Custom Ontology Augmentation for Semantic Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing semantic search systems struggle to provide customized search experiences without the computational expense of rerunning natural language processing (NLP) over entire corpora, as custom enrichment is costly and resource-intensive.
Innovation Solution
Integrating a user-specific custom ontology with a semantic search function to augment search queries, allowing users to view a commonly enriched corpus through their own ontological lens without requiring additional NLP processing, thereby providing a customized search experience without the need for rerunning NLP over the entire corpus.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If custom ontology enrichment is performed on the corpus, then the search experience is customized and more relevant to user needs, but the computational cost and processing time increase significantly
Solution Approach 1:
The system segments the ontology application process by separating the static corpus enrichment (performed once with a default ontology) from the dynamic query augmentation (performed per-user with custom ontology). This allows multiple users to apply different ontologies to the same enriched corpus without reprocessing the entire corpus for each user, significantly reducing computational overhead while maintaining customization.
Solution Approach 2:
The corpus is pre-enriched using a default ontology before deployment, so that when users apply their custom ontologies, they only need to augment queries rather than re-enrich the entire corpus. This preliminary enrichment action eliminates the need for costly repeated NLP processing while preserving the ability to provide customized search experiences.
2Adaptability or versatility
If a custom enriched corpus is created for each ontology, then each user gets an individualized search experience, but the resource consumption and processing time multiply with each custom corpus
Solution Approach 1:
A single enriched corpus serves multiple users with different ontologies simultaneously. The system augments queries with user-specific ontology information at search time rather than creating separate enriched corpora for each user. This universal approach allows one corpus to fulfill multiple customization needs, eliminating redundant processing and reducing time loss.
3Measurement precision
If natural language processing is rerun over the entire corpus for each custom ontology, then the search results are highly accurate and relevant, but the computational resources and energy consumption become prohibitively expensive
Solution Approach 1:
The system extracts only the necessary ontology-specific information from custom ontologies and applies it as query augmentation rather than reprocessing the entire corpus. This extraction approach maintains search accuracy by focusing computational resources on augmenting queries with relevant ontology concepts, synonyms, and relationships specific to each user, while avoiding the prohibitive cost of rerunning NLP on the whole corpus.
Data Source
AI summary
Techniques include updating a semantic search function with a custom ontology, the semantic search function initially supporting a separate ontology having been used to enrich a corpus. The custom ontology is used to augment input of a search query for the semantic search function, thereby providing a custom user experience for searching the corpus.