Our research sits at the intersection of machine learning, data visualization, and computational biology

We develop new machine-learning methods for understanding complex scientific datasets and apply them to biological problems in single-cell genomics and cellular imaging. A central goal of the lab is to build representations of high-dimensional data that allow researchers to discover patterns in large scientific datasets. We are also interested in making modern cellular, genomic, and protein foundation models more interpretable and transparent and using them to derive scientific insights. Some of the computational paradigms that we use and study are manifold learning, contrastive learning, self-supervised learning, and representation learning.