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 prevalence of Internet of things (IoT) devices and abundance of sensordata has created an increase in real-time data processing such as recognitionof speech, image, and video. While currently such processes are offloaded tothe computationally powerful cloud system, a localized and distributed approachis desirable because (i) it preserves the privacy of users and (ii) it omitsthe dependency on cloud services. However, IoT networks are usually composed ofresource-constrained devices, and a single device is not powerful enough toprocess real-time data. To overcome this challenge, we examine data and modelparallelism for such devices in the context of deep neural networks. We proposeMusical Chair to enable efficient, localized, and dynamic real-time recognitionby harvesting the aggregated computational power from the resource-constraineddevices in the same IoT network as input sensors. Musical chair adapts to theavailability of computing devices at runtime and adjusts to the inheritdynamics of IoT networks. To demonstrate Musical Chair, on a network ofRaspberry PIs (up to 12) each connected to a camera, we implement astate-of-the-art action recognition model for videos and two recognition modelsfor images. Compared to the Tegra TX2, an embedded low-power platform with asix-core CPU and a GPU, our distributed action recognition system achieves notonly similar energy consumption but also twice the performance of the TX2.Furthermore, in image recognition, Musical Chair achieves similar performanceand saves dynamic energy.

Musical Chair: Efficient Real-Time Recognition Using Collaborative IoT Devices