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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Although the convolutional neural networks (CNNs) have become popular forvarious image processing and computer vision task recently, it remains achallenging problem to reduce the storage cost of the parameters forresource-limited platforms. In the previous studies, tensor decomposition (TD)has achieved promising compression performance by embedding the kernel of aconvolutional layer into a low-rank subspace. However the employment of TD isnaively on the kernel or its specified variants. Unlike the conventionalapproaches, this paper shows that the kernel can be embedded into more generalor even random low-rank subspaces. We demonstrate this by compressing theconvolutional layers via randomly-shuffled tensor decomposition (RsTD) for astandard classification task using CIFAR-10. In addition, we analyze how thespatial similarity of the training data influences the low-rank structure ofthe kernels. The experimental results show that the CNN can be significantlycompressed even if the kernels are randomly shuffled. Furthermore, theRsTD-based method yields more stable classification accuracy than theconventional TD-based methods in a large range of compression ratios.

Low-Rank Embedding of Kernels in Convolutional Neural Networks under Random Shuffling