Content review method, apparatus, device, and medium

By segmenting and vectorizing documents and combining them with contextual information for dual verification, the problem of low efficiency and poor reliability in reviewing insurance terms on legal platforms has been solved, achieving more efficient and accurate content review.

CN122154695APending Publication Date: 2026-06-05PEOPLE'S INSURANCE COMPANY OF CHINA
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN Β· China
Patent Type
Applications(China)
Current Assignee / Owner
PEOPLE'S INSURANCE COMPANY OF CHINA
Filing Date
2026-01-13
Publication Date
2026-06-05

AI Technical Summary

Technical Problem

Existing technologies are inefficient in the process of reviewing insurance terms on legal platforms and cannot guarantee the reliability of the review results.

Method used

The document to be reviewed is divided into paragraph blocks, vectorized, and then similarity is retrieved from the vector database. A second matching is performed by combining the context information of the text blocks to determine the target vector and calculate the semantic similarity of the paragraphs, and finally, the content is reviewed.

Benefits of technology

It improves the efficiency and accuracy of content review, reduces misjudgments caused by differences in expression, enhances the ability to distinguish semantically similar but context-unrelated content, and achieves more precise and granular document content review.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122154695A_ABST
    Figure CN122154695A_ABST
Patent Text Reader

Abstract

The present disclosure provides a content review method, device, equipment and medium, the method comprises: determining the first document type of the to-be-reviewed document, and performing paragraph segmentation processing on the text content of the to-be-reviewed document according to the first document type to obtain a plurality of first text blocks; the first text block is subjected to vectorization processing to obtain a first vector, and the retrieval result of the first vector in the vector database is determined, the retrieval result includes: at least one candidate vector, and the similarity score of each candidate vector and the first vector; obtain the context information of the first vector in the to-be-reviewed document, and determine the content matching value between each candidate vector and the context information; determine the target vector from the plurality of candidate vectors according to the similarity score and the content matching value; determine the paragraph semantic similarity according to the first vector and the target vector; determine the content review result according to the paragraph semantic similarity corresponding to each first vector. Therefore, the content review efficiency and accuracy can be effectively improved.
Need to check novelty before this filing date? Find Prior Art