Improving energy efficiency and classification accuracy of neuromorphic chips by learning binary synaptic crossbars

TL;DR

This paper introduces a method for directly learning binary synaptic crossbars in neuromorphic chips, achieving high accuracy with less energy consumption and more stable results compared to previous probabilistic sampling approaches.

Contribution

The paper presents a novel approach to directly learn binary synaptic weights for neuromorphic chips, reducing energy use and improving stability over prior probabilistic methods.

Findings

Achieved 92.7% accuracy on MNIST with a small network, surpassing previous results.

Achieved 99.45% accuracy on larger networks, comparable to state-of-the-art.

Reduced ensemble size needed for similar or better accuracy, decreasing energy consumption.

Abstract

Deep Neural Networks (DNN) have achieved human level performance in many image analytics tasks but DNNs are mostly deployed to GPU platforms that consume a considerable amount of power. Brain-inspired spiking neuromorphic chips consume low power and can be highly parallelized. 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 has to be overcome. Previous work has achieved this by training a network to learn continuous probabilities and deployment to a neuromorphic architecture by random sampling these probabilities. An ensemble of sampled networks is needed to approximate the performance of the trained network. In the work presented in this paper, we have extended previous research by directly learning binary synaptic crossbars. Results on MNIST show that better performance can be achieved with a small network in one time step (92.7% maximum observed accuracy vs 95.98% accuracy in our work). Top results on a larger network are similar to previously published results (99.42% maximum observed accuracy vs 99.45% accuracy in our work). More importantly, in our work a smaller ensemble is needed to achieve similar or better accuracy than previous work, which translates into significantly decreased energy consumption for both networks. Results of our work are stable since they do not require random sampling.