How Machine Learning is Innovating Today's World

9Machine Learning Approaches to Catalysis

Sachidananda Nayak and Selvakumar Karuthapandi^*

Department of Chemistry, School of Advanced Sciences, VIT-AP University, Amaravati, Andhra Pradesh, India

Abstract

Data-driven research in chemistry has emerged as a new platform to identify potential molecules, examine dynamic reaction mechanisms, and extract knowledge from vast sets of data that are made possible by the use of rapidly growing machine learning (ML) approaches. The use of ML-based models can speed up computational algorithms and enhance computational chemistry findings to make chemical sciences more effective. This chapter provides a basic introduction to data collection, data processing, model validation approaches, basics of common ML models, and application of such models in catalysis. Finally, it discusses how computational chemistry and ML may be utilized to provide relevant predictions in the areas of atomistic understanding of catalysis.

Keywords: Machine learning (ML), catalysis, electrocatalysts, electrocatalysis, neural networks, structure activity predictions, CO₂ reduction, deep learning, predictive analysis

9.1 Introduction

Machine learning (ML) is a sophisticated artificial intelligence (AI) technique that has been extensively utilized for classification [1], numerical optimization [2], and pattern recognition [3]. ML is able to “learn” as its name suggests. Through the use of non-linear “black box” data processing, ML can give extremely complex correlations ...