2017

The CAPIO 2017 Conversational Speech Recognition System

Han, Kyu J., Chandrashekaran, Akshay, Kim, Jungsuk et al.

Understand

In this paper we show how we have achieved the state-of-the-art performance on the industry-standard NIST 2000 Hub5 English evaluation set.

We explore densely connected LSTMs, inspired by the densely connected convolutional networks recently introduced for image classification tasks.
We also propose an acoustic model adaptation scheme that simply averages the parameters of a seed neural network acoustic model and its adapted version.
This method was applied with the CallHome training corpus and improved individual system performances by on average 6.1% (relative) against the CallHome portion of the evaluation set with no performance loss on the Switchboard portion.

Built on

M. J. F. Gales, “Maximum likelihood linear transformations for HMM-based speech recognition,”
1997
Earlier work this paper cites.
J. F. William, W. M. Fisher, A. F. Martin, M. A. Przybocki, and D. S. Pallett, “NIST evaluation of conversational speech recognition over the telephone: English and mandarin performance results,” in
2000
Earlier work this paper cites.
P. Thompson and H. Nesi, “Research in progress, the British Academic Spoken English (BASE) corpus project.”
2001
Earlier work this paper cites.
A. Stolcke, “SRILM – An extensible language modeling toolkit,” in
2002
Earlier work this paper cites.
R. C. Simpson, S. L. Briggs, J. Ovens, and J. M. Swales, “The Michigan Corpus of Academic Spoken English.”
2002
Earlier work this paper cites.
A. Krizhevsky and G. Hinton, “Learning multiple layers of features from tiny images,” in
2009
Earlier work this paper cites.
S. Zagoruyko and N. Komodakis, “Wide residual networks,” in
Original
2009
Earlier work this paper cites.
B. Li and K. C. Sim, “Comparison of discriminative input and output transformations for speaker adaptation in the hybrid NN/HMM systems,” in
2010
Earlier work this paper cites.
N. Dehak, P. Kenny, R. Dehak, P. Dumouchel, and P. Ouellet, “Front-end factor analysis for speaker verification,”
2011
Earlier work this paper cites.
D. Povey, A. Ghoshal, G. Boulianne, L. Burget, O. Glembek, N. Goel, M. Hannemann, P. Motlicek, Y. Qian, P. Schwarz, J. Silovsky, G. Stemmer, and K. Vesely, “The Kaldi speech recognition toolkit,” in
2011
Earlier work this paper cites.
H. Xu, D. Povey, L. Mangu, and J. Zhu, “Minimum bayes risk decoding and system combination based on a recursion for edit distance,”
2011
Earlier work this paper cites.
J. Bergstra, R. Bardenet, Y. Bengio, and B. Kégl, “Algorithms for hyper-parameter optimization,” in
2011
Earlier work this paper cites.
A. Rousseau, P. Deléglise, and Y. Esteve, “TED-LIUM: an automatic speech recognition dedicated corpus.” in
2012
Earlier work this paper cites.
A. Rousseau, P. Deléglise, and Y. Estève, “TED-LIUM: an automatic speech recognition dedicated corpus,” in
2012
Earlier work this paper cites.

Similar

H. Liao, “Speaker adaptation of context dependent deep neural networks,” in
2013
Cited alongside, same era.
G. Saon, H. Soltau, D. Namahoo, and M. Picheny, “Speaker adaptation of neural network acoustic models using i-vectors,” in
2013
Cited alongside, same era.
V. Gupta, P. Kenny, P. Ouellet, and T. Stafylakis, “I-vector-based speaker adaptation of deep neural networks for French broadcast audio transcription,” in
2014
Cited alongside, same era.
Y. Miao, L. Jiang, H. Zhang, and F. Metze, “Towards speaker adaptive training of deep neural network acoustic models,” in
2014
Cited alongside, same era.
Y. Shi, W. Zhang, M. Cai, and J. Liu, “Variance regularization of RNNLM for speech recognition,” in
2014
Cited alongside, same era.
——, “Enhancing the TED-LIUM corpus with selected data for language modeling and more TED talks,” in
2014
Cited alongside, same era.
I. Himawan, P. Motlicek, M. F. Font, and S. Madikeri, “Towards utterance-based neural network adaptation in acoustic modeling,” in
2015
Cited alongside, same era.
R. K. Srivastava, K. Greff, and J. Schmidhuber, “Training very deep networks,” in
2015
Cited alongside, same era.
——, “Highway networks,” in
2015
Cited alongside, same era.
D. Povey, X. Zhang, and K. Sanjeev, “Parallel training of DNNs with natural gradient and paramter averaging,” in
2015
Cited alongside, same era.
V. Panayotov, G. Chen, D. Povey, and S. Khudanpur, “LibriSpeech: an ASR corpus based on public domain audio books,” in
2015
Cited alongside, same era.
W. Xiong, J. Droppo, X. Huang, F. Seide, M. Seltzer, and A. Stolcke, “Achieving human parity in conversational speech recognition,” in
2016
Cited alongside, same era.

Then

K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in
2016
Later among the works it cites.
Y. Zhang, G. Chen, D. Yu, K. Yao, S. Khudanpur, and J. Glass, “Highway long short-term memory RNNs for distant speech recognition,” in
2016
Later among the works it cites.
D. Povey, V. Peddinti, G. Galvez, P. Ghahrmani, V. Manohar, X. Na, Y. Wang, and S. Khudanpur, “Purely sequence-trained neural networks for asr based on lattice-free MMI,” in
2016
Later among the works it cites.
X. Chen, X. Liu, Y. Qian, M. J. F. Gales, and P. C. Woodland, “CUED-RNNLM - An open-source toolkit for efficient training and evaluation of recurrent neural network language models,” in
2016
Later among the works it cites.
G. Saon, G. Kurata, T. Sercu, K. Audhkhasi, S. Thomas, D. Dimitriadis, X. Cui, B. Ramabhadran, M. Picheny, L. Lim, B. Roomi, and P. Hall, “English conversational telephone speech recognition by humans and machines,” in
2017
Closest in time.
K. J. Han, S. Hahm, B. Kim, J. Kim, and I. Lane, “Deep learning-based telephony speech recognition in the wild,” in
2017
Closest in time.
W. Xiong, L. Wu, F. Alleva, J. Droppo, X. Huang, and A. Stolcke, “The Microsoft 2017 conversational speech recognition system,” in
2017
Closest in time.
G. Kurata, B. Ramabhadran, G. Saon, and A. Sethy, “Language modeling with highway LSTM,” in
2017
Closest in time.
A. Stolcke and J. Droppo, “Comparing human and machine errors in conversational speech transcription,” in
2017
Closest in time.
G. Pundak and T. Sainath, “Highway-LSTM and recurrent highway networks for speech recognition,” in
2017
Closest in time.
J. Kim, M. El-Khamy, and J. Lee, “Residual LSTM: Design of a deep recurrent architecture for distant speech recognition,” in
2017
Closest in time.
L. Huang, J. Sun, J. Xu, and Y. Yang, “An improved residual LSTM architecture for acoustic modeling,” in
2017
Closest in time.
G. Cheng, V. Peddinti, D. Povey, V. Manohar, S. Khudanpur, and Y. Yan, “An exploration of dropout with LSTMs,” in
2017
Closest in time.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…

Reading the bibliography…