# Improving classification accuracy of feedforward neural networks for spiking neuromorphic chips

## TL;DR

This paper introduces a new training algorithm for neural networks that directly configures neuromorphic hardware, significantly improving EEG classification accuracy while maintaining high performance on standard datasets.

## Contribution

It presents a novel learning method that trains binary crossbars compatible with TrueNorth neuromorphic chips, enabling better deployment of DNNs on energy-efficient hardware.

## Key findings

- Achieved 86% accuracy on EEG data, surpassing previous methods at 76%.
- Maintained state-of-the-art performance on MNIST dataset.
- Demonstrated compatibility with TrueNorth hardware constraints.

## Abstract

Deep Neural Networks (DNN) achieve human level performance in many image analytics tasks but DNNs are mostly deployed to GPU platforms that consume a considerable amount of power. New hardware platforms using lower precision arithmetic achieve drastic reductions in power consumption. More recently, brain-inspired spiking neuromorphic chips have achieved even lower power consumption, on the order of milliwatts, while still offering real-time processing. However, for deploying DNNs to energy efficient neuromorphic chips the incompatibility between continuous neurons and synaptic weights of traditional DNNs, discrete spiking neurons and synapses of neuromorphic chips need to be overcome. Previous work has achieved this by training a network to learn continuous probabilities, before it is deployed to a neuromorphic architecture, such as IBM TrueNorth Neurosynaptic System, by random sampling these probabilities. The main contribution of this paper is a new learning algorithm that learns a TrueNorth configuration ready for deployment. We achieve this by training directly a binary hardware crossbar that accommodates the TrueNorth axon configuration constrains and we propose a different neuron model. Results of our approach trained on electroencephalogram (EEG) data show a significant improvement with previous work (76% vs 86% accuracy) while maintaining state of the art performance on the MNIST handwritten data set.