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 Networks (DNN) have achieved human level performance in manyimage analytics tasks but DNNs are mostly deployed to GPU platforms thatconsume a considerable amount of power. Brain-inspired spiking neuromorphicchips consume low power and can be highly parallelized. However, for deployingDNNs to energy efficient neuromorphic chips the incompatibility betweencontinuous neurons and synaptic weights of traditional DNNs, discrete spikingneurons and synapses of neuromorphic chips has to be overcome. Previous workhas achieved this by training a network to learn continuous probabilities anddeployment to a neuromorphic architecture by random sampling theseprobabilities. An ensemble of sampled networks is needed to approximate theperformance of the trained network.  In the work presented in this paper, we have extended previous research bydirectly learning binary synaptic crossbars. Results on MNIST show that betterperformance can be achieved with a small network in one time step (92.7%maximum observed accuracy vs 95.98% accuracy in our work). Top results on alarger network are similar to previously published results (99.42% maximumobserved accuracy vs 99.45% accuracy in our work). More importantly, in ourwork a smaller ensemble is needed to achieve similar or better accuracy thanprevious work, which translates into significantly decreased energy consumptionfor both networks. Results of our work are stable since they do not requirerandom sampling.

Improving energy efficiency and classification accuracy of neuromorphic chips by learning binary synaptic crossbars