用于流式唤醒词检测的小足迹卷积循环网络
音频与语音处理
2020-11-26 v1
摘要
本工作中，我们提出小足迹卷积循环神经网络模型并应用于唤醒词检测问题，且以缩放点积注意力对其增强。我们发现，在250k参数预算下，与使用卷积神经网络的模型相比，采用CRNN可将错误接受率降低25%，同时参数规模减小10%；与词级稠密神经网络模型相比，在50k参数预算下参数规模减小75%可获得高达32%的改进。我们讨论了在流式音频上使用CRNN进行推理这一挑战性问题的解决方案，以及与CNN、DNN和DNN-HMM模型相比在起止索引误差与延迟方面的差异。
引用
@article{arxiv.2011.12941,
title = {Small Footprint Convolutional Recurrent Networks for Streaming Wakeword Detection},
author = {Mohammad Omar Khursheed and Christin Jose and Rajath Kumar and Gengshen Fu and Brian Kulis and Santosh Kumar Cheekatmalla},
journal= {arXiv preprint arXiv:2011.12941},
year = {2020}
}
备注
\c{opyright} 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works