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

Collaboration among industrial Internet of Things (IoT) devices and edgenetworks is essential to support computation-intensive deep neural network(DNN) inference services which require low delay and high accuracy. Samplingrate adaption which dynamically configures the sampling rates of industrial IoTdevices according to network conditions, is the key in minimizing the servicedelay. In this paper, we investigate the collaborative DNN inference problem inindustrial IoT networks. To capture the channel variation and task arrivalrandomness, we formulate the problem as a constrained Markov decision process(CMDP). Specifically, sampling rate adaption, inference task offloading andedge computing resource allocation are jointly considered to minimize theaverage service delay while guaranteeing the long-term accuracy requirements ofdifferent inference services. Since CMDP cannot be directly solved by generalreinforcement learning (RL) algorithms due to the intractable long-termconstraints, we first transform the CMDP into an MDP by leveraging the Lyapunovoptimization technique. Then, a deep RL-based algorithm is proposed to solvethe MDP. To expedite the training process, an optimization subroutine isembedded in the proposed algorithm to directly obtain the optimal edgecomputing resource allocation. Extensive simulation results are provided todemonstrate that the proposed RL-based algorithm can significantly reduce theaverage service delay while preserving long-term inference accuracy with a highprobability.

Accuracy-Guaranteed Collaborative DNN Inference in Industrial IoT via Deep Reinforcement Learning