Training Spiking Deep Networks for Neuromorphic Hardware

TL;DR

This paper introduces a scalable method for training spiking deep neural networks with leaky integrate-and-fire neurons, achieving state-of-the-art results on multiple datasets including ImageNet, and demonstrating potential for power-efficient neuromorphic hardware implementation.

Contribution

A novel scalable approach to train spiking neural networks by transforming deep artificial networks, enabling high performance and robustness on large datasets like ImageNet.

Findings

Achieved state-of-the-art results on five datasets including ImageNet.

Networks are more power-efficient on neuromorphic hardware compared to traditional hardware.

Method is scalable and compatible with various neural nonlinearities.

Abstract

We describe a method to train spiking deep networks that can be run using leaky integrate-and-fire (LIF) neurons, achieving state-of-the-art results for spiking LIF networks on five datasets, including the large ImageNet ILSVRC-2012 benchmark. Our method for transforming deep artificial neural networks into spiking networks is scalable and works with a wide range of neural nonlinearities. We achieve these results by softening the neural response function, such that its derivative remains bounded, and by training the network with noise to provide robustness against the variability introduced by spikes. Our analysis shows that implementations of these networks on neuromorphic hardware will be many times more power-efficient than the equivalent non-spiking networks on traditional hardware.