Thirty-Second Annual Robert J. Melosh Medal Competition: Deep Adversarial Learning for Uncertain Inverse Problems