About the Speaker:-

M. Natália D. S. Cordeiro is an Associate Professor of Chemistry at the Faculty of Sciences, University of Porto (Portugal) and now heads the Chemistry Material Modelling research group at the Associated Laboratory for Green Chemistry (https://laqv.requimte.pt/). Initially working on molecular simulations of ionic solutions and fluid interfaces, driven by challenges in electrochemistry, her current research primarily revolves around leveraging machine learning tools. This involves predicting property/reactivity and assessing the safety of chemicals/ nanomaterials, with direct applications in the fields of drug and materials design. She is the author of 350+ SCI works, of 25+ book chapters, and 150+ oral communications in national and international conferences.

Abstract:-

This talk explores two applications of supervised machine learning (ML) in chemistry:

Part 1: Predicting vaporisation enthalpy (D vap H m °) of volatile organic compounds (VOCs). We developed an explainable ML model using a large dataset of chemicals. This model makes accurate D vap H m ° predictions (validated with known VOCs and hold-out tests) and identifies key contributing factors. This opens doors to predicting other VOC properties.

Part 2: Optimizing vanadium-catalyzed epoxidation reactions. Here, the focus is to predict reaction yield for epoxidation of small alcohols and alkenes. Our ML model is able to uncover relevant chemical characteristics for catalyst design and substrate selection, offering a pathway for automated optimisation of epoxidation reactions.