一个机器学习和可解释的人工智能框架,通过对吸附物化学环境的有效表示来预测吸附能
Mohammadreza Karamad1,2, Aditya Biswas1
1Department of Chemistry, Simon Fraser University, 8888 University Drive, Burnaby, British Columbia V5A 1S6, Canada.
The Journal of chemical physics
|January 9, 2026
概括
机器学习模型使用高斯多极 (GMP) 特性预测吸附能量,以改进催化剂选. 可解释的人工智能 (XAI) 揭示了电子结构和元素特性是准确的结合能预测的关键.
科学领域:
- 计算化学和材料科学.
- 催化和表面科学.
- 机器学习在科学发现中的应用.
背景情况:
- 吸附能量对于理解异质催化和选新催化剂至关重要.
- 机器学习 (ML) 通过从计算数据中预测吸附能量来加速催化剂的发现.
- 准确的ML模型需要有效地表示原子化学环境,包括几何和电子属性.
研究的目的:
- 开发一个ML框架,使用先进的电子结构描述器来预测吸附能量.
- 为了将高斯多极 (GMP) 特征化纳入新元素表示.
- 通过提高预测准确性和模型可解释性来增强催化剂选和设计.
主要方法:
- 开发了一个ML模型,将高斯多极 (GMP) 特征与几何特征集成.
- 在多金属合金上预测的CO和H结合能.
- 应用沙普利增量解释 (XAI) 来解释模型预测.
- 利用聚类和t-SNE分析来评估模型的概括性.
主要成果:
- 获得的平均绝对误差为0.07 eV的CO和0.06 eV的H结合能.
- XAI分析发现吸附剂和第一近邻 (FNN) GMP特征是最有影响力的.
- 更广泛的元素属性 (沸点,组号,原子号) 证明比电子负性更具预测性.
- 证明类似的FNN环境与一致的结合能相关.
结论:
- 集成电子结构特征 (GMP) 和XAI的ML框架提高了吸附能量的预测准确性和可解释性.
- 这种方法为加速催化剂选和合理的催化剂设计提供了一个强大的策略.
- 高斯多极特征化在ML模型中为催化提供了一种新且有效的元素身份表示.
相关概念视频
Analyte Adsorption and Distribution
2.5K
In certain chromatographic separations, solutes transfer between the mobile phase and the stationary phase via sorption, which typically refers to the process of adsorption. For many chromatographic systems, the sorption process often depends on the polarity of the compounds—an expression of the overall dipole moment within the molecule. During the separation process, there is competition between the solute and solvent for adsorption to the stationary phase. Highly polar compounds and...
2.5K
Predicting Reaction Outcomes
10.0K
Kinetics describes the rate and path by which a reaction occurs. In contrast, thermodynamics deals with state functions and describes the properties, behavior, and components of a system. It is not concerned with the path taken by the process and cannot address the rate at which a reaction occurs. Although it does provide information about what can happen during a reaction process, it does not describe the detailed steps of what appears on an atomic or a molecular level. On the other hand,...
10.0K
Predicting Products: SN1 vs. SN2
15.9K
Nucleophilic substitution reactions of alkyl halides can proceed via an SN1 or an SN2 mechanism. While in SN2 reactions, the nucleophile attacks the substrate simultaneously as the leaving group departs, in SN1 reactions, the substrate first dissociates to give the carbocation intermediate. Various factors such as the structure of the substrate, the strength of the nucleophile, and the nature of the solvent promote one mechanism over the other.
With increased substitution on the alkyl halide,...
With increased substitution on the alkyl halide,...
15.9K
Potential-Energy Criterion for Equilibrium
905
Potential energy or potential function plays an essential role in determining the stability of a mechanical system. If a system is subjected to both gravitational and elastic forces, the potential function of the system can be expressed as the algebraic sum of gravitational and elastic potential energy. If the system is in equilibrium and is displaced by a small amount, then the work done on the system equals the negative of the change in the system's potential energy from the initial to the...
905
Adrenergic Agonists: Chemistry and Structure-Activity Relationship
3.8K
Adrenergic agonists' structure-activity relationship (SAR) determines their selectivity and efficacy. These agonists comprise a phenylethylamine moiety with an aromatic ring and an ethylamine side chain.
Aromatic ring substitutions: Substituting the aromatic ring with –OH groups at positions 3 and 4 yields catecholamines (e.g., epinephrine), which have a high affinity for adrenoceptors. Hydrogen bonding between –OH groups and receptors enhances adrenergic activity.
Separation of...
Aromatic ring substitutions: Substituting the aromatic ring with –OH groups at positions 3 and 4 yields catecholamines (e.g., epinephrine), which have a high affinity for adrenoceptors. Hydrogen bonding between –OH groups and receptors enhances adrenergic activity.
Separation of...
3.8K
Molecular Models
43.5K
Physical models representing molecular architectures of chemical compounds play essential roles in understanding chemistry. The use of molecular models makes it easier to visualize the structures and shapes of atoms and molecules.
43.5K
