通过等价局部表示和电荷平衡来学习非局部分子相互作用
Paul Fuchs1, Michał Sanocki1,2, Julija Zavadlav1,2
1Multiscale Modeling of Fluid Materials, Department of Engineering Physics and Computation, TUM School of Engineering and Design, Technical University of Munich, Munich, Germany.
图形神经网络的潜力通过长距离相互作用的电荷平衡层 (CELLI) 来增强,以建模非局部效应. 这种方法实现了局部模型的最新结果,改善了复杂系统的预测.
科学领域:
- 计算化学是一种计算化学.
- 材料科学是一种材料科学.
- 机器学习是机器学习.
背景情况:
- 图形神经网络 (GNN) 为化学模拟提供近量子力学精度.
- 目前的GNN主要是模拟本地交互,因为它们的信息传递性质.
- 模拟电荷转移和静电学等远程相互作用仍然是GNN面临的挑战.
研究的目的:
- 为GNN潜力开发一种新的架构,能够有效地建模远程交互.
- 将经典的电荷平衡 (Qeq) 方法概括为一个多功能GNN构建块.
- 提高GNN对复杂化学系统的可解释性和预测能力.
主要方法:
- 介绍长距离相互作用 (CELLI) 架构的电荷平衡层.
- 在GNN框架内对经典电荷平衡原理的概括.
- 开发一个与等价GNN潜力相容的模型无关层.
主要成果:
- 塞利成功地模拟了非局部交互,克服了严格局部GNN的局限性.
- 与现有的本地模型相比,在基准系统上实现了最先进的性能.
- 在各种数据集和大型分子结构中展示了概括能力.
结论:
- 塞利显著扩大了GNN潜力的适用性,使其适用于需要远程交互建模的系统.
- 由于CELLI对费用进行了明确的建模,从而提高了模型的解释性.
- 塞利为先进的分子模拟提供了一种计算效率高和强大的方法.
更多相关视频
08:54Vibrational Spectra of a N719-Chromophore/Titania Interface from Empirical-Potential Molecular-Dynamics Simulation, Solvated by a Room Temperature Ionic Liquid
Published on: January 25, 2020
08:04Excitonic Hamiltonians for Calculating Optical Absorption Spectra and Optoelectronic Properties of Molecular Aggregates and Solids
Published on: May 27, 2020
相关概念视频
Noncovalent Attractions in Biomolecules
Four types of noncovalent interactions are hydrogen bonds, van der Waals forces, ionic bonds, and hydrophobic interactions.
Hydrogen bonding results from the electrostatic attraction of a hydrogen atom covalently bonded to a strong-electronegative atom like oxygen,...
Noncovalent Attractions in Biomolecules
Inductive Effects on Chemical Shift: Overview
Resonance and Hybrid Structures
Resonance Structures and Resonance Hybrids
The Lewis structure of a nitrite anion (NO2−) may actually be drawn in two different ways, distinguished by the locations of the N–O and N=O bonds.
The Equilibrium Binding Constant and Binding Strength
The Equilibrium Binding Constant and Binding Strength
