Hello Edge: Keyword Spotting on Microcontrollers

TL;DR

This paper evaluates and optimizes neural network architectures for keyword spotting on microcontrollers, demonstrating that high accuracy can be achieved within strict resource constraints.

Contribution

It provides a comprehensive comparison of neural network architectures for KWS on microcontrollers and shows how to optimize them for accuracy and resource efficiency.

Findings

DS-CNN achieves 95.4% accuracy, outperforming similar models.

Neural networks can be optimized to fit microcontroller constraints.

Optimized architectures maintain high accuracy within limited memory and compute.

Abstract

Keyword spotting (KWS) is a critical component for enabling speech based user interactions on smart devices. It requires real-time response and high accuracy for good user experience. Recently, neural networks have become an attractive choice for KWS architecture because of their superior accuracy compared to traditional speech processing algorithms. Due to its always-on nature, KWS application has highly constrained power budget and typically runs on tiny microcontrollers with limited memory and compute capability. The design of neural network architecture for KWS must consider these constraints. In this work, we perform neural network architecture evaluation and exploration for running KWS on resource-constrained microcontrollers. We train various neural network architectures for keyword spotting published in literature to compare their accuracy and memory/compute requirements. We show that it is possible to optimize these neural network architectures to fit within the memory and compute constraints of microcontrollers without sacrificing accuracy. We further explore the depthwise separable convolutional neural network (DS-CNN) and compare it against other neural network architectures. DS-CNN achieves an accuracy of 95.4%, which is ~10% higher than the DNN model with similar number of parameters.