Knowledge Graph Data Management for Semantic Search Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional data search mechanisms in enterprise environments are limited in providing complete and accurate results, leading to inefficient use of computing resources and incomplete knowledge retrieval due to 'dark data' that is not easily searchable or analyzable.
Innovation Solution
A knowledge-enabled data management system that stores data as a knowledge graph using domain-specific ontologies, enabling semantic search queries and enriching factual data with human knowledge through automatic and manual means, and employing AI and machine learning to parse queries and return accurate results with probabilistic measures of relationships.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If string matching mechanisms are used for searching enterprise data, then the search process is simple to implement, but the ability to provide queried data is limited and results are incomplete
Solution Approach 1:
The patent introduces a knowledge graph as an intermediary layer between the search query and the enterprise data. The knowledge graph stores semantic relationships and contextual information that enable more accurate data retrieval without requiring complex search algorithms throughout the entire system. This intermediary structure resolves the contradiction by providing enhanced retrieval accuracy through semantic understanding while keeping the overall system architecture manageable.
Solution Approach 2:
The patent transforms the search approach by changing the parameter of data representation from simple string matches to structured knowledge entities with semantic relationships. By representing data in the knowledge graph with entities, properties, and relationships, the system achieves higher retrieval accuracy. This parameter change allows the system to move beyond basic string matching to semantic search capabilities.
2Productivity
If conventional knowledge query systems are used, then the system structure is simple, but computing resources are consumed inefficiently due to repeated queries returning inaccurate or incomplete results
Solution Approach 1:
The patent applies preliminary action by pre-processing and structuring enterprise data into a knowledge graph before queries are executed. During data ingestion, the system establishes semantic relationships, entities, and contextual information in advance. This preliminary structuring enables faster and more accurate query resolution without requiring repeated complex processing for each query, thereby improving productivity while reducing computing resource consumption.
Solution Approach 2:
The patent creates a knowledge graph copy or representation of the enterprise data that captures semantic relationships and contextual information. This copied structure serves as an optimized index that can be queried efficiently without repeatedly accessing and processing the original large-scale enterprise data, thus reducing computing resource consumption while maintaining search productivity.
3Ease of operation
If data is stored in traditional formats, then storage and access are straightforward, but most data becomes dark and not easily searchable or available for analytics
Solution Approach 1:
The patent changes the parameter of data storage from traditional flat formats to structured knowledge graph formats with entities, properties, and relationships. This transformation makes data more searchable and accessible by organizing it according to semantic meaning rather than storage convenience. The knowledge graph structure enables both easy operation through intuitive querying and reduced information loss through comprehensive data representation.
Solution Approach 2:
The patent adds another dimension to data storage by introducing semantic relationships and contextual information alongside the traditional data structure. This dimensional enhancement transforms data from isolated records to interconnected knowledge entities, making previously inaccessible 'dark data' searchable and analyzable while maintaining the original data's accessibility through multiple query pathways.
Data Source
AI summary
A knowledge enabled data management system ingests data and stores the data as an instance in a knowledge graph according to a domain specific ontology. The instance includes stored relationships of the entities in the instance. A query regarding the data may be parsed to derive a first query entity that is used to search the knowledge graph for a first graph entity corresponding to the first query entity. Results may be returned including a first identification of the first graph entity, a second identification of at least a second graph entity related to the first graph entity stored within the knowledge graph, and additional data corresponding to the first graph entity and to the second graph entity. The additional data may include a probabilistic measure of the relationship of the first graph entity to the second graph entity.