Global-Local Convolution with Spiking Neural Networks for Energy-efficient Keyword Spotting

TL;DR

This paper introduces a novel energy-efficient spiking neural network model for keyword spotting, combining global-local convolution and bottleneck processing to improve performance and reduce energy consumption on edge devices.

Contribution

The paper proposes a new end-to-end lightweight SNN model with innovative modules for speech feature extraction and signal processing, enhancing energy efficiency and accuracy.

Findings

Achieves competitive accuracy on Google Speech Commands Dataset

Uses fewer parameters than existing SNN-based KWS models

Demonstrates improved energy efficiency with novel modules

Abstract

Thanks to Deep Neural Networks (DNNs), the accuracy of Keyword Spotting (KWS) has made substantial progress. However, as KWS systems are usually implemented on edge devices, energy efficiency becomes a critical requirement besides performance. Here, we take advantage of spiking neural networks' energy efficiency and propose an end-to-end lightweight KWS model. The model consists of two innovative modules: 1) Global-Local Spiking Convolution (GLSC) module and 2) Bottleneck-PLIF module. Compared to the hand-crafted feature extraction methods, the GLSC module achieves speech feature extraction that is sparser, more energy-efficient, and yields better performance. The Bottleneck-PLIF module further processes the signals from GLSC with the aim to achieve higher accuracy with fewer parameters. Extensive experiments are conducted on the Google Speech Commands Dataset (V1 and V2). The results show our method achieves competitive performance among SNN-based KWS models with fewer parameters.