A risk assessment method and device based on a trusted data space and related equipment

Abstract

The application discloses a risk assessment method and device based on a trusted data space and related equipment, the method comprising: obtaining a digital contract, participant information, transaction data, environment data, and data space running data of the trusted data space; anchor point hash calculation to determine whether it is tampered with; obtaining a plurality of risk indicators and risk characteristic values; and obtaining the size of each risk level preset by the digital contract based on the global weight of each risk indicator and the risk characteristic value. The anchor point hash calculation can quickly determine whether the data information is tampered with, the preset large language model can determine the changes of the multi-source heterogeneous data such as the digital contract, the participant information, the transaction data, the environment data, and the data space running data in real time, and make a risk judgment in real time. In addition, when judging the risk based on the risk characteristic value of each risk indicator, the importance of each risk indicator in the global data contract is considered, so the judgment is more.

Description

Technical Field

[0001] This application relates to the field of big data, and in particular to a risk assessment method, apparatus and related equipment based on trusted data space. Background Technology

[0002] Currently, risk assessment for digital contracts in the trusted data space mainly relies on a combination of traditional rule engines and manual review. Traditional rule engines, based on preset fixed rules, have poor adaptability to the complex and ever-changing trusted data space environment. They cannot dynamically handle changes in multi-source heterogeneous data (such as time-series transaction data and unstructured environmental policy data) within the trusted data space, and their data tampering risk detection efficiency is low. Summary of the Invention

[0003] The main purpose of this application is to propose a risk assessment method, apparatus, and related equipment based on a trusted data space, aiming to solve the problems in the prior art where traditional rule engines have poor adaptability, cannot dynamically handle changes in multi-source heterogeneous data (such as time-series transaction data and unstructured environmental policy data) in a trusted data space, and have low efficiency in detecting data tampering risks.

[0004] To achieve the above objectives, this application proposes a risk assessment method based on trusted data space, comprising:

[0005] The system acquires digital contracts, participant information, transaction data, environmental data, and data space operation data from a trusted data space; these digital contracts, participant information, transaction data, environmental data, and data space operation data are multi-source heterogeneous data.

[0006] Anchor hash calculations are performed on the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data to determine whether the data has been tampered with.

[0007] If the digital contract has not been tampered with, the text semantics of the digital contract are analyzed by a preset large language model to obtain multiple risk indicators corresponding to each clause in the digital contract.

[0008] By using a pre-defined large language independent model corresponding to the participant information, transaction data, environmental data, and data space operation data respectively, semantic recognition and associated field matching are performed on each data or information to obtain the risk feature value of each risk indicator; the risk feature value of each risk indicator is obtained based on the field content of the corresponding risk indicator.

[0009] Based on the global weights and risk characteristic values ​​of each risk indicator, the magnitude of each preset risk level of the digital contract is obtained; wherein, the global weight of each risk indicator is obtained by analyzing a preset risk causal model based on the type of clause to which each risk indicator belongs, and the preset risk causal model is pre-trained based on a performance data set, which includes the digital contract clause type and performance result.

[0010] This application also proposes a risk assessment device based on trusted data space, comprising:

[0011] The acquisition module is used to acquire digital contracts, participant information, transaction data, environmental data, and data space operation data from the trusted data space; the digital contracts, participant information, transaction data, environmental data, and data space operation data are multi-source heterogeneous data;

[0012] The processing module is used to perform anchor hash calculations based on the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data to determine whether the data has been tampered with.

[0013] If the digital contract has not been tampered with, the text semantics of the digital contract are analyzed by a preset large language model to obtain multiple risk indicators corresponding to each clause in the digital contract.

[0014] By using a pre-defined large language independent model corresponding to the participant information, transaction data, environmental data, and data space operation data respectively, semantic recognition and associated field matching are performed on each data or information to obtain the risk feature value of each risk indicator; the risk feature value of each risk indicator is obtained based on the field content of the corresponding risk indicator.

[0015] Based on the global weights and risk characteristic values ​​of each risk indicator, the magnitude of each preset risk level of the digital contract is obtained; wherein, the global weight of each risk indicator is obtained by analyzing a preset risk causal model based on the type of clause to which each risk indicator belongs, and the preset risk causal model is pre-trained based on a performance data set, which includes the digital contract clause type and performance result.

[0016] This application also proposes a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the methods described in any of the above embodiments.

[0017] This application also proposes a computing device, including a processor, on which a computer program is stored, and the computer program, when executed, implements the method described in any of the above embodiments.

[0018] In this embodiment, when conducting risk assessment based on a trusted data space, the method first determines whether various data and information have been tampered with based on anchor hashing. Then, a pre-defined large language model is used to perform semantic parsing on the digital contract and construct multiple risk indicators for each clause in the digital contract. Using the pre-defined large language model corresponding to participant information, transaction data, environmental data, and data space operation data, semantic transformation is performed on these data to identify fields related to each risk indicator. Then, the risk characteristic value of each risk indicator is determined based on the fields corresponding to each risk indicator, and a risk judgment is made based on the risk characteristic value of each risk indicator. This method, using various pre-defined large language models, can determine changes in multi-source heterogeneous data such as digital contracts, participant information, transaction data, environmental data, and data space operation data in real time and make risk judgments in real time. Furthermore, when judging risk based on the risk characteristic value of each risk indicator, the importance of each risk indicator in the overall data contract is also considered, thus making the judgment more accurate. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the steps of a risk assessment method based on a trusted data space according to an embodiment of this application;

[0021] Figure 2 This is a block diagram of a risk assessment device based on a trusted data space according to an embodiment of this application;

[0022] Figure 3 This is a block diagram of a computer-readable storage medium in one embodiment of this application;

[0023] Figure 4 This is a block diagram of a computing device according to one embodiment of this application.

[0024] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0025] The principles and spirit of this application will now be described with reference to several exemplary embodiments. It should be understood that these embodiments are provided merely to enable those skilled in the art to better understand and implement this application, and are not intended to limit the scope of this application in any way. Rather, these embodiments are provided to make this application more thorough and complete, and to fully convey the scope of this application to those skilled in the art.

[0026] Those skilled in the art will understand that embodiments of this application can be implemented as a system, device, method, or computer program product. Therefore, this application can be specifically implemented in the following forms: entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0027] According to the embodiments of this application, a risk assessment method, apparatus and related equipment based on trusted data space are proposed.

[0028] In this article, it is important to understand that any number of elements in the accompanying figures is for illustrative purposes and not for limitation, and any naming is for distinction only and has no limiting meaning.

[0029] The principles and spirit of this application will be explained in detail below with reference to several representative embodiments.

[0030] Exemplary methods

[0031] like Figure 1 As shown in the embodiments of this application, the risk assessment method based on trusted data space includes the following steps S100-S600:

[0032] S100: Obtain digital contracts, participant information, transaction data, environmental data, and data space operation data from a trusted data space; the digital contracts, participant information, transaction data, environmental data, and data space operation data are multi-source heterogeneous data.

[0033] In this application embodiment, the Trusted Data Space (TDS) is a multi-party data sharing and interaction environment that ensures the authenticity, integrity, and confidentiality of data during its flow through data encryption, identity authentication, and access control, while also meeting compliance requirements (such as the General Data Protection Regulation (GDPR) and data security laws). It serves as the basic carrier for the performance of digital contracts.

[0034] In this embodiment, the digital contract is a core protocol that regulates data interaction behavior in a trusted data space. It clarifies the rights and obligations of data providers, users, and service providers, and serves as the rule basis for data collection, use, and transfer. For example, in this embodiment, the digital contract may include the following:

[0035] Contract Terms: In a trusted data space, contract terms must comply with laws and regulations such as the Data Security Law and the Personal Information Protection Law. They must also clearly define the data's usage scenarios (e.g., data analysis only, no resale), data processing methods (e.g., whether anonymized sharing is allowed, whether encrypted transmission is required), liability for breach of contract (e.g., compensation standards for exceeding the scope of use), and dispute resolution mechanisms (e.g., selection of arbitration institutions). For example, in a trusted data space for a manufacturing industry, the contract terms between Company A and Company B stipulate that "Company B may use Company A's production energy consumption data for energy efficiency optimization analysis. The data is valid for 3 months and may not be used for cooperative projects with Company C."

[0036] Signatory information includes the unique identifiers (such as digital certificate numbers) of all parties participating in the digital contract within the trusted data space, their entity type (enterprise, institution, individual), information of their legal representative / responsible person, and contact information, ensuring that the signatory's identity is traceable and verifiable. For example, the signatory information records "Enterprise X (Unified Social Credit Code XXX, Digital Certificate Number XXX), Institution Y (Public Institution Legal Person Certificate Number XXX)," and links it to the trusted data space's identity authentication system.

[0037] Performance Period: Clearly define the timeframe for data provision, use, and settlement, including the start / end date of data delivery, the effective / expired date of usage rights, and the fee settlement cycle. This timeframe is synchronously recorded by the trusted data space's timestamp system to ensure the time cannot be tampered with. For example: "From January 1, 2024 to December 31, 2024, Company A must provide Company B with real-time inventory data daily, and Company B must pay the data usage fee for the previous month before the 5th of each month."

[0038] Contract Amount: The fee standard and payment method for data transactions, including the unit price (e.g., 0.5 yuan per data item), total amount, payment cycle (e.g., per transaction, per month), payment channel (must be through a compliant payment interface within the trusted data space), and is linked to the contract performance progress (e.g., payment is automatically triggered after data delivery meets the standards). For example: "Based on the actual amount of logistics trajectory data transmitted each month, settlement is made at 0.1 yuan per item, and payment is completed through the bank interface within the data space before the 10th of each month."

[0039] In this embodiment, participant data refers to the basic information and credit-related data of all entities (such as enterprises, institutions, and individuals) accessing the trusted data space. This data is used to assess the compliance and credibility of the entities and serves as identity credentials for data space access and interaction. For example, in this embodiment, participant data may include the following:

[0040] Basic enterprise information includes the business license information (registered capital, business scope, registered address), industry classification (such as manufacturing, finance), organizational structure (such as contact departments and heads), and data security officer information submitted by the enterprise when registering with the Trusted Data Space. For example, the basic enterprise information of an e-commerce company records "business scope includes online sales, and the data security officer is Wang Moumou (contact information XXX)" and is linked to the enterprise qualification review record of the Trusted Data Space.

[0041] Credit record: An entity's credit performance within the trusted data space and external systems, including whether there are any data violations (such as using data beyond the permitted scope), administrative penalty records (such as being punished by regulatory authorities for data leakage), court default information, etc., which are dynamically updated by the trusted data space's credit rating system. For example: Company Z was marked in its credit record as "Violation in June 2023, credit rating downgraded to B, restricted from participating in highly sensitive data transactions" because it shared user data without authorization in 2023.

[0042] Historical performance record: This refers to the entity's performance in past digital contracts, including whether data was delivered on time, whether the data quality met standards (such as accuracy and completeness), and whether fees were paid as agreed. This information is used for risk assessment in future contract collaborations. For example, Company Y's historical performance record shows that "in 2022, contracts with 3 companies were 100% fulfilled on time, with a data accuracy rate of 98%; in 2023, there was one instance of delayed delivery due to a system failure, for which compensation was paid according to the contract."

[0043] In this embodiment, transaction data refers to records of data transactions and related activities occurring within a trusted data space. It serves as evidence of the data value transfer process, leaving a complete trace and being tamper-proof. For example, in this embodiment, transaction data may include the following:

[0044] Transaction log: Records key information for each data transaction, including transaction number (a unique identifier within the data space), transacting parties (related contracting party information), transaction data type (e.g., user profile data, sensor data), transaction quantity (e.g., 1000 records), transaction time (accurate to milliseconds, with associated timestamp), and transaction status (e.g., pending payment, completed, cancelled). For example: Transaction log record "Number T20240501001, Company A purchased 500 records of 'East China Home Appliance Sales Data' from Company B, transaction time 2024-05-01 09:30:15, status: completed".

[0045] Payment records: Records of fund transfers associated with transaction flows, including payment amount, payment time, payment channel (such as bank accounts or third-party payment platforms connected to the data space), payment vouchers (such as electronic receipt numbers), and payee account information (which must be filed in the data space), ensuring traceability of fund flows. For example: The payment record shows "Corresponding transaction T20240501001, Company B paid 500 yuan through the XX Bank interface at 09:35:20 on 2024-05-01, voucher number P20240501001."

[0046] In this embodiment, environmental data refers to external dynamic information that affects data transactions and decision-making within the trusted data space. It assists participants in assessing risks and optimizing strategies, serving as an external reference system for data interaction. For example, in this embodiment, environmental data may include the following:

[0047] Market conditions: Information on market supply and demand and price fluctuations related to data transactions, such as the market transaction volume, average price, and popular procurement directions (e.g., for product R&D) of a certain type of data (e.g., new energy vehicle user data), is synchronized in real time by DataSpace through industry database interfaces. For example, the environmental data records "In Q1 2024, the market transaction volume of energy consumption data in the manufacturing industry increased by 15% year-on-year, with an average price of 0.8 yuan / piece, and the main purchasers were energy-saving service companies."

[0048] Policy and regulatory changes: Updates to laws and regulations related to data security, privacy protection, and transaction compliance are automatically captured and pushed by Data Space's policy monitoring module. For example: "On June 1, 2024, the 'XX Province Data Transaction Management Measures' officially came into effect, adding a clause that 'sensitive data transactions must be filed,' and Data Space has updated its contract template accordingly."

[0049] Industry Dynamics: Information on technological developments and major events in the industry to which the data pertains, such as the emergence of new data security technologies in a particular industry (e.g., the application of federated learning in medical data), and data cooperation cases of leading companies, to help participants adjust their data strategies. For example: "In May 2024, the financial industry's trusted data space introduced a new 'data asset pledge' model, and a certain bank has completed its first data pledge loan through this model."

[0050] In this embodiment, data space operation data refers to records generated during the operation of the trusted data space system itself, used to ensure the stability, security, and auditability of the data space, and serves as a log archive for data space operation and maintenance. For example, in this embodiment, data space operation data may include the following:

[0051] Data storage status: Records the storage location of data within the trusted data space (e.g., distributed storage node number), storage format (e.g., encrypted file, structured database), storage capacity (used / remaining space), data backup status (e.g., whether off-site backup has been completed, backup time), and data lifecycle (e.g., creation time, expiration time). For example: "Company A's '2023 sales data' is stored on node S001, in AES encrypted JSON format, has completed 3 off-site backups, and is expected to expire on December 31, 2026."

[0052] Data transmission log: This log contains end-to-end information about the data flow within the space, including the initiator / receiver (identifier of associated participants), data ID (related transaction history), transmission time, transmission path (e.g., from node A to node B), transmission status (e.g., success, failure, number of retries), encryption method (e.g., TLS 1.3), ensuring data transmission traceability. For example: "2024-05-01 09:32:00, Enterprise A transmitted data ID D1001 to B, path S001→S003, encryption method TLS 1.3, transmission successful, time taken 2.5 seconds."

[0053] Access Log: Records of all entities accessing the data space and data, including the accessing entity (related participant information), access time, accessed object (such as a contract, a set of data), access operation (such as query, download, modification), access permission verification result (such as passed, denied and reason), terminal device information (such as IP address, device fingerprint), etc. For example: Example: "2024-05-01 10:00:15, Enterprise B accessed data D1001 through IP 113.XXX.XXX, operation: download, permission verification passed, device fingerprint: XXX".

[0054] S200: Based on the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data, perform anchor hash calculations to determine whether the data has been tampered with.

[0055] In this embodiment of the application, it can be determined whether the digital contract, participant information, transaction data, environmental data, and data space operation data have been tampered with based on the following steps S210-240:

[0056] S210: Determine the core fields of the digital contract, the participant information, the transaction data, the environmental data, the data space operation data, and the anchor clauses of each core field in the digital contract;

[0057] S220: Bind the core fields of the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data to their corresponding anchor terms to generate various binding results;

[0058] S230: Perform hash calculations on the binding results corresponding to the digital contract, the participant information, the transaction data, the environment data, and the data space operation data respectively to obtain the current contract anchor hash corresponding to the digital contract, the participant information, the transaction data, the environment data, and the data space operation data;

[0059] S240: Based on the current contract anchor hash corresponding to the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data, respectively, compare them with the initial contract anchor hash corresponding to the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data in the digital space to determine whether the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data have been tampered with.

[0060] In this embodiment of the application, the core fields of the digital contract can be key information that is directly linked to the validity of the data contract and the rights and obligations, such as: the amount of the performance bond, the data delivery timestamp, the identity of the contracting party, the authorized usage period, and the data usage scope definition field. If these fields are tampered with, they will directly undermine the legal validity of the digital contract.

[0061] In this embodiment, the anchor clause of the core field of the digital contract can be the unique identifier (ContractID) of the digital contract. Binding the core field of the digital contract to the unique identifier of the digital contract and then performing hash encryption can ensure the legal validity of the digital contract.

[0062] In this embodiment of the application, the core fields of the participant information can be key information related to the authenticity of identity and the validity of credit, such as: the unified social credit code of the enterprise, the ID number of the legal representative, the credit rating score, the number of historical default records and timestamps. If these core fields are tampered with, it may lead to the impersonation of the participant's identity and the distortion of the credit status, thereby affecting the eligibility to sign the data contract or the credibility of the performance.

[0063] In this embodiment of the application, the anchor clause for the core fields of the participant information in the digital contract can be the participant qualification clause. By binding the core fields of the participant information with the anchor clause and then performing hash encryption, the immutability of the participant's identity and credit information can be ensured.

[0064] In this embodiment of the application, the core fields of the transaction data can be information related to the authenticity of the transaction and the accuracy of the amount, such as: transaction amount, payment timestamp, account identifiers of both parties to the transaction, logistics waybill number, receipt confirmation timestamp, etc. If these fields are tampered with, it will lead to false transactions, incorrect fund settlement, and violation of the transaction authenticity guarantee clauses in the contract.

[0065] In this embodiment of the application, the anchor clauses of the core fields of the transaction data in the digital contract can be related terms for the transaction target and amount. By binding the core fields of the transaction data with the anchor clauses and then performing hash encryption, the traceability of the transaction chain can be ensured.

[0066] In this embodiment of the application, the core fields of environmental data can be key information related to the effective conditions of the data contract and risk warnings, such as: policy and regulation effective / expiration timestamps, industry benchmark prices, market supervision level identifiers, and records of sudden risk events. If these fields are tampered with, it may cause the external conditions on which the contract is based to be distorted, violating the environmental adaptability clause of the contract.

[0067] In this embodiment of the application, the anchor clause of the core field of environmental data in the digital contract can be an external condition reference clause. By binding the core field of environmental data with the corresponding anchor clause and then hashing and encrypting it, the validity of environmental data as a supplementary basis for the contract can be ensured.

[0068] In this application embodiment, the core fields of the data running in the data space can be key information related to data security and access control, such as: data visitor identity, access timestamp, access operation type (read / write / process), data transmission node IP, and storage location hash value. If these fields are tampered with, they may cover up unauthorized access, data leakage, and other behaviors, violating the data security protection clauses in the contract.

[0069] In this embodiment, the anchor clause of the data space operation data in the digital contract can be a clause related to the access permission level agreement. By binding the core fields of the data space operation data with the anchor clause and then hashing and encrypting them, it can be ensured that the data flow complies with the permission rules.

[0070] For example, for digital contracts, assuming the core field is determined to be F and the unique identifier of the digital contract is ContractID, the core field of the digital contract is bound to the digital contract identifier. For example, the core field value F and ContractID are structurally concatenated to obtain the binding result S.

[0071] Additionally, a separator can be added during concatenation to avoid confusion between field values ​​and IDs, for example:

[0072] S = F + "|" + ContractID;

[0073] The "|" symbol is used as a separator.

[0074] Once the binding result is obtained, it can be hashed and encrypted as follows:

[0075] Hanchor = SHA-256(S) = SHA - 256( F + "∣" + ContractID )

[0076] SHA-256 is a hash algorithm, and Hanchor is the hash value of the current contract anchor obtained by encrypting the binding result based on the current digital contract. It includes both the core fields of the digital contract and the anchor terms.

[0077] In addition, when a digital contract is initially generated, its initial contract anchor hash value is also stored. This initial contract anchor hash value is obtained by hashing the binding result of the core fields of the initial digital contract and its unique identifier. Therefore, by comparing the current contract anchor hash value with the initial contract anchor hash value, if they match, it means that the core fields and anchor terms in the current digital contract have not been tampered with. If they do not match, it means that the information fields or anchor terms in the current digital contract have been tampered with, which poses a risk.

[0078] Similarly, when initially generating participant information, transaction data, environmental data, and data space operation data, each core field and anchor clause is bound, and the binding result is hashed to obtain the initial contract hash value. Then, when subsequently obtaining participant information, transaction data, environmental data, and data space operation data, each current core field and current anchor clause are bound, and hashed to obtain the current contract anchor hash value. By comparing this current anchor hash value with the initial contract anchor hash value, it can be determined whether the core fields and anchor clauses of the participant information, transaction data, environmental data, and data space operation data have been tampered with.

[0079] In this application embodiment, when the current contract anchor hash value of digital contract, participant information, transaction data, environmental data, and data space operation data is different from their original contract anchor hash values, the clauses violated by the data tampering can be accurately located according to the bound anchor terms, and a breach reminder can be issued.

[0080] S300: Perform text semantic parsing on the digital contract using a preset large language model to obtain multiple risk indicators corresponding to each clause in the digital contract.

[0081] In this embodiment of the application, each risk indicator can be obtained based on the following steps S310-S320:

[0082] S310: Determine the clause type of each clause in the digital contract based on the preset large language model. The clause types include core clauses, auxiliary clauses, and external environment clauses.

[0083] S320: Based on the core clauses, the auxiliary clauses, and the external environment clauses, construct various risk indicators for each category of clauses.

[0084] In this application embodiment, the preset large language model may include a contract parsing model, which parses the data contract based on the contract parsing model, determines the priority of each clause according to the semantics of each clause in the data contract, and then divides the clauses into core clauses, auxiliary clauses, and external environment clauses.

[0085] In this embodiment of the application, the contract parsing model can be obtained by pre-training and fine-tuning based on the BERT (Bidirectional Encoder Representations from Transformers) model, specifically including a pre-training stage and a fine-tuning stage, as follows:

[0086] Pre-training phase

[0087] The training samples in the pre-training phase can be large-scale unlabeled text corpora, such as the multilingual, structured encyclopedic knowledge contained in Wikipedia, and the text content of various books contained in Books Corpus.

[0088] During training, the input format for each training sample can be such that each sample includes two sentences. The input sequence for each training sample can be constructed as follows:

[0089] 1) Add a special marker [CLS] at the beginning (for aggregate representation in subsequent classification tasks);

[0090] 2) Use the [SEP] marker to separate the two sentences;

[0091] 3) The vector of each token (vocabulary) consists of three superimposed parts: Token Embeddings, the embedding vector of the word itself; Segment Embeddings, which distinguishes the embeddings of two sentences (e.g., sentence A is 0 and sentence B is 1); Position Embeddings, which marks the position of the word in the sequence.

[0092] During pre-training, two self-supervised tasks can be used to train the model, such as the Masked Language Model (MLM) and the Next Sentence Prediction (LSP) model. The MLM model randomly masks 15% of the tokens in the input text; for example, 80% of the tokens are replaced with the special marker [MASK], 10% are replaced with random words from the vocabulary, and 10% remain unchanged. The model then predicts the masked tokens, thus forcing the learning of contextual semantic relationships. Additionally, in each training sample, 50% of the two sentences are consecutive true sentence pairs, and 50% are randomly concatenated sentence pairs. Based on the LSP model, the LSP output predicts whether the two sentences are consecutive, learning sentence-level semantic logic.

[0093] In addition, in the embodiments of this application, the loss function in the pre-training optimization process can be the sum of the MLM (Masked Language Modeling) loss function and the NSP (Next Sentence Prediction) loss function. The model parameters are iteratively optimized through a gradient descent algorithm (such as Adam) until the loss converges.

[0094] Fine-tuning stage

[0095] During fine-tuning, textual data (such as supplementary clauses and dispute records) in areas such as digital contracts and digital rights can be used to fine-tune the existing BERT pre-trained model using corpus data from the digital contract domain.

[0096] During fine-tuning, the parameters of the pre-trained model can be frozen, and only the top-level network can be updated. The model can be trained with the goal of classifying and matching tasks in the domain (such as judging the compliance of contract terms, priority judgment, core clause judgment, auxiliary clause judgment, external environment clause judgment, and risk indicator judgment), so that the model can more accurately understand the terminology in the data contract domain.

[0097] For example, in this embodiment of the application, through the fine-tuning stage, the contract parsing model is able to determine that the tiered performance clauses in the digital contract are the core clauses, and the corresponding risk indicators are the tiered clause default risk indicator, the threshold calculation error risk indicator, etc.

[0098] For example, through the fine-tuning phase, the contract parsing model can identify the performance milestone clauses in the digital contract as core clauses, and the corresponding risk indicators are milestone progress lag risk indicators, milestone data quality risk indicators, etc.

[0099] For example, through the fine-tuning phase, the contract parsing model can identify the data rights and scope of use clauses in digital contracts as core clauses, and the corresponding risk indicators are the risk indicators of exceeding the time limit for use and exceeding the scope of sub-licensing.

[0100] For example, through the fine-tuning phase, the contract parsing model can determine that the data availability clause in the digital contract is the core clause, and the corresponding risk indicators are the business risk indicators of availability timeout and the risk indicators of insufficient availability stability.

[0101] For example, through the fine-tuning phase, the contract parsing model can identify auxiliary clauses in digital contracts, and the corresponding risk indicators are such as report format error risk indicators and contact information update delay risk indicators.

[0102] For example, through the fine-tuning phase, the contract parsing model can determine whether the external condition reference clauses and contract environment adaptability clauses in the digital contract are external environment clauses, and the corresponding risk indicators are risk indicators such as the risk indicator of the scope of data use being restricted due to policy changes and the risk indicator of the impact of industry fluctuations on the performance progress.

[0103] It should be noted that the above core terms, auxiliary terms, external environment terms, and corresponding risk indicators are all illustrative. The specific core terms, auxiliary terms, and external environment terms, as well as the specific terms and risk indicators contained therein, are based on the specific digital contract.

[0104] S400: Using a preset large language independent model corresponding to the participant information, the transaction data, the environmental data, and the data space operation data respectively, semantic recognition and associated field matching are performed on each data or information to obtain the risk feature value of each risk indicator; the risk feature value of each risk indicator is obtained based on the field content of the corresponding risk indicator.

[0105] In this embodiment of the application, when semantically transforming the participant information, the transaction data, the environmental data, and the data space operation data to identify the participant information, the transaction data, the environmental data, and the fields in the data space corresponding to each risk indicator, different large language models can be used.

[0106] In this embodiment of the application, for participant information, a participant credit model can be used to semantically transform the obtained participant information, determine the fields in the participant information that correspond to each risk indicator, and then map the fields in the participant information that correspond to each risk indicator to the corresponding risk indicator to obtain the risk feature value of each risk indicator.

[0107] In this embodiment, the participant credit model can also be pre-trained and fine-tuned based on the BERT (Bidirectional Encoder Representations from Transformers) model. The training process in the pre-training stage is the same as that in the contract parsing model, and will not be described in detail here. In the fine-tuning stage, the parameters of the pre-trained model can be frozen, and only the top-level network can be updated. Fine-tuning training is performed with the goal of judging the fields corresponding to risk indicators related to participant information. This enables the participant credit model to more accurately understand the fields in the participant information that correspond to each risk indicator, such as credit record fields and historical performance fields, and also enables the participant credit model to more accurately understand the participants in the risk features corresponding to each risk indicator.

[0108] For example, for risk indicators such as report format error and contact information update delay, the participant information can be read based on the participant credit model, and the relevant fields in the participant information that correspond to the auxiliary clauses in the digital contract, such as the report format field and the contact information update field, can be determined. Then, based on the report format field, the risk characteristic value of the report format error risk indicator can be determined, and based on the contact information update field, the risk characteristic value of the contact information update delay risk indicator can be determined.

[0109] In this embodiment of the application, for transaction data, a transaction compliance model can be used to perform semantic transformation on the acquired transaction data, determine the fields in the transaction data corresponding to each risk indicator, and then map the fields in the transaction data corresponding to each risk indicator to each risk indicator to obtain the risk characteristic value of each risk indicator.

[0110] The transaction compliance model can also be pre-trained and fine-tuned based on the BERT (Bidirectional Encoder Representations from Transformers) model. The training process in the pre-training stage is the same as that in the contract parsing model, and will not be elaborated here. In the fine-tuning stage, the parameters of the pre-trained model can be frozen, and only the top-level network can be updated. The fine-tuning is carried out with the goal of judging the fields corresponding to risk indicators related to transaction data. This allows the transaction compliance model to more accurately understand the fields in the transaction data corresponding to various risk indicators, such as transaction history fields, payment record fields, payment amount fields, and logistics information fields.

[0111] For example, for milestone progress lag risk indicators and milestone data quality non-compliance risk indicators, the transaction data can be semantically transformed based on the transaction compliance model, and the fields corresponding to the performance milestone clauses in the digital contract can be determined, such as the performance progress field, time node field, quality field, and key field missing rate. Then, based on the performance progress field and time node field, the progress gap characteristics and time node deviation characteristics can be determined. Based on the progress gap characteristics and time node deviation characteristics, the risk characteristic value of the milestone progress lag risk indicator can be determined. Based on the quality field, the quality deviation characteristics can be determined. Based on the quality deviation characteristics and key field missing rate, the risk characteristic value of the milestone data quality non-compliance risk indicator can be determined.

[0112] For example, regarding the risk indicators for tiered clause default and threshold calculation error, the transaction data can be semantically transformed based on the transaction compliance model, and the fields corresponding to the tiered performance clauses in the digital contract can be determined, such as the condition threshold field, action execution field, clause indicator field, statistical indicator field, and threshold trigger status field. Then, based on the condition threshold field and action execution field, the risk characteristic value of the tiered clause default risk indicator can be determined, and based on the clause indicator field, statistical indicator field, and threshold trigger status field, the risk characteristic value of the threshold calculation error risk indicator can be determined.

[0113] In this embodiment of the application, for data space operation data, the operation monitoring model can be used to perform semantic transformation on the acquired data space operation data, determine the fields in the data space operation data corresponding to each risk indicator, and map the fields in the data space operation data corresponding to each risk indicator to each risk indicator to obtain the risk characteristic value of each risk indicator.

[0114] The operation monitoring model can also be pre-trained and fine-tuned based on the BERT (Bidirectional Encoder Representations from Transformers) model. The training process in the pre-training stage is the same as that in the contract parsing model, and will not be elaborated here. In the fine-tuning stage, the parameters of the pre-trained model can be frozen, and only the top-level network can be updated. The fine-tuning training is carried out with the fields corresponding to risk indicators related to the data space operation data as the target, so that the operation monitoring model can more accurately understand the fields in the data space operation data corresponding to various risk indicators, such as access log fields, data access timeout fields, and storage status fields.

[0115] For example, for the overdue use exceeding authority risk indicators and the sub-authorization exceeding the limit risk indicators, the data space operation data can be semantically transformed based on the operation monitoring model, and the fields corresponding to the data rights and usage scope clauses in the digital contract can be determined, such as the timestamp field, permission status field, domain identification field, and sub-authorization level field. Then, based on the timestamp field and permission status field, the risk characteristic value of the overdue use exceeding authority risk indicator can be determined, and based on the domain identification field and sub-authorization level field, the risk characteristic value of the sub-authorization exceeding the limit risk indicator can be determined.

[0116] For example, regarding the risk indicators of availability timeout impacting business and the risk indicators of insufficient availability stability, the operational data in the data space can be semantically transformed based on the operation monitoring model, and the fields corresponding to the availability clauses in the digital contract can be determined, such as the response time deviation field, the continuous timeout duration field, the success rate deviation field, and the failure frequency field. Then, based on the response time deviation field and the continuous timeout duration field, the risk characteristic value of the risk indicator of availability timeout impacting business can be determined, and based on the success rate deviation field and the failure frequency field, the risk characteristic value of the risk indicator of insufficient availability stability can be determined.

[0117] In this embodiment of the application, for environmental data, an environmental adaptation model can be used to perform semantic transformation on the environmental data, determine the fields in the environmental data corresponding to each risk indicator, and then map the fields in the environmental data corresponding to each risk indicator to each risk indicator to obtain the risk feature value of each risk indicator.

[0118] The environment adaptation model can also be pre-trained and fine-tuned based on the BERT (Bidirectional Encoder Representations from Transformers) model. The training process in the pre-training stage is the same as that in the contract parsing model, and will not be elaborated here. In the fine-tuning stage, the parameters of the pre-trained model can be frozen, and only the top-level network can be updated. Fine-tuning is then performed with the goal of judging the fields corresponding to risk indicators related to the environmental data. This allows the environment adaptation model to more accurately understand the fields in the environmental data corresponding to various risk indicators, such as policy fields and external reference clause fields. For example, for the risk indicator of policy changes leading to limited data usage, the environmental data can be semantically transformed based on the environment adaptation model, and the fields corresponding to external condition reference clauses in the digital contract, such as policy fields, can be determined. Based on the policy fields, the risk characteristic value of the risk indicator of policy changes leading to limited data usage can then be determined.

[0119] For example, regarding the risk indicator of industry fluctuations affecting contract performance, the environmental data can be semantically transformed based on the environmental adaptation model, and the corresponding fields of external conditions referenced in the digital contract, such as policy fields and industry dynamic fields, can be determined. Based on the policy fields, the risk characteristic value of the risk indicator of policy changes leading to limited data usage can be determined, and based on the industry dynamic fields, the risk characteristic value of the risk indicator of industry fluctuations affecting contract performance can be determined.

[0120] In this embodiment, based on the contract parsing model, participant credit model, transaction compliance model, operation monitoring model, and environment adaptation model, semantic transformation is performed on digital contracts, participant information, transaction data, environmental data, and data space operation data, respectively, thereby identifying the relevant fields in multimodal data corresponding to each risk indicator, and then determining the risk characteristic value of each defense indicator. Regardless of how the multimodal data changes, risk characteristics can be identified in real time, providing a theoretical basis for real-time judgment of risk magnitude.

[0121] S500: Based on the global weights and risk characteristic values ​​of each risk indicator, obtain the magnitude of each preset risk level of the digital contract; wherein, the global weight of each risk indicator is obtained by analyzing a preset risk causal model based on the type of clause to which each risk indicator belongs, and the preset risk causal model is pre-trained based on a performance data set, which includes the digital contract clause type and performance result.

[0122] In this embodiment of the application, the global weight of each risk indicator can be determined based on the following steps S510-S550:

[0123] S510: Based on the relative importance of the core clauses, the auxiliary clauses, and the external environment clauses, construct a first judgment matrix using the causal model.

[0124] In this embodiment of the application, the causal model can be pre-trained based on a performance data set, which includes digital contract term types and performance results.

[0125] In this embodiment, the causal model can be selected from models suitable for causal inference, such as structural equation modeling, causal forest modeling, and dual machine learning modeling. The training of the causal model can generally include the following stages:

[0126] Data collection and preprocessing stage:

[0127] This phase primarily involves collecting historical performance data sets, including the textual content of digital contracts (term details) and performance results (such as success / failure, performance rate, default losses, etc.).

[0128] After data collection, during the preprocessing stage, the terms of digital contracts in the performance data set can be classified. Based on expert experience, they can be classified into three categories: core terms, auxiliary terms, and external environment terms.

[0129] Training phase:

[0130] During training, different types of clauses can be transformed into features usable by the model. Interaction features can also be added between clauses, such as combinations of core clauses and auxiliary clauses. With the performance outcome as the dependent variable and the clause features as the independent variable, the causal model can estimate the relative importance of the three types of clauses.

[0131] In this embodiment, the relative importance of core terms, auxiliary terms, and external environment terms in a digital contract can be determined based on a causal model. For example, the causal model determines that the core terms are more important than the auxiliary terms and assigns an importance scale value of 5, where an importance scale value of 5 represents that the core terms are 5 times more important than the auxiliary terms; the core terms are more important than the external environment terms and assign an importance scale value of 3, where an importance scale value of 3 represents that the core terms are 3 times more important than the external environment terms; the external environment terms and auxiliary terms are equally important and assign an importance scale value of 1, thus yielding the following first judgment matrix:

[0132] C1 Core Clauses C2 Supplementary Clauses C3 External Environment Clauses

[0133] C1 Core Clause 153

[0134] C2 Supplementary Clause 1 / 511

[0135] C3 External Environment Clause 1 / 311

[0136] S520: Based on the first judgment matrix, determine the benchmark weights of the core clauses, the auxiliary clauses, and the external environment clauses.

[0137] In this embodiment of the application, when calculating the baseline weights of the core clause, auxiliary clause, and external environment clause based on the first judgment matrix, the eigenvector of the first judgment matrix can be calculated as follows:

[0138] First, sum each column of the first judgment matrix:

[0139] The sum of column 1: 1 + 1 / 5 + 1 / 3 ≈ 1 + 0.2 + 0.333 ≈ 1.533

[0140] The sum of column 2 is: 5 + 1 + 1 = 7

[0141] The sum of column 3 is: 3 + 1 + 1 = 5

[0142] Then, divide each element of the first judgment matrix by the sum of its column to obtain the standardized matrix, as follows:

[0143] C1 Core Clauses C2 Supplementary Clauses C3 External Environment Clauses

[0144] C1 core clause 0.652 0.714 0.6

[0145] C2 Supplementary Clauses 0.13 0.14 30.2

[0146] C3 External Environment Clause 0.217 0.143 0.2

[0147] Then, the average value of each row of the standard matrix is ​​calculated to obtain the feature vectors corresponding to the core clauses, auxiliary clauses, and external environment clauses, i.e., the baseline weights, as follows:

[0148] C1 Core Clause Baseline Weight = (0.652 + 0.714 + 0.6) / 3 = 0.655

[0149] C2 Supplementary Clause Baseline Weight = (0.130 + 0.143 + 0.2) / 3 = 0.158

[0150] C3 External Environment Benchmark Weight = (0.217 + 0.143 + 0.2) / 3 = 0.187

[0151] In this embodiment of the application, the core terms of the digital contract are taken as the most important principle. During risk assessment, they are transformed into a quantitative calculation basis with a weight of 65.5% for the core terms. In the subsequent fuzzy comprehensive evaluation, the impact of the risk indicators corresponding to the core terms on the final result must be greater than the impact of the risk indicators corresponding to the auxiliary terms and external environment terms on the final result, so that the final risk level determination can accurately reflect the rigid requirements of the digital contract.

[0152] S530: Based on the relative importance of various risk indicators under any type of clause, construct a second judgment matrix corresponding to each type of clause.

[0153] This application embodiment has three types of clauses: core clauses, auxiliary clauses, and external environment clauses. Each type of clause corresponds to multiple risk indicators. Therefore, in step S530, it is necessary to construct a second judgment matrix based on the risk indicators corresponding to each type of clause. Since there are three types of clauses, there are three second judgment matrices.

[0154] When constructing the judgment matrix corresponding to each type of clause, the relative importance of multiple risk indicators for each type can also be determined based on the method of assigning importance scale values ​​by experts or experience, which will not be elaborated here.

[0155] S540: Based on the second judgment matrix corresponding to each type of clause, determine the local weight of each risk indicator under each type of clause.

[0156] In this embodiment of the application, when calculating the weight (local weight) of each risk indicator corresponding to the second judgment matrix for each type of clause under that type, it can also be calculated based on the feature vector of the second judgment matrix. The specific calculation method is referred to step S520, which will not be described in detail here.

[0157] S550: Determine the global weight of each risk indicator based on its local weight and the benchmark weight corresponding to the type clause in which it belongs.

[0158] In step S520, the baseline weight corresponding to each type of clause is calculated, and in S540, the local weight of each risk indicator in the clause of its type is calculated. In this embodiment of the application, the product of the local weight of each risk indicator in the clause of its type and the baseline weight corresponding to the clause of its type is used as the global weight corresponding to the risk indicator.

[0159] In this embodiment, the weight calculation method using two levels (term type and risk indicators corresponding to each type of term) not only considers the importance of different types of terms, but also the importance of different risk indicators under the same type. The resulting global weight of each risk indicator avoids the distortion caused by simply distributing weights equally. Therefore, when conducting risk assessment based on the global weight of each risk indicator, it can more accurately reflect the true intent of the contract terms corresponding to each risk indicator.

[0160] After determining the global weights of each risk indicator, it is also necessary to determine the membership vector of each risk indicator under each preset risk level based on the risk characteristic values ​​of each risk indicator.

[0161] In the embodiments of this application, the preset risk levels of the digital contract may include three risk levels: low risk, medium risk, and high risk. In other embodiments, there may be more or fewer risk levels. This application does not impose a specific limit on the number of risk levels.

[0162] In step S400, various preset large language models can be used to determine the risk characteristic values ​​corresponding to each risk indicator based on participant information, transaction data, environmental data, and data space operation data. The risk characteristic values ​​are used to assess the risk level of the corresponding risk indicator.

[0163] For example, in this embodiment of the application, based on the comprehensive identification of various preset large language models, two risk indicators are identified:

[0164] Risk Indicator A: Payment Delay Risk (Core Clause);

[0165] Risk Indicator B: Risk of Report Format Errors (Supplementary Clause)

[0166] The payment delay risk corresponds to the payment delay risk indicator, and the report format error risk corresponds to the report format error risk indicator. Assuming the risk characteristic value of the payment delay risk indicator is 30 (which is a minor delay), the risk characteristic value of the report format error risk is 85 (the higher the score, the higher the risk).

[0167] In the embodiments of this application, the membership function is determined based on each risk feature value, which indicates the degree to which the risk indicator is at each risk level under that risk feature value.

[0168] Assuming that the membership function in this embodiment is as follows:

[0169] Low-risk function: When the risk characteristic value is <50, it belongs to low risk; when it is 50-70, the membership degree decreases linearly to 0.

[0170] Medium-risk function: When the risk characteristic value is between 50 and 70, the membership degree increases from 0 to 1; when it is between 70 and 90, it decreases from 1 to 0.

[0171] High-risk function: When the risk characteristic value is >90, it is completely high-risk (1); when the score is 70-90, the membership degree increases from 0 to 1.

[0172] For risk indicator A(30), its membership degree is as follows:

[0173] Low-risk membership: μ_L(30) ≈ 1.0 (far below 50)

[0174] Medium-risk membership: 0

[0175] High-risk membership: 0

[0176] Therefore, the membership vector of risk indicator A is [1, 0, 0], which represents the degree of low risk, medium risk, and high risk under the current risk characteristic value of risk indicator A, which are 1, 0, and 0, respectively.

[0177] For risk indicator B(85), its membership degree is as follows:

[0178] Low-risk membership: 0 (85 is much higher than 50)

[0179] Medium-risk membership degree: 85 falls within the decreasing range of 70-90, therefore, μ_M(85) = (90-85) / (90-70) = 0.25

[0180] High-risk membership degree: 85 points falls within the rising range of 70-90, μ_H(85) = (85-70) / (90-70) = 0.75

[0181] Therefore, the membership vector of risk indicator B is [0, 0.25, 0.75], which represents the degree of low risk, medium risk, and high risk under the current risk characteristic value of risk indicator B, which are 0, 0.25, and 0.75, respectively.

[0182] In this embodiment of the application, after determining the global weight of each risk indicator and the membership vector of each risk indicator feature value under each risk level, the risk level of each risk level can be determined based on the membership vector of each risk indicator under each risk level and the global weight of each risk indicator.

[0183] In this embodiment of the application, the risk level of each risk level can be determined based on the following steps S560-S570:

[0184] S560: Construct a fuzzy relation matrix based on the membership vectors of each risk indicator at each risk level.

[0185] Taking risk indicators A and B from step S500 as examples, the membership vector of risk indicator A is [1, 0, 0], and the membership vector of risk indicator B is [0, 0.25, 0.75]. Therefore, the following fuzzy relation matrix can be constructed;

[0186] Low risk, medium risk, high risk

[0187] Risk indicator A100

[0188] Risk indicator B: 0.25 0.75

[0189] S570: Use the global weights of each risk indicator as the weight vector of the fuzzy relation matrix to determine the risk level of each risk grade.

[0190] Assuming that, based on the calculation method of the global weights of each risk indicator in step S500, the global weight Wa of risk indicator A is determined to be 0.7 and the global weight Wb of risk indicator B is determined to be 0.3, then the global weight vectors of risk indicators A and B are [0.7, 0.3]. By performing a synthesis operation (such as weighted average synthesis) on the global weight vectors of the risk indicators and the fuzzy relation matrix, the risk level of each risk level can be determined.

[0191] In this embodiment of the application, the comprehensive membership vector can be calculated based on the following formula:

[0192] B(j) = (Wa * Raj) + (Wb * Rbj)

[0193] Where B(j) is the membership vector of risk level j, Wa is the global weight of risk indicator A, Raj is the membership of risk indicator A at risk level j, Wb is the global weight of risk indicator B, and Rbj is the membership of risk indicator B at risk level j.

[0194] Therefore, B (low risk) = (0.7 * 0) + (0.3 * 1.0) = 0.3

[0195] B (Medium Risk) = (0.7 * 0.25) + (0.3 * 0) = 0.175

[0196] B (High Risk) = (0.7 * 0.75) + (0.3 * 0) = 0.525

[0197] The comprehensive membership vector B = [0.3, 0.175, 0.525] represents the degree of low risk, medium risk, and high risk under the current risk characteristic values ​​of risk indicators A and B, which are 0.3, 0.175, and 0.525, respectively.

[0198] In this embodiment, when conducting risk assessment based on a trusted data space, the method first determines whether various data and information have been tampered with based on anchor hashing. Then, a pre-defined large language model is used to perform semantic parsing on the digital contract and construct multiple risk indicators for each clause in the digital contract. Using the pre-defined large language model corresponding to participant information, transaction data, environmental data, and data space operation data, semantic transformation is performed on these data to identify fields related to each risk indicator. Then, the risk characteristic value of each risk indicator is determined based on the fields corresponding to each risk indicator, and a risk judgment is made based on the risk characteristic value of each risk indicator. This method, using various pre-defined large language models, can determine changes in multi-source heterogeneous data such as digital contracts, participant information, transaction data, environmental data, and data space operation data in real time and make risk judgments in real time. Furthermore, when judging risk based on the risk characteristic value of each risk indicator, the importance of each risk indicator in the overall data contract is also considered, thus making the judgment more accurate.

[0199] In this embodiment of the application, after obtaining the comprehensive membership vector, the final risk can be determined based on the magnitude of the membership vectors of each risk level in the comprehensive membership vector. For example, the final risk can be determined as the risk level with the largest membership vector. For instance, in the comprehensive membership vector B = [0.3, 0.175, 0.525], the membership vector of the high-risk level is 0.525, which is greater than the membership vector of the low-risk level (0.3) and also greater than the membership vector of the medium-risk level (0.175). In this case, the final risk can be determined as the high-risk level.

[0200] In this embodiment of the application, after obtaining the digital contract, participant information, transaction data, environmental data, and data space operation data of the trusted data space, the method further includes:

[0201] Determine the dynamic performance terms in the digital contract;

[0202] Based on the transaction data and the data space operation data, obtain the time-series data corresponding to the dynamic performance terms;

[0203] Based on the aforementioned time-series data, a sliding window is constructed;

[0204] Based on the sliding window, the participating parties' ability to perform their obligations is determined.

[0205] In this embodiment of the application, the dynamic performance terms in the digital contract are, for example, terms related to quarterly growth rate, monthly transaction volume, quarterly transaction volume, etc., and the dynamic performance terms can be determined by the contract parsing model.

[0206] For time-series data in transaction data, it contains a large number of records that change frequently over time, which can directly reflect the dynamic performance process of dynamic performance terms in digital contracts. Therefore, the dynamic performance terms related to transactions in data contracts can be evaluated based on the time-series data in transaction data.

[0207] For example, in the embodiments of this application, the contract parsing model determines the dynamic performance terms in the digital contract as: "a minimum monthly order volume of 500 orders, a quarterly growth rate of ≥10%, and a quarterly payment growth of no less than 5%", and the quarterly delivery volume needs to increase by 5% week by week.

[0208] In this embodiment of the application, the transaction flow time series in the transaction data includes the occurrence time, amount, and transaction status of each transaction, such as "500 transactions at 09:00 on 2024-01-01" and "480 transactions at 10:30 on 2024-01-02". In this case, the transaction flow time series of each window can be obtained based on the transaction compliance model with a week or a month as a sliding window, thereby determining the weekly or monthly transaction volume, and thus judging the participant's ability to fulfill the "minimum monthly amount of 500" in the dynamic performance terms. In addition, the weekly growth rate, monthly growth rate, and quarterly growth rate can also be calculated based on the transaction flow time series of each window, thereby judging the participant's ability to fulfill the "quarterly growth rate ≥ 10%" in the dynamic performance terms.

[0209] In this embodiment of the application, the transaction data also includes a payment record time series, which includes the payment amount and the payment timestamp, such as "the monthly payment amounts from January 2024 to March 2024 are 100,000, 120,000, and 110,000 respectively". Therefore, based on the transaction compliance model, the payment record time series of each window can be obtained with a week or a month as a sliding window, and the payment amount growth rate can be analyzed and calculated to determine the participant's ability to fulfill the "quarterly payment growth of not less than 5%" in the dynamic performance terms.

[0210] In this embodiment of the application, the transaction data also includes logistics and delivery time series, which includes the time and quantity of data delivery, such as "100 data items delivered every Monday for 12 weeks". Therefore, based on the transaction compliance model, the logistics and delivery time series of each window can be obtained with a week or a month as a sliding window, and the cumulative delivery volume can be analyzed to determine the participant's ability to fulfill the "quarterly delivery volume needs to increase by 5% every week" in the dynamic performance terms.

[0211] In this application embodiment, the time-series data in the data space operation data records the state of the system changing over time. Therefore, the dynamic performance terms related to "data availability and access compliance" in the data contract can be evaluated based on the time-series data in the data space operation data.

[0212] For example, in this embodiment of the application, it is assumed that the contract parsing model determines that the dynamic terms in the digital contract include "data access exceeds 5 seconds for 5 consecutive minutes, and the quarterly data transmission volume growth rate is ≥10%", and the daily access shall not be less than 5 times during the contract's life.

[0213] In this embodiment of the application, the data space operation data includes access response time series, which includes the average response time per hour / minute, such as "2024-06-01 08:00 response time 3 seconds, 09:00 response time 4.5 seconds". Based on the operation monitoring model, the access response time series of each window can be obtained with a one-minute sliding window, and then the indicator "data access exceeds 5 seconds for 5 consecutive minutes" in the dynamic performance terms can be determined.

[0214] In this embodiment of the application, the data space operation data also includes the data transmission volume time series, which includes: cross-domain data transmission volume per hour / day, such as "the monthly transmission volume in 2024-Q1 is 100GB, 105GB, and 108GB". Then, based on the operation monitoring model, the data transmission volume time series of each window can be obtained with a month as a sliding window, and then it can be determined whether the "quarterly data transmission volume growth rate ≥ 10%" in the dynamic performance terms is met.

[0215] In this embodiment of the application, the data space operation data also includes access log time series. The access log time series includes the time, frequency and operation type of user accessing data. For example, "User A accesses the data 10 times a day from May 1 to May 31, 2024". Then, based on the operation monitoring model, the access log time series of each window can be obtained with one day as a sliding window, and then the indicator "the daily access shall not be less than 5 times during the contract period" in the dynamic performance terms can be determined.

[0216] In this embodiment of the application, in addition to obtaining time-series data from transaction data and the data space operation data, time data can also be obtained from participant information to evaluate some dynamic indicators in the digital contract.

[0217] For example, in this embodiment of the application, the monthly performance completion rate sequence of the participants can be obtained from the participant information based on the participant credit model. For example, the performance completion rates of a certain enterprise from January to April 2024 were 95%, 90%, 85%, and 80%, respectively. Based on this, the subsequent trend can be predicted by the LSTM (Long Short-Term Memory) model. If the predicted performance rate is lower than the "minimum completion rate of 80%" required by the digital contract for three consecutive months, then the "insufficient performance capability warning" will be triggered.

[0218] For example, in this embodiment of the application, a credit score time series can be obtained from the participant information based on the participant credit model, such as "credit score dropped from 90 to 75 in Q1 2024". Based on this, the credit maintenance clause in the digital contract, such as "credit score shall not be lower than 80", can be combined to analyze whether credit deterioration will lead to future default.

[0219] In the embodiments of this application, the contract terms in the digital contract (such as a minimum guaranteed quantity of 500 orders and a growth rate of 10%) are static evaluation benchmarks that do not change over time. Therefore, they do not generate time-series data and are only used as target thresholds for time-series analysis.

[0220] In the embodiments of this application, although environmental data such as market conditions and policies and regulations in the environmental data have a time dimension, the frequency of change is low (such as policies being updated monthly / quarterly), and they are more used for "trend judgment" rather than "high-frequency dynamic performance indicator assessment". Therefore, they are only used as auxiliary references in special scenarios (such as the impact of policy changes on quarterly performance) and do not constitute core time series data.

[0221] In this embodiment of the application, after obtaining the digital contract of the trusted data space, the method further includes:

[0222] Determine the lifecycle type of the digital contract, which includes trial period contracts, long-term contracts, and short-term contracts;

[0223] If the digital contract is a trial period contract, the risk indicator weight corresponding to the initial performance terms shall be increased;

[0224] If the digital contract is a long-term contract, the weight of the risk indicator corresponding to the long-term impact clause shall be increased.

[0225] If the digital contract is a short-term contract, the weight of the risk indicator corresponding to the short-term goal completion clause is increased.

[0226] In this embodiment, after pre-training and fine-tuning, the contract parsing model, participant credit model, transaction compliance model, and operation monitoring model can identify the common characteristics of five types of data, such as transaction amount, response time, and credit score. Based on the common characteristics of the five types of data, they can learn the general rules of performance behavior and default risk. However, the default risk assessment at this time does not take into account the life cycle of digital contracts. For example, digital contracts can be divided into trial period contracts, long-term contracts, and short-term contracts according to their life cycle, and the core risks of digital contracts with different life cycles are different.

[0227] For example, for trial period contracts (the first 3 months), the trial period is a critical period for both parties to establish trust. Its core risk characteristics are the initial performance behavior, such as whether the first delivery meets the standards and whether the first month's payment is made on time. The initial performance behavior is highly predictive of long-term cooperation risks. For example, if a contract is breached during the trial period, the probability of long-term breach is 3 times that of a regular contract.

[0228] For example, long-term contracts (≥1 year) are greatly affected by external environmental fluctuations, such as policy changes and market cycle fluctuations. Long-term contracts often include force majeure clauses, such as "exemption from liability for difficulties in performance due to sudden changes in industry policies." Therefore, the core risk of long-term contracts should focus on long-term trend risks.

[0229] For example, for short-term contracts (3-12 months), the fulfillment nodes are relatively dense, such as data delivery once a month. The risks mainly come from the completion of nodes rather than long-term trends. Moreover, the contracts rarely contain complex force majeure clauses. Therefore, the core risks of short-term contracts should focus on the density of fulfillment nodes and short-term goals.

[0230] Therefore, in this embodiment of the application, based on the pre-trained and fine-tuned contract parsing model, participant credit model, transaction compliance model and operation monitoring model, an adaptation layer can be added to each model. The adaptation layer is used to adjust each model according to the life cycle of the digital contract, so that the core risks of each life cycle contract can be given priority consideration when assessing risk.

[0231] For example, for trial period contracts, the adaptation layer is used to strengthen the predictive weight of initial behavior. For instance, the weight of performance characteristics in the first three months (such as first-month transaction volume and initial data quality compliance rate) is increased by 20% compared to the basic models. For example, the weight of the first-month behavior in the basic model is increased from 0.1 to 0.12 after adjustment by the adaptation layer.

[0232] In addition to increasing the predictive weight of initial behavior, the input of long-term behavior characteristics (such as changes in industry policies) can be temporarily weakened, and the input of initial behavior characteristics such as the completion of key initial nodes (such as whether qualification documents were submitted in the first week of the trial period, and whether the minimum transaction volume was reached in the first month) can be emphasized.

[0233] For example, in this application embodiment, if the trial period contract stipulates a minimum of 500 orders in the first month and the basic model predicts a default probability of 30%, after the adaptation layer increases the prediction weight of the behavioral characteristics in the first month, if the actual transaction volume in the first week is only 100 orders (far lower than the weekly average of 125 orders), the predicted default probability is adjusted to 50%, thus making it more sensitive to capturing initial risks.

[0234] For long-term contracts, the adaptation layer is used to increase the weight of external environmental features while retaining high weights on core terms (such as payment and data availability). For example, the weight of policy factors in the base model is increased from 0.05 to 0.0575. Additionally, it is used for feature set expansion, such as adding long-term impact features beyond the base model inputs, like macroeconomic indices over the past six months, the frequency of industry regulatory policy updates, and the annual revenue fluctuations of participating parties.

[0235] For example, in an embodiment of this application, a long-term contract stipulates that "the annual data transmission volume growth rate is ≥10%". The basic model only predicts an 80% probability of meeting the target based on historical transmission data. After the adaptation layer adds the feature that "tightening industry policies lead to a 30% increase in transmission costs", the predicted probability of meeting the target is reduced to 50%, which is suitable for judging long-term risks.

[0236] For short-term contracts, the fulfillment nodes are dense (e.g., data is delivered once a month), and the risks mainly come from the completion rate of nodes rather than long-term trends. In addition, the contracts rarely contain complex force majeure clauses. Therefore, the adaptation layer is used to remove long-term features (e.g., industry data of more than 5 years), retain only short-term node features (e.g., the number of remaining nodes, the delay rate of completed nodes), and increase the prediction weight of node time deviation features. The prediction weight of time deviation features is higher in short-term contracts than in long-term contracts.

[0237] For example, in an embodiment of this application, a short-term contract of 3 months requires "reconciliation to be completed on the last day of each month". The basic model predicts an overall default probability of 20%. Based on the characteristics of dense nodes, the adaptation layer increases the weight of monthly reconciliation delay by 10%. If the first month's reconciliation is delayed by 2 days, the predicted default probability for the following two months will rise to 35%.

[0238] In this embodiment, the risks of digital contracts with different lifecycles are common. For example, "payment delay" is a high risk in digital contracts of any lifecycle. The basic model can capture the general pattern. On the basis of the basic model, an adaptation layer is added. Only minor adjustments are needed to cover the differences between digital contracts with different lifecycles. There is no need to train an independent model for each type of digital contract, nor is there a need for repeated training, which can reduce maintenance costs.

[0239] In addition, some digital contracts may contain clauses that allow for the conversion of a trial period to a long-term contract, such as automatic renewal for two years after the trial period is completed. In this case, the method of adding an adaptation layer to the basic model in the embodiments of this application can quickly switch from the trial period mode to the long-term mode without changing the model, and can dynamically adapt to the needs.

[0240] In this embodiment, when conducting risk assessment based on a trusted data space, the method first determines whether various data and information have been tampered with based on anchor hashing. Then, a pre-defined large language model is used to perform semantic parsing on the digital contract and construct multiple risk indicators for each clause in the digital contract. Using the pre-defined large language model corresponding to participant information, transaction data, environmental data, and data space operation data, semantic transformation is performed on these data to identify fields related to each risk indicator. Then, the risk characteristic value of each risk indicator is determined based on the fields corresponding to each risk indicator, and a risk judgment is made based on the risk characteristic value of each risk indicator. This method, using various pre-defined large language models, can determine changes in multi-source heterogeneous data such as digital contracts, participant information, transaction data, environmental data, and data space operation data in real time and make risk judgments in real time. Furthermore, when judging risk based on the risk characteristic value of each risk indicator, the importance of each risk indicator in the overall data contract is also considered, thus making the judgment more accurate.

[0241] Exemplary device

[0242] This application also proposes a risk assessment device 100 for a trusted data space, such as... Figure 2 As shown in the embodiments of this application, the risk assessment setting 100 of the trusted data space includes:

[0243] The acquisition module 110 is used to acquire digital contracts, participant information, transaction data, environmental data, and data space operation data of the trusted data space; the digital contracts, participant information, transaction data, environmental data, and data space operation data are multi-source heterogeneous data;

[0244] Processing module 120,

[0245] This is used to perform anchor hash calculations based on the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data to determine whether the data has been tampered with.

[0246] If the digital contract has not been tampered with, the text semantics of the digital contract are analyzed by a preset large language model to obtain multiple risk indicators corresponding to each clause in the digital contract.

[0247] By using a pre-defined large language independent model corresponding to the participant information, transaction data, environmental data, and data space operation data respectively, semantic recognition and associated field matching are performed on each data or information to obtain the risk feature value of each risk indicator; the risk feature value of each risk indicator is obtained based on the field content of the corresponding risk indicator.

[0248] Based on the global weights and risk characteristic values ​​of each risk indicator, the magnitude of each preset risk level of the digital contract is obtained; wherein, the global weight of each risk indicator is obtained by analyzing a preset risk causal model based on the type of clause to which each risk indicator belongs, and the preset risk causal model is pre-trained based on a performance data set, which includes the digital contract clause type and performance result.

[0249] In this embodiment of the application, the processing module 120 is further configured to:

[0250] Identify the core fields of the digital contract, the participant information, the transaction data, the environmental data, the data space operation data, and the anchor clauses of each core field in the digital contract;

[0251] The core fields of the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data are bound to their corresponding anchor terms to generate various binding results;

[0252] Hash operations are performed on the binding results corresponding to the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data respectively to obtain the current contract anchor hash value corresponding to the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data.

[0253] Based on the current contract anchor hash value corresponding to the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data, respectively, compare it with the initial contract anchor hash value corresponding to the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data in the digital space to determine whether the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data have been tampered with.

[0254] In this embodiment of the application, the processing module 120 is further configured to:

[0255] The clause types in the digital contract are identified based on the preset large language model. The clause types include core clauses, auxiliary clauses, and external environment clauses.

[0256] Based on the core clauses, the auxiliary clauses, and the external environment clauses, risk indicators are constructed for each category of clauses.

[0257] In this embodiment of the application, the processing module 120 is further configured to:

[0258] Based on the relative importance of the core clauses, auxiliary clauses, and external environment clauses using the aforementioned risk causal model, a first judgment matrix is ​​constructed.

[0259] Based on the first judgment matrix, the benchmark weights of the core clauses, the auxiliary clauses, and the external environment clauses are determined.

[0260] Based on the relative importance of various risk indicators under any type of clause, a second judgment matrix is ​​constructed for each type of clause.

[0261] Based on the second judgment matrix corresponding to each type of clause, the local weights of each risk indicator under each type of clause are determined;

[0262] The global weight of each risk indicator is determined based on its local weight and the benchmark weight corresponding to the type clause in which it belongs.

[0263] In this embodiment of the application, the processing module 120 is further configured to:

[0264] A fuzzy relation matrix is ​​constructed based on the membership vectors of each risk indicator at each risk level.

[0265] The risk level of each risk level is determined by using the global weight of each risk indicator as the weight vector of the fuzzy relation matrix.

[0266] In this embodiment of the application, the processing module 120 is further configured to:

[0267] Determine the dynamic performance terms in the digital contract;

[0268] Based on the transaction data and the data space operation data, obtain the time-series data corresponding to the dynamic performance terms;

[0269] Based on the aforementioned time-series data, a sliding window is constructed;

[0270] Based on the sliding window, the participating parties' ability to perform their obligations is determined.

[0271] In this embodiment of the application, the processing module 120 is further configured to:

[0272] Determine the lifecycle type of the digital contract, which includes trial period contracts, long-term contracts, and short-term contracts;

[0273] If the digital contract is a trial period contract, the risk indicator weight corresponding to the initial performance terms shall be increased;

[0274] If the digital contract is a long-term contract, the weight of the risk indicator corresponding to the long-term impact clause shall be increased.

[0275] If the digital contract is a short-term contract, the weight of the risk indicator corresponding to the short-term goal completion clause is increased.

[0276] In this embodiment, when conducting risk assessment based on a trusted data space, the processing module 120 first determines whether various data and information have been tampered with based on anchor hashes. Then, it uses a preset large language model to perform semantic parsing on the digital contract and constructs multiple risk indicators for each clause in the digital contract. Using the preset large language model corresponding to participant information, transaction data, environmental data, and data space operation data, it performs semantic transformation on the participant information, transaction data, environmental data, and data space operation data respectively, identifying fields related to each risk indicator. Then, it determines the risk characteristic value of each risk indicator based on the fields corresponding to each risk indicator, and makes a risk judgment based on the risk characteristic value of each risk indicator. This method, using various preset large language models, can determine changes in multi-source heterogeneous data such as digital contracts, participant information, transaction data, environmental data, and data space operation data in real time and make risk judgments in real time. Furthermore, when judging risk based on the risk characteristic value of each risk indicator, it also considers the importance of each risk indicator in the overall data contract, thus making the judgment more accurate.

[0277] Exemplary media

[0278] After introducing the methods, media, and systems of exemplary embodiments of this application, the following references are made. Figure 3 The computer-readable storage medium of exemplary embodiments of this application will be described, please refer to... Figure 3 The computer-readable storage medium shown is an optical disc 70, on which a computer program (i.e., a program product) is stored. When the computer program is run by a processor, it implements the steps described in the above method embodiments, for example,

[0279] Obtain digital contracts, participant information, transaction data, environmental data, and data space operation data from a trusted data space;

[0280] By performing textual semantic parsing on the digital contract using a pre-set large language model, multiple risk indicators corresponding to each clause in the digital contract are constructed.

[0281] The participant information, transaction data, environmental data, and data space operation data are semantically transformed to identify the fields in the participant information, transaction data, environmental data, and data space that correspond to each risk indicator. Each field is then mapped to its corresponding risk indicator to obtain the risk characteristic value of each risk indicator.

[0282] Based on the relative importance of each risk indicator, the global weight of each risk indicator is determined.

[0283] Based on the risk characteristic values ​​of each risk indicator, the membership vector of each risk indicator under each preset risk level is determined.

[0284] Based on the membership vectors of each risk indicator at each risk level and the global weights of each risk indicator, the risk level of each risk level is determined. The specific implementation methods of each step will not be repeated here. It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory, or other optical and magnetic storage media, which will not be elaborated upon here.

[0285] Exemplary computing device

[0286] After introducing the methods, systems, and media of exemplary embodiments of this application, the following references are made. Figure 4 Computing devices according to exemplary embodiments of this application.

[0287] Figure 4 A block diagram is shown of an exemplary computing device 80 suitable for implementing embodiments of the present application. The computing device 80 may be a computer system or a server. Figure 4 The computing device 80 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0288] like Figure 4 As shown, the components of the computing device 80 may include, but are not limited to: one or more processors or processing units 801, system memory 802, and bus 803 connecting different system components (including system memory 802 and processing unit 801).

[0289] The computing device 80 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the computing device 80, including volatile and non-volatile media, removable and non-removable media.

[0290] System storage 802 may include computer system readable media in the form of volatile storage, such as random access memory (RAM) 8021 and / or cache storage 8022. Computing device 80 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, ROM 8023 may be used to read and write non-removable, non-volatile magnetic media (…). Figure 4 (Not shown in the image, usually referred to as "hard drive"). Although not shown in... Figure 4 The diagram illustrates that a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) can be provided. In these cases, each drive can be connected to bus 803 via one or more data media interfaces. System storage 802 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of this application.

[0291] A program / utility 8025 having a set (at least one) of program modules 8024 may be stored, for example, in system storage 802, and such program modules 8024 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment. Program modules 8024 typically perform the functions and / or methods described in the embodiments of this application.

[0292] The computing device 80 can also communicate with one or more external devices 804 (such as a keyboard, pointing device, display, etc.). This communication can be performed through the I / O interface 805. Furthermore, the computing device 80 can also communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 806. Figure 4 As shown, network adapter 806 communicates with other modules of computing device 80 (such as processing unit 801) via bus 803. It should be understood that, although... Figure 4 As not shown, it can be used in conjunction with computing device 80 with other hardware and / or software modules.

[0293] Processing unit 801 executes various functional applications and data processing by running programs stored in system memory 802, for example,

[0294] Obtain digital contracts, participant information, transaction data, environmental data, and data space operation data from a trusted data space;

[0295] By performing textual semantic parsing on the digital contract using a pre-set large language model, multiple risk indicators corresponding to each clause in the digital contract are constructed.

[0296] The participant information, transaction data, environmental data, and data space operation data are semantically transformed to identify the fields in the participant information, transaction data, environmental data, and data space that correspond to each risk indicator. Each field is then mapped to its corresponding risk indicator to obtain the risk characteristic value of each risk indicator.

[0297] Based on the relative importance of each risk indicator, the global weight of each risk indicator is determined.

[0298] Based on the risk characteristic values ​​of each risk indicator, the membership vector of each risk indicator under each preset risk level is determined.

[0299] Based on the membership vectors of each risk indicator at each risk level and the global weights of each risk indicator, the risk level of each risk level is determined. The specific implementation methods of each step will not be repeated here. It should be noted that although several units / modules or sub-units / sub-modules of the risk assessment device for the trusted data space are mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0300] Finally, it should be noted that the above-described embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The protection scope of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the technical scope disclosed in this application. Such modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be covered within the protection scope of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.

[0301] Furthermore, although the operations of the method of this application are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

Claims

1. A risk assessment method based on trusted data space, comprising: Obtain digital contracts, participant information, transaction data, environmental data, and data space operation data from a trusted data space; The digital contract, the participant information, the transaction data, the environmental data, and the data space operation data are multi-source heterogeneous data; Anchor hash calculations are performed on the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data to determine whether the data has been tampered with. If the digital contract has not been tampered with, the text semantics of the digital contract are analyzed by a preset large language model to obtain multiple risk indicators corresponding to each clause in the digital contract. By using a pre-defined large language independent model corresponding to the participant information, the transaction data, the environmental data, and the data space operation data respectively, semantic recognition and related field matching are performed on each data or information to obtain the risk feature value of each risk indicator; The risk characteristic value of each risk indicator is obtained based on the field content of the corresponding risk indicator; Based on the global weights and risk characteristic values ​​of each risk indicator, the magnitude of each preset risk level of the digital contract is obtained; wherein, the global weight of each risk indicator is obtained by analyzing a preset risk causal model based on the type of clause to which each risk indicator belongs, and the preset risk causal model is pre-trained based on a performance data set, which includes the digital contract clause type and performance result.

2. The risk assessment method based on a trusted data space as described in claim 1, wherein the step of performing anchor hash calculations on the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data respectively to determine whether they have been tampered with includes: Identify the core fields of the digital contract, the participant information, the transaction data, the environmental data, the data space operation data, and the anchor clauses of each core field in the digital contract; The core fields of the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data are bound to their corresponding anchor terms to generate various binding results; Hash operations are performed on the binding results corresponding to the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data respectively to obtain the current contract anchor hash value corresponding to the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data. Based on the current contract anchor hash value corresponding to the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data, respectively, compare it with the initial contract anchor hash value corresponding to the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data in the trusted data space to determine whether the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data have been tampered with.

3. The risk assessment method based on trusted data space as described in claim 1, wherein the step of performing text semantic parsing on the digital contract using a preset large language model to obtain multiple risk indicators corresponding to each clause in the digital contract includes: The clause types in the digital contract are identified based on the preset large language model. The clause types include core clauses, auxiliary clauses, and external environment clauses. Based on the core clauses, the auxiliary clauses, and the external environment clauses, risk indicators are constructed for each category of clauses.

4. The risk assessment method based on trusted data space as described in claim 3, wherein the global weights of each risk indicator are obtained based on the following method: Based on the relative importance of the core clauses, auxiliary clauses, and external environment clauses using the aforementioned risk causal model, a first judgment matrix is ​​constructed. Based on the first judgment matrix, the benchmark weights of the core clauses, the auxiliary clauses, and the external environment clauses are determined. Based on the relative importance of various risk indicators under any type of clause, a second judgment matrix is ​​constructed for each type of clause. Based on the second judgment matrix corresponding to each type of clause, the local weights of each risk indicator under each type of clause are determined; The global weight of each risk indicator is determined based on its local weight and the benchmark weight corresponding to the type clause in which it belongs.

5. The risk assessment method based on trusted data space as described in claim 1, wherein obtaining the magnitude of each preset risk level of the digital contract based on the global weight and risk characteristic value of each risk indicator includes: A fuzzy relation matrix is ​​constructed based on the membership vectors of each risk indicator at each risk level. The risk level of each risk level is determined by using the global weight of each risk indicator as the weight vector of the fuzzy relation matrix.

6. The risk assessment method based on a trusted data space as described in claim 1, further comprising, after acquiring the digital contracts, participant information, transaction data, environmental data, and data space operation data of the trusted data space: Determine the dynamic performance terms in the digital contract; Based on the transaction data and the data space operation data, obtain the time-series data corresponding to the dynamic performance terms; Based on the aforementioned time-series data, a sliding window is constructed; Based on the sliding window, the participating parties' ability to perform their obligations is determined.

7. The risk assessment method based on trusted data space as described in claim 1, wherein after obtaining the digital contract of the trusted data space, the method further includes: Determine the lifecycle type of the digital contract, which includes trial period contracts, long-term contracts, and short-term contracts; If the digital contract is a trial period contract, the risk indicator weight corresponding to the initial performance terms shall be increased; If the digital contract is a long-term contract, the weight of the risk indicator corresponding to the long-term impact clause shall be increased. If the digital contract is a short-term contract, the weight of the risk indicator corresponding to the short-term goal completion clause is increased.

8. A risk assessment device based on a trusted data space, comprising: The acquisition module is used to acquire digital contracts, participant information, transaction data, environmental data, and data space operation data from the trusted data space. The digital contract, the participant information, the transaction data, the environmental data, and the data space operation data are multi-source heterogeneous data; The processing module is used to perform anchor hash calculations based on the digital contract, the participant information, the transaction data, the environmental data, and the data space operation data to determine whether the data has been tampered with. If the digital contract has not been tampered with, the text semantics of the digital contract are analyzed by a preset large language model to obtain multiple risk indicators corresponding to each clause in the digital contract. By using a pre-defined large language independent model corresponding to the participant information, the transaction data, the environmental data, and the data space operation data respectively, semantic recognition and related field matching are performed on each data or information to obtain the risk feature value of each risk indicator; The risk characteristic value of each risk indicator is obtained based on the field content of the corresponding risk indicator; Based on the global weights and risk characteristic values ​​of each risk indicator, the magnitude of each preset risk level of the digital contract is obtained; wherein, the global weight of each risk indicator is obtained by analyzing a preset risk causal model based on the type of clause to which each risk indicator belongs, and the preset risk causal model is pre-trained based on a performance data set, which includes the digital contract clause type and performance result.

9. A computer-readable storage medium comprising instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1-7.

10. A computing device comprising a processor having a computer program stored thereon, the computer program executing to implement the method as described in any one of claims 1-7.

Images