Towards sustainable chemical and biocatalytic synthesis through machine learning-supported experimental decision-making

Machine learning (ML) has gained considerable attention across scientific fields, including organic synthesis and biocatalysis. With the potential to achieve more resource-efficient and sustainable reaction development through reducing experimental burden, the motivation to embrace ML to support decision making is high. To enhance ML’s utility, the dataset and ML approach need to be designed to align well with the goal of a given experiment. This article reviews recent advances in ML for small molecule methods and biocatalysis by categorizing the types of problems in organic synthesis. By comparing the forefront of both fields, we identify opportunities, such as dataset design strategies, as well as research directions to further facilitate the incorporation of ML in both small molecule reactions and biocatalysis.