Infusing theory into deep learning for interpretable reactivity prediction.

Infusing theory into deep learning for interpretable reactivity prediction.
复制标题

DOI:
10.1038/s41467-021-25639-8
复制
发表时间:
2021-09-06
影响因子:
16.6
通讯作者:
Xin H
Xin H
中科院分区:
综合性期刊1区
文献类型:
--
作者:
Wang SH;Pillai HS;Wang S;Achenie LEK;Xin H

文献摘要

参考文献

被引文献

相似文献

尽管数据采集和算法开发最近取得了进展,但机器学习(ML)在实际催化剂设计中采用仍面临巨大挑战,这主要是由于其通用性有限且可解释性差。在此,我们开发了一种理论注入神经网络 (TinNet) 方法,它将深度学习算法与成熟的 d 带化学吸附理论相结合,用于过渡金属表面的反应性预测。使用活性位点集合中的简单吸附物(例如 *OH、*O 和 *N)作为代表性描述符物种,我们证明 TinNet 在预测性能方面与纯数据驱动的 ML 方法相当,同时具有固有的可解释性。将物理相互作用的科学知识纳入数据学习中,可以进一步阐明化学键的本质,并为机器学习发现具有所需催化特性的新基序开辟新途径。由于其黑盒性质,机器学习在催化剂设计方面面临挑战。在这里,作者开发了一种注入理论的神经网络方法,将深度学习算法与成熟的 d 带化学吸附理论相结合,用于过渡金属表面的反应性预测。
Despite recent advances of data acquisition and algorithms development, machine learning (ML) faces tremendous challenges to being adopted in practical catalyst design, largely due to its limited generalizability and poor explainability. Herein, we develop a theory-infused neural network (TinNet) approach that integrates deep learning algorithms with the well-established d-band theory of chemisorption for reactivity prediction of transition-metal surfaces. With simple adsorbates (e.g., *OH, *O, and *N) at active site ensembles as representative descriptor species, we demonstrate that the TinNet is on par with purely data-driven ML methods in prediction performance while being inherently interpretable. Incorporation of scientific knowledge of physical interactions into learning from data sheds further light on the nature of chemical bonding and opens up new avenues for ML discovery of novel motifs with desired catalytic properties. Machine learning faces challenges in catalyst design due to its black-box nature. Here, the authors develop a theory-infused neural network approach that integrates deep learning algorithms with the well-established d-band theory of chemisorption for reactivity prediction of transition-metal surfaces.
DOI: 10.1016/j.electacta.2017.06.012
发表时间: 2017-08-10
影响因子: 6.6
作者:
Amin, Hatem M. A.;Molls, Christoph.;Baltruschat, Helmut
通讯作者: Baltruschat, Helmut
DOI: 10.1038/s41557-018-0125-5
发表时间: 2018-10-01
期刊: NATURE CHEMISTRY
影响因子: 21.8
作者:
Greiner, M. T.;Jones, T. E.;Schloegl, R.
通讯作者: Schloegl, R.
DOI: 10.1021/acs.jpclett.9b01428
发表时间: 2019-08-01
影响因子: 5.7
作者:
Back, Seoin;Yoon, Junwoong;Ulissi, Zachary W.
通讯作者: Ulissi, Zachary W.
DOI: 10.1103/physrevlett.93.156801
发表时间: 2004-10-08
影响因子: 8.6
作者:
Kitchin, JR;Norskov, JK;Chen, JG
通讯作者: Chen, JG
DOI: 10.1021/acsami.9b06889
发表时间: 2019-09-18
影响因子: 9.5
作者:
Li, Guang-Fu;Divinagracia, Maricor;Chuang, Po-Ya Abel
通讯作者: Chuang, Po-Ya Abel