Feasibility Estimation for Molecular Feature Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current information technology methods for material discovery are inefficient, consuming significant time and resources due to the inability to effectively filter out infeasible feature vectors in the search for new materials with target properties.
Innovation Solution
A computer-implemented method is introduced that generates a target structure vector from a molecule candidate's feature vector, using information about partial structures, atoms, and rings to determine feasibility, thereby identifying and eliminating infeasible feature vectors, reducing computational resources and time needed for material discovery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If comprehensive feature vector search is performed to find molecules with target properties, then the likelihood of finding suitable materials increases, but the computational time and resources increase significantly
Solution Approach 1:
The patent performs preliminary feasibility checking of feature vectors before conducting comprehensive material discovery searches. By generating target structure vectors from feature vectors and checking feasibility conditions (such as valence rules, aromaticity, and structural constraints) in advance, the system eliminates impossible molecular structures before they consume computational resources for detailed analysis, thus resolving the contradiction between thorough search and computational efficiency
Solution Approach 2:
The patent extracts and separates the feasibility verification step from the main material discovery process. By isolating the feasibility check as a distinct preliminary operation that filters out invalid feature vectors, the system removes the burden of processing infeasible structures from the computational workflow, maintaining high reliability while reducing overall computational time
2Reliability
If comprehensive feature vector search is performed to find molecules with target properties, then the likelihood of finding suitable materials increases, but the computational resources required increase significantly
Solution Approach 1:
The patent performs preliminary feasibility checking of feature vectors before conducting comprehensive material discovery searches. By generating target structure vectors from feature vectors and checking feasibility conditions (such as valence rules, aromaticity, and structural constraints) in advance, the system eliminates impossible molecular structures before they consume computational resources for detailed analysis, thus resolving the contradiction between thorough search and computational efficiency
Solution Approach 2:
The patent extracts and separates the feasibility verification step from the main material discovery process. By isolating the feasibility check as a distinct preliminary operation that filters out invalid feature vectors, the system removes the burden of processing infeasible structures from the computational workflow, maintaining high reliability while reducing overall computational time
3Productivity
If feasibility checking is performed for all feature vectors, then infeasible structures are eliminated, but the processing time increases
Solution Approach 1:
The patent applies partial feasibility checking by focusing on critical structural features (such as valence rules, aromaticity, and basic structural constraints) rather than performing exhaustive validation on all possible molecular properties. This selective approach checks only the most important feasibility conditions that can be verified quickly, eliminating clearly infeasible structures without investing excessive time in comprehensive validation of every feature vector
Data Source
AI summary
Feature vector feasibility is estimated by generating a target structure vector that represents numbers of a plurality of partial structures, from a feature vector of a molecule candidate, determining whether a molecule structure of the molecule candidate is feasible by using at least the target structure vector.