Inference with Artificial Neural Networks on Analog Neuromorphic Hardware

TL;DR

This paper explores the use of the BrainScaleS-2 neuromorphic hardware for neural network inference, demonstrating calibration, optimization, and high accuracy classification of MNIST digits with energy-efficient analog in-memory computing.

Contribution

It presents calibration and optimization strategies for analog neuromorphic hardware and demonstrates near-software accuracy on MNIST classification.

Findings

Achieved 98.0% accuracy on MNIST classification.

Demonstrated the effectiveness of hardware-in-the-loop training.

Showcased high energy efficiency of analog in-memory computing.

Abstract

The neuromorphic BrainScaleS-2 ASIC comprises mixed-signal neurons and synapse circuits as well as two versatile digital microprocessors. Primarily designed to emulate spiking neural networks, the system can also operate in a vector-matrix multiplication and accumulation mode for artificial neural networks. Analog multiplication is carried out in the synapse circuits, while the results are accumulated on the neurons' membrane capacitors. Designed as an analog, in-memory computing device, it promises high energy efficiency. Fixed-pattern noise and trial-to-trial variations, however, require the implemented networks to cope with a certain level of perturbations. Further limitations are imposed by the digital resolution of the input values (5 bit), matrix weights (6 bit) and resulting neuron activations (8 bit). In this paper, we discuss BrainScaleS-2 as an analog inference accelerator and present calibration as well as optimization strategies, highlighting the advantages of training with hardware in the loop. Among other benchmarks, we classify the MNIST handwritten digits dataset using a two-dimensional convolution and two dense layers. We reach 98.0% test accuracy, closely matching the performance of the same network evaluated in software.