Machine Learning for Molecules

The research in the Eli Weinstein Group bridges fundamental machine learning methodology with applications to molecular science. The group develops probabilistic approaches that enable predictive understanding of molecular properties and guide the design of new molecules with desired properties. A central aim is to advance both the theoretical foundations of machine learning and its practical use for molecular systems.

Research

Our activities focus on two main research lines. The first is the co-design of experiments and inference algorithms, where new ML methods are created to actively steer chemical synthesis, optimize high-throughput screening, and improve model training. This tight integration of experimental data generation and algorithmic development accelerates molecular discovery.

The second line of research explores learning from “natural experiments” that occur outside the laboratory, such as large-scale evolutionary processes and patient-derived data. Here, we develop ML techniques that can extract patterns from complex, heterogeneous datasets to reveal functional relationships at the molecular level.

Together, these activities provide both methodological innovations and practical tools for understanding and designing molecules, contributing to a deeper theoretical and applied foundation for machine learning in molecular systems.

Contact

Eli Nathan Weinstein

Eli Nathan Weinstein Assistant Professor

Vacancies

Ongoing research projects involve mechanistic studies to uncover the origins of selectivity in CO₂ hydrogenation, as well as the development of high-entropy oxide catalysts for the reverse water-gas shift reaction.

Our group is also involved in Reformable, a pre-commercial DTU startup that deals with catalytic dry reforming of methane and CO2.

Group members