Meta Learning via Learned Loss

Sarah Bechtle, Artem Molchanov, Yevgen Chebotar, Edward Grefenstette, Ludovic Righetti, Gaurav Sukhatme, Franziska Meier

TL;DR: Meta learning a loss function for optimisation for supervised learning tasks as well as reinforcement learning

Abstract: We present a meta-learning method for learning parametric loss functions that can generalize across different tasks and model architectures. We develop a pipeline for training such loss functions, targeted at maximizing the performance of model learn- ing with them. We observe that the loss landscape produced by our learned losses significantly improves upon the original task-specific losses in both supervised and reinforcement learning tasks. Furthermore, we show that our meta-learning framework is flexible enough to incorporate additional information at meta-train time. This information shapes the learned loss function such that the environment does not need to provide this information during meta-test time.

Keywords: Meta Learning, Reinforcement Learning, Loss Learning