# Switchable Normalization for Learning-to-Normalize Deep Representation

arXiv: https://arxiv.org/abs/1907.10473 · HTML: https://mlsanity.com/paper/1907.10473/

## Key facts

- Cited by 16 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 22 Jul 2019.

## Built on by (most cited first, 16 of 16)

- [Rebalancing Batch Normalization for Exemplar-based Class-Incremental Learning](https://mlsanity.com/paper/2201.12559/) (2022)
- [Dynamic Token Normalization Improves Vision Transformers](https://mlsanity.com/paper/2112.02624/) (2021)
- [Logit Normalization for Long-tail Object Detection](https://mlsanity.com/paper/2203.17020/) (2022)
- [Normalization Perturbation: A Simple Domain Generalization Method for Real-World Domain Shifts](https://mlsanity.com/paper/2211.04393/) (2022)
- [Towards Generalizable Person Re-identification with a Bi-stream Generative Model](https://mlsanity.com/paper/2206.09362/) (2022)
- [To Fold or Not to Fold: a Necessary and Sufficient Condition on Batch-Normalization Layers Folding](https://mlsanity.com/paper/2203.14646/) (2022)
- [An Open-World, Diverse, Cross-Spatial-Temporal Benchmark for Dynamic Wild Person Re-Identification](https://mlsanity.com/paper/2403.15119/) (2024)
- [An Enhanced Harmonic Densely Connected Hybrid Transformer Network Architecture for Chronic Wound Segmentation Utilising Multi-Colour Space Tensor Merging](https://mlsanity.com/paper/2410.03359/) (2024)
- [Systematic Analysis and Removal of Circular Artifacts for StyleGAN](https://mlsanity.com/paper/2103.01090/) (2021)
- [Federated Learning for Surgical Vision in Appendicitis Classification: Results of the FedSurg EndoVis 2024 Challenge](https://mlsanity.com/paper/2510.04772/) (2025)
- [SIRe-Networks: Convolutional Neural Networks Architectural Extension for Information Preservation via Skip/Residual Connections and Interlaced Auto-Encoders](https://mlsanity.com/paper/2110.02776/) (2021)
- [A simple normalization technique using window statistics to improve the out-of-distribution generalization on medical images](https://mlsanity.com/paper/2207.03366/) (2022)
- [KL Regularized Normalization Framework for Low Resource Tasks](https://mlsanity.com/paper/2212.11275/) (2022)
- [LayerAct: Advanced Activation Mechanism for Robust Inference of CNNs](https://mlsanity.com/paper/2306.04940/) (2023)
- [Enhancing Neural Network Representations with Prior Knowledge-Based Normalization](https://mlsanity.com/paper/2403.16798/) (2024)
- [Exploring the Efficacy of Group-Normalization in Deep Learning Models for Alzheimer's Disease Classification](https://mlsanity.com/paper/2404.00946/) (2024)

## Abstract (arXiv)

We address a learning-to-normalize problem by proposing Switchable Normalization (SN), which learns to select different normalizers for different normalization layers of a deep neural network. SN employs three distinct scopes to compute statistics (means and variances) including a channel, a layer, and a minibatch. SN switches between them by learning their importance weights in an end-to-end manner. It has several good properties. First, it adapts to various network architectures and tasks. Second, it is robust to a wide range of batch sizes, maintaining high performance even when small minibatch is presented (e.g. 2 images/GPU). Third, SN does not have sensitive hyper-parameter, unlike group normalization that searches the number of groups as a hyper-parameter. Without bells and whistles, SN outperforms its counterparts on various challenging benchmarks, such as ImageNet, COCO, CityScapes, ADE20K, MegaFace, and Kinetics. Analyses of SN are also presented to answer the following three questions: (a) Is it useful to allow each normalization layer to select its own normalizer? (b) What impacts the choices of normalizers? (c) Do different tasks and datasets prefer different normalizers? We hope SN will help ease the usage and understand the normalization techniques in deep learning. The code of SN has been released at https://github.com/switchablenorms.