In this work we present simple grapheme-based system for low-resource speech recognition using Babel data for Turkish spontaneous speech (80 hours). We have investigated different neural network architectures performance, including fully-convolutional, recurrent and ResNet with GRU. Different features and normalization techniques are compared as well. We also proposed CTC-loss modification using segmentation during training, which leads to improvement while decoding with small beam size. Our best model achieved word error rate of 45.8%, which is the best reported result for end-to-end systems using in-domain data for this task, according to our knowledge.
@article{arxiv.1807.00868,
title = {Exploring End-to-End Techniques for Low-Resource Speech Recognition},
author = {Vladimir Bataev and Maxim Korenevsky and Ivan Medennikov and Alexander Zatvornitskiy},
journal= {arXiv preprint arXiv:1807.00868},
year = {2018}
}
Comments
Accepted for Specom 2018, 20th International Conference on Speech and Computer