Infusing Theory into Deep Learning for Interpretable Reactivity Prediction

TL;DR

This paper introduces a theory-infused neural network that combines deep learning with chemisorption theory to improve interpretability and generalizability in predicting the reactivity of transition-metal surfaces for catalyst design.

Contribution

The study presents a novel neural network approach that integrates physical theory with machine learning, enhancing interpretability without sacrificing predictive accuracy.

Findings

TinNet matches pure ML methods in prediction accuracy.

Incorporating physical theory improves interpretability.

Method enables discovery of new catalytic motifs.

Abstract

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. Here, 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.