Unified Knowledge Graphs via Type-Specific Subgraph Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As graph databases grow in entity and relationship count, discovering relationships becomes increasingly complex and computationally intensive, making it difficult for organizations to leverage their graph databases effectively for business insights or service improvements.
Innovation Solution
The method involves receiving entity data, forming type-specific groups, disambiguating entity data within these groups, creating relationships, and exporting them to type-specific subgraphs to form a unified knowledge graph, which reduces redundancy and improves processing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Difficulty of detecting and measuring
If simple matching algorithms are used to discover relationships in graph databases, then relationship discovery is performed, but computational load becomes intractable with hundreds of millions of entities
Solution Approach 1:
The patent segments the massive graph database into multiple type-specific subgraphs based on entity types (e.g., customers, products, services). This segmentation reduces the search space for relationship discovery algorithms, allowing them to operate on smaller, manageable subsets rather than the entire graph, thereby maintaining computational efficiency while still discovering relationships.
Solution Approach 2:
The patent introduces an intermediary layer of type-specific subgraphs that act as mediators between the raw graph data and the relationship discovery process. These subgraphs organize entities by type and pre-compute certain relationships, serving as an intermediate structure that simplifies subsequent relationship discovery queries on the full graph.
2Measurement precision
If manual review is used to confirm relationship matches, then relationship accuracy is improved, but the time required becomes impossibly long
Solution Approach 1:
The system implements self-service relationship verification through automated confidence scoring and validation mechanisms. Type-specific subgraphs enable automated verification of relationships within each entity type context, reducing or eliminating the need for manual review while maintaining high accuracy through algorithmic validation rules specific to each entity type.
3Quantity of substance
If the graph database grows in entity and relationship count, then data comprehensiveness is improved, but complexity of the graph increases rapidly
Solution Approach 1:
The patent divides the complex graph database into multiple type-specific subgraphs, each containing entities of a particular type and their relationships. This segmentation reduces the apparent complexity of the overall graph by organizing it into manageable, homogeneous subsets, making it easier to query, maintain, and analyze even as the total data volume grows.
Solution Approach 2:
The patent adds a dimensional organization to the graph data by introducing type-specific dimensions or layers. Instead of a single flat graph structure, the data is organized across multiple dimensional layers corresponding to different entity types, allowing the system to handle large volumes of data while maintaining structural clarity through this additional organizational dimension.
4Loss of information
If conventional graph database structures are used, then all entity relationships are maintained, but processing efficiency decreases with massive datasets
Solution Approach 1:
The patent segments the graph into type-specific subgraphs that maintain complete relationship information within each entity type context. This segmentation allows processing operations to be performed efficiently on smaller subsets while preserving the completeness of relationships relevant to each entity type, thereby improving processing efficiency without losing critical relationship information.
Data Source
AI summary
Certain aspects of the present disclosure provide techniques for generating a unified knowledge graph. In one example, a method includes receiving entity data from a data source comprising a plurality of entities; forming a plurality of type-specific groups of entity data based on the received entity data; for each respective type-specific group of entity data of the plurality of type-specific groups of entity data; disambiguating the entity data within the respective type-specific group of entity data; creating a plurality of entity relationships based on the disambiguated entity data; and exporting the plurality of entity relationships to a type-specific subgraph; and forming a unified knowledge graph based on a plurality of type-specific subgraphs, wherein each type-specific subgraph of the plurality of type-specific subgraphs is associated with a single type-specific group of entity data of the plurality of type-specific groups of entity data.