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

Despite the recent success of deep learning models in numerous applications,their widespread use on mobile devices is seriously impeded by storage andcomputational requirements. In this paper, we propose a novel networkcompression method called Adaptive Dimension Adjustment Tucker decomposition(ADA-Tucker). With learnable core tensors and transformation matrices,ADA-Tucker performs Tucker decomposition of arbitrary-order tensors.Furthermore, we propose that weight tensors in networks with proper order andbalanced dimension are easier to be compressed. Therefore, the high flexibilityin decomposition choice distinguishes ADA-Tucker from all previous low-rankmodels. To compress more, we further extend the model to Shared Core ADA-Tucker(SCADA-Tucker) by defining a shared core tensor for all layers. Our methodsrequire no overhead of recording indices of non-zero elements. Without loss ofaccuracy, our methods reduce the storage of LeNet-5 and LeNet-300 by ratios of691 times and 233 times, respectively, significantly outperforming state of theart. The effectiveness of our methods is also evaluated on other threebenchmarks (CIFAR-10, SVHN, ILSVRC12) and modern newly deep networks (ResNet,Wide-ResNet).

ADA-Tucker: Compressing Deep Neural Networks via Adaptive Dimension Adjustment Tucker Decomposition