Material Property Search System Using Extracted Relational Formulas
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing search systems for material property parameters are impractical due to limitations in handling huge numbers of parameters and lack of efficient methods for extracting quantitative relations, requiring extensive manpower and expertise for database creation, and failing to process relations among multiple parameters effectively.
Innovation Solution
A method and system for extracting relational formulas from textbook documents using natural language processing and deep learning, storing them in a material property relationship database, and using these formulas to search for quantitative relations among material property parameters, enabling the creation of a comprehensive database without manual expertise and reducing manual effort.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual extraction and storage of relational formulas is performed, then database accuracy and reliability are improved, but labor cost and time consumption increase significantly
Solution Approach 1:
The system enables automatic extraction of relational formulas from textbook documents using natural language processing and deep learning technologies. The extraction unit autonomously identifies and extracts formulas without human intervention, while the storage unit automatically stores them in the database, making the entire database creation process self-service oriented and eliminating manual labor
Solution Approach 2:
The patent replaces the mechanical manual process of extracting and storing relational formulas with an automated information processing system. Natural language processing and deep learning algorithms substitute human cognitive and manual operations, transforming the mechanical process into an automated computational system that extracts formulas directly from document images or text
2Reliability
If comprehensive manual expertise is required for database creation, then data quality is improved, but ease of manufacture deteriorates
Solution Approach 1:
The system eliminates the need for manual expertise by implementing self-service extraction through natural language processing and deep learning. The extraction unit automatically processes textbook documents and identifies relational formulas without requiring domain experts or manual verification, making database creation accessible to anyone with basic technical knowledge
Solution Approach 2:
The patent introduces natural language processing and deep learning technologies as intermediaries between the textbook documents and the database. These intermediary systems bridge the gap between unstructured document content and structured database entries, automatically transforming raw information into quality-assured data without requiring human expertise in the intermediate processing stage
3Adaptability or versatility
If quantitative relations among multiple material property parameters are processed, then search capability is improved, but device complexity increases
Solution Approach 1:
The search unit is designed with multi-functionality to handle various types of searches including single parameter searches, multi-parameter relational searches, and quantitative relationship analyses. The same search unit processes different query types by interpreting search conditions and executing appropriate retrieval operations, eliminating the need for separate specialized systems for each search type
Solution Approach 2:
The patent introduces a relational formula database as an intermediary layer between the search unit and the material property data. This intermediary database stores pre-extracted relational formulas that encode quantitative relationships among multiple parameters, allowing the search unit to efficiently process complex multi-parameter queries without directly computing relationships, thereby reducing computational complexity
Data Source
AI summary
Provided is a method of extracting a relational formula that relates two material property parameters from a textbook document using a computer and storing the extracted relational formula in a material property relationship database, which method enables search in consideration of a quantitative relationship between material properties. The information processing method according to the present invention comprises inputting a relational formula representing a relation between a pair of material property parameters in a material property relationship database storing pairs of mutually related material property parameters. Equation information representing a relational formula is extracted from read input data, and multiple variables constituting a relational formula and a relational formula specifying that relation are extracted from the equation information. Description defining each of the extracted variables is extracted from the input data, and each variable is associated with a material property parameter with reference to the material property relationship database. The extracted relational formula is input in the material property relationship database in association with the pair of material property parameters corresponding to two of the multiple variables constituting the relational formula.