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

Deep neural network (DNN) latency characterization is a time-consumingprocess and adds significant cost to Neural Architecture Search (NAS) processeswhen searching for efficient convolutional neural networks for embedded visionapplications. DNN Latency is a hardware dependent metric and requires directmeasurement or inference on target hardware. A recently introduced latencyestimation technique known as MAPLE predicts DNN execution time on previouslyunseen hardware devices by using hardware performance counters. Leveragingthese hardware counters in the form of an implicit prior, MAPLE achievesstate-of-the-art performance in latency prediction. Here, we propose MAPLE-Xwhich extends MAPLE by incorporating explicit prior knowledge of hardwaredevices and DNN architecture latency to better account for model stability androbustness. First, by identifying DNN architectures that exhibit a similarlatency to each other, we can generate multiple virtual examples tosignificantly improve the accuracy over MAPLE. Secondly, the hardwarespecifications are used to determine the similarity between training and testhardware to emphasize training samples captured from comparable devices(domains) and encourages improved domain alignment. Experimental results usinga convolution neural network NAS benchmark across different types of devices,including an Intel processor that is now used for embedded vision applications,demonstrate a 5% improvement over MAPLE and 9% over HELP. Furthermore, weinclude ablation studies to independently assess the benefits of virtualexamples and hardware-based sample importance.

MAPLE-X: Latency Prediction with Explicit Microprocessor Prior Knowledge