Is Neuromorphic MNIST Neuromorphic? Analyzing the Discriminative Power of Neuromorphic Datasets in the Time Domain.

Laxmi R Iyer,Yansong Chua,Haizhou Li

doi:10.3389/fnins.2021.608567

Abstract

A major characteristic of spiking neural networks (SNNs) over conventional artificial neural networks (ANNs) is their ability to spike, enabling them to use spike timing for coding and efficient computing. In this paper, we assess if neuromorphic datasets recorded from static images are able to evaluate the ability of SNNs to use spike timings in their calculations. We have analyzed N-MNIST, N-Caltech101 and DvsGesture along these lines, but focus our study on N-MNIST. First we evaluate if additional information is encoded in the time domain in a neuromorphic dataset. We show that an ANN trained with backpropagation on frame-based versions of N-MNIST and N-Caltech101 images achieve 99.23 and 78.01% accuracy. These are comparable to the state of the art—showing that an algorithm that purely works on spatial data can classify these datasets. Second we compare N-MNIST and DvsGesture on two STDP algorithms, RD-STDP, that can classify only spatial data, and STDP-tempotron that classifies spatiotemporal data. We demonstrate that RD-STDP performs very well on N-MNIST, while STDP-tempotron performs better on DvsGesture. Since DvsGesture has a temporal dimension, it requires STDP-tempotron, while N-MNIST can be adequately classified by an algorithm that works on spatial data alone. This shows that precise spike timings are not important in N-MNIST. N-MNIST does not, therefore, highlight the ability of SNNs to classify temporal data. The conclusions of this paper open the question—what dataset can evaluate SNN ability to classify temporal data?

Highlights

The remarkable performance and efficiency of the brain have prompted scientists to build systems that mimic it—for studying biological function as well as improving engineering systems
In contrast in DvsGesture a neuromorphic dataset not derived from static images, our artificial neural networks (ANNs) has an accuracy of 71.01% which is far less than the 96.49% accuracy obtained by spiking neural networks (SNN), showing that SNN is preferred over ANN in datasets where additional temporal information contained in the timing of spikes is present
The ANN obtains comparable to state-of-the-art results on Neuromorphic MNIST (N-MNIST) (99.23%) and N-Caltech101 (78.01%)

Summary

Introduction

The remarkable performance and efficiency of the brain have prompted scientists to build systems that mimic it—for studying biological function as well as improving engineering systems. Networks of the first and second generations do not have neurons that spike. These networks, known as artificial neural networks (ANNs) have real-valued outputs and can be seen as time averaged firing rates of neurons. Third generation networks are more mathematically accurate models of biological neurons. A neuron of the third generation network receives incoming spikes through its synapses and fires a spike when its membrane potential exceeds a threshold. Such a neuron can use spike time coding, described below. Before we describe spike time coding, we will first enumerate the different definitions of firing rate currently used

Objectives

Methods

Results

Discussion

Conclusion