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

Deep learning models have introduced various intelligent applications to edgedevices, such as image classification, speech recognition, and augmentedreality. There is an increasing need of training such models on the devices inorder to deliver personalized, responsive, and private learning. To addressthis need, this paper presents a new solution for deploying and trainingstate-of-the-art models on the resource-constrained devices. First, the paperproposes a novel filter-pruning-based model compression method to createlightweight trainable models from large models trained in the cloud, withoutmuch loss of accuracy. Second, it proposes a novel knowledge transfer method toenable the on-device model to update incrementally in real time or near realtime using incremental learning on new data and enable the on-device model tolearn the unseen categories with the help of the in-cloud model in anunsupervised fashion. The results show that 1) our model compression method canremove up to 99.36% parameters of WRN-28-10, while preserving a Top-1 accuracyof over 90% on CIFAR-10; 2) our knowledge transfer method enables thecompressed models to achieve more than 90% accuracy on CIFAR-10 and retain goodaccuracy on old categories; 3) it allows the compressed models to convergewithin real time (three to six minutes) on the edge for incremental learningtasks; 4) it enables the model to classify unseen categories of data (78.92%Top-1 accuracy) that it is never trained with.

Enabling Deep Learning on Edge Devices through Filter Pruning and Knowledge Transfer