Boosting keyword spotting through on-device learnable user speech characteristics

arXiv:2403.07802v1 Announce Type: cross
Abstract: Keyword spotting systems for always-on TinyML-constrained applications require on-site tuning to boost the accuracy of offline trained classifiers when deployed in unseen inference conditions. Adapting to the speech peculiarities of target users requires many in-domain samples, often unavailable in real-world scenarios. Furthermore, current on-device learning techniques rely on computationally intensive and memory-hungry backbone update schemes, unfit for always-on, battery-powered devices. In this work, we propose a novel on-device learning architecture, composed of a pretrained backbone …

abstract accuracy applications arxiv boost boosting classifiers cs.lg cs.sd current domain eess.as inference offline on-device learning samples speech systems through tinyml type world