This presentation discusses three key aspects of the application of machine learning (ML) in organic chemistry: (i) the automated curation of reactions data sets from the literature and electronic laboratory notebooks (ELN), (ii) the featurizations using a combination of RDKit and electronic structure calculations in automated workflows and (iii) use of these data sets to predict the reaction outcome. The capabilities and pitfalls of these approaches were investigated in the context of a series of pericyclic reactions where the rules that determine reaction outcome are well understood. Progress towards ML models that learn these rules from physically meaningful features and predict the reaction outcomes will be discussed.<br/>In the second part of the presentation, the application of ML models, including ML surrogate models, to predict the reaction outcome of a series of transition metal catalyzed reaction will presented. The importance use training the models on the data suitable for the problem at hand, as opposed to large but generalized datasets, is discussed.

Machine learning in organic synthesis: Opportunities and challenges

Fall 2022 abstracts, posters, and presentations will launch in mid-September.

ACS Fall 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.