Knowledge Graph Search for Compositional Queries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search engines struggle to effectively handle compositional queries that involve multiple entity types and relationships, such as finding American banks near Japanese restaurants, as they typically rely on single, fixed location or temporal criteria, failing to provide nuanced results.
Innovation Solution
A computer-implemented method using a knowledge graph to identify and compare entity references of different types based on specified relationships, determining resultant entities that meet the query criteria by analyzing attributes and relationships within the graph, and presenting these results to users through a user interface.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If search engines use single, fixed location or temporal criteria, then the search system is simple and easy to operate, but the search results lack precision and relevance for compositional queries
Solution Approach 1:
The patent segments the search query into multiple entity types (e.g., person, location, organization) and processes each entity type separately through dedicated entity recognition modules. This allows the system to handle complex compositional queries by breaking them down into manageable components while maintaining operational simplicity for users.
Solution Approach 2:
The patent introduces a knowledge graph as an intermediary structure that stores pre-defined relationships between entities. This mediator enables the system to resolve complex queries by querying pre-computed relationships rather than performing complex real-time analysis, thus improving precision without proportionally increasing system complexity.
2Measurement precision
If search engines process compositional queries with multiple entity types, then the search results become more relevant and accurate, but the processing time and computational resources increase
Solution Approach 1:
The patent performs preliminary actions by pre-processing and structuring entity relationships in a knowledge graph before queries are submitted. Entity types and their relationships are identified and organized in advance, allowing the system to quickly retrieve pre-computed results rather than performing complex analysis for each query, thus reducing processing time while maintaining high accuracy.
Solution Approach 2:
The patent applies different processing strategies to different entity types based on their specific characteristics. Each entity type (person, location, organization) has dedicated recognition and processing rules tailored to its unique attributes, enabling efficient and accurate processing without uniformly applying complex algorithms to all entities.
3Loss of information
If search engines return comprehensive results for compositional queries, then the information completeness improves, but the ease of operation and result presentation deteriorates
Solution Approach 1:
The patent segments search results by entity type and presents them in an organized, hierarchical format. Instead of returning a flat list of comprehensive results, the system groups results by entity category (e.g., persons, locations, organizations) with clear headings and structure, making comprehensive information easier to navigate and understand.
Solution Approach 2:
The patent inverts the traditional search result presentation by organizing results around entity types and their relationships rather than presenting a flat list of matching documents. This inversion allows users to see the structured relationships between entities and access comprehensive information through an intuitive, category-based interface.
Data Source
AI summary
A technique for providing search results may include determining a first entity type, a second entity type, and a relationship type based on a compositional query. The technique may also include identifying nodes of a knowledge graph corresponding to entity references of the first entity type and entity references of the second entity type. The technique may also include determining from the knowledge graph an attribute value corresponding to the relationship type for each entity reference of the first entity type and for each entity reference of the second entity type. The technique may also include comparing the attribute value of each entity reference of the first entity type with the attribute value of each entity reference of the second entity type. The technique may also include determining one or more resultant entity references from the entity references of the first entity type based on the comparing.