Vector Acquisition for Compound Word Classification Accuracy

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing technologies face challenges in accurately calculating the similarity between words when corresponding vectors are not held by a learned model or when the classification accuracy of these vectors is insufficient.

Innovation Solution

A vector acquisition method that uses a learned model to input text containing compound words, output adjusted vectors for the individual words, and acquire a compound vector by combining these adjusted vectors, thereby improving classification accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If vectors held by a learned model in advance are used to calculate the degree of similarity between words, then the calculation can be performed easily, but the classification accuracy is insufficient when the vectors are not held by the learned model or when the classification accuracy of the vectors held by the learned model is insufficient

Engineering Contradiction:

Improvecalculation efficiencyVSAvoidclassification accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The compound word is divided into multiple constituent words, and vectors are acquired for each constituent word separately. These individual vectors are then combined to form a compound vector, enabling accurate representation of compound words that may not be directly held in the learned model.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Vectors corresponding to multiple constituent words obtained by dividing a compound word are combined to create a compound vector. This merging process allows the system to generate accurate representations for compound words by integrating information from their individual components.

Inventive Principle:

Principle #5Merging (Combining)

2Ease of operation

If vectors held by a learned model in advance are used, then the degree of similarity between words can be calculated easily, but the accuracy is insufficient when corresponding vectors are not held or classification accuracy is insufficient

Engineering Contradiction:

Improveease of calculationVSAvoidclassification accuracy

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The system pre-processes compound words by dividing them into constituent words and acquiring vectors for each component before performing similarity calculations. This preliminary action ensures that accurate vectors are available when needed, maintaining both ease of operation and high accuracy.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

Constituent words serve as intermediaries between the compound word and the vector representation. By acquiring vectors for the constituent words and combining them, the system creates an intermediate representation that accurately reflects the compound word's meaning while maintaining computational efficiency.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS12235881B2Vector acquisition method, vector acquisition apparatus, and recording medium

Publication Date: 2025.02.25 SCREEN HOLDINGS CO LTD

AI summary

A vector acquisition method includes: inputting, into a learned model, at least one piece of text including at least two of a plurality of words obtained by dividing a compound, the compound being a word divisible into the plurality of words; outputting, from the learned model, an adjusted vector corresponding to at least one of the words obtained by dividing the compound in the input piece of text; and acquiring a compound vector corresponding to the compound using the adjusted vector output from the learned model. Classification accuracy of vectors corresponding to words can thereby be enhanced.