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

Recent breakthroughs in Deep Neural Networks (DNNs) have fueled atremendously growing demand for bringing DNN-powered intelligence into mobileplatforms. While the potential of deploying DNNs on resource-constrainedplatforms has been demonstrated by DNN compression techniques, the currentpractice suffers from two limitations: 1) merely stand-alone compressionschemes are investigated even though each compression technique only suit forcertain types of DNN layers; and 2) mostly compression techniques are optimizedfor DNNs' inference accuracy, without explicitly considering otherapplication-driven system performance (e.g., latency and energy cost) and thevarying resource availability across platforms (e.g., storage and processingcapability). To this end, we propose AdaDeep, a usage-driven, automated DNNcompression framework for systematically exploring the desired trade-offbetween performance and resource constraints, from a holistic system level.Specifically, in a layer-wise manner, AdaDeep automatically selects the mostsuitable combination of compression techniques and the correspondingcompression hyperparameters for a given DNN. Thorough evaluations on sixdatasets and across twelve devices demonstrate that AdaDeep can achieve up to$18.6\times$ latency reduction, $9.8\times$ energy-efficiency improvement, and$37.3\times$ storage reduction in DNNs while incurring negligible accuracyloss. Furthermore, AdaDeep also uncovers multiple novel combinations ofcompression techniques.

AdaDeep: A Usage-Driven, Automated Deep Model Compression Framework for Enabling Ubiquitous Intelligent Mobiles