Inferring Constitutive Models of Antarctic Glacial Ice via Physics-Informed Deep Learning

Explore a conference talk that presents cutting-edge research on understanding Antarctic glacial ice behavior through advanced computational methods. Learn how physics-informed deep learning techniques can be applied to infer constitutive models that describe the mechanical properties and flow behavior of glacial ice in Antarctica. Discover the intersection of machine learning, glaciology, and climate science as the speaker demonstrates how these innovative approaches can improve our understanding of ice sheet dynamics and their response to warming conditions. Gain insights into the mathematical frameworks and computational strategies used to model complex ice behavior, and understand how these methods contribute to better predictions of ice sheet evolution in a changing climate. The presentation addresses key challenges in polar climate research and showcases how interdisciplinary approaches combining physics-based modeling with machine learning can advance our knowledge of critical climate system components.