Sentence Structure Vectorization via Partial Graph Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sentence structure vectorization methods face challenges in efficiently representing the semantic structure of sentences, particularly due to the exponential increase in calculations required for vectorizing text structures as the size of dependency graphs grows, and fail to adequately express parallel structures.
Innovation Solution
A sentence structure vectorization device and method that performs morphological analysis and dependency parsing to generate a dependence structure graph, extracting partial structure information and converting it into a numerical sequence of vector values, where each vector value represents the number of appearances of morpheme structures, reducing the number of calculations by excluding non-existent dependency relationships and merging morpheme structures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If partial graphs are extracted from the dependency graph structure to express semantic structure, then the ability to represent sentence meaning is improved, but the number of calculations increases exponentially
Solution Approach 1:
The patent extracts only the essential partial graphs from the dependency graph structure that are necessary for representing sentence meaning, rather than processing all possible partial graphs. This selective extraction approach maintains semantic representation accuracy while avoiding exponential calculation growth.
Solution Approach 2:
The patent segments the dependency graph into manageable partial structures that can be processed independently. By dividing the complex graph into smaller, meaningful segments, the system can extract features efficiently without having to process the entire graph structure at once, thus reducing computational complexity.
2Loss of information
If dependency parsing is performed on the entire dependency graph structure, then complete sentence structure information is obtained, but computational complexity increases
Solution Approach 1:
The patent extracts only the necessary partial graph structures from the complete dependency graph that are relevant for sentence meaning representation. This selective extraction maintains information completeness for the intended purpose while reducing processing complexity by excluding unnecessary graph elements.
3Measurement precision
If more partial graphs are extracted to improve vectorization accuracy, then sentence meaning representation is enhanced, but the number of calculations increases
Solution Approach 1:
The patent extracts only the essential partial graphs that contribute most to sentence meaning representation, rather than extracting all possible partial graphs. This selective approach maintains high vectorization accuracy while avoiding the exponential increase in calculations that would result from comprehensive extraction.
Data Source
AI summary
A sentence structure vectorization device includes processing circuitry to generate a plurality of morphemes by performing morphological analysis on an input sentence; to generate a dependence structure graph regarding the plurality of morphemes by performing dependency parsing on the plurality of morphemes; and to generate a sentence structure vector by extracting a plurality of pieces of partial structure information from the dependence structure graph and converting a morpheme string corresponding to the plurality of pieces of partial structure information into a numerical sequence.