TaskNorm: Rethinking Batch Normalization for Meta-Learning

John Bronskill, Jonathan Gordon, James Requeima, Sebastian Nowozin, Richard Turner
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:1153-1164, 2020.

Abstract

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.

Cite this Paper


BibTeX
@InProceedings{pmlr-v119-bronskill20a, title = {{T}ask{N}orm: Rethinking Batch Normalization for Meta-Learning}, author = {Bronskill, John and Gordon, Jonathan and Requeima, James and Nowozin, Sebastian and Turner, Richard}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {1153--1164}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/bronskill20a/bronskill20a.pdf}, url = {http://proceedings.mlr.press/v119/bronskill20a.html}, abstract = {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.} }
Endnote
%0 Conference Paper %T TaskNorm: Rethinking Batch Normalization for Meta-Learning %A John Bronskill %A Jonathan Gordon %A James Requeima %A Sebastian Nowozin %A Richard Turner %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-bronskill20a %I PMLR %P 1153--1164 %U http://proceedings.mlr.press/v119/bronskill20a.html %V 119 %X 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.
APA
Bronskill, J., Gordon, J., Requeima, J., Nowozin, S. & Turner, R.. (2020). TaskNorm: Rethinking Batch Normalization for Meta-Learning. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:1153-1164 Available from http://proceedings.mlr.press/v119/bronskill20a.html.

Related Material