Entity Relation Categorization Using Relation and Feature Vectors

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing techniques for categorizing relations between entities are inaccurate due to reliance on syntactic structure of sentence expressions, failing to consider entity features.

Innovation Solution

Generate relation vectors and feature vectors from selected sentences using neural networks to represent relations and entity features, and update parameters to enhance similarity, enabling accurate categorization.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If relation categorization is based on syntactic structure of sentence expressions, then the categorization process is simple, but the categorization accuracy deteriorates because entity features are not considered

Engineering Contradiction:

Improvecategorization accuracyVSAvoidprocessing complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent merges syntactic structure analysis with entity feature analysis into a unified relation categorization framework. The system simultaneously processes both the grammatical structure of sentences and the semantic features of entities to generate comprehensive relation vectors, thereby improving categorization accuracy without excessive complexity increase

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The patent transforms the categorization approach by changing from purely syntactic parameters to include semantic parameters. Entity features such as type, attributes, and contextual information are converted into vector representations that are combined with syntactic features, enabling more accurate relation categorization through parameter expansion

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If relation categorization considers entity features, then categorization accuracy improves, but the processing complexity increases due to additional feature extraction requirements

Engineering Contradiction:

Improvecategorization accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary extraction and representation of entity features before relation categorization. Entity features are pre-processed into vector forms and stored for efficient retrieval during relation analysis, reducing the time penalty associated with feature extraction while maintaining improved categorization accuracy

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12572580B2Information processing device, information processing method, and program

Publication Date: 2026.03.10 NEC CORP

AI summary

In order to more accurately categorize a relation between a plurality of entities, an information processing apparatus (1) includes a relation vector generation section (11), a feature vector generation section (12), and a relation categorization section (13). The relation vector generation section (11) generates a relation vector representing a relation between a plurality of entities of interest from at least one sentence which has been selected from a sentence set and in which the plurality of entities of interest occur. The feature vector generation section (12) generates, for each entity of interest, a feature vector representing a feature of that entity of interest from at least one sentence which has been selected from the sentence set and in which that entity of interest occurs. The relation categorization section (13) categorizes a relation between the plurality of entities of interest with use of a relation vector and feature vectors.