Hierarchical Material Prediction Workflows for Flexible Model Composition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods lack a systematic approach to predict the performance, properties, and structures of materials by freely combining processing means, structure, and performance, as highlighted in Non-Patent Document 1.
Innovation Solution
A prediction management system that includes a descriptor storage unit, prediction model storage unit, workflow storage unit, and execution unit, which manage descriptors, prediction models, and workflows in hierarchical layers, allowing for the input and output of parameters related to processing means, structure, and performance, enabling flexible model construction and prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a systematic approach for predicting material properties by freely combining processing means, structure, and performance is implemented, then the adaptability and versatility of the prediction system is improved, but the device complexity increases due to the need to manage multiple prediction models, descriptors, and workflows in hierarchical layers
Solution Approach 1:
The prediction management system is segmented into four hierarchical layers: descriptor layer, prediction model layer, workflow layer, and execution result layer. Each layer manages specific types of information independently, allowing the system to handle complex prediction tasks while maintaining organizational clarity. Descriptors store material parameters, prediction models define input-output relationships, workflows sequence multiple models, and execution results capture outcomes, creating a modular structure that balances versatility with manageability.
Solution Approach 2:
The patent introduces a hierarchical dimensional structure to organize prediction information. By arranging descriptors, models, workflows, and results in four distinct hierarchical layers with unique identifiers, the system transforms the complexity management from a flat structure into a multi-dimensional framework. This dimensional organization allows flexible combination of processing means, structure, and performance predictions while maintaining systematic control over the increasing complexity.
2Productivity
If multiple prediction models are connected via workflows to enable free combination of processing means, structure, and performance predictions, then the productivity of material prediction is improved, but the difficulty of detecting and measuring the relationships among models increases
Solution Approach 1:
The workflow acts as an intermediary layer that connects multiple prediction models through descriptors. Each workflow defines explicit input-output relationships between models using standardized descriptor interfaces, making the connections detectable and measurable. The execution unit traverses these workflows systematically, tracking how input descriptors flow through sequential models to produce output descriptors, thereby reducing the difficulty of detecting model relationships while maintaining high prediction productivity.
Solution Approach 2:
Descriptors serve as universal interfaces that enable multiple prediction models to connect and interact. Each descriptor represents a standardized material parameter that can be both an input to one model and an output from another, creating a flexible network of predictions. This universal descriptor interface simplifies the detection and measurement of model relationships, as the same descriptor type can be tracked across different models and workflows, enhancing both productivity and traceability.
Data Source
AI summary
System that makes a prediction regarding a material, including: storing descriptors each describing a parameter regarding processing means, structure, property, or performance of the material; storing prediction models each describing an input and output relationship among the descriptors, wherein each prediction model inputs one of at least two of the processing means, structure, property, and performance, and outputs another one thereof; storing workflows each describing that at least two of the prediction models are connected via a descriptor, wherein the output of one prediction model is accepted as the input of another prediction model; giving an input to each of the workflows, executing the workflow including execution of each prediction model included in the workflow, and storing an execution result including an output result of each prediction model included in the workflow; managing execution results, workflows, prediction models, and descriptors in four hierarchical layers by assigning a unique identifier.