We present a new back propagation based training algorithm for discrete-time spiking neural networks (SNN). Inspired by recent deep learning algorithms on binarized neural networks, binary activation with a straight-through gradient estimator is used to model the leaky integrate-fire spiking neuron, overcoming the difficulty in training SNNs using back propagation. Two SNN training algorithms are proposed: (1) SNN with discontinuous integration, which is suitable for rate-coded input spikes, and (2) SNN with continuous integration, which is more general and can handle input spikes with temporal information. Neuromorphic hardware designed in 40nm CMOS exploits the spike sparsity and demonstrates high classification accuracy (>98% on MNIST) and low energy (48.4-773 nJ/image).
@article{arxiv.1709.06206,
title = {Algorithm and Hardware Design of Discrete-Time Spiking Neural Networks Based on Back Propagation with Binary Activations},
author = {Shihui Yin and Shreyas K. Venkataramanaiah and Gregory K. Chen and Ram Krishnamurthy and Yu Cao and Chaitali Chakrabarti and Jae-sun Seo},
journal= {arXiv preprint arXiv:1709.06206},
year = {2017}
}
Comments
2017 IEEE Biomedical Circuits and Systems (BioCAS)