Dimensionality reduction for speech recognition using neighborhood components analysis
Dimensionality reduction for speech recognition using neighborhood components analysis
复制标题
使用邻域成分分析进行语音识别降维
DOI：
10.21437/interspeech.2007-376
复制
发表时间：
2007
影响因子：
4.6
通讯作者：
Timothy J. Hazen
中科院分区：
文献类型：
--
作者：
Natasha Singh;M. Collins;Timothy J. Hazen
先前的工作已经考虑了用于学习高维声学表示到低维空间的投影的方法。本文将邻域成分分析（NCA）[2]方法应用于语音识别器的声学建模。NCA学习声学向量的投影，优化与最近邻分类器的分类精度密切相关的标准。我们将正则化引入到该方法中，进一步提高了性能。我们描述了一个演讲转录任务的实验，比较了使用NCA和HLDA学习的投影[1]。正则化的NCA使WER相对于HLDA绝对降低0.7%，相当于相对降低1.9%。索引词：语音识别，声学建模，降维
Previous work has considered methods for learning projections of high-dimensional acoustic representations to lower dimensional spaces. In this paper we apply the neighborhoodcomponentsanalysis (NCA) [2] method to acoustic modeling in a speech recognizer. NCA learns a projection of acoustic vectors that optimizes a criterion that is closely related to the classification accuracy of a nearest-neighbor classifier. We introduce regularization into this method, giving further improvements in performance. We describe experiments on a lecture transcription task, comparing projections learned using NCA and HLDA [1] . Regularized NCA gives a 0.7% absolute reduction in WER over HLDA, which corresponds to a relative reduction of 1.9%. Index Terms: speech recognition, acoustic modeling, dimensionality reduction