A legal consultation system construction method based on multi-model collaborative reasoning

By constructing a dynamic legal knowledge base and a multi-model collaborative reasoning mechanism, the problems of illusion, insufficient logical reasoning, and lagging knowledge updates in existing legal consultation systems have been solved, thus achieving the reliability and logical rigor of legal answers and supporting the accurate handling of complex legal issues.

CN120030130BActive Publication Date: 2026-06-16EAST CHINA NORMAL UNIV
0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
EAST CHINA NORMAL UNIV
Filing Date
2025-02-28
Publication Date
2026-06-16

AI Technical Summary

⚠Technical Problem

Existing legal consultation systems suffer from problems such as illusion, lack of ability to construct logical steps for complex problems, insufficient ability to calculate legal time, lagging knowledge updates, and insufficient multi-task collaboration capabilities. As a result, the legal answers generated lack clear chains of evidence, have unclear logical reasoning, and cannot respond to legal changes in a timely manner.

⚗Method used

A dynamic legal knowledge base is constructed using Retrieval Enhanced Generation (RAG) technology. Combined with a multi-model collaborative reasoning mechanism, and through the integration of a task analysis-specific model and external tools, a structured execution plan is generated to ensure the rigor and accuracy of legal dialogue.

🎯Benefits of technology

It significantly reduces model illusion, improves the reliability of legal solutions and logical reasoning ability, ensures the timeliness of the knowledge base, and can accurately handle complex legal issues and support the application needs of key legal scenarios.

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

Abstract

The application discloses a legal consultation system construction method based on multi-model collaborative reasoning, which has the following characteristics: first, by constructing a dynamically updatable legal knowledge base, combined with search enhancement generation technology, it ensures that the answers generated by the system are always based on the latest laws and regulations. And this method adopts a two-stage reasoning mechanism: in the first stage, the task analysis large language model deeply analyzes the problem based on relevant laws and available external tool libraries, generating a structured execution plan containing reasoning steps and tool calling schemes; in the second stage, the content generation large language model performs multi-step reasoning according to the execution plan, and finally generates an answer content conforming to the legal professional standard. Compared with the prior art, the application divides complex legal problems into executable sub-task sequences through a multi-model collaborative working mechanism, making up for the limitations of single models in logical reasoning and knowledge coverage, and significantly enhancing the accuracy and credibility of the legal dialogue generation content.
Need to check novelty before this filing date? Find Prior Art