October 29, 2021

About the speaker

Dr. Craig Gin is a Postdoctoral Research Scholar from North Carolina State University. He obtained his PhD in mathematics at Texas A&M University under Prabir Daripa. He studied the stability of multi-layer Hele-Shaw and porous media flows. He completed a postdoc in the applied math department at the University of Washington. He is currently doing another postdoc in the department of population health and pathobiology at North Carolina State University.

Science-informed Mechanic Learning

Machine learning is becoming ubiquitous in the sciences. Incorporating scientific domain knowledge into machine learning algorithms can significantly increase their effectiveness. I will present two examples to demonstrate how scientific domain-specific knowledge can inform the choices we make when devising machine learning strategies. The first example uses deep learning to find coordinate transformations that linearize nonlinear partial differential equations. The second example harnesses properties of microbiome data in order to better use the microbiome to predict disease states.