Organic chemical reactions are widely employed in various scientific research and industrial processes. However, two critical yet underexplored aspects are their reproducibility and scalability. Reproducibility refers to the sensitivity of chemical reactions to general key conditions, while scalability pertains to the sensitivity of chemical reactions to the key parameters of the reactor. Both factors are vital for the progression and practical application of organic reactions. However, comprehensive studies in this area are limited to date. This research seeks to fill this knowledge gap by rigorously investigating the reproducibility and scalability of organic chemical reactions. To solve this task we propose a novel machine-learning model for classification on imbalanced datasets.

SuperUROP is a fantastic opportunity to engage in research, which is something I started doing in my freshman year. Then, I became interested in using statistics to solve problems in drug discovery. I believe that theoretical and applied research are often closely related and through this project, I’d like to deepen my understanding of both. I’m looking forward to designing explainable machine-learning models to facilitate molecular discovery.