Adaptation of Multilingual Stacked Bottle-neck Neural Network Structure for New Language

feature extraction, Bottle-neck features, neural network adaptation, multilingual neural networks, Stacked Bottle- Neck structure

In this paper a multilingual training of Stacked Bottle- Neck neural network structure for feature extraction is addressed. While for languages with plentiful resources, the optimal approach is to train the BN-NN on the target data, limited resources call for re-using data from other languages.

The neural network based features became an inseparable part of state-of-the-art LVCSR systems. In order to perform well, the network has to be trained on a large amount of in-domain data. With the increasing emphasis on fast development of ASR system on limited resources, there is an effort to alleviate the need of in-domain data. To evaluate the effectiveness of other resources, we have trained the Stacked Bottle-Neck neural networks structure on multilingual data investigating several training strategies while treating the target language as the unseen one. Further, the systems were adapted to the target language by re-training. Finally, we evaluated the effect of adaptation of individual NNs in the Stacked Bottle-Neck structure to find out the optimal adaptation strategy. We have shown that the adaptation can significantly improve system performance over both, the multilingual network and network trained only on target data. The experiments were performed on Babel Year 1 data.