LLM Data Object Extraction and Visualization for Ontology Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current processes for updating an ontology are complex, unintegrated, and time-consuming, often requiring fragmented steps that are not easily repeatable or accurate.
Innovation Solution
A data extraction system utilizing machine learning, natural language processing, and interactive visualization techniques to automate the aggregation of information into databases as data objects, allowing users to efficiently update the ontology through an LLM-based unstructured data processing pipeline and UI-based functionality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated LLM-based processing is implemented, then productivity and time efficiency improve, but system complexity increases
Solution Approach 1:
The system divides ontology updating into distinct modular components: an LLM-based natural language processing module for extracting information from unstructured data, a structured data generation module for creating standardized triples, and an ontology integration module for incorporating new knowledge. This segmentation enables automated high-speed processing while managing complexity through clear module boundaries and specialized functions.
Solution Approach 2:
The patent introduces structured data (triples in subject-predicate-object format) as an intermediary representation between unstructured natural language input and the ontology knowledge base. This intermediary layer translates diverse unstructured data into a standardized format that the ontology system can process efficiently, bridging the gap between natural language processing and knowledge representation while maintaining system manageability.
2Loss of time
If manual ontology updating processes are used, then system complexity remains low, but time consumption and labor requirements increase
Solution Approach 1:
The system enables self-service automated ontology updating by allowing the LLM to autonomously extract information from unstructured data sources, generate structured triples, and integrate new knowledge into the ontology without requiring manual intervention for each update cycle. This self-service capability dramatically reduces time consumption and labor requirements while maintaining operational simplicity through automated workflows.
3Reliability
If fragmented manual steps are used for ontology updates, then accuracy may be maintained through human review, but repeatability and consistency deteriorate
Solution Approach 1:
The patent transforms the ontology updating process from manual parameter adjustment to automated parameter-driven processing by using the LLM to automatically extract entities, relationships, and attributes from unstructured data. The structured triple format standardizes all parameters (subject, predicate, object) ensuring consistent representation. This parameter standardization maintains reliability through consistent data quality while enabling high-level automation with repeatable results across different update cycles.
Data Source
AI summary
Computer-implemented systems and methods are disclosed, including systems and methods utilizing language models for generating data objects and/or updating an ontology. A computer-implemented method may include: employing one or more large language models (“LLMs”) to generate at least a data triple and a classified triple; executing, using the classified triple, a similarity search with reference to an ontology to determine that the classified triple at least partially matches one or more data object types defined in the ontology; in response to the determination, adding into a first database at least a first data object of a first data object type that represents a first entity in the data triple and a second data object of a second data object type that represents a second entity in the data triple.