Inverse methods are commonly used to infer unobservable properties of ice sheets, and this talk presents the benefits and shortcomings of classical methods as well as future directions for addressing these shortcomings using Bayesian methods and machine learning.
