Title : Predictive modelling for catalysis using supervised machine learning
Abstract:
The field of catalysis is undergoing a paradigm shift with the advent of data-driven methodologies that enable rapid prediction, screening, and optimization of catalytic processes. Supervised Machine Learning (ML), in particular, offers powerful tools to model complex, nonlinear relationships between catalyst features and performance metrics, transforming how researchers approach catalyst design and reaction engineering.
This keynote will explore the fundamentals and frontiers of supervised learning techniques applied to catalysis, showcasing how historical experimental data can be harnessed to build robust predictive models. From regression models estimating turnover frequencies to classification algorithms identifying optimal catalyst compositions, the session will cover a spectrum of ML approaches tailored to chemical engineering challenges. Through real-world case studies and research insights, the talk will highlight the integration of domain knowledge, feature engineering, and model validation strategies essential for creating interpretable and accurate models. Attendees will gain a deeper understanding of how machine learning can augment traditional catalysis workflows, reduce development timelines, and drive innovation in sustainable chemical processes. The session aims to provide researchers, engineers, and industry stakeholders with practical strategies and scientific insights to effectively apply supervised machine learning in catalytic research, thereby fostering faster innovation and more sustainable chemical processes.