Atomic Structure Simulation Model Generation for Accuracy-Atom Count Trade-off
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for simulating atomic structures face a trade-off between accuracy and the number of atoms that can be simulated, with high accuracy methods limited to small numbers of atoms and low accuracy methods capable of simulating larger systems but with reduced precision.
Innovation Solution
An information processing apparatus that determines parameters for a second model based on the types of atoms to be analyzed, generating a model capable of outputting analysis results for atomic structures, thereby enabling a balance between simulation accuracy and the number of atoms that can be simulated.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional simulation methods with high accuracy (CCSD, DFT, machine learning potential) are used, then the accuracy of atomic structure simulation is improved, but the number of atoms that can be simulated is limited
Solution Approach 1:
The patent segments the simulation problem by dividing atoms into groups based on their types and spatial distributions. Instead of treating all atoms individually with high-accuracy methods, the system creates multiple simulation regions with different accuracy levels, allowing high accuracy for specific regions of interest while using lower accuracy methods for other regions, thereby enabling simulation of larger numbers of atoms.
Solution Approach 2:
The patent applies local quality by assigning different simulation accuracy levels to different spatial regions. High-accuracy simulations are concentrated in specific regions where detailed atomic structure information is critical, while other regions use lower-accuracy methods. This allows the system to maintain high measurement precision where needed while expanding the overall number of atoms that can be simulated.
2Quantity of substance
If conventional simulation methods with low accuracy (classical potential) are used, then the number of atoms that can be simulated is increased, but the accuracy of simulation is reduced
Solution Approach 1:
The patent segments the simulation domain into multiple regions with different accuracy requirements. By dividing the system into zones that can be simulated with appropriate accuracy levels, the method enables simulation of large numbers of atoms using lower-accuracy classical potentials in most regions, while reserving high-accuracy methods for specific critical regions.
Solution Approach 2:
The patent applies partial action by using high-accuracy simulation methods for only the necessary portions of the system rather than the entire system. This allows the majority of atoms to be simulated with computationally efficient lower-accuracy methods, thereby increasing the total number of atoms that can be simulated while maintaining sufficient accuracy for the regions where it matters most.
Data Source
AI summary
An information processing apparatus according to an embodiment includes at least one processor and at least one memory. The at least one processor determines, based on types of atoms to be analyzed by a second model, at least part of parameters of a first model having been trained with first data. The at least one processor generates, by using the at least part of parameters of the first model, the second model different from the first model. The second model is a model from which an analysis result of an atomic structure is output by inputting the atomic structure.


