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

L2 regularization for weights in neural networks is widely used as a standardtraining trick. However, L2 regularization for gamma, a trainable parameter ofbatch normalization, remains an undiscussed mystery and is applied in differentways depending on the library and practitioner. In this paper, we study whetherL2 regularization for gamma is valid. To explore this issue, we consider twoapproaches: 1) variance control to make the residual network behave likeidentity mapping and 2) stable optimization through the improvement ofeffective learning rate. Through two analyses, we specify the desirable andundesirable gamma to apply L2 regularization and propose four guidelines formanaging them. In several experiments, we observed the increase and decrease inperformance caused by applying L2 regularization to gamma of four categories,which is consistent with our four guidelines. Our proposed guidelines werevalidated through various tasks and architectures, including variants ofresidual networks and transformers.

Guidelines for the Regularization of Gammas in Batch Normalization for Deep Residual Networks