Abstract
This paper describes minimum word error (MWE) training of recurrent neural network language models (RNNLMs) for speech recognition. RNNLMs are usually trained to minimize a cross entropy of estimated word probabilities against the correct word sequence, which corresponds to maximum likelihood criterion. However, this training does not necessarily maximize a performance measure in a target task, i.e. it does not minimize word error rate (WER) explicitly in speech recognition. To solve such a problem, several discriminative training methods have already been proposed for n-gram language models, but those for RNNLMs have not sufficiently investigated. In this paper, we propose a MWE training method for RNNLMs, and report significant WER reductions when we applied the MWE method to a standard Elman-type RNNLM and a more advanced model, a Long Short-Term Memory (LSTM) RNNLM. We also present efficient MWE training with N-best lists on Graphics Processing Units (GPUs).
| Original language | English |
|---|---|
| Title of host publication | 2016 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2016 - Proceedings |
| Publisher | Institute of Electrical and Electronics Engineers Inc. |
| Pages | 5990-5994 |
| Number of pages | 5 |
| Volume | 2016-May |
| ISBN (Electronic) | 9781479999880 |
| DOIs | |
| Publication status | Published - 2016 May 18 |
| Externally published | Yes |
| Event | 41st IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2016 - Shanghai, China Duration: 2016 Mar 20 → 2016 Mar 25 |
Other
| Other | 41st IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2016 |
|---|---|
| Country | China |
| City | Shanghai |
| Period | 16/3/20 → 16/3/25 |
Fingerprint
Keywords
- Long short-term memory
- Minimum word error training
- Recurrent neural network language model
- Speech recognition
ASJC Scopus subject areas
- Signal Processing
- Software
- Electrical and Electronic Engineering
Cite this
Minimum word error training of long short-term memory recurrent neural network language models for speech recognition. / Hori, Takaaki; Hori, Chiori; Watanabe, Shinji; Hershey, John R.
2016 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2016 - Proceedings. Vol. 2016-May Institute of Electrical and Electronics Engineers Inc., 2016. p. 5990-5994 7472827.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Minimum word error training of long short-term memory recurrent neural network language models for speech recognition
AU - Hori, Takaaki
AU - Hori, Chiori
AU - Watanabe, Shinji
AU - Hershey, John R.
PY - 2016/5/18
Y1 - 2016/5/18
N2 - This paper describes minimum word error (MWE) training of recurrent neural network language models (RNNLMs) for speech recognition. RNNLMs are usually trained to minimize a cross entropy of estimated word probabilities against the correct word sequence, which corresponds to maximum likelihood criterion. However, this training does not necessarily maximize a performance measure in a target task, i.e. it does not minimize word error rate (WER) explicitly in speech recognition. To solve such a problem, several discriminative training methods have already been proposed for n-gram language models, but those for RNNLMs have not sufficiently investigated. In this paper, we propose a MWE training method for RNNLMs, and report significant WER reductions when we applied the MWE method to a standard Elman-type RNNLM and a more advanced model, a Long Short-Term Memory (LSTM) RNNLM. We also present efficient MWE training with N-best lists on Graphics Processing Units (GPUs).
AB - This paper describes minimum word error (MWE) training of recurrent neural network language models (RNNLMs) for speech recognition. RNNLMs are usually trained to minimize a cross entropy of estimated word probabilities against the correct word sequence, which corresponds to maximum likelihood criterion. However, this training does not necessarily maximize a performance measure in a target task, i.e. it does not minimize word error rate (WER) explicitly in speech recognition. To solve such a problem, several discriminative training methods have already been proposed for n-gram language models, but those for RNNLMs have not sufficiently investigated. In this paper, we propose a MWE training method for RNNLMs, and report significant WER reductions when we applied the MWE method to a standard Elman-type RNNLM and a more advanced model, a Long Short-Term Memory (LSTM) RNNLM. We also present efficient MWE training with N-best lists on Graphics Processing Units (GPUs).
KW - Long short-term memory
KW - Minimum word error training
KW - Recurrent neural network language model
KW - Speech recognition
UR - http://www.scopus.com/inward/record.url?scp=84973341477&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84973341477&partnerID=8YFLogxK
U2 - 10.1109/ICASSP.2016.7472827
DO - 10.1109/ICASSP.2016.7472827
M3 - Conference contribution
AN - SCOPUS:84973341477
VL - 2016-May
SP - 5990
EP - 5994
BT - 2016 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2016 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
ER -