通过主动学习微调图形神经网络:释放在非水性系统上训练的图形神经网络的潜力,以减少水性CO22
Zihao Jiao1,2, Yu Mao2, Ruihu Lu2
1International Research Center for Renewable Energy, State Key Laboratory of Multiphase Flow in Power Engineering, Xi'an Jiaotong University, Xi'an, Shaanxi 710049, China.
Journal of chemical theory and computation
|March 14, 2025
概括
本研究引入了一种主动学习方法,以适应图形神经网络 (GNN) 用于水性催化. 这种方法加速了模拟,并准确地预测了水中的反应路径,这对于现实的催化机制预测至关重要.
科学领域:
- 计算化学的计算化学
- 材料科学 材料科学 材料科学
- 化学工程是化学工程的重要组成部分.
背景情况:
- 图形神经网络 (GNN) 在模拟化学相互作用方面表现出色,但由于水的复杂性,与水系统作斗争.
- 适应非水性GNN模型用于水基催化是实现现实的模拟的必要条件.
研究的目的:
- 开发一种基于活跃学习的微调方法,用于水性催化中的 GNNs.
- 为了加快几何优化和过渡状态搜索的计算成本,同时保持 DFT 精度.
- 研究催化路径和金属合金在水环境中的性能.
主要方法:
- 基于主动学习的GNN对水系统的微调.
- 几何优化和过渡状态搜索工作流程,具有放松的强度值和DFT改进.
- 基于干扰的力量和能量训练,用于初始分子动力学模拟.
主要成果:
- 在二氧化碳减排反应几何优化的2-3倍加速.
- 确定*CO-*COH作为水性Cu和Cu合金系统中最有利的途径.
- 证明了强大的布伦斯特德-埃文斯-波兰尼关系和水中的各种金属合金的可比合效率.
结论:
- 提出的积极学习方法有效地将GNN的适用性扩展到水性催化系统.
- 新的工作流加速了在现实的溶解条件下对催化机制的预测.
- 在水性环境中,GNN显示出模拟水性环境中复杂动态的巨大潜力.
相关概念视频
Titration in Nonaqueous Solvents
705
Most acid-base titrations are performed in an aqueous medium. In aqueous titrations, water competes with weaker acids or bases for proton donation or acceptance, leading to ambiguous endpoints in the titration curve. Water also affects the partial ionization of weak acids or bases. For example, water accepts a proton from acetic acid to form hydronium and acetate ions. The hydronium ion formed is a stronger acid than acetic acid, and the acetate ion is a stronger base than water. As a result,...
705
Calculating Equilibrium Concentrations
47.2K
Being able to calculate equilibrium concentrations is essential to many areas of science and technology—for example, in the formulation and dosing of pharmaceutical products. After a drug is ingested or injected, it is typically involved in several chemical equilibria that affect its ultimate concentration in the body system of interest. Knowledge of the quantitative aspects of these equilibria is required to compute a dosage amount that will solicit the desired therapeutic effect.
A more...
A more...
47.2K


