Knowledge Graph Indirect Entity Search via Vector Embeddings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems struggle to identify indirect entities and stakeholders impacted by a project, especially those not explicitly mentioned in project-specific documents, which can lead to unforeseen negative consequences.
Innovation Solution
A computer-implemented method using a knowledge graph to search for indirect entities by generating embedding entities with vectorial representations, identifying similar entities based on similarity thresholds, and returning new entities not explicitly found in project-specific documents as indirect entities in search results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional search methods are used to identify entities in project documents, then the search process is simple and fast, but indirect entities and stakeholders not explicitly mentioned in documents cannot be identified
Solution Approach 1:
The patent introduces a knowledge graph as an intermediary structure that connects project entities with general entities through embedding spaces. This mediator enables the system to infer indirect entities by finding similar entities in the embedding space, resolving the contradiction between identification completeness and system complexity
Solution Approach 2:
The patent replaces traditional text-based search mechanisms with a vector-based embedding system. By transforming entities into vector representations and using similarity calculations, the system can identify indirect entities without relying on explicit text mentions, thus improving completeness while maintaining manageable complexity through mathematical operations
2Reliability
If knowledge graph with embedding entities is used to identify indirect entities, then completeness of entity identification is improved, but computational complexity and processing time increase
Solution Approach 1:
The patent performs preliminary actions by pre-computing embedding vectors for entities and organizing them in a knowledge graph structure before actual search operations. This preprocessing enables faster similarity calculations during entity identification, reducing processing time while maintaining identification completeness
Solution Approach 2:
The patent implements a similarity threshold mechanism that filters embedding entities based on a predefined threshold. By only processing entities that exceed this threshold, the system avoids unnecessary computational overhead while ensuring that all relevant indirect entities are identified, thus optimizing the balance between completeness and processing time
Data Source
AI summary
Provided are techniques for searching for indirect entities using a knowledge graph. A search request for indirect entities for one or more project specific input documents is received. A knowledge graph is generated with entities and impacts in the one or more project specific input document and with entities and impacts in one or more general input documents. Embedding entities are generated for the entities in the knowledge graph, where each of the embedding entities has a corresponding vectorial representation that represents a position in an embedding space. Embedding entities are identified based on similarity values of the corresponding vectorial representations exceeding a similarity threshold. In response to identifying at least one new entity from the identified embedding entities not explicitly found in the one or more project specific input documents, the at least one new entity is returned as an indirect entity in a search result.