Detail publikace

Multilingual Sequence-to-Sequence Speech Recognition: Architecture, Transfer Learning, and Language Modeling

CHO, J. BASKAR, M. LI, R. WIESNER, M. MALLIDI, S. YALTA, N. KARAFIÁT, M. WATANABE, S. HORI, T.

Originální název

Multilingual Sequence-to-Sequence Speech Recognition: Architecture, Transfer Learning, and Language Modeling

Typ

článek ve sborníku ve WoS nebo Scopus

Jazyk

angličtina

Originální abstrakt

Sequence-to-sequence (seq2seq) approach for low-resource ASR is a relatively new direction in speech research. The approach benefits by performing model training without using lexicon and alignments. However, this poses a new problem of requiring more data compared to conventional DNN-HMM systems. In this work, we attempt to use data from 10 BABEL languages to build a multilingual seq2seq model as a prior model, and then port them towards 4 other BABEL languages using transfer learning approach. We also explore different architectures for improving the prior multilingual seq2seq model. The paper also discusses the effect of integrating a recurrent neural network language model (RNNLM) with a seq2seq model during decoding. Experimental results show that the transfer learning approach from the multilingual model shows substantial gains over monolingual models across all 4 BABEL languages. Incorporating an RNNLM also brings significant improvements in terms of %WER, and achieves recognition performance comparable to the models trained with twice more training data.

Klíčová slova

Automatic speech recognition (ASR), sequence to sequence, multilingual setup, transfer learning, language modeling

Autoři

CHO, J.; BASKAR, M.; LI, R.; WIESNER, M.; MALLIDI, S.; YALTA, N.; KARAFIÁT, M.; WATANABE, S.; HORI, T.

Vydáno

18. 12. 2018

Nakladatel

IEEE Signal Processing Society

Místo

Athens

ISBN

978-1-5386-4334-1

Kniha

Proceedings of 2018 IEEE WORKSHOP ON SPOKEN LANGUAGE TECHNOLOGY (SLT 2018)

Strany od

521

Strany do

527

Strany počet

7

URL

https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8639655

BibTex

@inproceedings{BUT163489,
  author="CHO, J. and BASKAR, M. and LI, R. and WIESNER, M. and MALLIDI, S. and YALTA, N. and KARAFIÁT, M. and WATANABE, S. and HORI, T.",
  title="Multilingual Sequence-to-Sequence Speech Recognition: Architecture, Transfer Learning, and Language Modeling",
  booktitle="Proceedings of 2018 IEEE WORKSHOP ON SPOKEN LANGUAGE TECHNOLOGY (SLT 2018)",
  year="2018",
  pages="521--527",
  publisher="IEEE Signal Processing Society",
  address="Athens",
  doi="10.1109/SLT.2018.8639655",
  isbn="978-1-5386-4334-1",
  url="https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8639655"
}

Dokumenty

cho_slt2018_08639655.pdf