Deep Recurrent Neural Networks for Acoustic Modelling

Abstract

We present a novel deep Recurrent Neural Network (RNN) model for acoustic modelling in Automatic Speech Recognition (ASR). We term our contribution as a TC-DNN-BLSTM-DNN model, the model combines a Deep Neural Network (DNN) with Time Convolution (TC), followed by a Bidirectional Long Short-Term Memory (BLSTM), and a final DNN. The first DNN acts as a feature processor to our model, the BLSTM then generates a context from the sequence acoustic signal, and the final DNN takes the context and models the posterior probabilities of the acoustic states. We achieve a 3.47 WER on the Wall Street Journal (WSJ) eval92 task or more than 8% relative improvement over the baseline DNN models.

References in corpus (2)

• Sequence Transduction with Recurrent Neural Networks
• Show and Tell: A Neural Image Caption Generator

Cited by in corpus (10)

• A Comprehensive Study of Deep Bidirectional LSTM RNNs for Acoustic Modeling in Speech Recognition
• Fully Convolutional Speech Recognition
• Transferring Knowledge from a RNN to a DNN
• Letter-Based Speech Recognition with Gated ConvNets
• Who Needs Words? Lexicon-Free Speech Recognition
• IMS-Speech: A Speech to Text Tool
• Simplified End-to-End MMI Training and Voting for ASR
• On Training Recurrent Networks with Truncated Backpropagation Through Time in Speech Recognition
• KoSpeech: Open-Source Toolkit for End-to-End Korean Speech Recognition
• Toward Optimal Run Racing: Application to Deep Learning Calibration