TaskNorm: Rethinking Batch Normalization for Meta-Learning

John Bronskill · Jonathan Gordon · James Requeima · Sebastian Nowozin · Richard E Turner


Keywords: [ Algorithms ] [ Meta-learning and Automated ML ] [ Transfer, Multitask and Meta-learning ]

[ Abstract ]
[ Slides
Tue 14 Jul noon PDT — 12:45 p.m. PDT
Wed 15 Jul 1 a.m. PDT — 1:45 a.m. PDT


Modern meta-learning approaches for image classification rely on increasingly deep networks to achieve state-of-the-art performance, making batch normalization an essential component of meta-learning pipelines. However, the hierarchical nature of the meta-learning setting presents several challenges that can render conventional batch normalization ineffective, giving rise to the need to rethink normalization in this setting. We evaluate a range of approaches to batch normalization for meta-learning scenarios, and develop a novel approach that we call TaskNorm. Experiments on fourteen datasets demonstrate that the choice of batch normalization has a dramatic effect on both classification accuracy and training time for both gradient based- and gradient-free meta-learning approaches. Importantly, TaskNorm is found to consistently improve performance. Finally, we provide a set of best practices for normalization that will allow fair comparison of meta-learning algorithms.

Chat is not available.