Recurrent neural networks (RNNs) are a powerful model for sequential data. End-to-end training methods such as Connectionist Temporal Classification make it possible to train RNNs for sequence labelling problems where the input-output alignment is unknown. The combination of these methods with the Long Short-term Memory RNN architecture has proved particularly fruitful, delivering state-of-the-art results in cursive handwriting recognition. However RNN performance in speech recognition has so far been disappointing, with better results returned by deep feedforward networks. This paper investigates deep recurrent neural networks, which combine the multiple levels of representation that have proved so effective in deep networks with the flexible use of long range context that empowers RNNs. When trained end-to-end with suitable regularisation, we find that deep Long Short-term Memory RNNs achieve a test set error of 17.7% on the TIMIT phoneme recognition benchmark, which to our knowledge is the best recorded score.

Speech recognition with deep recurrent neural networks

Recurrent neural networks (RNNs) are a powerful model for sequential data. End-to-end training methods such as Connectionist Temporal Classification make it possible to train RNNs for sequence labelling problems where the input-output alignment is unknown. The combination of these methods with the Long Short-term Memory RNN architecture has proved particularly fruitful, delivering state-of-the-art results in cursive handwriting recognition. However RNN performance in speech recognition has so far been disappointing, with better results returned by deep feedforward networks. This paper investigates \emph{deep recurrent neural networks}, which combine the multiple levels of representation that have proved so effective in deep networks with the flexible use of long range context that empowers RNNs. When trained end-to-end with suitable regularisation, we find that deep Long Short-term Memory RNNs achieve a test set error of 17.7% on the TIMIT phoneme recognition benchmark, which to our knowledge is the best recorded score.

Speech Recognition with Deep Recurrent Neural Networks

Recurrent neural networks are powerful sequence learners. They are able to incorporate context information in a flexible way, and are robust to localised distortions of the input data. These properties make them well suited to sequence labelling, where input sequences are transcribed with streams of labels. The aim of this thesis is to advance the state-of-the-art in supervised sequence labelling with recurrent networks. Its two main contributions are (1) a new type of output layer that allows recurrent networks to be trained directly for sequence labelling tasks where the alignment between the inputs and the labels is unknown, and (2) an extension of the long short-term memory network architecture to multidimensional data, such as images and video sequences.

https://mediatum.ub.tum.de/doc/673554/file.pdf

Supervised Sequence Labelling with Recurrent Neural Networks

Recently, the hybrid deep neural network (DNN)- hidden Markov model (HMM) has been shown to significantly improve speech recognition performance over the conventional Gaussian mixture model (GMM)-HMM. The performance improvement is partially attributed to the ability of the DNN to model complex correlations in speech features. In this paper, we show that further error rate reduction can be obtained by using convolutional neural networks (CNNs). We first present a concise description of the basic CNN and explain how it can be used for speech recognition. We further propose a limited-weight-sharing scheme that can better model speech features. The special structure such as local connectivity, weight sharing, and pooling in CNNs exhibits some degree of invariance to small shifts of speech features along the frequency axis, which is important to deal with speaker and environment variations. Experimental results show that CNNs reduce the error rate by 6%-10% compared with DNNs on the TIMIT phone recognition and the voice search large vocabulary speech recognition tasks.

/pdf/convolutional-neural-networks-for-speech-recognition-59lskxzlgf.pdf

Convolutional neural networks for speech recognition

Gaussian mixture models are currently the dominant technique for modeling the emission distribution of hidden Markov models for speech recognition. We show that better phone recognition on the TIMIT dataset can be achieved by replacing Gaussian mixture models by deep neural networks that contain many layers of features and a very large number of parameters. These networks are first pre-trained as a multi-layer generative model of a window of spectral feature vectors without making use of any discriminative information. Once the generative pre-training has designed the features, we perform discriminative fine-tuning using backpropagation to adjust the features slightly to make them better at predicting a probability distribution over the states of monophone hidden Markov models.

/pdf/acoustic-modeling-using-deep-belief-networks-3vvqg8rhp3.pdf

Acoustic Modeling Using Deep Belief Networks

Hidden Markov modeling is extended to speaker-independent phone recognition. Using multiple codebooks of various linear-predictive-coding (LPC) parameters and discrete hidden Markov models (HMMs) the authors obtain a speaker-independent phone recognition accuracy of 58.8-73.8% on the TIMIT database, depending on the type of acoustic and language models used. In comparison, the performance of expert spectrogram readers is only 69% without use of higher level knowledge. The authors introduce the co-occurrence smoothing algorithm, which enables accurate recognition even with very limited training data. Since the results were evaluated on a standard database, they can be used as benchmarks to evaluate future systems. >

Speaker-independent phone recognition using hidden Markov models

A description is given of SPHINX, a system that demonstrates the feasibility of accurate, large-vocabulary, speaker-independent, continuous speech recognition. SPHINX is based on discrete hidden Markov models (HMMs) with LPC- (linear-predictive-coding) derived parameters. To provide speaker independence, knowledge was added to these HMMs in several ways: multiple codebooks of fixed-width parameters, and an enhanced recognizer with carefully designed models and word-duration modeling. To deal with coarticulation in continuous speech, yet still adequately represent a large vocabulary, two new subword speech units are introduced: function-word-dependent phone models and generalized triphone models. With grammars of perplexity 997, 60, and 20, SPHINX attained word accuracies of 71, 94, and 96%, respectively, on a 997-word task. >

H.-W. Hon

Papers

Speaker-independent phone recognition using hidden Markov models

An overview of the SPHINX speech recognition system