The CAPIO 2017 conversational speech recognition system

KJ Han, A Chandrashekaran, J Kim, I Lane - arXiv preprint arXiv …, 2017 - arxiv.org
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 …

Neural network language modeling with letter-based features and importance sampling

H Xu, K Li, Y Wang, J Wang, S Kang… - … on acoustics, speech …, 2018 - ieeexplore.ieee.org
In this paper we describe an extension of the Kaldi software toolkit to support neural-based
language modeling, intended for use in automatic speech recognition (ASR) and related …

Recurrent neural network language model training with noise contrastive estimation for speech recognition

X Chen, X Liu, MJF Gales… - 2015 IEEE International …, 2015 - ieeexplore.ieee.org
In recent years recurrent neural network language models (RNNLMs) have been
successfully applied to a range of tasks including speech recognition. However, an …

Efficient training and evaluation of recurrent neural network language models for automatic speech recognition

X Chen, X Liu, Y Wang, MJF Gales… - … /ACM Transactions on …, 2016 - ieeexplore.ieee.org
Recurrent neural network language models (RNNLMs) are becoming increasingly popular
for a range of applications including automatic speech recognition. An important issue that …

Scalable multi corpora neural language models for asr

A Raju, D Filimonov, G Tiwari, G Lan… - arXiv preprint arXiv …, 2019 - arxiv.org
Neural language models (NLM) have been shown to outperform conventional n-gram
language models by a substantial margin in Automatic Speech Recognition (ASR) and other …

Lattice rescoring strategies for long short term memory language models in speech recognition

S Kumar, M Nirschl, D Holtmann-Rice… - 2017 IEEE Automatic …, 2017 - ieeexplore.ieee.org
Recurrent neural network (RNN) language models (LMs) and Long Short Term Memory
(LSTM) LMs, a variant of RNN LMs, have been shown to outperform traditional N-gram LMs …

Automatic speech recognition with very large conversational finnish and estonian vocabularies

S Enarvi, P Smit, S Virpioja… - IEEE/ACM Transactions …, 2017 - ieeexplore.ieee.org
Today, the vocabulary size for language models in large vocabulary speech recognition is
typically several hundreds of thousands of words. While this is already sufficient in some …

[PDF][PDF] Deep Learning-Based Telephony Speech Recognition in the Wild.

KJ Han, S Hahm, BH Kim, J Kim, IR Lane - Interspeech, 2017 - isca-archive.org
In this paper, we explore the effectiveness of a variety of Deep Learning-based acoustic
models for conversational telephony speech, specifically TDNN, bLSTM and CNN-bLSTM …

Improving the training and evaluation efficiency of recurrent neural network language models

X Chen, X Liu, MJF Gales… - 2015 IEEE International …, 2015 - ieeexplore.ieee.org
Recurrent neural network language models (RNNLMs) are becoming increasingly popular
for speech recognition. Previously, we have shown that RNNLMs with a full (non-classed) …

[PDF][PDF] Densely Connected Networks for Conversational Speech Recognition.

KJ Han, A Chandrashekaran, J Kim, IR Lane - INTERSPEECH, 2018 - isca-archive.org
In this paper we show how we have achieved the state-of-theart performance on the industry-
standard NIST 2000 Hub5 English evaluation set. We propose densely connected LSTMs …