Characterisation of speech diversity using self-organising maps
Computation and Language
2017-02-08 v1 Neural and Evolutionary Computing
Sound
Abstract
We report investigations into speaker classification of larger quantities of unlabelled speech data using small sets of manually phonemically annotated speech. The Kohonen speech typewriter is a semi-supervised method comprised of self-organising maps (SOMs) that achieves low phoneme error rates. A SOM is a 2D array of cells that learn vector representations of the data based on neighbourhoods. In this paper, we report a method to evaluate pronunciation using multilevel SOMs with /hVd/ single syllable utterances for the study of vowels, for Australian pronunciation.
Keywords
Cite
@article{arxiv.1702.02092,
title = {Characterisation of speech diversity using self-organising maps},
author = {Tom A. F. Anderson and David M. W. Powers},
journal= {arXiv preprint arXiv:1702.02092},
year = {2017}
}
Comments
16th Speech Science and Technology Conference (SST2016)