fMLLR based feature-space speaker adaptation of DNN acoustic models

Abstract

We investigate the problem of speaker adaptation of DNN acoustic models in two settings: the traditional unsupervised adaptation and a supervised adaptation (SuA) where a few minutes of transcribed speech is available. SuA presents additional difficulties when a test speaker’s adaptation information does not match the registered speaker’s information… (More)

Topics

3 Figures and Tables

Statistics

010020020162017
Citations per Year

Citation Velocity: 80

Averaging 80 citations per year over the last 2 years.

Learn more about how we calculate this metric in our FAQ.

Cite this paper

@inproceedings{Parthasarathi2015fMLLRBF, title={fMLLR based feature-space speaker adaptation of DNN acoustic models}, author={Sree Hari Krishnan Parthasarathi and Bj{\"o}rn Hoffmeister and Spyridon Matsoukas and Arindam Mandal and Nikko Strom and Sri Garimella}, booktitle={INTERSPEECH}, year={2015} }