Geometric Deep Learning at the 2024 Dusty Plasma Workshop

PhD Candidate Matt Raymond gives a talk and a poster presentation (not shown) on his work at the 2024 Dusty Plasma Workshop in Minneapolis, MN. Matt described his use of Geometric Deep Learning (GDL) to predict nanoparticle growth in nonthermal plasma. This presentation demonstrated that combining GDL with molecular dynamics simulations can save ~50% of the computational cost of nonthermal plasma simulations, and that GDL models can make accurate predictions for nanostructures that were unseen during training. The abstract can be seen here.

This work is supported by Army Research Office MURI Grant W911NF-18-1-0240 (PI Prof. Kortshagen)