Relational Semantic Structures for Descriptor Standardization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating ontologies and knowledge maps struggle with accurately identifying relationships in large volumes of text data due to varying terminology across different authors and time periods, leading to inefficiencies and inaccuracies.
Innovation Solution
A machine learning-based approach using a semantic standardization model that leverages embedding vectors to perform sequence detection and pattern recognition, enabling the generation of relational semantic structures through a preliminary model and a first semantic standardization model to correlate descriptors with embeddings, and generate a relational semantic structure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional methods are used to generate ontologies and knowledge maps, then the process is simpler, but the accuracy and reliability of identifying relationships in text data deteriorates
Solution Approach 1:
The patent replaces traditional mechanical/algorithmic relationship extraction methods with machine learning models, specifically transformer-based architectures that use embedding vectors to capture semantic relationships. This substitution enables the system to achieve higher accuracy in identifying relationships across large volumes of text data with varying terminology.
Solution Approach 2:
The patent changes the fundamental parameters of relationship extraction by using embedding vectors instead of traditional keyword matching or syntactic analysis. The embedding space transformation allows for more accurate capture of semantic relationships, enabling the model to identify connections between concepts that use different terminology across various contexts.
2Measurement precision
If traditional keyword matching methods are used, then the processing is faster, but the ability to discern intricate patterns and sequences in text data deteriorates
Solution Approach 1:
The patent performs preliminary action by pre-training embedding models on large datasets before actual relationship extraction. This pre-computation of embeddings allows the system to quickly query and compare semantic relationships during processing, achieving both high precision pattern recognition and efficient processing speeds for large volumes of text.
Solution Approach 2:
The patent introduces embedding vectors as an intermediary representation between raw text and relationship extraction. This intermediary layer captures semantic meaning and enables efficient pattern matching, allowing the system to discern intricate patterns in text data faster than traditional methods while maintaining high precision.
3Productivity
If manual ontology construction is used, then the quality control is better, but the productivity and scalability deteriorates
Solution Approach 1:
The patent implements self-service by enabling the machine learning model to automatically extract relationships from text data without requiring extensive manual annotation or curation. The model uses its trained embedding representations to autonomously identify relationships, significantly improving productivity while maintaining quality through the robustness of the pre-trained models.
Solution Approach 2:
The patent incorporates feedback mechanisms where the system can validate and refine its relationship extractions. The model continuously learns from the data it processes, improving its accuracy over time. This feedback loop enables the system to maintain high reliability in relationship identification while scaling to process large volumes of text data efficiently.
Data Source
AI summary
Systems and methods for generating relational semantic structures. In some aspects, the system retrieves a training dataset for the generation of a first semantic standardization model, comprising a set of concepts and a set of descriptors. The system trains a preliminary model based on the training dataset to correlate the set of descriptors with a set of embeddings. Based on the set of embeddings, the system trains the first semantic standardization model to find patterns and relations between embeddings. Using the first semantic standardization model, the system generates a relational semantic structure comprising a relation from each concept in the set of concepts to one or more descriptors in the set of descriptors. Based on the relational semantic structure, the system generates one or more notifications to update and standardize descriptors used to describe concepts in the set of concepts.