A Closer Look at the Training Strategy for Modern Meta-Learning

From a stability perspective, we analyze the generalization error bound of generic meta-learning algorithms trained with such strategy.