Risk identification method, system and equipment based on electronic contract text and medium

Abstract

The invention discloses a risk identification method, system and device based on an electronic contract text and a medium, and relates to the technical field of artificial intelligence, and the method comprises the steps: carrying out the preprocessing of the electronic contract text, and obtaining a to-be-analyzed contract text; using a deep learning model to obtain semantic representation of the contract text according to the to-be-analyzed contract text; performing first evaluation processing on the semantic representation by utilizing a constructed vocabulary library to obtain a first evaluation result of the semantic representation; and performing second evaluation processing on the first evaluation result based on a target word vector model and a Drools rule engine to obtain a target evaluation result of the to-be-analyzed contract text. The method is suitable for a contract analysis management system, and the automation and intelligence of the contract analysis management system are improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to risk identification methods, systems, devices and media based on electronic contract texts. Background Art

[0002] With the rapid development of e-commerce and the advancement of digital transformation in various industries, electronic contracts, as an indispensable part of commercial transactions, are becoming increasingly popular and applied. However, with the increase in the number and complexity of electronic contracts, it has become a challenge to manage and analyze whether their content is reasonable and legal and whether they comply with specific rules.

[0003] In practice, with the rapid development and digital transformation of e-commerce, the use of electronic contracts has exploded. The content of the contract involves a wide range of businesses and projects. There may be risk issues in the contract. If they are not discovered and handled in time, they may cause financial losses and legal risks to the company or individual. At present, there are various electronic contract management systems on the market, but they are designed to provide contract storage, tracking and management functions. The analysis of whether its content is reasonable and legal still relies on manual operation, lacks the ability of automated and intelligent analysis, and has the problem of low efficiency and accuracy in the analysis of contract content.

[0004] It can be seen that the existing risk identification scheme based on electronic contract texts has the following shortcomings: the rationality of the electronic contract text depends on manual review by the company's legal department, which has high labor costs, and the contract cannot be properly handled within a reasonable time frame. The time cost is high and the number of contracts is large and the content is rich. Manual analysis is prone to omissions and has a high error rate. Summary of the invention

[0005] In view of this, the present application provides a risk identification method, system, device and medium based on electronic contract text, the main purpose of which is to solve the technical problems of high labor cost, high time cost and high error rate in the existing electronic contract risk analysis.

[0006] According to one aspect of the present application, a risk identification method based on an electronic contract text is provided, the method comprising:

[0007] Preprocess the electronic contract text to obtain the contract text to be analyzed;

[0008] Using a deep learning model, a semantic representation of the contract text is obtained according to the contract text to be analyzed;

[0009] Performing a first evaluation process on the semantic representation using the constructed vocabulary library to obtain a first evaluation result of the semantic representation;

[0010] Based on the target word vector model and the Drools rule engine, a second evaluation process is performed on the first evaluation result to obtain a target evaluation result of the contract text to be analyzed.

[0011] According to another aspect of the present application, a risk identification system based on an electronic contract text is provided, the system comprising:

[0012] A preprocessing module is used to preprocess the electronic contract text to obtain the contract text to be analyzed;

[0013] A semantic analysis module, used to obtain a semantic representation of the contract text according to the contract text to be analyzed by using a deep learning model;

[0014] A first evaluation module, used for performing a first evaluation process on the semantic representation using the constructed vocabulary library to obtain a first evaluation result of the semantic representation;

[0015] The second evaluation module is used to perform a second evaluation process on the first evaluation result based on the target word vector model and the Drools rule engine to obtain a target evaluation result of the contract text to be analyzed.

[0016] According to another aspect of the present application, a computer storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned risk identification method based on electronic contract text is implemented.

[0017] According to another aspect of the present application, a computer device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor implements the above-mentioned risk identification method based on electronic contract text when executing the program.

[0018] By means of the above technical scheme, the risk identification method, system, device and medium based on the electronic contract text provided by the present application are compared with the technical scheme of the electronic contract management system with high labor cost, high time cost and high error rate for the existing electronic contract risk analysis. The present application pre-processes the electronic contract text to obtain the contract text to be analyzed; uses the deep learning model to obtain the semantic representation of the contract text according to the contract text to be analyzed; uses the constructed vocabulary library to perform a first evaluation process on the semantic representation to obtain the first evaluation result of the semantic representation; based on the target word vector model and the Drools rule engine, performs a second evaluation process on the first evaluation result to obtain the target evaluation result of the contract text to be analyzed. It can be seen that by performing a hierarchical evaluation process on the contract text to be analyzed, the automation and intelligent capabilities of the reasonable and legal analysis of the contract text to be analyzed are improved, the lack of automated analysis in the existing electronic contract management system and the demand for intelligent processing of contract content are solved, so as to effectively identify and manage the potential risks in the electronic contract, and provide a more efficient and safer solution for the contract management of the enterprise.

[0019] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0021] Figure 1 A schematic diagram of a process of risk identification method based on electronic contract text provided in an embodiment of the present application is shown;

[0022] Figure 2 A schematic diagram of a process of another risk identification method based on an electronic contract text provided in an embodiment of the present application is shown;

[0023] Figure 3 A schematic diagram of the structure of a risk identification system based on electronic contract text provided by an embodiment of the present application is shown;

[0024] Figure 4 A structural schematic diagram of another risk identification system based on electronic contract text provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0025] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other without conflict.

[0026] In view of the technical problems of high labor cost, high time cost and high error rate in the existing electronic contract risk analysis, this embodiment provides a risk identification method based on electronic contract text, which improves the automation and intelligent ability of reasonable and legal analysis of the contract text to be analyzed by performing hierarchical evaluation and processing, and provides a more efficient and safer solution for the contract management of enterprises. Figure 1 As shown, the above method comprises the following steps:

[0027] Step 101: pre-process the electronic contract text to obtain the contract text to be analyzed.

[0028] In this embodiment, the electronic contract text in the contract analysis and management system is preprocessed. The contract analysis and management system obtains the electronic contract text from the internal contract archiving system of the enterprise and the third-party contract management platform, which can ensure the integrity, accuracy and security of the electronic contract text. In the preprocessing stage, the natural language processing technology NLP is used to tokenize, segment and stem the electronic contract text so as to convert the electronic contract text into a form that the system can understand and process. In addition, the preprocessing can also clean and normalize the electronic contract text, remove unnecessary special characters, stop words and noise data, so as to reduce interference and errors in the subsequent analysis process.

[0029] Step 102: Utilize a deep learning model to obtain a semantic representation of the contract text based on the contract text to be analyzed.

[0030] In this embodiment, the Transformer model is used for semantic analysis. Natural language processing technology NLP can be used to establish the semantic representation of the electronic contract text first, and then the semantic representation and contextual semantic representation of the electronic contract text obtained by sequence-to-sequence semantic analysis can be used to perform intelligent content analysis and risk assessment later. According to the needs of the actual application scenario, the contract analysis management system of this embodiment adopts the Transformer model to ensure the accuracy and efficiency of the analysis. It should be noted that the natural language processing technology NLP here is the natural language processing in the trained Transformer model, which is different from the natural language processing used for tokenization, word segmentation and stemming in step 101, thereby effectively improving the accuracy of the semantic analysis of the electronic contract text.

[0031] Step 103: Perform a first evaluation process on the semantic representation using the constructed vocabulary library to obtain a first evaluation result of the semantic representation.

[0032] In this embodiment, a vocabulary library is constructed for performing a first evaluation process on the semantic representation, and the semantic representation includes a semantic representation of the electronic contract text and a contextual semantic representation. The vocabulary library may include one or more. According to the needs of the actual application scenario, a first vocabulary library for representing the untrustworthy blacklist can be constructed, and the first vocabulary library includes multiple first semantic representations for representing the untrustworthy blacklist, and a second vocabulary library for representing the second semantic representation of sensitive semantics can be constructed, and the first vocabulary library includes multiple second semantic representations for representing sensitive semantics. It should be noted that the vocabulary library here can be updated regularly, and the number of vocabulary libraries can be increased, and the semantic representations in the vocabulary libraries can be increased, which are not specifically limited here.

[0033] Furthermore, the constructed vocabulary library is used to perform a first evaluation process on the semantic representation and contextual semantic representation of the electronic contract text, and a first evaluation result of the contract text to be analyzed is obtained by identifying whether the semantic representation and contextual semantic representation of the electronic contract text contain a first semantic representation used to characterize the default blacklist, and a second semantic representation used to characterize sensitive semantics. The first evaluation result may include a first recognition result of the first semantic representation and / or the second semantic representation contained in the semantic representation and contextual semantic representation of the electronic contract text, and a first evaluation result of filtering the first recognition result containing the first semantic representation and / or the second semantic representation.

[0034] According to the needs of the actual application scenario, the first evaluation process may also include marking the first semantic representation and / or second semantic representation identified in the semantic representation of the electronic contract text and the contextual semantic representation to obtain a first evaluation result including the marking process result. The first evaluation process and the first evaluation result are not specifically limited here.

[0035] Step 104: Based on the target word vector model and the Drools rule engine, perform a second evaluation process on the first evaluation result to obtain a target evaluation result of the contract text to be analyzed.

[0036] In this embodiment, a second evaluation process is performed based on the first evaluation result. The second evaluation process may be based on an irrationality analysis of legal terms and a risk level assessment. The irrationality analysis may include identifying whether there are ambiguous terms, unequal terms, risky terms, etc. in the first evaluation result. After the contract analysis management system of this embodiment performs a preliminary evaluation using step 103, it can further evaluate the electronic contract text to identify ambiguous terms, unreasonable conditions or potential legal risk points in the electronic contract text, thereby effectively avoiding unnecessary problems and disputes encountered by enterprises during the contract execution process, and effectively improving the automation and intelligence of the contract analysis management system.

[0037] For this embodiment, the electronic contract text can be preprocessed according to the above scheme to obtain the contract text to be analyzed; the deep learning model is used to obtain the semantic representation of the contract text according to the contract text to be analyzed; the semantic representation is first evaluated by using the constructed vocabulary library to obtain the first evaluation result of the semantic representation; the first evaluation result is second evaluated based on the target word vector model to obtain the target evaluation result of the contract text to be analyzed. Compared with the technical solutions of the existing electronic contract risk analysis electronic contract management system with high labor cost, high time cost and high error rate, the hierarchical evaluation of the contract text to be analyzed is performed to improve the automation and intelligent capabilities of the reasonable and legal analysis of the contract text to be analyzed, solve the lack of automated analysis in the existing electronic contract management system, and solve the need for intelligent processing of contract content, so as to effectively identify and manage potential risks in electronic contracts and provide a more efficient and safer solution for enterprise contract management.

[0038] Furthermore, as a refinement and expansion of the specific implementation of the above embodiment, in order to fully illustrate the specific implementation process of this embodiment, another risk identification method based on electronic contract text is provided. This embodiment can be applied to the contract analysis management system. This embodiment can improve the automation and intelligent capabilities of reasonable and legal analysis of the contract text to be analyzed by performing hierarchical evaluation processing on the contract text to be analyzed. Figure 2 As shown, the method includes:

[0039] Step 201: pre-process the electronic contract text to obtain the contract text to be analyzed.

[0040] Step 202: Utilize a deep learning model to obtain a semantic representation of the contract text based on the contract text to be analyzed.

[0041] In implementation, step 201 is consistent with step 101 , and step 202 is consistent with step 102 .

[0042] Step 203: construct a first vocabulary library for identifying whether the semantic representation includes a first semantic representation for representing a dishonest blacklist.

[0043] Step 204: construct a second vocabulary library for identifying whether the semantic representation includes a second semantic representation for representing sensitive semantics.

[0044] Step 205: Perform a first evaluation process on the semantic representation using the constructed vocabulary library to obtain a first evaluation result of the semantic representation.

[0045] During implementation, the first evaluation result is the first semantic representation and / or the second semantic representation contained in the semantic representation of the contract text. Construct a blacklist library (first vocabulary library) and a sensitive vocabulary library (second vocabulary library). Specifically, the blacklist library can also include multiple blacklist sub-libraries, for example, a blacklist sub-library for untrustworthy enterprises: by calling the "National Enterprise Credit Information Publicity System" interface, the untrustworthy enterprise blacklist is obtained in real time, and added to the untrustworthy enterprise blacklist sub-library; the untrustworthy individual blacklist sub-library: by calling the "China Enforcement Information Disclosure Network" interface, the untrustworthy person blacklist is obtained in real time, and added to the untrustworthy individual blacklist sub-library; the sensitive vocabulary library includes a variety of sensitive vocabulary sub-libraries, and the content of the sensitive vocabulary sub-library is not specifically limited here; according to the needs of the actual application scenario, manual maintenance can also be performed by manual collection to reduce cooperation risks.

[0046] Step 206: Build multiple contract element libraries based on the business attributes of the legal terms.

[0047] To illustrate the specific implementation of step 206, as a preferred embodiment, step 206 includes: constructing a first contract element library for identifying whether the semantic representation contains a first contract element for representing ambiguous clauses; constructing a second contract element library for identifying whether the semantic representation contains a second contract element for representing unequal clauses; and constructing a third contract element library for identifying whether the semantic representation contains a third contract element for representing risky clauses.

[0048] During implementation, a first contract element library for characterizing ambiguous clauses is established: natural language processing or manual methods can be used to collect and organize vocabulary and phrases of ambiguous clauses from contract cases, legal documents and legal research based on past experience and cases, which can specifically include: ambiguous quality standard clauses: the quality standard of the product or service is not specified; ambiguous performance time clauses: the time or period of performance is not clearly stipulated; ambiguous price clauses: the price clauses are vague and the specific amount or calculation method is not clear; ambiguous breach of contract liability clauses: the description of breach of contract liability is unclear and the scope of compensation or liability is not specified; ambiguous dispute resolution method clauses: the description of the dispute resolution mechanism is not specific, and the specific procedures for mediation, arbitration or litigation in dispute resolution are not clear; ambiguous confidentiality clauses: the scope or requirements of confidential information are unclear; ambiguous termination clauses: the conditions for termination of the contract are unclear.

[0049] Accordingly, the second contract element library used to characterize unequal terms can specifically include: unfair terms: unilateral exemption clauses, unequal liability clauses, unfair price clauses; unfair performance clauses: overly harsh performance requirements, unreasonable liquidated damages, overly strict performance periods; unfair overly restrictive clauses: overly restrictive competition clauses (non-competition clauses), unreasonable transfer clauses, that is, unreasonable restrictions on the transfer of contract rights and obligations, which affect the inability of one party to the contract to effectively perform the contract; unfair confidentiality and intellectual property clauses: too wide a confidentiality scope, unclear ownership of intellectual property rights; unfair dispute resolution clauses: unreasonable dispute resolution mechanisms, that is, the selection of designated arbitration institutions or courts. Accordingly, the third contract element library used to characterize risky clauses can specifically include: potential legal risk points of the contract subject, that is, the contract party has a dishonest person or dishonest enterprise; potential legal risk points of breach of contract liability and compensation clauses; potential legal risk points of contract performance period; potential legal risk points of payment clauses, that is, unclear payment clauses, and unclear amount, time, and method; potential legal risk points of contract modification and termination clauses.

[0050] Step 207: Based on the multiple contract element libraries, a word vector model is trained to perform a second evaluation process on the first evaluation result.

[0051] In order to illustrate the specific implementation method of step 207, as a preferred embodiment, step 207 includes: constructing a contract text sample set and an initialization word vector model; training the initialization word vector model according to the contract text sample set and the multiple contract element libraries to obtain a target word vector model for identifying whether the contract text sample set contains corresponding contract elements in the multiple contract element libraries; wherein the contract text sample set includes a contract text training set, a contract text verification set and a contract text test set.

[0052] During implementation, a sample set of contract texts is constructed, and the contract elements in the sample set are respectively marked as belonging to the first contract element of ambiguous clauses, the second contract element of unequal clauses, or the third contract element of potential legal risk points (as well as the different risk levels corresponding to the contract elements in the electronic contract text). Further, the sample set of contract texts is divided into a 70% training set, a 15% validation set, and a 15% test set. The training set is used to train the word vector model, the validation set is used to further adjust and optimize the parameters of the word vector model, and the test set is used to evaluate the trained word vector model and obtain the final target word vector model.

[0053] Furthermore, the training process of the initialization word vector model specifically includes: converting the electronic contract text into Unicode text format, removing stop words and punctuation marks through word segmentation processing to obtain the semantic representation of the electronic contract text, identifying whether the electronic contract text contains words or phrases in the blacklist and the sensitive word library through the first evaluation processing method, and using the initialization word vector model Word2Vec to obtain the semantic relationship between the semantic representations by mapping the first evaluation result to the position in the high-dimensional vector space to characterize the words or phrases in the electronic contract text. Each semantic representation is taken as a vector. In the vector space, semantically similar words or phrases will have similar vector representations. Therefore, the semantic representation in the electronic contract text is determined by calculating the cosine similarity with the first contract element belonging to the ambiguous clause, the second contract element of the unequal clause or the third contract element of the potential legal risk point, as well as the semantic similarity between different risk levels.

[0054] Step 208: Based on the target word vector model and the Drools rule engine, perform a second evaluation process on the first evaluation result to obtain a target evaluation result of the contract text to be analyzed.

[0055] In order to illustrate the specific implementation method of step 208, as a preferred embodiment, step 208 also includes: using the target word vector model to perform a first risk identification process on the first assessment result to obtain a first risk identification result in the target assessment result; using the constructed Drools rule engine to perform a second risk identification process on the first risk identification result to obtain a second risk identification result in the target assessment result.

[0056] In implementation, the target word vector model obtained based on training can identify the first risk identification result of the first assessment result, that is, the first risk identification result includes whether the electronic contract text contains the first contract element belonging to the ambiguous clause, the second contract element of the unequal clause, or the third contract element of the potential legal risk point (as well as the different risk levels corresponding to the contract elements in the electronic contract text). It should be noted that the different risk levels corresponding to the contract elements here are different from the risk level judgment conditions in which different rules correspond to different risk levels. Different rules correspond to different risk levels, which may include multiple contract elements corresponding to one rule condition, that is, using the constructed Drools rule engine, the low, medium, and high risk levels of the electronic contract text are determined based on the rule conditions.

[0057] To illustrate the specific implementation of step 208, as a preferred embodiment, the step of using the constructed Drools rule engine to perform a second risk identification process on the first risk identification result to obtain the second risk identification result in the target assessment result includes: using the risk level rule table of contract elements in the contract element library contained in the Drools rule engine, and obtaining the risk level of the corresponding contract element in the first risk identification result through the second risk identification process.

[0058] During implementation, the Drools rule engine framework is integrated into the contract analysis management system, and the Drools rule engine is used to convert rule conditions into rule files (risk level rule library), that is, the risk level corresponding to each rule condition is defined, so that the first risk identification result output by the target word vector model is sent to the Drools rule engine, and the Drools rule engine further performs a second risk identification process on the first risk identification result based on the predefined risk level rule library to obtain the risk level of the electronic contract text.

[0059] It should be noted that the first risk identification result based on the output of the word vector model may also include different risk levels corresponding to multiple contract elements. It can be obtained by training the word vector model, or by the Drools rule engine. It can also be based on the word vector model After the risk level result is obtained, the Drools rule engine is used to verify, modify and supplement the risk level result to obtain the risk level of the electronic contract text. The training process of the word vector model and the combination of the word vector model and the Drools rule engine are not specifically limited here.

[0060] According to the needs of actual application scenarios, a risk level rule library is set based on the contract elements in the contract element library, and different rules correspond to different risk levels. It should be noted that the rules and the contract element library and contract elements do not correspond one to one. For example, multiple contract elements meet the conditions of a rule to establish an association between the two. The association between the rule conditions and the risk level specifically includes:

[0061] ① The risk level of the completeness of the contract terms: the low risk level is: the contract terms are complete, and the key terms (information of the two parties to the contract, the purpose and scope of the contract, the quality standards of the subject matter of the contract, the terms of performance, the price and payment terms, the liability for breach of contract, the validity and entry into force information of the contract) are clearly stipulated, but other terms (contract change and termination instructions, guarantees and warranties, dispute resolution mechanism) are not reflected in the contract; the medium risk level is: the key terms of the contract are missing, but the missing amount is not large; the high risk level is: the key terms of the contract are incomplete, and a large number of key terms are missing.

[0062] ② Risk level of clarity and accuracy of contract terms: low risk level means: the terms are clear, accurate and unambiguous; medium risk level means: some terms are not clearly stated or may cause ambiguity; high risk level means: the terms are vague and ambiguous, which may lead to misunderstanding or legal disputes.

[0063] ③ Risk level of breach of contract liability and compensation clauses: low risk level: breach of contract liability and compensation clauses are clear, specific, fair and reasonable. Medium risk level: breach of contract liability and compensation clauses are basically clear, but may need to be further refined. High risk level: breach of contract liability and compensation clauses are unclear, or the clauses are unfair to one party, which may lead to disputes.

[0064] ④ Risk level of contract performance period: low risk level: the contract performance period is clear and both parties have a common understanding of the timetable. Medium risk level: the contract performance period is basically clear, but there may be some ambiguity. High risk level: the performance period is unclear or there is great uncertainty, which may lead to performance problems.

[0065] ⑤ Risk level of payment terms: low risk level: payment terms are clear, with the amount, time and method all specified. Medium risk level: payment terms are basically clear, but some details may need to be further determined. High risk level: payment terms are unclear or unreasonable, which may affect the performance of the contract.

[0066] ⑥ Risk level of the change and termination clauses: low risk level: the change and termination clauses are clear and reasonable, and both parties have clear rights and obligations. Medium risk level: the change and termination clauses are basically clear, but may need further refinement. High risk level: the change and termination clauses are unclear or unreasonable, which may lead to difficulties in contract performance.

[0067] ⑦ Risk level of qualifications and reputation of both parties: low risk level: both parties have good qualifications and reputation, and strong performance ability. Medium risk level: both parties have basically good qualifications and reputation, but there are certain risks. High risk level: both parties have problems with qualifications and reputation, and their performance ability may be insufficient.

[0068] ⑧ Risk level of unreasonable clauses in the contract: low risk level: no unreasonable clauses (unfair clauses, unreasonable performance clauses, unreasonable overly restrictive clauses, unreasonable confidentiality and intellectual property clauses, unreasonable dispute resolution clauses) in the contract. Medium risk level: a small number of clauses are unreasonable, and there are certain risks. High risk level: more than half of the unreasonable clauses may affect the fairness of the contract.

[0069] According to the needs of actual application scenarios, it is also possible to extract an evaluation module from the target word vector model for performing a second evaluation on the first evaluation result, integrate the evaluation module into the constructed Drools rule engine, and perform a second evaluation on the first risk identification result in combination with the risk level rule library preset in the Drools rule engine to obtain the target evaluation result of the contract text to be analyzed, thereby reducing the computational pressure of the contract analysis management system to a certain extent.

[0070] Furthermore, based on the target evaluation results of the contract text to be analyzed, a result visualization report is generated through data visualization technology. The result visualization report can include contract content analysis and risk assessment results, including interactive reports, charts, and heat maps, and displays the risk distribution and key issues (contract elements with high risk levels) of the electronic contract text, presenting the analysis results in an intuitive and easy-to-understand manner. According to the needs of the actual application scenario, it can also include the evaluation summary of the electronic contract text, recommended improvement measures, and legal advice content, which can not only help managers quickly understand the overall situation of contract management, but also provide support and basis for corporate decision-making.

[0071] Furthermore, the contract analysis management system can monitor newly added or modified contract texts in real time, and can also perform manual review operations on electronic contract texts with a high risk ratio exceeding 50%. It can also divide the urgency of contract signing into four levels: severe urgency, high urgency, moderate urgency and low urgency according to the start date of contract signing, and create review work orders of corresponding levels, and configure corresponding staff approval processes to achieve real-time monitoring, continuous optimization and system model updates, and maintain the accuracy and practicality of the analysis results.

[0072] Furthermore, the system can learn from user feedback and actual applications, continuously optimize analysis algorithms and models, and enable them to have efficient data processing and model updating capabilities. This can also be achieved by introducing incremental learning and continuous integration and deployment (CI / CD) methods to ensure that the system can remain flexible and efficient in the face of ever-changing contract management needs.

[0073] By applying the technical solution of this embodiment, the contract content can be analyzed intelligently, quickly and efficiently through an integrated mechanism, which can free up most of the manual review of contracts and reduce labor costs; through the blacklist mechanism and sensitive word mechanism in the vocabulary library, it is possible to avoid dishonest partners, reduce contract risks, and quickly filter unreasonable and illegal content with high accuracy, few errors and low error rates; through the word vector model, the content is quickly identified and analyzed, saving time costs for urgent contracts. It can be seen that by performing hierarchical evaluation and processing on the contract text to be analyzed, the automation and intelligent capabilities of the reasonable and legal analysis of the contract text to be analyzed are improved, the lack of automated analysis in the existing electronic contract management system and the demand for intelligent processing of contract content are solved, thereby effectively identifying and managing potential risks in electronic contracts, and providing a more efficient and secure solution for the contract management of enterprises. Compared with the technical solutions of the existing electronic contract management system with high labor cost, high time cost and high error rate for risk analysis of electronic contracts, the technical challenges of real-time monitoring and continuous optimization, from electronic contract collection and preprocessing to semantic analysis, irrationality analysis and risk level assessment, to result visualization and report generation, are overcome, so that contracts can be evaluated in an automated manner, indicating the potential degree of risk, providing enterprises with efficient and accurate contract management solutions, effectively reducing legal risks and improving business efficiency.

[0074] Further, as Figure 1 The specific implementation of the method, the embodiment of the present application provides a risk identification system based on electronic contract text, such as Figure 3 As shown, the system includes: a preprocessing module 33, a semantic analysis module 34, a first evaluation module 35, and a second evaluation module 36.

[0075] The preprocessing module 33 is used to preprocess the electronic contract text to obtain the contract text to be analyzed.

[0076] The semantic analysis module 34 is used to use a deep learning model to obtain a semantic representation of the contract text based on the contract text to be analyzed.

[0077] The first evaluation module 35 is used to perform a first evaluation process on the semantic representation using the constructed vocabulary library to obtain a first evaluation result of the semantic representation.

[0078] The second evaluation module 36 is used to perform a second evaluation process on the first evaluation result based on the target word vector model and the Drools rule engine to obtain a target evaluation result of the contract text to be analyzed.

[0079] In specific application scenarios, such as Figure 4 As shown, the system further includes a first building module 31 .

[0080] The first construction module 31 is used to construct a first vocabulary library for identifying whether the semantic representation includes a first semantic representation for representing a default blacklist; and to construct a second vocabulary library for identifying whether the semantic representation includes a second semantic representation for representing sensitive semantics; wherein the first evaluation result is the first semantic representation and / or the second semantic representation contained in the semantic representation of the contract text.

[0081] In a specific application scenario, the system further includes a second building module 32 .

[0082] The second construction module 32 is used to construct multiple contract element libraries based on the business attributes of the legal terms; and, based on the multiple contract element libraries, train a word vector model for performing a second evaluation process on the first evaluation result.

[0083] In a specific application scenario, the step of constructing multiple contract element libraries based on the business attributes of legal terms includes: constructing a first contract element library to identify whether the semantic representation contains a first contract element for representing ambiguous terms; and, constructing a second contract element library to identify whether the semantic representation contains a second contract element for representing unequal terms; and, constructing a third contract element library to identify whether the semantic representation contains a third contract element for representing risky terms.

[0084] In a specific application scenario, the step of training a target word vector model for performing a second evaluation on a first evaluation result based on the multiple contract element libraries includes: constructing a contract text sample set and an initialization word vector model; and, based on the contract text sample set and the multiple contract element libraries, training the initialization word vector model to obtain a target word vector model for identifying whether the contract text sample set contains corresponding contract elements in the multiple contract element libraries; wherein the contract text sample set includes a contract text training set, a contract text verification set, and a contract text test set.

[0085] In a specific application scenario, the second evaluation module 36 includes: a first identification submodule 361 and a second identification submodule 362 .

[0086] The first identification submodule 361 is used to perform a first risk identification process on the first assessment result using a target word vector model to obtain a first risk identification result in the target assessment result.

[0087] The second identification submodule 362 is used to use the constructed Drools rule engine to perform a second risk identification process on the first risk identification result to obtain a second risk identification result in the target assessment result.

[0088] In a specific application scenario, the second identification submodule 362 is specifically used to utilize the risk level rule table of contract elements in the contract element library contained in the Drools rule engine to obtain the risk level of the corresponding contract element in the first risk identification result through a second risk identification process.

[0089] In a specific application scenario, the step of locating the abnormal position of the abnormal load device and / or the cable according to the abnormal element value in the target matrix includes: determining the first abnormal error index of the corresponding row number in the first matrix by traversing the number of rows corresponding to the abnormal element value in the target matrix, and the first abnormal error index is used to characterize the abnormal load device; determining the second abnormal error index by traversing the second matrix, and the second abnormal error index is used to characterize the abnormal cable position.

[0090] It should be noted that for other corresponding descriptions of the functional units involved in the risk identification system based on electronic contract text provided in the embodiment of the present application, please refer to Figure 1 and Figure 2 The corresponding description in will not be repeated here.

[0091] Based on the above Figure 1 and Figure 2 The method shown in the embodiment of the present application accordingly provides a computer storage medium on which a computer program is stored, and when the program is executed by a processor, the above-mentioned Figure 1 and Figure 2 The risk identification method based on electronic contract text is shown.

[0092] Based on this understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each implementation scenario of the present application.

[0093] Based on the above Figure 1 , Figure 2 The method shown, and Figure 3 , Figure 4 In order to achieve the above-mentioned purpose, the embodiment of the present application further provides a computer device, which can be a personal computer, a server, a network device, etc. The physical device includes a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to achieve the above-mentioned Figure 1 and Figure 2 The risk identification method based on electronic contract text is shown.

[0094] Optionally, the computer device may further include a user interface, a network interface, a camera, a radio frequency (RF) circuit, a sensor, an audio circuit, a WI-FI module, etc. The user interface may include a display, an input unit such as a keyboard, etc., and the optional user interface may also include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface, a wireless interface (such as a Bluetooth interface, a WI-FI interface), etc.

[0095] Those skilled in the art will appreciate that the computer device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or a combination of certain components, or different component arrangements.

[0096] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of a computer device and supports the operation of information processing programs and other software and / or programs. The network communication module is used to achieve communication between components within the storage medium and with other hardware and software in the physical device.

[0097] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform, or by hardware. By applying the technical solution of the present application, compared with the technical solution of the existing electronic contract risk analysis electronic contract management system with high labor cost, high time cost, and high error rate, this embodiment has technical challenges from electronic contract collection and preprocessing to semantic analysis, irrationality analysis and risk level assessment, and then to result visualization and report generation, real-time monitoring and continuous optimization, so as to evaluate the contract in an automated manner, indicate the potential risk level, and provide enterprises with efficient and accurate contract management solutions, effectively reduce legal risks and improve business efficiency.

[0098] Those skilled in the art will appreciate that the accompanying drawings are only schematic diagrams of a preferred implementation scenario, and the modules or processes in the accompanying drawings are not necessarily necessary for implementing the present application. Those skilled in the art will appreciate that the modules in the system in the implementation scenario can be distributed in the system of the implementation scenario according to the implementation scenario description, or can be changed accordingly and located in one or more systems different from the present implementation scenario. The modules of the above-mentioned implementation scenario can be combined into one module, or can be further split into multiple submodules.

[0099] The above serial numbers of this application are only for description and do not represent the advantages and disadvantages of the implementation scenarios. The above disclosure is only a few specific implementation scenarios of this application, but this application is not limited to them, and any changes that can be thought of by technicians in this field should fall within the scope of protection of this application.

Claims

1. A risk identification method based on electronic contract text, characterized in that: include: Preprocess the electronic contract text to obtain the contract text to be analyzed; Using a deep learning model, a semantic representation of the contract text is obtained according to the contract text to be analyzed; Performing a first evaluation process on the semantic representation using the constructed vocabulary library to obtain a first evaluation result of the semantic representation; Based on the target word vector model and the Drools rule engine, a second evaluation process is performed on the first evaluation result to obtain a target evaluation result of the contract text to be analyzed.

2. The method according to claim 1, characterized in that: Before the step of performing a first evaluation process on the semantic representation using the constructed vocabulary library to obtain a first evaluation result of the semantic representation, the method further includes: Constructing a first vocabulary library for identifying whether the semantic representation includes a first semantic representation for representing a dishonest blacklist; Constructing a second vocabulary library for identifying whether the semantic representation includes a second semantic representation for representing sensitive semantics; The first evaluation result is the first semantic representation and / or the second semantic representation contained in the semantic representation of the contract text.

3. The method according to claim 1, characterized in that Before the step of performing a second evaluation process on the first evaluation result based on the target word vector model to obtain the target evaluation result of the contract text to be analyzed, the method further includes: Build multiple contract element libraries based on the business attributes of legal clauses; Based on the multiple contract element libraries, a word vector model is trained to perform a second evaluation process on the first evaluation result.

4. The method according to claim 3, characterized in that The step of constructing multiple contract element libraries based on the business attributes of legal clauses includes: Constructing a first contract element library for identifying whether the semantic representation contains a first contract element for representing an ambiguous clause; Constructing a second contract element library for identifying whether the semantic representation contains a second contract element for representing unequal terms; A third contract element library is constructed to identify whether the semantic representation contains a third contract element for representing risk clauses.

5. The method according to claim 3 or 4, characterized in that: The step of training a target word vector model for performing a second evaluation process on the first evaluation result based on the multiple contract element libraries includes: Construct a sample set of contract texts and initialize the word vector model; Training an initialization word vector model according to the contract text sample set and the multiple contract element libraries to obtain a target word vector model for identifying whether the contract text sample set contains corresponding contract elements in the multiple contract element libraries; Among them, the contract text sample set includes a contract text training set, a contract text verification set and a contract text test set.

6. The method according to claim 1, characterized in that The step of performing a second evaluation process on the first evaluation result based on the target word vector model and the Drools rule engine to obtain a target evaluation result of the contract text to be analyzed includes: Performing a first risk identification process on the first assessment result using a target word vector model to obtain a first risk identification result in the target assessment result; The constructed Drools rule engine is used to perform a second risk identification process on the first risk identification result to obtain a second risk identification result in the target assessment result.

7. The method according to claim 6, characterized in that The step of performing a second risk identification process on the first risk identification result by using the constructed Drools rule engine to obtain the second risk identification result in the target assessment result includes: The risk level rule table of the contract elements in the contract element library contained in the Drools rule engine is used to obtain the risk level of the corresponding contract element in the first risk identification result through the second risk identification process.

8. A risk identification system based on electronic contract text, characterized in that: include: A preprocessing module is used to preprocess the electronic contract text to obtain the contract text to be analyzed; A semantic analysis module, used to obtain a semantic representation of the contract text according to the contract text to be analyzed by using a deep learning model; A first evaluation module, used for performing a first evaluation process on the semantic representation using the constructed vocabulary library to obtain a first evaluation result of the semantic representation; The second evaluation module is used to perform a second evaluation process on the first evaluation result based on the target word vector model and the Drools rule engine to obtain a target evaluation result of the contract text to be analyzed.

9. A computer storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the risk identification method based on the electronic contract text as described in any one of claims 1 to 7 is implemented.

10. A computer device comprising a storage medium, a processor, and a computer program stored in the storage medium and executable on the processor, characterized in that: When the processor executes the program, the risk identification method based on the electronic contract text described in any one of claims 1 to 7 is implemented.

Images