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 Open

Presents the pdf analysis as premium feature

PDF Analysis

Insights

Presents the summarisation as premium feature

Summarisation

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

With the popularity of Internet of Things (IoT), edge computing and cloudcomputing, more and more stream analytics applications are being developedincluding real-time trend prediction and object detection on top of IoT sensingdata. One popular type of stream analytics is the recurrent neural network(RNN) deep learning model based time series or sequence data prediction andforecasting. Different from traditional analytics that assumes data areavailable ahead of time and will not change, stream analytics deals with datathat are being generated continuously and data trend/distribution could change(a.k.a. concept drift), which will cause prediction/forecasting accuracy todrop over time. One other challenge is to find the best resource provisioningfor stream analytics to achieve good overall latency. In this paper, we studyhow to best leverage edge and cloud resources to achieve better accuracy andlatency for stream analytics using a type of RNN model called long short-termmemory (LSTM). We propose a novel edge-cloud integrated framework for hybridstream analytics that supports low latency inference on the edge and highcapacity training on the cloud. To achieve flexible deployment, we studydifferent approaches of deploying our hybrid learning framework includingedge-centric, cloud-centric and edge-cloud integrated. Further, our hybridlearning framework can dynamically combine inference results from an LSTM modelpre-trained based on historical data and another LSTM model re-trainedperiodically based on the most recent data. Using real-world and simulatedstream datasets, our experiments show the proposed edge-cloud deployment is thebest among all three deployment types in terms of latency. For accuracy, theexperiments show our dynamic learning approach performs the best among alllearning approaches for all three concept drift scenarios.

An Edge-Cloud Integrated Framework for Flexible and Dynamic Stream Analytics