Neural Networks for Keyword Spotting on IoT Devices
arXiv:2101.00693v1
Originality Synthesis-oriented
AI Analysis
This work addresses the problem of deploying keyword spotting models on resource-constrained IoT devices for manufacturers and users of such devices.
This paper explores the use of Neural Networks for keyword spotting on IoT devices. The authors propose a Convolutional Neural Network design that aims to reduce the number of multiplications and model parameters for execution on memory and computation-constrained hardware.
We explore Neural Networks (NNs) for keyword spotting (KWS) on IoT devices like smart speakers and wearables. Since we target to execute our NN on a constrained memory and computation footprint, we propose a CNN design that. (i) uses a limited number of multiplies. (ii) uses a limited number of model parameters.