Debdeep Pati

My research involves developing Bayesian methods for complex objects including high-dimensional sparse vectors, matrices, shapes of non-Euclidean objects and large graphs. I am also interested in studying Bayesian model selection consistency under complex settings. Modeling the distributions of objects contained within images motivated some of my collaborative work, e.g., in applications of tumor tracking in targeted radiation therapy. More recently, I have become interested in building models for discovering patterns in large networks and to predict cognition from connectomics data.