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

Cloud-edge collaborative inference approach splits deep neural networks(DNNs) into two parts that run collaboratively on resource-constrained edgedevices and cloud servers, aiming at minimizing inference latency andprotecting data privacy. However, even if the raw input data from edge devicesis not directly exposed to the cloud, state-of-the-art attacks targetingcollaborative inference are still able to reconstruct the raw private data fromthe intermediate outputs of the exposed local models, introducing seriousprivacy risks. In this paper, a secure privacy inference framework forcloud-edge collaboration is proposed, termed CIS, which supports adaptivelypartitioning the network according to the dynamically changing networkbandwidth and fully releases the computational power of edge devices. Tomitigate the influence introduced by private perturbation, CIS provides a wayto achieve differential privacy protection by adding refined noise to theintermediate layer feature maps offloaded to the cloud. Meanwhile, with a giventotal privacy budget, the budget is reasonably allocated by the size of thefeature graph rank generated by different convolution filters, which makes theinference in the cloud robust to the perturbed data, thus effectively trade-offthe conflicting problem between privacy and availability. Finally, we constructa real cloud-edge collaborative inference computing scenario to verify theeffectiveness of inference latency and model partitioning onresource-constrained edge devices. Furthermore, the state-of-the-art cloud-edgecollaborative reconstruction attack is used to evaluate the practicalavailability of the end-to-end privacy protection mechanism provided by CIS.

Privacy-preserving Security Inference Towards Cloud-Edge Collaborative Using Differential Privacy