期刊
2021 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP 2021)
卷 -, 期 -, 页码 2830-2834出版社
IEEE
DOI: 10.1109/ICASSP39728.2021.9414848
关键词
Keyword spotting; neural architecture search
This paper proposes a method using neural architecture search to discover more efficient networks, which can effectively reduce power consumption of smart audio devices.
Smart audio devices are gated by an always-on lightweight keyword spotting program to reduce power consumption. It is however challenging to design models that have both high accuracy and low latency for accurate and fast responsiveness. Many efforts have been made to develop end-to-end neural networks, in which depthwise separable convolutions, temporal convolutions, and LSTMs are adopted as building units. Nonetheless, these networks designed with human expertise may not achieve an optimal trade-off in an expansive search space. In this paper, we propose to leverage recent advances in differentiable neural architecture search to discover more efficient networks. Our searched model attains 97.2% top-1 accuracy on Google Speech Command Dataset v1 with only nearly 100K parameters.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据