BriefGPT.xyz
Jul, 2018
基于对抗学习的无监督领域自适应技术，用于提高语音识别的鲁棒性
Unsupervised Domain Adaptation by Adversarial Learning for Robust Speech Recognition
HTML
PDF
Pavel Denisov, Ngoc Thang Vu, Marc Ferras Font
TL;DR
利用对抗性学习进行无监督适应性，采用神经网络和未分类的适应性数据解决了远程语音识别问题，相对于没有适应的模型，相对词错误率下降了19.8%。
Abstract
In this paper, we investigate the use of
adversarial learning
for
unsupervised adaptation
to unseen recording conditions, more specifically, single microphone far-field speech. We adapt
→