TY - JOUR
T1 - An Introduction to Probabilistic Spiking Neural Networks
T2 - Probabilistic Models, Learning Rules, and Applications
AU - Jang, Hyeryung
AU - Simeone, Osvaldo
AU - Gardner, Brian
AU - Gruning, Andre
N1 - Funding Information:
This work was supported in part by the European Research Council under the European Union’s Horizon 2020 research and innovation program under grant 725731 and by the U.S. National Science Foundation under grant ECCS 1710009. André Grüning (partly) and Brian Gardner (fully) are supported by the European Union’s Horizon 2020 Framework Programme for Research and Innovation under the specific grant agreement 785907 (Human Brain Project SGA2).
Publisher Copyright:
© 1991-2012 IEEE.
PY - 2019/11
Y1 - 2019/11
N2 - Spiking neural networks (SNNs) are distributed trainable systems whose computing elements, or neurons, are characterized by internal analog dynamics and by digital and sparse synaptic communications. The sparsity of the synaptic spiking inputs and the corresponding event-driven nature of neural processing can be leveraged by energy-efficient hardware implementations, which can offer significant energy reductions as compared to conventional artificial neural networks (ANNs). The design of training algorithms for SNNs, however, lags behind hardware implementations: most existing training algorithms for SNNs have been designed either for biological plausibility or through conversion from pretrained ANNs via rate encoding.
AB - Spiking neural networks (SNNs) are distributed trainable systems whose computing elements, or neurons, are characterized by internal analog dynamics and by digital and sparse synaptic communications. The sparsity of the synaptic spiking inputs and the corresponding event-driven nature of neural processing can be leveraged by energy-efficient hardware implementations, which can offer significant energy reductions as compared to conventional artificial neural networks (ANNs). The design of training algorithms for SNNs, however, lags behind hardware implementations: most existing training algorithms for SNNs have been designed either for biological plausibility or through conversion from pretrained ANNs via rate encoding.
UR - http://www.scopus.com/inward/record.url?scp=85075013745&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85075013745&partnerID=8YFLogxK
U2 - 10.1109/MSP.2019.2935234
DO - 10.1109/MSP.2019.2935234
M3 - Article
AN - SCOPUS:85075013745
SN - 1053-5888
VL - 36
SP - 64
EP - 77
JO - IEEE Signal Processing Magazine
JF - IEEE Signal Processing Magazine
IS - 6
M1 - 8891810
ER -