English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

In this paper, we analyze batch normalization from the perspective ofdiscriminability and find the disadvantages ignored by previous studies: thedifference in $l_2$ norms of sample features can hinder batch normalizationfrom obtaining more distinguished inter-class features and more compactintra-class features. To address this issue, we propose a simple yet effectivemethod to equalize the $l_2$ norms of sample features. Concretely, we$l_2$-normalize each sample feature before feeding them into batchnormalization, and therefore the features are of the same magnitude. Since theproposed method combines the $l_2$ normalization and batch normalization, wename our method $L_2$BN. The $L_2$BN can strengthen the compactness ofintra-class features and enlarge the discrepancy of inter-class features. The$L_2$BN is easy to implement and can exert its effect without any additionalparameters or hyper-parameters. We evaluate the effectiveness of $L_2$BNthrough extensive experiments with various models on image classification andacoustic scene classification tasks. The results demonstrate that the $L_2$BNcan boost the generalization ability of various neural network models andachieve considerable performance improvements.

$L_2$BN: Enhancing Batch Normalization by Equalizing the $L_2$ Norms of Features