Large model fine tuning method based on data asset security knowledge graph

CN120910892APending Publication Date: 2025-11-07FUJIAN FUJITSU COMM SOFTWARE CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202510769642.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-06-10
Publication Date
2025-11-07

AI Technical Summary

Technical Problem

Existing technologies struggle to effectively analyze and manage relationships between large-scale data assets, particularly in terms of data asset access relationships and security tracing, resulting in inefficient data security audits and an inability to meet complex security audit standards assessment requirements.

Method used

Construct a data asset security knowledge graph and use large model fine-tuning technology to enable the model to understand and process complex data asset relationships, including access patterns and security attributes. Use the fine-tuned model for real-time monitoring and analysis to generate security policies to prevent security incidents.

Benefits of technology

It improves the accuracy and efficiency of data security analysis, enables the provision of personalized data analysis services, and enhances the precision of security prediction and risk management capabilities.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120910892A_ABST
    Figure CN120910892A_ABST
Patent Text Reader

Abstract

The invention discloses a large model fine tuning method based on a data asset security knowledge graph. The method comprises the following steps: acquiring data asset information and constructing a data asset knowledge graph; the large model creates a fine adjustment data set based on the data asset knowledge graph, and the fine adjustment data set is utilized to train the large model to obtain the fine-adjusted large model, so that the large model accurately understands and analyzes the security of the data assets; performing real-time monitoring and analysis by using the access behavior of the data assets of the large model after fine tuning, and judging whether each data access request meets a security auditing standard or not so as to identify potential security risks to form assessment; and a corresponding security policy is generated according to the evaluation result, and timely adjustment is performed to prevent the occurrence of security incidents. According to the method, the large model can master the complex relationship between the data assets and respective security classification, so that the accuracy and efficiency of data security analysis are improved.
Need to check novelty before this filing date? Find Prior Art