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

Spiking neural networks are efficient computation models for low-powerenvironments. Spike-based BP algorithms and ANN-to-SNN (ANN2SNN) conversionsare successful techniques for SNN training. Nevertheless, the spike-base BPtraining is slow and requires large memory costs. Though ANN2NN provides alow-cost way to train SNNs, it requires many inference steps to mimic thewell-trained ANN for good performance. In this paper, we propose a SNN-to-ANN(SNN2ANN) framework to train the SNN in a fast and memory-efficient way. TheSNN2ANN consists of 2 components: a) a weight sharing architecture between ANNand SNN and b) spiking mapping units. Firstly, the architecture trains theweight-sharing parameters on the ANN branch, resulting in fast training and lowmemory costs for SNN. Secondly, the spiking mapping units ensure that theactivation values of the ANN are the spiking features. As a result, theclassification error of the SNN can be optimized by training the ANN branch.Besides, we design an adaptive threshold adjustment (ATA) algorithm to addressthe noisy spike problem. Experiment results show that our SNN2ANN-based modelsperform well on the benchmark datasets (CIFAR10, CIFAR100, and Tiny-ImageNet).Moreover, the SNN2ANN can achieve comparable accuracy under 0.625x time steps,0.377x training time, 0.27x GPU memory costs, and 0.33x spike activities of theSpike-based BP model.

SNN2ANN: A Fast and Memory-Efficient Training Framework for Spiking Neural Networks