A core aspect of intelligence is the ability to quickly learn new tasks by drawing upon prior experience from related tasks. Recent work has studied how meta-learning algorithms [51, 55, 41] can acquire such a capability by learning to efficiently learn a range of tasks, thereby enabling learning of a new task with as little as a single example [50, 57, 15].

Meta learning has demonstrated tremendous success in few-shot learning with limited supervised data. In those settings, the meta model is usually overparameterized.

Meta learning has demonstrated tremendous success in few-shot learning with limited supervised data. In those settings, the meta model is usually overparameterized.

This appendix contains the supplementary material for the main text. In Appendix B, we provide details of the derivation of the implicit gradients in Eqs. We first establish here this result. This negative result illustrates the need for alternative estimators. In Appendix A.2.3, we analyze the behavior of Lemma 1 Alternatively, we can follow the approach described in Section 4.2 to estimate both From Eq. (10), we follow the same derivation as Eq.

Neural Information Processing Systems http://nips.cc/

However, one major contribution is that Table 1 can be extended with more columns (e.g. one more