Abstract
In this paper, we explore several new schemes to train a seq2seq model to integrate a pre-trained language model (LM). Our proposed fusion methods focus on the memory cell state and the hidden state in the seq2seq decoder long short-term memory (LSTM), and the memory cell state is updated by the LM unlike the prior studies. This means the memory retained by the main seq2seq would be adjusted by the external LM. These fusion methods have several variants depending on the architecture of this memory cell update and the use of memory cell and hidden states which directly affects the final label inference. We performed the experiments to show the effectiveness of the proposed methods in a mono-lingual ASR setup on the Librispeech corpus and in a transfer learning setup from a multilingual ASR (MLASR) base model to a low-resourced language. In Librispeech, our best model improved WER by 3.7%, 2.4% for test clean, test other relatively to the shallow fusion baseline, with multilevel decoding. In transfer learning from an MLASR base model to the IARPA Babel Swahili model, the best scheme improved the transferred model on eval set by 9.9%, 9.8% in CER, WER relatively to the 2-stage transfer baseline.
| Original language | English |
|---|---|
| Title of host publication | 2019 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 - Proceedings |
| Publisher | Institute of Electrical and Electronics Engineers Inc. |
| Pages | 6191-6195 |
| Number of pages | 5 |
| ISBN (Electronic) | 9781479981311 |
| DOIs | |
| Publication status | Published - 2019 May 1 |
| Event | 44th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 - Brighton, United Kingdom Duration: 2019 May 12 → 2019 May 17 |
Publication series
| Name | ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings |
|---|---|
| Volume | 2019-May |
| ISSN (Print) | 1520-6149 |
Conference
| Conference | 44th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 |
|---|---|
| Country | United Kingdom |
| City | Brighton |
| Period | 19/5/12 → 19/5/17 |
Fingerprint
Keywords
- Automatic speech recognition (ASR)
- cold fusion
- deep fusion
- language model
- sequence to sequence
- shallow fusion
ASJC Scopus subject areas
- Software
- Signal Processing
- Electrical and Electronic Engineering
Cite this
Language Model Integration Based on Memory Control for Sequence to Sequence Speech Recognition. / Cho, Jaejin; Watanabe, Shinji; Hori, Takaaki; Baskar, Murali Karthick; Inaguma, Hirofumi; Villalba, Jesus; Dehak, Najim.
2019 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. p. 6191-6195 8683380 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings; Vol. 2019-May).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Language Model Integration Based on Memory Control for Sequence to Sequence Speech Recognition
AU - Cho, Jaejin
AU - Watanabe, Shinji
AU - Hori, Takaaki
AU - Baskar, Murali Karthick
AU - Inaguma, Hirofumi
AU - Villalba, Jesus
AU - Dehak, Najim
PY - 2019/5/1
Y1 - 2019/5/1
N2 - In this paper, we explore several new schemes to train a seq2seq model to integrate a pre-trained language model (LM). Our proposed fusion methods focus on the memory cell state and the hidden state in the seq2seq decoder long short-term memory (LSTM), and the memory cell state is updated by the LM unlike the prior studies. This means the memory retained by the main seq2seq would be adjusted by the external LM. These fusion methods have several variants depending on the architecture of this memory cell update and the use of memory cell and hidden states which directly affects the final label inference. We performed the experiments to show the effectiveness of the proposed methods in a mono-lingual ASR setup on the Librispeech corpus and in a transfer learning setup from a multilingual ASR (MLASR) base model to a low-resourced language. In Librispeech, our best model improved WER by 3.7%, 2.4% for test clean, test other relatively to the shallow fusion baseline, with multilevel decoding. In transfer learning from an MLASR base model to the IARPA Babel Swahili model, the best scheme improved the transferred model on eval set by 9.9%, 9.8% in CER, WER relatively to the 2-stage transfer baseline.
AB - In this paper, we explore several new schemes to train a seq2seq model to integrate a pre-trained language model (LM). Our proposed fusion methods focus on the memory cell state and the hidden state in the seq2seq decoder long short-term memory (LSTM), and the memory cell state is updated by the LM unlike the prior studies. This means the memory retained by the main seq2seq would be adjusted by the external LM. These fusion methods have several variants depending on the architecture of this memory cell update and the use of memory cell and hidden states which directly affects the final label inference. We performed the experiments to show the effectiveness of the proposed methods in a mono-lingual ASR setup on the Librispeech corpus and in a transfer learning setup from a multilingual ASR (MLASR) base model to a low-resourced language. In Librispeech, our best model improved WER by 3.7%, 2.4% for test clean, test other relatively to the shallow fusion baseline, with multilevel decoding. In transfer learning from an MLASR base model to the IARPA Babel Swahili model, the best scheme improved the transferred model on eval set by 9.9%, 9.8% in CER, WER relatively to the 2-stage transfer baseline.
KW - Automatic speech recognition (ASR)
KW - cold fusion
KW - deep fusion
KW - language model
KW - sequence to sequence
KW - shallow fusion
UR - http://www.scopus.com/inward/record.url?scp=85069000387&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85069000387&partnerID=8YFLogxK
U2 - 10.1109/ICASSP.2019.8683380
DO - 10.1109/ICASSP.2019.8683380
M3 - Conference contribution
T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
SP - 6191
EP - 6195
BT - 2019 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
ER -