门控循环神经网络的门激活信号分析及其与音素边界的相关性
声音
2017-09-01 v2 计算与语言
机器学习
摘要
本文分析了门控循环神经网络内部的门激活信号，发现此类信号的时间结构与音素边界高度相关。这一相关性进一步通过一组音素分割实验得到验证，在该实验中获得了优于标准方法的结果。
引用
@article{arxiv.1703.07588,
title = {Gate Activation Signal Analysis for Gated Recurrent Neural Networks and Its Correlation with Phoneme Boundaries},
author = {Yu-Hsuan Wang and Cheng-Tao Chung and Hung-yi Lee},
journal= {arXiv preprint arXiv:1703.07588},
year = {2017}
}
备注
5 pages, The code is available at https://github.com/allyoushawn/timit_gas.git