Graph-Based Text Ranking for Concept Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing text retrieval systems are largely extractive, making it difficult for humans to decipher the meaning of summarized text, and rely on word-based scoring methods like tf*idf, which are not effective in identifying concepts of interest in a corpus of text.
Innovation Solution
A method and system that builds a graph covering the textual material, scoring nodes and sub-graph portions to identify graph fragments, which are then used to generate a textual output that represents the concepts of interest, using a graph ranking algorithm and natural language processing to reorder and rephrase text for improved coherence and readability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If extractive summarization is used to identify text fragments, then the system maintains simplicity in text processing, but the coherence and readability of the generated summary deteriorates
Solution Approach 1:
The patent transitions from word-based scoring (1D) to graph-based scoring (2D/3D), where nodes represent words/concepts and edges represent relationships. This dimensional change allows the system to capture contextual relationships while maintaining computational feasibility, resolving the contradiction between simplicity and coherence.
Solution Approach 2:
The patent introduces a graph structure as an intermediary between raw text and summary generation. This graph acts as a mediator that preserves semantic relationships while enabling more sophisticated scoring, thus improving coherence without directly complicating the summarization process.
2Productivity
If word-based scoring methods like tf*idf are used, then the system maintains computational efficiency, but the ability to identify concepts of interest deteriorates
Solution Approach 1:
The patent merges multiple scoring signals (term frequency, graph structure, node centrality) into a unified graph-based ranking system. This combination allows the system to maintain computational efficiency while significantly improving concept identification accuracy through multi-factor evaluation.
Solution Approach 2:
The graph structure serves multiple functions simultaneously: it represents lexical relationships, enables path-based scoring, captures contextual information, and facilitates concept identification. This multi-functionality allows the system to achieve high measurement precision without proportionally increasing computational complexity.
3Measurement precision
If graph-based ranking is implemented to improve concept identification, then the precision of identifying concepts of interest improves, but the device complexity increases
Solution Approach 1:
The patent segments the graph processing into distinct phases: graph construction from text, node/edge scoring, and summary generation. This segmentation allows each component to be optimized independently, reducing overall system complexity while maintaining high precision in concept identification.
Data Source
AI summary
The present invention is a method and system for identifying words, text fragments, or concepts of interest in a corpus of text. A graph is built which covers the corpus of text. The graph includes nodes and links, where nodes represent a word or a concept and links between the nodes represent directed relation names. A score is then computed for each node in the graph. Scores can also be computed for larger sub-graph portions of the graph (such as tuples) The scores are used to identify desired sub-graph portions of the graph, those sub-graph portions being referred to as graph fragments.