The Peculiar Optimization and Regularization Challenges...Meta-Learning - Chelsea Finn