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 have been proven to be highly effective tools in variousdomains, yet their computational and memory costs restrict them from beingwidely deployed on portable devices. The recent rapid increase of edgecomputing devices has led to an active search for techniques to address theabove-mentioned limitations of machine learning frameworks. The quantization ofartificial neural networks (ANNs), which converts the full-precision synapticweights into low-bit versions, emerged as one of the solutions. At the sametime, spiking neural networks (SNNs) have become an attractive alternative toconventional ANNs due to their temporal information processing capability,energy efficiency, and high biological plausibility. Despite being driven bythe same motivation, the simultaneous utilization of both concepts has yet tobe thoroughly studied. Therefore, this work aims to bridge the gap betweenrecent progress in quantized neural networks and SNNs. It presents an extensivestudy on the performance of the quantization function, represented as a linearcombination of sigmoid functions, exploited in low-bit weight quantization inSNNs. The presented quantization function demonstrates the state-of-the-artperformance on four popular benchmarks, CIFAR10-DVS, DVS128 Gesture,N-Caltech101, and N-MNIST, for binary networks (64.05\%, 95.45\%, 68.71\%, and99.43\% respectively) with small accuracy drops and up to 31$\times$ memorysavings, which outperforms existing methods.

Low Precision Quantization-aware Training in Spiking Neural Networks with Differentiable Quantization Function