Contract review closed-loop optimization method and system based on user weak feedback

Abstract

The invention discloses a contract review closed-loop optimization method and system based on user weak feedback, and the key points of the technical scheme are that the system comprises a user input interface; a contract analysis module; a review logic initialization module; an examination model module; an error history retrieval module; a result correction module; a user feedback storage module; a closed loop optimization module; according to the method, the accuracy and efficiency of contract review are improved through an intelligent review process, risk analysis is carried out on the contract by automatically analyzing the contract text, matching the review list and applying dynamically optimized review logic in combination with a large language model, and the system can identify potential problems and improve the contract review efficiency. According to the method, the review logic can be continuously adjusted through weak feedback of the user to form closed-loop optimization, the review rule can be flexibly adjusted according to contract characteristics, industry requirements and regulation changes, so that more accurate and personalized review services are provided for the user, and the review effect can be continuously improved in the process of continuously accumulating the feedback.

Description

Technical Field

[0001] The present invention relates to the technical field of automated review, and particularly to a method and system for optimizing the closed-loop of contract review based on weak user feedback. Background Art

[0002] With the continuous development of enterprises, contracts play an increasingly important role in daily operations. Especially in large enterprises and multinational companies, the number of contracts is huge, and the content is complex, involving diverse legal terms, regulations, and business terms. The management and review of contracts have become extremely important. To ensure the compliance, legality, and risk control of contracts, enterprises usually adopt contract management systems to track and manage the entire life cycle of contracts, covering aspects such as contract drafting, negotiation, review, signing, performance, and archiving.

[0003] Currently, existing contract management systems mostly rely on predefined review checklists and fixed review logics. Usually, during the contract review process, legal personnel set review criteria and review the contract content accordingly. The review mechanisms of these traditional systems rely on manually set rules and have the following significant problems:

[0004] 1. Existing contract review systems usually rely on pre-set review checklists and fixed review logics;

[0005] 2. Most existing systems only record user feedback errors and lack an effective mechanism to collect and respond to user feedback information in a timely manner;

[0006] 3. Current contract review systems lack automated optimization functions based on machine learning and artificial intelligence;

[0007] The root cause of these deficiencies is that traditional contract management systems fail to fully integrate modern artificial intelligence technologies, especially large language models, and lack the capabilities of real-time optimization, dynamic update, and intelligent feedback. To solve the above problems, we propose a method and system for optimizing the closed-loop of contract review based on weak user feedback. Summary of the Invention

[0008] In view of the deficiencies of the prior art, the present invention provides a method and system for optimizing the closed-loop of contract review based on weak user feedback to solve the problems raised in the background art.

[0009] The above technical objectives of the present invention are achieved through the following technical solutions:

[0010] A method for optimizing the closed-loop of contract review based on weak user feedback, the method comprising the following steps:

[0011] S1. The user uploads the contract text to be reviewed to the platform through the system interface. The system preprocesses the contract text, generates basic contract information, and provides data support for subsequent reviews.

[0012] S2. The system parses the uploaded contract text, automatically matches an appropriate review checklist according to the contract type, content, and terms, and generates preliminary review points to ensure the pertinence and accuracy of the review.

[0013] S3. The system loads relevant review logics from the predefined review logic library, and automatically iteratively optimizes the loaded review logics in combination with the large language model. The user can, according to the contract characteristics, industry requirements, or the latest regulations, select to apply new logics, reject new logics, or update the review logics according to needs through the "weak feedback" method. The information fed back by the user is recorded and used as the basis for system adjustment.

[0014] S4. The system executes the contract review through the large language model based on the confirmed review checklist, automatically analyzes the contract text, identifies potential risks, and outputs preliminary review results.

[0015] S5. After the review is completed, the system generates preliminary review results and displays them to the user, allowing the user to provide immediate feedback.

[0016] S6. The system retrieves historical error review data based on the current review logic context, contract context, and machine review results, analyzes and identifies potential misjudgments or errors in similar situations, and provides reference data for subsequent corrections.

[0017] S7. If similar error data is retrieved, the system generates correction suggestions through the large language model, and adjusts the current review logic or review results according to the suggestions. In this step, the system performs intelligent correction on the error results, improves the review accuracy, and avoids repeated errors.

[0018] S8. The corrected review results are displayed to the user, and the user's feedback is collected. The user can submit "weak feedback" through the interface, such as confirmation, rejection, or suggestions for further modification of the review logic.

[0019] S9. The user feedback and review results are stored in the review result index library together, forming a quadruple (Pi, Ci, RMi, RHi) containing the review logic context, contract context, machine review results, and user feedback. This data is used for subsequent optimization and closed-loop updates.

[0020] S10. The system analyzes and applies the user feedback in real time through the closed-loop optimization module, analyzes the historical feedback data in combination with the large language model, and automatically optimizes the review logic. If the error feedback reaches a specific threshold, the closed-loop optimization module triggers a dynamic update of the review logic library, and the system automatically adjusts the review logic and applies it to the next review.

[0021] S11. All feedback information is processed and applied to the review logic, and finally this review process is completed and the next round of review is entered.

[0022] Further, the "weak feedback" includes the following operations:

[0023] The user confirms to apply the new review logic;

[0024] The user rejects the new review logic;

[0025] The user updates the review logic according to actual needs to adapt to contract characteristics, regulatory changes or business requirements.

[0026] Further, the initialization step of the review logic includes:

[0027] The system loads the predefined review logic designed by professional legal personnel and algorithm engineers;

[0028] The system intelligently updates the loaded review logic through a large language model and makes dynamic adjustments according to business requirements, regulatory changes and user feedback.

[0029] Further, the closed-loop optimization module automatically optimizes according to user feedback, historical data and machine review results, and the optimization process includes:

[0030] Update the review logic based on user feedback;

[0031] Automatically correct the system to improve the accuracy of contract review;

[0032] If the cumulative error information reaches a predetermined threshold, the update of the review logic library is automatically triggered to ensure that the review model adapts to the latest business rules and regulatory requirements.

[0033] Further, the error review history retrieval module identifies potential misjudgments in similar review logic contexts and contract contexts based on historical review data and generates corrective suggestions.

[0034] Further, the user feedback information storage module is used to store the weak feedback of the user and generate a quadruple (Pi, Ci, RMi, RHi) according to the feedback for subsequent optimization use.

[0035] Further, the feedback information library is used to record and analyze the results of user interactions, and perform logical correction according to the consistency between user and machine feedback to continuously improve the accuracy of review results.

[0036] The present invention also provides a contract review closed-loop optimization system based on user weak feedback, and the system includes:

[0037] A user input interface for receiving the contract text uploaded by the user and providing a preprocessing function for the basic contract information.

[0038] A contract parsing module for parsing the contract text, automatically identifying the contract type, terms, and matching the review checklist.

[0039] A review logic initialization module for loading a predefined review logic library and dynamically optimizing the loaded review logic in combination with a large language model, allowing users to adjust the logic according to the feedback.

[0040] A review model module for performing contract review, conducting risk analysis using a large language model, and outputting preliminary review results.

[0041] An error history retrieval module for retrieving historical review errors and generating corrective suggestions through a large language model to improve the review accuracy.

[0042] A result correction module for automatically correcting the review results, generating suggestions based on a large language model, and making logical adjustments.

[0043] A user feedback storage module for storing and managing user feedback information and generating quadruple records for use by the closed-loop optimization module.

[0044] A closed-loop optimization module for continuously optimizing the review logic by real-time analyzing user feedback and machine review results. If the error information accumulates to a certain threshold, the system will automatically update the review logic.

[0045] Furthermore, the system has a dynamically updated review logic library, which supports intelligent updates based on user feedback to ensure the continuous optimization of the review logic.

[0046] Furthermore, the closed-loop optimization module makes automated logical adjustments based on user feedback and historical data to ensure the accuracy and adaptability of the contract review results.

[0047] In summary, the present invention mainly has the following beneficial effects:

[0048] 1. By introducing the user "weak feedback" mechanism, the review logic of the present invention can be dynamically adjusted according to the contract content, business requirements, and regulatory changes. Users can update or customize the review logic in real time during the review process to ensure that the system can quickly adapt to different types and contents of contracts, thus avoiding the lag and misjudgment of the system. The system is no longer limited to fixed review points but can flexibly adjust the review checklist and strategies, enabling the system to quickly respond to changes in contract terms and business requirements, significantly improving the dynamic adaptability of the review.

[0049] 2. The present invention establishes a real-time feedback information database that can record and analyze user feedback in a timely manner. It not only records error logs but also dynamically updates the review logic based on the feedback information. Users can provide instant feedback after each review, and the system can automatically apply this feedback for optimization. By continuously collecting user feedback information, the system can continuously learn and improve, thereby enhancing the accuracy and efficiency of the review. The weak feedback from users becomes an important data source for the system's self-optimization, significantly improving the response speed and application efficiency of the feedback mechanism.

[0050] 3. The present invention automatically updates and optimizes the review logic during the system review process by introducing a large language model, with the ability of self-learning and adjustment. The system not only corrects errors based on historical data but also can identify complex terms in the contract through an intelligent model, improving the level of review intelligence. By introducing a closed-loop optimization module, the system can continuously optimize the review logic according to user feedback and historical review results, reducing manual intervention and enhancing the level of automation and intelligence. This automated optimization mechanism greatly improves the efficiency and accuracy of the review. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is a flowchart of the method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0052] In order to make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the described embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present invention.

[0053] The following embodiments are used to illustrate the present invention but cannot be used to limit the scope of protection of the present invention. The conditions in the embodiments can be further adjusted according to specific conditions. Any simple improvement to the method of the present invention under the premise of the concept of the present invention belongs to the scope of protection required by the present invention.

[0054] Embodiment 1

[0055] Refer to Figure 1 , a contract review closed-loop optimization method based on weak user feedback, the method comprising the following steps:

[0056] S1. The user uploads the contract text to be reviewed to the platform through the system interface, and the system preprocesses the contract text to generate basic contract information and provides data support for subsequent reviews;

[0057] S2. The system parses the uploaded contract text, automatically matches an appropriate review checklist according to the contract type, content, and terms, and generates preliminary review points to ensure the pertinence and accuracy of the review;

[0058] S3. The system loads relevant review logics from a predefined review logic library, and automatically iteratively optimizes the loaded review logics in combination with a large language model. Users can, according to the contract characteristics, industry requirements, or the latest regulations, select to apply new logics, reject new logics, or update the review logics according to their needs through the "weak feedback" method. The information fed back by users is recorded and used as the basis for system adjustment;

[0059] S4. The system performs contract review through a large language model based on the confirmed review checklist, automatically analyzes the contract text, identifies potential risks, and outputs preliminary review results;

[0060] S5. After the review is completed, the system generates preliminary review results and displays them to the user, allowing the user to provide immediate feedback;

[0061] S6. The system retrieves historical error review data based on the current review logic context, contract context, and machine review results, analyzes and identifies potential misjudgments or errors in similar situations, and provides reference data for subsequent corrections;

[0062] S7. If similar error data is retrieved, the system generates correction suggestions through a large language model, and adjusts the current review logic or review results according to the suggestions. In this step, the system performs intelligent correction on the error results, improves the review accuracy, and avoids repeated errors;

[0063] S8. The corrected review results are displayed to the user, and the user's feedback is collected. Users can submit "weak feedback" through the interface, such as confirmation, rejection, or suggestions for further modification of the review logic;

[0064] S9. The user feedback and review results are stored in the review result index library together, forming a quadruple (Pi, Ci, RMi, RHi) containing the review logic context, contract context, machine review results, and user feedback. This data is used for subsequent optimization and closed-loop update;

[0065] S10. The system analyzes and applies user feedback in real time through a closed-loop optimization module, analyzes historical feedback data in combination with a large language model, and automatically optimizes the review logic. If the error feedback reaches a specific threshold, the closed-loop optimization module triggers a dynamic update of the review logic library, and the system automatically adjusts the review logic and applies it to the next review;

[0066] S11. All feedback information is processed and applied to the review logic, and finally this review process is completed and the next round of review is entered.

[0067] The present invention also provides a closed-loop optimization system for contract review based on weak user feedback. The system includes:

[0068] A user input interface for receiving the contract text uploaded by the user and providing a preprocessing function for basic contract information;

[0069] A contract parsing module for parsing the contract text, automatically identifying the contract type, terms, and matching the review checklist;

[0070] A review logic initialization module for loading a predefined review logic library and dynamically optimizing the loaded review logic in combination with a large language model, allowing the user to adjust the logic according to the feedback;

[0071] A review model module for performing contract review, conducting risk analysis using a large language model, and outputting a preliminary review result;

[0072] An error history retrieval module for retrieving historical review errors and generating corrective suggestions through a large language model to improve review accuracy;

[0073] A result correction module for automatically correcting the review result, generating suggestions based on a large language model, and making logical adjustments;

[0074] A user feedback storage module for storing and managing user feedback information and generating quadruple records for use by the closed-loop optimization module;

[0075] A closed-loop optimization module for continuously optimizing the review logic by real-time analyzing user feedback and machine review results. If the error information accumulates to a certain threshold, the system will automatically update the review logic.

[0076] Embodiment 2

[0077] This embodiment describes how to implement a closed-loop optimization method for contract review based on weak user feedback. This method optimizes the contract review process through specific steps and dynamically adjusts the review logic to ensure the accuracy and adaptability of the review results. The steps are as follows:

[0078] Step 1, the user uploads the contract text: The user uploads the contract text to be reviewed to the platform through the system interface. After the platform receives the contract, it performs preliminary preprocessing on the text, extracts the basic information in the contract (such as contract number, signatory, signing date, etc.), and prepares data support for subsequent review;

[0079] Step 2, contract parsing and review checklist generation: The system parses the contract text and automatically generates a matching review checklist according to elements such as contract type and clause content. The system identifies the key clauses of the contract (such as payment clauses, default clauses, etc.) and provides basic data for subsequent review points;

[0080] Step 3, Loading Review Logic and User Feedback: The system loads relevant review logic from a predefined review logic library. Subsequently, based on the user's requirements and contract characteristics, the system preliminarily optimizes the review logic in combination with a large language model. During this process, the user can provide "weak feedback" on the review logic provided by the system, such as confirming, rejecting, or adjusting the review logic. The system will record these feedbacks and use them as the basis for subsequent optimization.

[0081] Step 4, Contract Review and Preliminary Risk Identification: The system performs contract review based on the review checklist confirmed by the user. The system comprehensively analyzes the contract text through a large language model, automatically identifies potential legal risks (such as unreasonable terms, regulatory conflicts, etc.), and generates preliminary review results.

[0082] Step 5, User Feedback Collection and Display: After the review is completed, the system displays the preliminary review results to the user. The user can provide feedback on the results, such as confirming that the review is correct, raising rejection opinions, or suggesting modifying the review logic. At this time, the user's "weak feedback" will affect the update of the review logic and the accuracy of subsequent review results.

[0083] Step 6, Historical Data Retrieval and Misjudgment Analysis: The system analyzes historical review data through an error history retrieval module to identify potential misjudgments or errors in similar situations. The system combines historical data and the current review logic to provide correction suggestions to help improve the review accuracy.

[0084] Step 7, Review Result Correction and Feedback Update: The system generates correction suggestions based on historical data and misjudgment analysis results, and adjusts the current review logic or review results. The user can provide feedback again based on the adjusted review results, and all feedbacks will be recorded and stored for subsequent optimization.

[0085] Step 8, Closed-loop Optimization and Review Logic Update: The system analyzes and applies user feedback in real time through a closed-loop optimization module, combines a large language model to analyze historical feedback data, and automatically optimizes the review logic. If the error feedback reaches a certain threshold, the system will trigger a dynamic update of the review logic library, and the optimized review logic will be applied to the next review.

[0086] Step 9, Continuous Optimization and Next Round of Review: After each review is completed, user feedback and review results will be stored and used for the optimization of the next round of review logic. Through closed-loop optimization, the system continuously improves the accuracy and efficiency of the review, ensuring that each review can be dynamically adjusted according to user feedback and historical data.

[0087] In summary, the present invention provides a method and system for optimizing the closed-loop of contract review based on weak user feedback, aiming to improve the accuracy and efficiency of contract review through an intelligent review process. By automatically parsing contract texts, matching review checklists, applying dynamically optimized review logic, and combining large language models for risk analysis of contracts, the system can not only identify potential problems but also continuously adjust the review logic through weak user feedback to form a closed-loop optimization. This method can flexibly adjust review rules according to contract characteristics, industry requirements, and regulatory changes, thereby providing more accurate and personalized review services for users and continuously improving the review effect in the process of accumulating feedback.

[0088] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that unless otherwise defined, the technical terms or scientific terms used in the present invention should have the ordinary meaning understood by those with ordinary skills in the field to which the present invention belongs. The terms "including" or "comprising" and similar words used in the present invention mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents.

[0089] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A closed-loop optimization method for contract review based on user weak feedback, characterized in that: The method comprises the following steps: S1. The user uploads the contract text to be reviewed to the platform through the system interface. The system pre-processes the contract text, generates basic contract information, and provides data support for subsequent review; S2. The system parses the uploaded contract text, automatically matches the appropriate review checklist according to the contract type, content and terms, and generates preliminary review points to ensure the pertinence and accuracy of the review; S3. The system loads relevant review logic from the predefined review logic library and automatically iterates and optimizes the loaded review logic in combination with the large language model. Users can choose to apply new logic, reject new logic, or update the review logic as needed through "weak feedback" according to contract characteristics, industry requirements or the latest regulations. User feedback information is recorded and used as a basis for system adjustment; S4. Based on the confirmed review list, the system performs contract review through a large language model, automatically analyzes the contract text, identifies potential risks, and outputs preliminary review results; S5. After completing the review, the system generates preliminary review results and displays them to the user, allowing the user to provide immediate feedback; S6. Based on the current review logic context, contract context and machine review results, the system retrieves historical error review data, analyzes and identifies potential misjudgments or errors in similar situations, and provides reference data for subsequent corrections; S7. If similar erroneous data is retrieved, the system generates correction suggestions through the large language model and adjusts the current review logic or review results according to the suggestions. In this step, the system intelligently corrects the erroneous results to improve the review accuracy and avoid repeated errors; S8. The corrected review results are displayed to the user and user feedback is collected. The user can submit "weak feedback" through the interface, such as confirmation, rejection, or suggestion for further review logic modification; S9. User feedback and review results are stored together in the review result index library to form a four-tuple (Pi, Ci, RMi, RHi) containing review logic context, contract context, machine review results and user feedback. This data is used for subsequent optimization and closed-loop update. S10. The system uses a closed-loop optimization module to analyze and apply user feedback in real time, and combines the large language model to analyze historical feedback data to automatically optimize the review logic. If the error feedback reaches a certain threshold, the closed-loop optimization module triggers a dynamic update of the review logic library, and the system automatically adjusts the review logic and applies it to the next review; S11. All feedback information is processed and applied to the review logic, finally completing the review process and entering the next round of review.

2. The closed-loop optimization method for contract review based on user weak feedback according to claim 1 is characterized in that: The "weak feedback" includes the following operations: User confirmation to apply the new review logic; Users reject the new review logic; Users can update the review logic based on actual needs to adapt to contract characteristics, regulatory changes or business needs.

3. The closed-loop optimization method for contract review based on user weak feedback according to claim 1 is characterized in that: The review logic initialization step includes: The system is loaded with predefined review logic designed by professional legal personnel and algorithm engineers; The system intelligently updates the loaded review logic through a large language model, and makes dynamic adjustments based on business needs, regulatory changes, and user feedback.

4. The closed-loop optimization method for contract review based on user weak feedback according to claim 1 is characterized in that: The closed-loop optimization module automatically optimizes based on user feedback, historical data, and machine review results. The optimization process includes: Update review logic based on user feedback; Automatically calibrate the system to improve the accuracy of contract review; If the accumulation of error information reaches a predetermined threshold, the update of the review logic library will be automatically triggered to ensure that the review model adapts to the latest business rules and regulatory requirements.

5. The closed-loop optimization method for contract review based on user weak feedback according to claim 1 is characterized in that: The error review history retrieval module identifies potential misjudgments in similar review logic contexts and contract contexts based on historical review data, and generates correction suggestions.

6. The closed-loop optimization method for contract review based on user weak feedback according to claim 1 is characterized in that: The user feedback information storage module is used to store the user's weak feedback and generate a four-tuple (Pi, Ci, RMi, RHi) according to the feedback for subsequent optimization.

7. The closed-loop optimization method for contract review based on user weak feedback according to claim 1 is characterized in that: The feedback information library is used to record and analyze user interaction results, perform logical corrections based on the consistency of user and machine feedback, and continuously improve the accuracy of review results.

8. A closed-loop optimization system for contract review based on user weak feedback, characterized in that: The system comprises: User input interface, used to receive the contract text uploaded by the user and provide basic contract information preprocessing function; Contract parsing module, used to parse contract text, automatically identify contract type, terms and match the review list; The review logic initialization module is used to load the predefined review logic library and dynamically optimize the loaded review logic in combination with the large language model, allowing users to adjust the logic based on feedback; The review model module is used to perform contract review, use the large language model to perform risk analysis, and output preliminary review results; Error history retrieval module, which is used to retrieve historical review errors and generate correction suggestions through a large language model to improve review accuracy; The result correction module is used to automatically correct the review results, generate suggestions based on the large language model, and make logical adjustments; A user feedback storage module is used to store and manage user feedback information and generate four-tuple records for use by the closed-loop optimization module; The closed-loop optimization module is used to continuously optimize the review logic by analyzing user feedback and machine review results in real time. If the error information accumulates to a certain threshold, the system will automatically update the review logic.

9. The closed-loop optimization system for contract review based on user weak feedback according to claim 8 is characterized in that: The system has a dynamically updated review logic library that supports intelligent updates based on user feedback to ensure continuous optimization of the review logic.

10. The closed-loop optimization system for contract review based on user weak feedback according to claim 8, characterized in that: The closed-loop optimization module performs automated logic adjustments based on user feedback and historical data to ensure the accuracy and adaptability of contract review results.

Images