A Spike in Performance: Training Hybrid-Spiking Neural Networks with Quantized Activation Functions

TL;DR

This paper introduces a novel training method for hybrid Spiking Neural Networks that uses quantized activation functions, achieving state-of-the-art accuracy with high energy efficiency.

Contribution

It presents a new approach to train hybrid SNNs by interpolating between non-spiking and spiking regimes, outperforming existing recurrent models in accuracy.

Findings

Hybrid SNN outperforms SotA recurrent architectures in accuracy.

Activities are reduced to at most 3.74 bits on average.

Achieves high energy efficiency with 1.26 significant bits per weight.

Abstract

The machine learning community has become increasingly interested in the energy efficiency of neural networks. The Spiking Neural Network (SNN) is a promising approach to energy-efficient computing, since its activation levels are quantized into temporally sparse, one-bit values (i.e., "spike" events), which additionally converts the sum over weight-activity products into a simple addition of weights (one weight for each spike). However, the goal of maintaining state-of-the-art (SotA) accuracy when converting a non-spiking network into an SNN has remained an elusive challenge, primarily due to spikes having only a single bit of precision. Adopting tools from signal processing, we cast neural activation functions as quantizers with temporally-diffused error, and then train networks while smoothly interpolating between the non-spiking and spiking regimes. We apply this technique to the Legendre Memory Unit (LMU) to obtain the first known example of a hybrid SNN outperforming SotA recurrent architectures -- including the LSTM, GRU, and NRU -- in accuracy, while reducing activities to at most 3.74 bits on average with 1.26 significant bits multiplying each weight. We discuss how these methods can significantly improve the energy efficiency of neural networks.