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

Regularization is crucial to the success of many practical deep learningmodels, in particular in a more often than not scenario where there are only afew to a moderate number of accessible training samples. In addition to weightdecay, data augmentation and dropout, regularization based on multi-brancharchitectures, such as Shake-Shake regularization, has been proven successfulin many applications and attracted more and more attention. However, beyondmodel-based representation augmentation, it is unclear how Shake-Shakeregularization helps to provide further improvement on classification tasks,let alone the baffling interaction between batch normalization and shaking. Inthis work, we present our investigation on Shake-Shake regularization, drawingconnections to the vicinal risk minimization principle and discriminativefeature learning in verification tasks. Furthermore, we identify a strongresemblance between batch normalized residual blocks and batch normalizedrecurrent neural networks, where both of them share a similar convergencebehavior, which could be mitigated by a proper initialization of batchnormalization. Based on the findings, our experiments on speech emotionrecognition demonstrate simultaneously an improvement on the classificationaccuracy and a reduction on the generalization gap both with statisticalsignificance.

Normalization Before Shaking Toward Learning Symmetrically Distributed Representation Without Margin in Speech Emotion Recognition