Multilingual Training and Cross-lingual Adaptation on CTC-based Acoustic Model

TL;DR

This paper explores multilingual CTC-based acoustic models for speech recognition, focusing on adaptation and regularization techniques like LHUC and dropout to improve performance on under-resourced languages.

Contribution

It introduces a universal IPA-based CTC model with adaptation methods that enhance multilingual speech recognition, especially with limited data.

Findings

LHUC improves language adaptation performance

Dropout during adaptation reduces overfitting

Achieves competitive results with DNN/HMM systems

Abstract

Multilingual models for Automatic Speech Recognition (ASR) are attractive as they have been shown to benefit from more training data, and better lend themselves to adaptation to under-resourced languages. However, initialisation from monolingual context-dependent models leads to an explosion of context-dependent states. Connectionist Temporal Classification (CTC) is a potential solution to this as it performs well with monophone labels. We investigate multilingual CTC in the context of adaptation and regularisation techniques that have been shown to be beneficial in more conventional contexts. The multilingual model is trained to model a universal International Phonetic Alphabet (IPA)-based phone set using the CTC loss function. Learning Hidden Unit Contribution (LHUC) is investigated to perform language adaptive training. In addition, dropout during cross-lingual adaptation is also studied and tested in order to mitigate the overfitting problem. Experiments show that the performance of the universal phoneme-based CTC system can be improved by applying LHUC and it is extensible to new phonemes during cross-lingual adaptation. Updating all the parameters shows consistent improvement on limited data. Applying dropout during adaptation can further improve the system and achieve competitive performance with Deep Neural Network / Hidden Markov Model (DNN/HMM) systems on limited data.