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

Implementing large-scale deep neural networks with high computationalcomplexity on low-cost IoT devices may inevitably be constrained by limitedcomputation resource, making the devices hard to respond in real-time. Thisdisjunction makes the state-of-art deep learning algorithms, i.e. CNN(Convolutional Neural Networks), incompatible with IoT world. We present alow-bit (range from 8-bit to 1-bit) scheme with our local quantization regionalgorithm. We use models in Caffe model zoo as our example tasks to evaluatethe effect of our low precision data representation scheme. With the availableof local quantization region, we find implementations on top of those schemescould greatly retain the model accuracy, besides the reduction of computationalcomplexity. For example, our 8-bit scheme has no drops on top-1 and top-5accuracy with 2x speedup on Intel Edison IoT platform. Implementations based onour 4-bit, 2-bit or 1-bit scheme are also applicable to IoT devices withadvances of low computational complexity. For example, the drop on our task isonly 0.7% when using 2-bit scheme, a scheme which could largely savetransistors. Making low-bit scheme usable here opens a new door for furtheroptimization on commodity IoT controller, i.e. extra speed-up could be achievedby replacing multiply-accumulate operations with the proposed table look-upoperations. The whole study offers a new approach to relief the challenge ofbring advanced deep learning algorithm to resource constrained low-cost IoTdevice.

Deploy Large-Scale Deep Neural Networks in Resource Constrained IoT Devices with Local Quantization Region