Hudson Smith

I am an Assistant Professor in the Department of Mathematics and Statistical Sciences at Clemson University. I develop techniques for incorporating prior knowledge into deep learning systems for data-constrained applications.

The synthesis of established domain knowledge with flexible learning-based methods leads to better anomaly detection rates, improved sample efficiency, and explainable inference, without sacrificing the expressive power of data-driven techniques like deep learning. See my publications

I have a PhD in theoretical atomic Physics from Ohio State university. As a physicist, I learned the power of approaching problems from first principles. As a machine learning scientist, I have seen the power of flexible, data-driven models. This dual experience allows me to integrate the strengths of both approaches. See my physics-based visualizations.