Over the last decade, machine learning has revolutionized how we model materials at the atomic scale: With the help of powerful algorithms, researchers can quickly model new materials or simulate interactions that would have taken months or even years of work using previous techniques.
Even though machine learning approaches at the atomic level have matured over the past few decades, that isn’t true at larger scales; the same caliber of tools and techniques are not available for larger materials building blocks, like colloids, or large molecules held in suspension. “Working at both the nanoscale and the atomistic scale has shown me the disparity between machine learning applications and methodologies between the two scales,” says Rose Cersonsky, who started as an assistant professor in the Department of Chemical and Biological Engineering in January 2023. “What my lab is going to focus on first and foremost is learning what can we apply from the atomistic community to accelerate our understanding and our abilities at the colloidal and nanoscale.”
Cersonsky grew up in Connecticut and received a bachelor’s degree in materials science and a minor in computer science from the University of Connecticut. She earned her PhD at the University of Michigan studying macromolecular science and engineering. There, Cersonsky earned renown for her work studying photonic structures, crystallographic patterns responsible for the beautiful colors in chameleons and many species of butterflies that can be targeted with nanoparticle self-assembly. Her thesis, titled “Designing Nanoparticles for the Self-Assembly of Novel Materials,” won numerous top University of Michigan distinctions and the Victor K. LaMer Award from the American Chemical Society Colloids Division.
After her PhD, she joined the Laboratory of Computational Science and Modeling at the École Polytechnique Fédérale de Lausanne, Switzerland, as a postdoctoral researcher, working on machine learning methods for atomistic simulations for the last three years. During her postdoctoral work, Cersonsky developed and implemented multiple machine learning methodologies and applied them to further the understanding of molecular crystallization, an important process in innumerable applications, including drug stabilization and biological processes.
In Madison, she hopes to bring all of her research interests together and investigate new directions. “Colloids and nanoparticles are not the same as atoms and often do not behave like atoms,” she says. “We will pursue how can we apply a mathematically rigorous machine learning infrastructure to large particles in a way that helps us to understand them better and to design them better. My group will first focus on figuring out how we represent shapes and particle design in these models.”
Cersonsky says her work, which is primarily computational, will illuminate the properties of many macromolecules, including polymers (extremely long, complex molecular chains that make up things like plastic) and biomolecules. She believes this work will complement research on plastic sustainability and nanoscale drug delivery systems by other faculty in CBE. “We are in a climate crisis and we really need to think about the ways in which we produce and use materials,” she says.
Cersonsky is also excited about other types of collaboration. At Connecticut and Michigan, she led public outreach programs to foster inclusivity in science for grade school, high school and college students. She hopes to continue those efforts at UW-Madison, and also hopes to get back into her personal hobby: musical theater. “I love the intricacy and the cooperation of a music production. It just takes so many pieces,” she says. When you have an amazing set design and cast and music director and everything works together … it is almost transcendental.”