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

The current trend of pushing CNNs deeper with convolutions has created apressing demand to achieve higher compression gains on CNNs where convolutionsdominate the computation and parameter amount (e.g., GoogLeNet, ResNet and WideResNet). Further, the high energy consumption of convolutions limits itsdeployment on mobile devices. To this end, we proposed a simple yet effectivescheme for compressing convolutions though applying k-means clustering on theweights, compression is achieved through weight-sharing, by only recording $K$cluster centers and weight assignment indexes. We then introduced a novelspectrally relaxed $k$-means regularization, which tends to make hardassignments of convolutional layer weights to $K$ learned cluster centersduring re-training. We additionally propose an improved set of metrics toestimate energy consumption of CNN hardware implementations, whose estimationresults are verified to be consistent with previously proposed energyestimation tool extrapolated from actual hardware measurements. We finallyevaluated Deep $k$-Means across several CNN models in terms of both compressionratio and energy consumption reduction, observing promising results withoutincurring accuracy loss. The code is available athttps://github.com/Sandbox3aster/Deep-K-Means

Deep $k$-Means: Re-Training and Parameter Sharing with Harder Cluster Assignments for Compressing Deep Convolutions