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 use of Deep Learning hardware algorithms for embedded applications ischaracterized by challenges such as constraints on device power consumption,availability of labeled data, and limited internet bandwidth for frequenttraining on cloud servers. To enable low power implementations, we considerefficient bitwidth reduction and pruning for the class of Deep Learningalgorithms known as Discriminative Deep Belief Networks (DDBNs) forembedded-device classification tasks. We train DDBNs with both generative anddiscriminative objectives under an approximate computing framework and analyzetheir power-at-performance for supervised and semi-supervised applications. Wealso investigate the out-of-distribution performance of DDBNs when theinference data has the same class structure yet is statistically different fromthe training data owing to dynamic real-time operating environments. Based onour analysis, we provide novel insights and recommendations for choice oftraining objectives, bitwidth values, and accuracy sensitivity with respect tothe amount of labeled data for implementing DDBN inference with minimum powerconsumption on embedded hardware platforms subject to accuracy tolerances.

Generative and Discriminative Deep Belief Network Classifiers: Comparisons Under an Approximate Computing Framework