Contract review report generation method based on large model integrated optimization

Through the integrated optimization technology of large-scale models, combined with contract classification, intent identification and risk point review database, the problem of insufficient flexibility of existing tools in complex contract review is solved, and more accurate and efficient contract review is achieved, improving the quality of review and robustness.

CN118940732AActive Publication Date: 2024-11-12GUIZHOU UNIV
7 Cites 12 Cited by

Patent Information

Application Number
CN202411436259.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-15
Publication Date
2024-11-12
Estimated Expiration
2044-10-15

AI Technical Summary

Technical Problem

Existing rules-based approaches to contract review tools are not flexible enough to accurately capture text semantics while dealing with complex or non-standardized contracts, resulting in inaccurate evaluations and the need for frequent updates of rules, increasing costs and time.

Method used

Using a method based on large-scale integrated optimization, through contract classification, intent identification and risk point review database, weighted voting is carried out in combination with multiple pre-trained language models, a comprehensive review report is generated to improve the accuracy and efficiency of review.

Benefits of technology

It improves the accuracy and efficiency of contract review, enhances the robust processing capability of complex contracts, reduces deviations from individual models, and ensures the reliability and quality of review results.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to the technical field of intelligent identification model application, in particular to a contract review report generation method based on large model integration optimization, and solves the technical problem that a method for improving contract review efficiency and accuracy through an automatic process to deeply analyze and review contract texts is lacked in the prior art. The method comprises a risk point review library, a contract classification process, an intention recognition process, an obtained contract type and a contract intention, traversing each review problem in the risk point review library, transmitting the review problems as input to a plurality of pre-trained large-scale language models, and integrating outputs of the plurality of pre-trained large-scale language models to obtain a risk point review result. A voting device is used for summarizing and evaluating results, and audit results of all risk points are integrated to form a comprehensive review report, so that the understanding ability of the model for specific parts of legal documents can be greatly improved, and potential risks and obligations can be identified. According to the method, the accuracy and pertinence of the model can be improved, and higher examination quality is ensured.
Need to check novelty before this filing date? Find Prior Art