Learnable Loss Geometries with Mirror Descent for Scalable and Convergent Meta-Learning

Open in new window