MimicNorm: Weight Mean and Last BN Layer Mimic the Dynamic of Batch Normalization

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Fei, Wen, Dai, Wenrui, Li, Chenglin, Zou, Junni, Xiong, Hongkai
Natura: Preprint
Pubblicazione: 2020
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913566418796544
author Fei, Wen
Dai, Wenrui
Li, Chenglin
Zou, Junni
Xiong, Hongkai
author_facet Fei, Wen
Dai, Wenrui
Li, Chenglin
Zou, Junni
Xiong, Hongkai
contents Substantial experiments have validated the success of Batch Normalization (BN) Layer in benefiting convergence and generalization. However, BN requires extra memory and float-point calculation. Moreover, BN would be inaccurate on micro-batch, as it depends on batch statistics. In this paper, we address these problems by simplifying BN regularization while keeping two fundamental impacts of BN layers, i.e., data decorrelation and adaptive learning rate. We propose a novel normalization method, named MimicNorm, to improve the convergence and efficiency in network training. MimicNorm consists of only two light operations, including modified weight mean operations (subtract mean values from weight parameter tensor) and one BN layer before loss function (last BN layer). We leverage the neural tangent kernel (NTK) theory to prove that our weight mean operation whitens activations and transits network into the chaotic regime like BN layer, and consequently, leads to an enhanced convergence. The last BN layer provides autotuned learning rates and also improves accuracy. Experimental results show that MimicNorm achieves similar accuracy for various network structures, including ResNets and lightweight networks like ShuffleNet, with a reduction of about 20% memory consumption. The code is publicly available at https://github.com/Kid-key/MimicNorm.
format Preprint
id arxiv_https___arxiv_org_abs_2010_09278
institution arXiv
publishDate 2020
record_format arxiv
spellingShingle MimicNorm: Weight Mean and Last BN Layer Mimic the Dynamic of Batch Normalization
Fei, Wen
Dai, Wenrui
Li, Chenglin
Zou, Junni
Xiong, Hongkai
Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
Substantial experiments have validated the success of Batch Normalization (BN) Layer in benefiting convergence and generalization. However, BN requires extra memory and float-point calculation. Moreover, BN would be inaccurate on micro-batch, as it depends on batch statistics. In this paper, we address these problems by simplifying BN regularization while keeping two fundamental impacts of BN layers, i.e., data decorrelation and adaptive learning rate. We propose a novel normalization method, named MimicNorm, to improve the convergence and efficiency in network training. MimicNorm consists of only two light operations, including modified weight mean operations (subtract mean values from weight parameter tensor) and one BN layer before loss function (last BN layer). We leverage the neural tangent kernel (NTK) theory to prove that our weight mean operation whitens activations and transits network into the chaotic regime like BN layer, and consequently, leads to an enhanced convergence. The last BN layer provides autotuned learning rates and also improves accuracy. Experimental results show that MimicNorm achieves similar accuracy for various network structures, including ResNets and lightweight networks like ShuffleNet, with a reduction of about 20% memory consumption. The code is publicly available at https://github.com/Kid-key/MimicNorm.
title MimicNorm: Weight Mean and Last BN Layer Mimic the Dynamic of Batch Normalization
topic Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2010.09278