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/

Zendy Plus

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

Zendy Tools

Zendy Open

The increasing demand for on-device deep learning services calls for a highlyefficient manner to deploy deep neural networks (DNNs) on mobile devices withlimited capacity. The cloud-based solution is a promising approach to enablingdeep learning applications on mobile devices where the large portions of a DNNare offloaded to the cloud. However, revealing data to the cloud leads topotential privacy risk. To benefit from the cloud data center without theprivacy risk, we design, evaluate, and implement a cloud-based framework ARDENwhich partitions the DNN across mobile devices and cloud data centers. A simpledata transformation is performed on the mobile device, while theresource-hungry training and the complex inference rely on the cloud datacenter. To protect the sensitive information, a lightweight privacy-preservingmechanism consisting of arbitrary data nullification and random noise additionis introduced, which provides strong privacy guarantee. A rigorous privacybudget analysis is given. Nonetheless, the private perturbation to the originaldata inevitably has a negative impact on the performance of further inferenceon the cloud side. To mitigate this influence, we propose a noisy trainingmethod to enhance the cloud-side network robustness to perturbed data. Throughthe sophisticated design, ARDEN can not only preserve privacy but also improvethe inference performance. To validate the proposed ARDEN, a series ofexperiments based on three image datasets and a real mobile application areconducted. The experimental results demonstrate the effectiveness of ARDEN.Finally, we implement ARDEN on a demo system to verify its practicality.

Not Just Privacy: Improving Performance of Private Deep Learning in Mobile Cloud