BS4NN: Binarized Spiking Neural Networks with Temporal Coding and Learning

TL;DR

BS4NN introduces a binarized spiking neural network leveraging temporal coding, achieving high accuracy on MNIST datasets with reduced memory and computational requirements, and outperforming simple binary neural networks.

Contribution

The paper presents BS4NN, a novel binarized spiking neural network that operates in the time domain, combining binary weights with temporal coding for improved efficiency and accuracy.

Findings

Achieved 97.0% on MNIST with minimal accuracy loss.

Outperformed simple BNNs on benchmark datasets.

Demonstrated the effectiveness of temporal coding in spiking networks.

Abstract

We recently proposed the S4NN algorithm, essentially an adaptation of backpropagation to multilayer spiking neural networks that use simple non-leaky integrate-and-fire neurons and a form of temporal coding known as time-to-first-spike coding. With this coding scheme, neurons fire at most once per stimulus, but the firing order carries information. Here, we introduce BS4NN, a modification of S4NN in which the synaptic weights are constrained to be binary (+1 or -1), in order to decrease memory (ideally, one bit per synapse) and computation footprints. This was done using two sets of weights: firstly, real-valued weights, updated by gradient descent, and used in the backward pass of backpropagation, and secondly, their signs, used in the forward pass. Similar strategies have been used to train (non-spiking) binarized neural networks. The main difference is that BS4NN operates in the time domain: spikes are propagated sequentially, and different neurons may reach their threshold at different times, which increases computational power. We validated BS4NN on two popular benchmarks, MNIST and Fashion-MNIST, and obtained reasonable accuracies for this sort of network (97.0% and 87.3% respectively) with a negligible accuracy drop with respect to real-valued weights (0.4% and 0.7%, respectively). We also demonstrated that BS4NN outperforms a simple BNN with the same architectures on those two datasets (by 0.2% and 0.9% respectively), presumably because it leverages the temporal dimension. The source codes of the proposed BS4NN are publicly available at https://github.com/SRKH/BS4NN.