A multi-type rule extraction and verification method with full-process automation

By using the WICD paradigm and large model to automatically generate verification logic, the problem of automating the extraction and verification of multiple types of rules has been solved. This has achieved unified rule format and automatic generation of verification logic, improving processing efficiency and accuracy.

CN122242709APending Publication Date: 2026-06-19HUA DATA TECH (SHANGHAI) CO LTD
0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
HUA DATA TECH (SHANGHAI) CO LTD
Filing Date
2026-05-21
Publication Date
2026-06-19

AI Technical Summary

⚠Technical Problem

Existing technologies are difficult to adapt to the automatic extraction of multiple types of rules. The rule formats are not uniform and need to be manually organized. The verification logic relies on manual presets, which is not flexible enough and cannot achieve full-process automation.

⚗Method used

The WICD paradigm is used to parse and format the rule text, generate a verification reasoning path library, and automatically generate verification logic using a large model. This enables automatic rule extraction, unified formatting, and automatic generation of verification logic.

🎯Benefits of technology

It enables automatic extraction and standardized formatting of multiple rule types, improving processing efficiency and accuracy, solving the problem of disconnect between rule extraction and verification processes, and achieving fully automated connection of the entire process.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
Patent Text Reader

Abstract

This invention provides a fully automated method for extracting and verifying multiple types of rules, relating to the field of intelligent document verification technology. The fully automated method includes: parsing regulations and rules documents; extracting rules from the rule text based on the WICD paradigm; classifying rules into numerical, image, and text categories; setting an initial verification reasoning path based on the WICD paradigm, and then obtaining effective reasoning paths through adjustment and filtering; using these as an aid, and combining them with the specific verification reasoning paths for each type of rule to generate a specific verification path; finally, performing verification and outputting the results. This invention achieves fully automated integration of rule processing and verification, solving the core problems of low efficiency, poor accuracy, and poor process integration in existing technologies.
Need to check novelty before this filing date? Find Prior Art