Fast and Accurate Recurrent Neural Network Acoustic Models for Speech Recognition
This work addresses incremental improvements in speech recognition models for large vocabulary applications.
The paper tackles improving LSTM RNN acoustic models for speech recognition by introducing techniques like frame stacking and reduced frame rate, resulting in more accurate models and faster decoding, with CD phone modeling providing further gains.
We have recently shown that deep Long Short-Term Memory (LSTM) recurrent neural networks (RNNs) outperform feed forward deep neural networks (DNNs) as acoustic models for speech recognition. More recently, we have shown that the performance of sequence trained context dependent (CD) hidden Markov model (HMM) acoustic models using such LSTM RNNs can be equaled by sequence trained phone models initialized with connectionist temporal classification (CTC). In this paper, we present techniques that further improve performance of LSTM RNN acoustic models for large vocabulary speech recognition. We show that frame stacking and reduced frame rate lead to more accurate models and faster decoding. CD phone modeling leads to further improvements. We also present initial results for LSTM RNN models outputting words directly.