An electronic document auxiliary reading method, system, device and storage medium based on reading blockage identification and intra-text dependency back chaining

By constructing a heterogeneous intra-text dependency graph and multi-hop back chain search, combined with user interaction event flow, the problems of coarse-grained blocking positioning and excessive redundant information in electronic document reading are solved, achieving precise positioning and adaptive optimization, thus improving the effectiveness of the reading assistance system.

CN122334232APending Publication Date: 2026-07-03韩泽嵘
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
韩泽嵘
Filing Date
2026-04-10
Publication Date
2026-07-03

AI Technical Summary

Technical Problem

Existing technologies struggle to accurately pinpoint reading blockages in electronic document reading. They lack structured dependency modeling within documents, restrict search paths, and lack user feedback self-optimization mechanisms, resulting in coarse blockage location granularity, excessive redundant information, and insufficient relevance of auxiliary results.

Method used

A heterogeneous intratext dependency graph is constructed, and content unit nodes are connected through various relational edges. User interaction event streams are collected, reading blocking evidence vectors are constructed, and multi-hop backlink search is performed. The optimal content block is selected using an objective function, and dynamic updates are performed in combination with user behavior to achieve targeted interpretation and self-optimization.

Benefits of technology

It achieves fine-grained localization of reading blockage, directional interpretation based on structured dependencies, significantly reduces redundant information loading, improves the relevance and stability of auxiliary results, and has the ability to self-optimize user behavior.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122334232A_ABST
    Figure CN122334232A_ABST
Patent Text Reader

Abstract

This invention discloses an electronic document assisted reading method, system, device, and storage medium based on reading blockage identification and in-text dependency backlinks. The method includes: performing structured parsing of the electronic document to construct a heterogeneous in-text dependency graph containing multiple types of nodes and relational edges; collecting user interaction event streams, constructing reading blockage evidence vectors and projecting them onto visible nodes to locate the blockage starting point; statistically analyzing the distribution of incoming edge types to determine the blockage source category, selecting a subset of relational edges to perform constrained multi-hop backlink search; constructing an objective function based on coverage, redundancy, path cost, and historical validity to filter supporting content blocks, and performing a two-layer structured output; and dynamically updating path weights based on user feedback. This invention achieves fine-grained localization and targeted source tracing of reading blockages, reduces redundant loading, and improves auxiliary relevance and system stability.
Need to check novelty before this filing date? Find Prior Art