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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Deep Learning is moving to edge devices, ushering in a new age of distributedArtificial Intelligence (AI). The high demand of computational resourcesrequired by deep neural networks may be alleviated by approximate computingtechniques, and most notably reduced-precision arithmetic with coarselyquantized numerical representations. In this context, Bonseyes comes in as aninitiative to enable stakeholders to bring AI to low-power and autonomousenvironments such as: Automotive, Medical Healthcare and Consumer Electronics.To achieve this, we introduce LPDNN, a framework for optimized deployment ofDeep Neural Networks on heterogeneous embedded devices. In this work, we detailthe quantization engine that is integrated in LPDNN. The engine depends on afine-grained workflow which enables a Neural Network Design Exploration and asensitivity analysis of each layer for quantization. We demonstrate the enginewith a case study on Alexnet and VGG16 for three different techniques fordirect quantization: standard fixed-point, dynamic fixed-point and k-meansclustering, and demonstrate the potential of the latter. We argue that using aGaussian quantizer with k-means clustering can achieve better performance thanlinear quantizers. Without retraining, we achieve over 55.64\% saving forweights' storage and 69.17\% for run-time memory accesses with less than 1\%drop in top5 accuracy in Imagenet.

QUENN: QUantization Engine for low-power Neural Networks