or other optimization algorithms to minimize the error between the model's predicted contribution quantization value and the historically verified contribution, while adding regularization constraints to ensure the model's robustness under different data scales and task contexts. In each iteration, the mapping output of the contribution commitment value under the model is simulated through encrypted operations to generate an initial quantization value. Then, a hash constraint is applied to the quantization value (e.g., calculating H(quantization value || context)) and compared with historical verification records to ensure the output is immutable and verifiable. Simultaneously, a zero-knowledge proof protocol is used to verify that the model output satisfies preset logical rules without revealing plaintext. After training, the model can automatically generate quantized contribution values ​​by inputting new contribution commitment values ​​and public parameters in the encrypted domain. At the same time, it maintains compatibility with zero-knowledge proof protocols, ensuring that the generated quantized contributions can be used for subsequent contribution proof generation and incentive allocation. Meanwhile, the original data and intermediate calculations are encrypted throughout the process, meeting the requirements for end-to-end privacy protection and verifiability.

[0101] This solution ensures the authenticity and immutability of contributions, and also supports end-to-end privacy protection and verifiable incentive allocation, providing a reliable basis for subsequent smart contract execution and auditing.

[0102] Figure 2 A flowchart illustrating the contribution evaluation method for federal models used in government-enterprise collaboration provided in this disclosure embodiment. The method may include the following steps: S201, obtain the first model update data generated by the government entity when training the federated model based on the first local data, and the second model update data generated by the enterprise entity when training the federated model based on the second local data, and encrypt the first model update data and the second model update data using a verifiable encryption method to obtain the first encrypted differential data and the second encrypted differential data.

[0103] S202, within a preset encryption domain, perform encrypted feature fusion processing on the first encrypted differential data and the second encrypted differential data to obtain the first encrypted feature result of the government entity and the second encrypted feature result of the enterprise entity.

[0104] S203, within a preset encryption domain, the encryption performance impact assessment and fair allocation rule quantification of the first and second encrypted feature results are performed to generate the first encryption contribution result of the government entity and the second encryption contribution result of the enterprise entity.

[0105] S204, obtain the first public parameter corresponding to the first encrypted contribution result and the second public parameter corresponding to the second encrypted contribution result, generate the first contribution certificate of the government entity based on the first encrypted contribution result and the first public parameter, and generate the second contribution certificate of the enterprise entity based on the second encrypted contribution result and the second public parameter.

[0106] S205, the first contribution proof and the second contribution proof are submitted to the blockchain network respectively to obtain the first on-chain evidence record and the second on-chain evidence record stored in the distributed ledger.

[0107] A blockchain network can refer to a decentralized distributed computing network composed of multiple nodes. The nodes maintain the consistency of the ledger based on consensus algorithms (such as PoS, PBFT, Raft, etc.) to ensure that the written data is immutable and traceable. It typically uses cryptographic hashes, digital signatures, and chained data structures to achieve secure storage and verification.

[0108] A distributed ledger can refer to a data storage structure in a blockchain network that is jointly maintained by multiple nodes. Its characteristics are that a copy of the ledger is stored synchronously on all participating nodes, any addition or modification must be confirmed through consensus, and tamper-proof and full traceability are achieved through the timestamps and hash pointers of the blocks.

[0109] The first on-chain evidence record can refer to the on-chain data record generated after the first contribution proof submitted by the government entity is packaged into a block by the blockchain network and confirmed by consensus. This record includes the hash digest of the contribution proof, the submission timestamp, the identity of the submitter, and the unique block index number generated by the blockchain.

[0110] The second on-chain evidence record can refer to the on-chain data record generated after the second contribution proof submitted by the enterprise entity is packaged into a block by the blockchain network and confirmed by consensus.

[0111] First, the first and second contribution proofs are hashed separately, and corresponding hash digests are generated using secure hash algorithms (such as SHA-256 or SM3) for subsequent on-chain verification of data integrity. Next, a digital signature of the submitter is appended to each proof and its hash digest. The digital signature is generated using the submitter's private key combined with an elliptic curve signature algorithm (such as ECDSA or SM2) to ensure the authenticity and non-repudiation of the proof's source. Subsequently, the system packages the signed data into a blockchain transaction request. The transaction request structure includes: the transaction initiator's identity, transaction timestamp, contribution proof hash, contribution proof ciphertext digest, and digital signature. This transaction request is propagated to the entire blockchain network via node broadcasting. After receiving the transaction, all nodes verify its validity using a preset consensus algorithm (such as PBFT or Raft). The verification steps include: checking the legality of the transaction signature, the uniqueness of the contribution proof hash, and whether the submission time is within a preset time limit. Verified transactions are packaged into a new block, along with a block timestamp, a pointer to the previous block's hash, and a Merkle root hash, ensuring the integrity and immutability of the blockchain structure. After the new block is constructed, the system generates a unique transaction index number for each transaction within the block and establishes a hash mapping between the transaction index number and the transaction content for rapid subsequent retrieval and verification. For transaction requests containing a first contribution proof submitted by a government entity, its entry in the new block forms the first on-chain evidence record; for transaction requests containing a second contribution proof submitted by a business entity, its entry in the new block forms the second on-chain evidence record. Each on-chain evidence record is stored in the distributed ledger as an immutable data entry, containing, but not limited to: transaction index number, submitter's identity, contribution proof hash digest, contribution proof ciphertext digest, digital signature verification result, block height, block timestamp, and Merkle path information. The transaction index number uniquely identifies the notarized entry; the submitter's identity ensures the traceability of the notarized source; the contribution proof hash digest and ciphertext digest are used for subsequent integrity verification; the digital signature verification result indicates the authenticity status of the notarized entry; the block height and block timestamp establish the ledger position and time sequence of the notarized entry; and the Merkle path information guarantees the verifiability of the transaction within the block. Once a new block is officially written into the distributed ledger, the notarized records on the first and second chains are permanently fixed in the blockchain data structure. Any subsequent node can verify the integrity, authenticity, and immutability of the notarized records based on the hash chain, Merkle tree verification path, and digital signature.

[0112] S206, based on the first digest information of the first on-chain evidence record and the second digest information of the second on-chain evidence record, perform a secure aggregation operation to obtain a joint commitment value, and write the joint commitment value into the distributed ledger to obtain a joint evidence record.

[0113] The first digest information can refer to the set of digest fields extracted from the first on-chain evidence record for subsequent aggregation calculations, including but not limited to: the hash digest, ciphertext digest, transaction index number, and Merkle path digest of the first contribution proof. Its purpose is to provide a lightweight representation that allows for aggregation calculations and verifiability without exposing the complete contribution proof content.

[0114] Second Digest Information: This refers to the set of digest fields extracted from the second chain's on-chain evidence records for subsequent aggregation calculations, including but not limited to: hash digest of the second contribution proof, ciphertext digest, transaction index number, and Merkle path digest.

[0115] A joint commitment value refers to a unified commitment result generated within a preset encrypted domain after performing a secure aggregation operation on the first digest information and the second digest information. This commitment value is usually calculated through homomorphic encryption addition operations (such as Paillier homomorphic encryption) or commitment superposition algorithms (such as Pedersen commitment aggregation). It possesses immutability and zero-knowledge characteristics, and can uniformly bind the evidence storage digests of government entities and enterprise entities without disclosing the original digest information, for use in subsequent trusted verification of joint evidence storage.

[0116] Within a preset encrypted domain, first and second digest information can be extracted from the first and second on-chain evidence records, respectively. The extracted content includes the hash digest, ciphertext digest, transaction index number, and corresponding Merkle path digest of each contribution proof. To ensure the security of the digest information during aggregation computation, a second digest value needs to be calculated again using a secure hash function (such as SHA-256 or SM3) after extraction to ensure the uniqueness and collision resistance of the input data.

[0117] Next, secure aggregation operations are performed within the encrypted domain. Specifically: if the Paillier homomorphic encryption algorithm is used, the system first encrypts the first and second digest information using a public encryption key to obtain two encrypted digests; then, homomorphic addition is performed in the ciphertext space, directly adding the encrypted first and second digest information to obtain a joint commitment value of the encrypted states; finally, the system partially decrypts or performs zero-knowledge verification on the joint commitment value of the encrypted states using a private key or a decentralized decryption authorization mechanism to generate a publicly verifiable joint commitment value. Another approach is an aggregation algorithm based on Pedersen commitments: unilateral commitment values ​​are generated using the digest information of each party and independent random blind factors, and then the two commitment values ​​are merged through group operations (such as addition on elliptic curve groups) to obtain a joint commitment value. This value inherently possesses binding and concealment properties, ensuring that the original digest information cannot be forged or deduced.

[0118] After generating the joint commitment value, the system constructs a new blockchain transaction request. This request includes the joint commitment value, a timestamp indicating its generation, a set of participant identities, an aggregation algorithm identifier, and a random factor hash for verification. The transaction request is broadcast to the blockchain network by nodes, and all nodes verify its validity using a consensus algorithm (such as PBFT or Raft). Verification steps include checking if the joint commitment value was generated from valid digest information, if the random factor hash matches, and if the submitter identity set is valid. Verified transactions are packaged into a new block, appended with the block timestamp, the previous block hash pointer, and the Merkle root hash, ensuring data immutability and chain consistency. Finally, the new transaction entry written to the distributed ledger constitutes the joint notarization record.

[0119] S207, based on the first contribution proof, a preset incentive allocation smart contract is invoked to generate a first incentive settlement record for the government entity, and based on the second contribution proof, a preset incentive allocation smart contract is invoked to generate a second incentive settlement record for the enterprise entity.

[0120] A pre-defined incentive distribution smart contract refers to a contract program pre-deployed in a blockchain network, written in a smart contract programming language (such as Solidity or Rust), and running on a blockchain virtual machine (such as Ethereum EVM or SubstrateVM). This contract has built-in contribution calculation rules, reward distribution functions, settlement trigger conditions, and exception handling mechanisms. Upon receiving proof of contribution, it automatically calculates and settles the incentive amount, points, or resources for participants according to predefined algorithmic rules.

[0121] The first incentive settlement record can refer to the on-chain settlement result generated by the government entity through the incentive allocation smart contract based on its first contribution proof. The record includes information such as the government entity's identity identifier, the corresponding contribution proof hash, the incentive amount or points value, the generation timestamp, and the contract transaction hash.

[0122] The second incentive settlement record can refer to the on-chain settlement result generated by the enterprise entity through the incentive allocation smart contract based on its second contribution proof. The record includes information such as the enterprise entity's identity identifier, the corresponding contribution proof hash, the incentive amount or points value, the generation timestamp, and the contract transaction hash.

[0123] Before submitting the contribution proof to the smart contract, the system first performs hash verification on the proof. A secure hash algorithm (such as SHA-256 or SM3) is used to calculate the proof digest, which is then compared with the digest record stored on the blockchain to ensure the proof content has not been tampered with. Simultaneously, the system verifies the submitter's digital signature. The signature is verified using an elliptic curve signature algorithm (such as ECDSA or SM2) combined with the public key to ensure the authenticity of the proof's origin. The system invokes the incentive allocation function of the smart contract via a blockchain transaction request. The transaction request structure includes: the contribution proof hash, the submitter's identity identifier, the transaction timestamp, and the signature credential. The transaction request is broadcast across the blockchain network to all nodes, where the contract function is executed by the virtual machine environment of each node. During execution, the smart contract first reads the commitment value and contribution parameters from the contribution proof and then invokes the incentive allocation rules defined within the contract. The rules can include: a contribution weighting function: calculating a weighted average based on the participant's contribution and the task's weight; a reward mapping function: mapping the weighted contribution value to a specific reward amount (e.g., tokens, points, or resource shares); and a threshold checking mechanism: ensuring that participants with contributions below a minimum threshold are not allocated rewards or receive fixed compensation. After the contract calculation is complete, the settlement result is output, including the participant's identity, reward amount, and the hash binding of the corresponding contribution proof. After executing the reward calculation, the contract automatically generates an on-chain transaction result, which is the incentive settlement record. For government entities, the generated result is the first incentive settlement record, containing the government entity's identity identifier, contribution proof hash, reward amount, settlement timestamp, and transaction hash; for enterprises, the generated result is the second incentive settlement record, containing the enterprise entity's identity identifier, contribution proof hash, reward amount, settlement timestamp, and transaction hash. After all nodes execute the smart contract, they confirm the validity of the transaction results using a consensus algorithm (such as PBFT or Raft), package the settlement record into a new block, and attach a block timestamp, Merkle root hash, and a pointer to the previous block's hash to ensure that the settlement record is immutable and traceable in the distributed ledger. Ultimately, the government entity and the enterprise entity respectively obtain the first incentive settlement record and the second incentive settlement record, which are bound to their identity and contribution proof. These records serve as trusted on-chain credentials.

[0124] In this embodiment, independent on-chain storage, secure aggregation verification, and automatic smart contract incentives can be achieved for government and enterprise contributions, balancing privacy protection and collaborative trust, and improving the transparency, security, and automation of cross-organizational collaboration.

[0125] Based on the above technical solution, optionally, after generating the second incentive settlement record for the enterprise entity, the method further includes: In the pre-set incentive allocation smart contract, the built-in rule engine performs incentive decision analysis on the first contribution index of the first contribution proof and the second contribution index of the second contribution proof to determine the first incentive parameters for the government entity and the second incentive parameters for the enterprise entity. Based on the preset incentive allocation smart contract, the incentive allocation operation is performed according to the first incentive parameter and the second incentive parameter to generate the first on-chain incentive certificate for the government entity and the second on-chain incentive certificate for the enterprise entity.

[0126] In this solution, the built-in rule engine refers to a decision-making computation component pre-installed in the incentive allocation smart contract, used for parameter parsing and rule matching of the input contribution proof. Its specific implementation is typically based on rule matching algorithms (such as conditional decision trees, Boolean logic tables, or rule matching computation based on weight matrices), capable of automatically analyzing the contribution indicators of different entities and outputting corresponding incentive parameters.

[0127] The primary contribution metric can refer to the set of core parameters extracted from the proof of primary contribution, used to measure the actual contribution of the government entity. Examples include: number of data samples, data quality score, task completion rate, and training accuracy improvement rate.

[0128] The second contribution metric can refer to the set of core parameters extracted from the second contribution proof, used to measure the actual contribution of the enterprise entity, such as: the amount of computing resources provided, data coverage, and the improvement rate of model convergence speed.

[0129] The first incentive parameter can refer to the numerical or weighted parameter generated by the built-in rule engine after performing analysis and rule matching on the first contribution indicator, which can be used for incentive allocation, such as reward points, token quantity, or resource allocation share.

[0130] The second incentive parameter can refer to the numerical or weight parameters generated by the built-in rule engine after performing analysis and rule matching on the second contribution metric, which can be used for incentive allocation.

[0131] The first on-chain incentive certificate can refer to the on-chain incentive record generated by the government entity in the preset incentive allocation smart contract based on its first contribution proof and the smart contract's built-in rule engine. The content includes the government entity's identity identifier, the hash value of the corresponding first contribution proof, the value of the first incentive parameter, the incentive allocation amount or points value, the transaction timestamp, and the smart contract transaction hash, etc.

[0132] The second on-chain incentive certificate can refer to the on-chain incentive record generated by the enterprise entity in the preset incentive allocation smart contract based on its second contribution proof and the smart contract's built-in rule engine. The content includes the enterprise entity's identity identifier, the hash value of the corresponding second contribution proof, the value of the second incentive parameter, the incentive allocation amount or points value, the transaction timestamp, and the smart contract transaction hash, etc.

[0133] In the pre-defined incentive allocation smart contract, the first contribution indicator of the first contribution proof and the second contribution indicator of the second contribution proof are first parsed into a data format recognizable by the smart contract. Within the contract, a built-in rule engine is invoked to perform rule matching and logical reasoning on the parsed indicators. Specifically, this includes: using Boolean logic and conditional judgments to compare each contribution indicator with a pre-defined incentive threshold; combining multi-dimensional scoring functions (such as weighted integral calculation or standardized contribution ratio calculation) to generate a quantitative incentive decision basis; simultaneously using priority queues and weighted scheduling algorithms to handle conflicts or deviations between different contribution dimensions, ensuring that the contribution value of each entity can be comprehensively evaluated according to fairness, efficiency, and security; based on the evaluation results, the first and second contribution indicators are mapped to corresponding first and second incentive parameters. The incentive parameters may include incentive amount, point value, or token allocation ratio, and are accompanied by timestamps, random blind factors, and smart contract call records to ensure on-chain immutability and traceability. After determining the incentive parameters, the smart contract uses a signature mechanism within the cryptographic domain to perform integrity verification on each incentive allocation: the incentive parameters are bound to the participant's identity, contribution proof hash, and current blockchain state hash to generate a transaction object, and the participant's public key is used to encrypt the signature to ensure the authenticity and non-repudiation of the transaction; subsequently, the smart contract constructs an on-chain transaction based on the incentive parameters and bound data, and broadcasts the transaction to blockchain network nodes. The nodes use consensus algorithms (such as PBFT or Raft) to verify the legality of the transaction, the validity of the signature, and the consistency of the contribution proof hash. After successful verification, the smart contract writes the transaction data into a new block and generates corresponding first and second on-chain incentive certificates. Each certificate contains the subject's identity, contribution proof hash, incentive parameter value, allocation amount or points value, generation timestamp, transaction hash, and contract call record, thereby forming a traceable, transparent, and tamper-proof on-chain incentive record in the distributed ledger.

[0134] This solution enables transparent, fair, and traceable incentive allocation, generates highly efficient and reliable on-chain credentials, and protects the privacy of contribution data.

[0135] Based on the above technical solution, optionally, after generating the first on-chain incentive certificate for the government entity and the second on-chain incentive certificate for the enterprise entity, the method further includes: Extract the first incentive parameters from the first chain incentive certificate, perform a difference analysis on the first incentive parameters and the first contribution index, and obtain the first difference analysis result. Extract the second incentive parameters from the incentive certificate on the second chain, perform a difference analysis on the second incentive parameters and the second contribution index, and obtain the second difference analysis results. Based on the results of the first and second difference analyses, the parameters of the built-in rule engine in the preset incentive allocation smart contract are adjusted.

[0136] In this scheme, the first incentive parameter can refer to the parameter value extracted from the first on-chain incentive certificate of the government entity to quantify the incentive allocation amount of the government entity. It usually includes incentive amount, points or other reward quantification indicators, which are used to reflect the actual incentive allocation of the government entity in the contribution proof.

[0137] The first difference analysis result can refer to the quantitative result of the difference obtained by comparing and analyzing the first incentive parameter with the corresponding first contribution index. It is used to evaluate the deviation between the actual incentive allocation and the expected contribution, including information such as absolute difference, relative deviation or standardized error.

[0138] The second incentive parameter can refer to the parameter value extracted from the incentive certificate on the second chain of the enterprise entity, which is used to quantify the incentive allocation amount of the enterprise entity.

[0139] The second difference analysis result can refer to the quantitative result of the difference obtained by comparing and analyzing the second incentive parameter with the corresponding second contribution index.

[0140] The parameters of the built-in rules engine can refer to the adjustable values ​​or weights in the preset incentive allocation smart contract used to control the incentive decision-making logic and allocation strategy, including parameters such as threshold settings, contribution weight, proportional coefficient, and incentive upper / lower limits. By adjusting these parameters, the incentive allocation strategy can be optimized and incentive deviations can be corrected.

[0141] In the pre-defined incentive allocation smart contract, the first incentive parameter is parsed and extracted from the first on-chain incentive certificate. Specifically, this involves reading the smart contract transaction data stored on-chain, including the government entity's identity identifier, the corresponding contribution proof hash, and the incentive allocation numerical field. The incentive amount or points value field is then mapped to a standardized numerical vector to obtain the quantified first incentive parameter. Subsequently, a difference analysis is performed between the first incentive parameter and the first contribution indicator corresponding to the government entity. This is done by first normalizing both the first incentive parameter and the first contribution indicator to eliminate dimensional differences; then, the difference is calculated using difference calculation or relative deviation formulas, such as calculating the absolute difference or relative error for each indicator dimension, forming the first difference analysis result, which quantifies the deviation between the actual incentive allocation and the expected contribution. Similarly, the second incentive parameter is parsed and extracted from the second on-chain incentive certificate. The enterprise entity's identity identifier, contribution proof hash, and incentive allocation numerical field are read. After mapping the incentive value to a standardized vector, a difference analysis is performed with the corresponding second contribution indicator. The second difference analysis result is obtained through normalization, difference calculation, and relative deviation calculation. After obtaining the results of the first and second difference analyses, the parameters of the built-in rule engine in the preset incentive allocation smart contract are adjusted. Specifically, this involves: identifying incentive deviation patterns based on the difference analysis results; calculating the correction coefficient for each rule weight or threshold; mapping the absolute difference or relative error to the adjustment amount of the contribution weight, proportional coefficient, or incentive upper and lower limits in the rule engine; and updating the parameter values ​​stored in the smart contract. This achieves adaptive optimization of the incentive allocation strategy, making future incentive allocations closer to the actual contributions of each participant, ensuring fairness and system sustainability. The entire process is executed in an encrypted environment of the blockchain ledger and smart contract, ensuring the integrity of on-chain data and the verifiability of operations.

[0142] This solution can dynamically adjust incentive allocation rules, making incentive parameters more accurately match actual contributions, improving allocation fairness and transparency, while maintaining on-chain data security and verifiability.

[0143] Figure 3 This disclosure provides a schematic block diagram of a federal model contribution evaluation system for government-enterprise collaboration, as provided in an embodiment. The system comprises: The data encryption module 301 is used to obtain the first model update data generated by the government entity when training the federated model based on the first local data, and the second model update data generated by the enterprise entity when training the federated model based on the second local data. The first model update data and the second model update data are encrypted using a verifiable encryption method to obtain the first encrypted differential data and the second encrypted differential data. The encrypted feature fusion module 302 is used to perform encrypted feature fusion processing on the first encrypted differential data and the second encrypted differential data within a preset encrypted domain to obtain the first encrypted feature result of the government entity and the second encrypted feature result of the enterprise entity. The encryption contribution result generation module 303 is used to evaluate the encryption performance impact of the first encryption state feature result and the second encryption state feature result and quantify the fair allocation rules within a preset encryption domain, and generate the first encryption contribution result of the government entity and the second encryption contribution result of the enterprise entity. The contribution proof generation module 304 is used to obtain the first public parameter corresponding to the first encrypted contribution result and the second public parameter corresponding to the second encrypted contribution result, generate the first contribution proof of the government entity based on the first encrypted contribution result and the first public parameter, and generate the second contribution proof of the enterprise entity based on the second encrypted contribution result and the second public parameter.

[0144] like Figure 4 As shown, this application embodiment also provides an electronic device 400, including a processor 401, a memory 402, and a program or instructions stored in the memory 402 and executable on the processor 401. When the program or instructions are executed by the processor 401, they implement the various processes of the above-described embodiment of the federal model contribution evaluation method for government-enterprise collaboration and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0145] It should be noted that the electronic devices in the embodiments of this application include the mobile electronic devices and non-mobile electronic devices described above.

[0146] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described embodiment of the federal model contribution evaluation system for government-enterprise collaboration and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0147] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0148] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element. Furthermore, it should be noted that the scope of the methods and systems in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in reverse order, depending on the functions involved.

[0149] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0150] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

[0151] The above description is merely a preferred embodiment and the technical principles employed in this application. This application is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions that can be made by those skilled in the art will not depart from the scope of protection of this application. Therefore, although this application has been described in detail through the above embodiments, this application is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of this application, the scope of which is determined by the scope of the claims.

Claims

1. A method for evaluating the contribution of a federal model for government-enterprise collaboration, characterized in that, The method includes: The system obtains the first model update data generated by the government entity when training the federated model based on the first local data, and the second model update data generated by the enterprise entity when training the federated model based on the second local data. The first model update data and the second model update data are encrypted using a verifiable encryption method to obtain the first encrypted differential data and the second encrypted differential data. Within a preset encryption domain, the first encrypted differential data and the second encrypted differential data are subjected to encrypted feature fusion processing to obtain the first encrypted feature result of the government entity and the second encrypted feature result of the enterprise entity. Within a preset encryption domain, the encryption performance impact of the first and second encrypted feature results is evaluated and the fair allocation rules are quantified to generate the first encryption contribution result of the government entity and the second encryption contribution result of the enterprise entity. Obtain the first public parameter corresponding to the first encrypted contribution result and the second public parameter corresponding to the second encrypted contribution result; generate a first contribution certificate for the government entity based on the first encrypted contribution result and the first public parameter; and generate a second contribution certificate for the enterprise entity based on the second encrypted contribution result and the second public parameter.

2. The method according to claim 1, characterized in that, in, Within a preset encryption domain, the first encrypted differential data and the second encrypted differential data undergo cryptographic feature fusion processing to obtain the first cryptographic feature result for the government entity and the second cryptographic feature result for the enterprise entity, including: A shallow nonlinear transformation is performed on the first encrypted differential data within a preset encryption domain to obtain the first preliminary encrypted feature representation of the government entity; and a shallow nonlinear transformation is performed on the second encrypted differential data within a preset encryption domain to obtain the second preliminary encrypted feature representation of the enterprise entity. Within a preset encryption domain, an affine transformation is performed on the first preliminary encrypted feature representation and the second preliminary encrypted feature representation based on preset global statistical parameters to obtain the first normalized encrypted feature representation of the government entity and the second normalized encrypted feature representation of the enterprise entity. Within a preset encryption domain, the first normalized encrypted feature representation and the second normalized encrypted feature representation are projected onto a unified encrypted semantic space through a preset linear mapping matrix, respectively, to obtain the first mapped encrypted feature representation of the government entity and the second mapped encrypted feature representation of the enterprise entity. Within a preset encryption domain, the first mapping cryptographic feature representation and the second mapping cryptographic feature representation are subjected to dimensional coordination processing to obtain the first aligned cryptographic feature representation of the government entity and the second aligned cryptographic feature representation of the enterprise entity. Within a preset encrypted domain, the first aligned encrypted feature representation and the second aligned encrypted feature representation are weighted and combined according to a preset fusion weighting rule to obtain the first encrypted feature result of the government entity and the second encrypted feature result of the enterprise entity.

3. The method according to claim 1, characterized in that, in, Within a preset encryption domain, the encryption performance impact of the first and second encrypted feature results is evaluated, and fair allocation rules are quantified to generate the first encryption contribution result for government entities and the second encryption contribution result for enterprise entities, including: Within a preset encryption domain, the encryption performance impact of the first and second encrypted feature results is evaluated based on preset performance evaluation indicators, resulting in a first performance impact score for the government entity and a second performance impact score for the enterprise entity. Within a preset encrypted domain, the first performance impact score, the second performance impact score, and the preset fair allocation rules are input into a preset game theory allocation model to generate the first encrypted contribution result for the government entity and the second encrypted contribution result for the enterprise entity.

4. The method according to claim 1, characterized in that, in, Before generating a first contribution certificate for the government entity based on the first encrypted contribution result and the first public parameter, and before generating a second contribution certificate for the enterprise entity based on the second encrypted contribution result and the second public parameter, the method further includes: The system obtains the first identity identifier, the first participation data scale, and the first task context information of the government entity, performs a consistency check on the first identity identifier, the first participation data scale, and the first task context information, and obtains a first check result. The system also obtains the second identity identifier, the second participation data scale, and the second task context information of the enterprise entity, performs a consistency check on the second identity identifier, the second participation data scale, and the second task context information, and obtains a second check result. Accordingly, a first contribution certificate for the government entity is generated based on the first encrypted contribution result and the first public parameter, and a second contribution certificate for the enterprise entity is generated based on the second encrypted contribution result and the second public parameter, including: If the first verification result and the second verification result meet the preset verification conditions, a first contribution commitment value bound to the government entity is generated based on the first encrypted contribution result and the first public parameter, and a second contribution commitment value bound to the enterprise entity is generated based on the second encrypted contribution result and the second public parameter. Based on a preset contribution calculation model, a first contribution commitment value, and a first public parameter, a first verifiable proof is generated by calling a preset zero-knowledge proof protocol; and based on a preset contribution calculation model, a second contribution commitment value, and a second public parameter, a second verifiable proof is generated by calling a preset zero-knowledge proof protocol. Perform integrity encapsulation on the first verifiable proof and the second verifiable proof to obtain the first contribution proof of the government entity and the second contribution proof of the enterprise entity.

5. The method according to claim 1, characterized in that, in, After generating the second contribution proof of the enterprise entity based on the second encrypted contribution result and the second public parameters, the method further includes: The first contribution proof and the second contribution proof are submitted to the blockchain network respectively to obtain the first on-chain evidence record and the second on-chain evidence record stored in the distributed ledger. Based on the first digest information of the first on-chain evidence record and the second digest information of the second on-chain evidence record, a secure aggregation operation is performed to obtain a joint commitment value, and the joint commitment value is written into the distributed ledger to obtain a joint evidence record; Based on the first contribution proof, a preset incentive allocation smart contract is invoked to generate a first incentive settlement record for the government entity, and based on the second contribution proof, a preset incentive allocation smart contract is invoked to generate a second incentive settlement record for the enterprise entity.

6. The method according to claim 5, characterized in that, in, After generating the second incentive settlement record for the corporate entity, the method further includes: In the pre-set incentive allocation smart contract, the built-in rule engine performs incentive decision analysis on the first contribution index of the first contribution proof and the second contribution index of the second contribution proof to determine the first incentive parameters for the government entity and the second incentive parameters for the enterprise entity. Based on the preset incentive allocation smart contract, the incentive allocation operation is performed according to the first incentive parameter and the second incentive parameter to generate the first on-chain incentive certificate for the government entity and the second on-chain incentive certificate for the enterprise entity.

7. The method according to claim 6, characterized in that, in, After generating the first on-chain incentive certificate for the government entity and the second on-chain incentive certificate for the enterprise entity, the method further includes: Extract the first incentive parameters from the first chain incentive certificate, perform a difference analysis on the first incentive parameters and the first contribution index, and obtain the first difference analysis result. Extract the second incentive parameters from the incentive certificate on the second chain, perform a difference analysis on the second incentive parameters and the second contribution index, and obtain the second difference analysis results. Based on the results of the first and second difference analyses, the parameters of the built-in rule engine in the preset incentive allocation smart contract are adjusted.

8. A contribution evaluation system for a federal model of government-enterprise collaboration, characterized in that, The system includes: The data encryption module is used to obtain the first model update data generated by the government entity when training the federated model based on the first local data, and the second model update data generated by the enterprise entity when training the federated model based on the second local data. The first model update data and the second model update data are encrypted using a verifiable encryption method to obtain the first encrypted differential data and the second encrypted differential data. The encrypted feature fusion module is used to perform encrypted feature fusion processing on the first encrypted differential data and the second encrypted differential data within a preset encrypted domain to obtain the first encrypted feature result of the government entity and the second encrypted feature result of the enterprise entity. The encryption contribution result generation module is used to evaluate the encryption performance impact of the first encryption state feature result and the second encryption state feature result and quantify the fair allocation rules within a preset encryption domain, and generate the first encryption contribution result of the government entity and the second encryption contribution result of the enterprise entity. The contribution proof generation module is used to obtain the first public parameter corresponding to the first encrypted contribution result and the second public parameter corresponding to the second encrypted contribution result, generate the first contribution proof of the government entity based on the first encrypted contribution result and the first public parameter, and generate the second contribution proof of the enterprise entity based on the second encrypted contribution result and the second public parameter.

9. An electronic device, characterized in that, It includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the federal model contribution assessment method for government-business collaboration as described in any one of claims 1-7.

10. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the federal model contribution assessment method for government-enterprise collaboration as described in any one of claims 1-7.

Images