Personalized nuclear power management system question and answer method and device

By constructing a nuclear power policy document knowledge base and combining user characteristics with a large-scale language model, the personalization problem of the nuclear power management policy Q&A system was solved, generating personalized, accurate and compliant answers to nuclear power management policy questions and answers to meet the needs of different users.

CN122240764APending Publication Date: 2026-06-19CHINA NUCLEAR POWER OPERATION TECH CORP
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
CHINA NUCLEAR POWER OPERATION TECH CORP
Filing Date
2026-02-27
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

The existing nuclear power management system's question-and-answer system lacks personalization and cannot effectively combine user characteristics with a large language model, resulting in a large gap between the answers and the user's subjective intent. Furthermore, the knowledge graph construction and maintenance costs are high, making it difficult to adapt to rapidly updated management systems.

Method used

A nuclear power regulatory document knowledge base is constructed. The documents are structured through OCR parsing and domain-optimized embedding models. Personalized answers are generated by combining user characteristics and large-scale language models, including semantic unit vectorization, user modeling, enhanced prompt word construction, and compliance checks, to ensure the accuracy and compliance of the answers.

Benefits of technology

It optimizes text filtering and answer generation based on user characteristics, generating personalized, accurate, and compliant answers to nuclear power management regulations, meeting the needs of different users, and improving the personalization and efficiency of the question-and-answer system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122240764A_ABST
    Figure CN122240764A_ABST
Patent Text Reader

Abstract

This disclosure belongs to the field of nuclear power technology, specifically relating to a personalized question-and-answer method and apparatus for nuclear power management regulations. This disclosure aims to address the impact of user personalization characteristics on question-and-answer results, proposing a personalized question-and-answer method for nuclear power management regulations based on the collaboration of user characteristics and a large language model. The method of this disclosure sorts the retrieved text information based on a feature association module of user characteristics and semantic units. Then, the relevant text and questions are input into a large language model, which generates personalized answers based on the user's personalized characteristics. This allows for optimized text filtering by combining user characteristics and integrating retrieval enhancement generation with a large language model to generate personalized answers, more accurately meeting the needs of different users.
Need to check novelty before this filing date? Find Prior Art