Meta-Learning with Versatile Loss Geometries for Fast Adaptation Using Mirror Descent

Open in new window