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

This paper considers a setting where embedded devices are used to acquire andclassify images. Because of limited computing capacity, embedded devices relyon a parsimonious classification model with uneven accuracy. When localclassification is deemed inaccurate, devices can decide to offload the image toan edge server with a more accurate but resource-intensive model. Resourceconstraints, e.g., network bandwidth, however, require regulating suchtransmissions to avoid congestion and high latency. The paper investigates thisoffloading problem when transmissions regulation is through a token bucket, amechanism commonly used for such purposes. The goal is to devise a lightweight,online offloading policy that optimizes an application-specific metric (e.g.,classification accuracy) under the constraints of the token bucket. The paperdevelops a policy based on a Deep Q-Network (DQN), and demonstrates both itsefficacy and the feasibility of its deployment on embedded devices. Of note isthe fact that the policy can handle complex input patterns, includingcorrelation in image arrivals and classification accuracy. The evaluation iscarried out by performing image classification over a local testbed usingsynthetic traces generated from the ImageNet image classification benchmark.Implementation of this work is available athttps://github.com/qiujiaming315/edgeml-dqn.

Adaptive Edge Offloading for Image Classification Under Rate Limit