- Keywords: ES, MAML, evolution, strategies, meta, learning, gaussian, perturbation, reinforcement, learning, adaptation
- TL;DR: We provide a new framework for MAML in the ES/blackbox setting, and show that it allows deterministic and linear policies, better exploration, and non-differentiable adaptation operators.
- Abstract: We introduce ES-MAML, a new framework for solving the model agnostic meta learning (MAML) problem based on Evolution Strategies (ES). Existing algorithms for MAML are based on policy gradients, and incur significant difficulties when attempting to estimate second derivatives using backpropagation on stochastic policies. We show how ES can be applied to MAML to obtain an algorithm which avoids the problem of estimating second derivatives, and is also conceptually simple and easy to implement. Moreover, ES-MAML can handle new types of nonsmooth adaptation operators, and other techniques for improving performance and estimation of ES methods become applicable. We show empirically that ES-MAML is competitive with existing methods and often yields better adaptation with fewer queries.
- Original Pdf: pdf