دانلود رایگان مقاله انگلیسی شبکه های عصبی عمیق با اسپایک وزن دار - الزویر 2018

Abstract

Spiking neural networks are being regarded as one of the promising alternative techniques to overcome the high energy costs of artificial neural networks. It is supported by many researches showing that a deep convolutional neural network can be converted into a spiking neural network with near zero accuracy loss. However, the advantage on energy consumption of spiking neural networks comes at a cost of long classification latency due to the use of Poissondistributed spike trains (rate coding), especially in deep networks. In this paper, we propose to use weighted spikes, which can greatly reduce the latency by assigning a different weight to a spike depending on which time phase it belongs. Experimental results on MNIST, SVHN, CIFAR-10, and CIFAR-100 show that the proposed spiking neural networks with weighted spikes achieve significant reduction in classification latency and number of spikes, which leads to faster and more energy-efficient spiking neural networks than the conventional spiking neural networks with rate coding. We also show that one of the state-of-theart networks the deep residual network can be converted into spiking neural network without accuracy loss.

8. Conclusion

We proposed a novel spiking neural network with weighted spikes to overcome the slow classification problem of the conventional SNNs with rate coding. The key idea is assigning different weights to spikes depending on the time phase within a period to encode more information within less number of spikes. The proposed SNN-WS reduces the classification latency as well as number of spikes significantly. The approach can be applied to various types of networks including deep residual networks and various datasets without accuracy loss. We also proposed an early decision algorithm to further reduce latency and number of spikes at the expense of small accuracy loss.