Semantic Entity Search Using Vector Space for Data Integration

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing data management systems face challenges in integrating disparate data sources due to the lack of a central model that links and integrates physical data elements, leading to issues such as data duplication, inaccurate analytics, and difficulty in searching and exploring data relationships, especially in relational and graph databases.

Innovation Solution

A knowledge graph server is implemented to create a central nervous system across data sources, utilizing a vector space model and machine learning models to harmonize and deduplicate data, enabling semantic searching and exploration of data relationships through a unified data model.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If data is stored in multiple different tables with different formats and column labels, then data can be collected from various sources, but it becomes difficult to search and integrate the data without knowing all the different formats, tables, and column labels

Engineering Contradiction:
Improvedata integration capabilityVSAvoidsearch difficulty
Core Design Contradiction:
Adaptability or versatilityVSEase of operation

Solution Approach 1:

The patent introduces a central data model as an intermediary layer between physical data sources and users. This central model provides a unified view of data elements across multiple tables and formats, allowing users to search and access data without needing to know the underlying physical storage details. The central model acts as a mediator that translates diverse physical data representations into a consistent logical structure.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent segments the data management system into distinct layers: physical data sources with various formats, a central data model that organizes data elements logically, and access interfaces. This segmentation allows each layer to handle its specific concerns independently, improving both adaptability to different data sources and ease of data access.

Inventive Principle:
Principle #1Segmentation

2Device complexity

If traditional data management systems are used without a central model, then data storage is simple, but data duplication occurs and analytics accuracy deteriorates

Engineering Contradiction:
Improvesystem structureVSAvoidanalytics accuracy
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The central data model serves as a mediator that coordinates data access and management across the system. It maintains a unified view of data elements, preventing duplication by providing a single source of truth for data definitions and relationships. This intermediary structure ensures that analytics operations work with consistent, accurate data regardless of the underlying physical storage variations.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Adaptability or versatility

If disparate data sources are integrated without a unified model, then data from multiple sources can be accessed, but resource consumption increases and efficiency decreases

Engineering Contradiction:
Improvedata source compatibilityVSAvoiddata exploration efficiency
Core Design Contradiction:
Adaptability or versatilityVSProductivity

Solution Approach 1:

The central data model provides a universal interface for accessing data from multiple sources. It defines a common structure and semantics that works across different physical data sources, allowing the same data access operations to function efficiently regardless of the underlying source. This multi-functionality enables efficient data exploration without requiring separate handling for each data source.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentUS12450241B2Semantic entity search using vector space
Publication Date: 2025.10.21 WELLS FARGO BANK NA
  • US12450241B2 patent drawing
  • US12450241B2 patent drawing
  • US12450241B2 patent drawing

AI summary

A method may include storing a vector space representation of a set of concepts in a semantic ontology, wherein locations of the set of concepts in the vector space are based on weights of a neural network trained using triples from the semantic ontology; presenting a semantic search user interface including: a text input portion to receive a search query from an input device; a suggested search portion; and a search results portion; receiving, from a user computing device, the search query; converting, using at least one processor, the search query into a vector; computing, using the at least one processor, a set of possible concepts related to the search query based on a distance between the vector and other concepts in the vector space representation; and presenting the set of possible concepts related to the search query in the suggested search portion of the semantic search user interface.