Knowledge Graph Modeling for Chemical Data Automation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In chemical engineering ecosystems, the complexity of managing relationships and market data across hundreds of thousands of chemicals and millions of relationships leads to inefficiencies, resulting in delayed and outdated information for customers, as subject matter experts spend excessive time creating and updating models using spreadsheets.
Innovation Solution
A computer-implemented method that identifies knowledge graphs for components in a production environment and trains machine learning models to predict attributes, using correlation analysis to select and combine relevant attributes, thereby reducing the time and effort required for data processing and model updates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If subject matter experts manually create and update models using spreadsheets to manage relationships and market data, then model accuracy and completeness can be maintained, but the time required for data processing and model updates increases significantly
Solution Approach 1:
The patent replaces the manual mechanical process of spreadsheet-based model creation with an automated machine learning system. The ML model automatically ingests unstructured data, extracts relationships, and generates structured models without human intervention in the data processing phase, thereby reducing time while maintaining accuracy through systematic analysis.
Solution Approach 2:
The system enables self-service by allowing the machine learning model to autonomously process data, identify relationships, and update models without requiring continuous human supervision. The automated pipeline performs data ingestion, processing, and model generation independently, freeing subject matter experts from routine tasks.
2Quantity of substance
If the chemical engineering ecosystem expands to include hundreds of thousands of chemicals and millions of relationships, then the comprehensiveness of market data improves, but the complexity of managing and processing this data increases
Solution Approach 1:
The patent replaces complex manual data management processes with automated machine learning systems that can handle large-scale unstructured data. The ML model automatically processes millions of relationships and chemicals, extracting meaningful patterns without requiring proportional increases in human management resources.
Solution Approach 2:
The machine learning system serves multiple functions simultaneously: data ingestion, relationship extraction, model generation, and prediction. This multi-functional approach consolidates what would otherwise require separate manual processes for each task, reducing overall system complexity despite the scale of data.
3Loss of time
If subject matter experts continuously supervise and update models to keep market data current, then the timeliness of information provided to customers improves, but the labor requirements and operational effort increase
Solution Approach 1:
The system achieves self-service through automated continuous monitoring and updating by the machine learning model. The model autonomously ingests new data, reprocesses relationships, and updates predictions without requiring subject matter experts to continuously supervise or manually update models, thereby maintaining timeliness while reducing operational effort.
Solution Approach 2:
The patent implements continuous automated data processing and model updating by the machine learning system. Unlike manual processes that occur intermittently when experts are available, the ML model operates continuously to ingest new data and update predictions, ensuring constant timeliness without proportional increases in human effort.
Data Source
AI summary
A method, apparatus, system, and computer product for modeling a production environment. A computer system identifies a knowledge graph for a component in the production environment. The computer system trains a machine learning model to predict a set of attributes for the component using the knowledge graph.