通过结合从初始状态到平衡状态的位移来增强玻璃动力学的预测
Xiao Jiang1, Zean Tian1, Yikun Hu1
1College of Computer Science and Electronic Engineering, Hunan University, Changsha 410012, China.
The journal of physical chemistry. B
|March 6, 2025
概括
这项研究引入了一种新的机器学习方法,使用矢量位移来预测玻璃动态. 以等差约束的不变图神经网络 (EIGNN) 改善了对结构动力学相关性的理解.
科学领域:
- 凝聚物质物理学 凝聚物质物理学
- 材料科学 材料科学 材料科学
- 计算化学计算化学
背景情况:
- 玻璃过渡是一个复杂的现象.
- 机器学习 (ML) 模型通过考虑粒子移位来增强玻璃动态的预测.
- 在子内的粒子振动的方向方面已经被忽视了.
研究的目的:
- 将矢量位移纳入ML模型以预测玻璃动态.
- 引入等效约束不变图神经网络 (EIGNN),以改进结构编码.
- 为了证明粒子位移方向在玻璃动力学中的重要性.
主要方法:
- 利用从初始到平衡状态的向量位移作为ML模型的结构输入.
- 开发和应用等效约束不变图神经网络 (EIGNN).
- 在GlassBench数据集中的3DKob-Andersen系统上验证EIGNN.
主要成果:
- EIGNN显著提高了对玻璃系统中结构动力学相关性的理解.
- 该模型证明了强大的温度可转移性.
- 一个简化的模型 (EIGNN++) 显示了位移参数代表了局部债券定向顺序.
结论:
- 矢量位移是预测玻璃动态的一个关键因素.
- 子动态的方向在改善预测模型方面发挥着至关重要的作用.
- EIGNN为研究玻璃过渡提供了一个强大的框架.
相关概念视频
Solution Equilibrium and Saturation
18.3K
Imagine adding a small amount of sugar to a glass of water, stirring until all the sugar has dissolved, and then adding a bit more. You can repeat this process until the sugar concentration of the solution reaches its natural limit, a limit determined primarily by the relative strengths of the solute-solute, solute-solvent, and solvent-solvent attractive forces. You can be certain that you have reached this limit because, no matter how long you stir the solution, undissolved sugar remains. The...
18.3K
Recrystallization: Solid–Solution Equilibria
1.0K
Recrystallization is a purification technique used to separate impurities from solid compounds. In this technique, no chemical reactions occur. Instead, it exploits physical properties only, specifically, the solubility differences between the desired compound and impurities, either at a single temperature or at different temperatures, and under other selected conditions. The solid-solution equilibrium (solubility equilibrium) of each component in the solution represents a binary phase...
1.0K


