Deep Mountain Glaciers

Schematic of a Physics Informed Neural Network (PINN) taken from L.Lu et. al (2019)

Course advisor - Dr. Rachel Ward

Based on recent developments in Physics-Informed Machine Learing, we attempt to use a neural network for the following use cases:

• Emulate a highly non-linear, diffusive mountain glacier model
• Infer the coefficients of the mountain glacier model from artificially generated data

The code and the presentation can be found here.