WaveSense: Efficient Temporal Convolutions with Spiking Neural Networks for Keyword Spotting

TL;DR

WaveSense introduces a neuromorphic-compatible spiking neural network inspired by WaveNet, achieving high keyword spotting accuracy with low power consumption suitable for edge devices.

Contribution

The paper proposes WaveSense, a novel spiking neural network architecture that mimics dilated temporal convolutions, optimized for neuromorphic hardware and improved keyword spotting performance.

Findings

Outperforms existing spiking neural networks in keyword spotting

Achieves near state-of-the-art accuracy of CNNs and LSTMs

Designed for efficient neuromorphic implementation

Abstract

Ultra-low power local signal processing is a crucial aspect for edge applications on always-on devices. Neuromorphic processors emulating spiking neural networks show great computational power while fulfilling the limited power budget as needed in this domain. In this work we propose spiking neural dynamics as a natural alternative to dilated temporal convolutions. We extend this idea to WaveSense, a spiking neural network inspired by the WaveNet architecture. WaveSense uses simple neural dynamics, fixed time-constants and a simple feed-forward architecture and hence is particularly well suited for a neuromorphic implementation. We test the capabilities of this model on several datasets for keyword-spotting. The results show that the proposed network beats the state of the art of other spiking neural networks and reaches near state-of-the-art performance of artificial neural networks such as CNNs and LSTMs.