• Publications
  • Influence
Maximum likelihood linear regression for speaker adaptation of continuous density hidden Markov models
TLDR
A method of speaker adaptation for continuous density hidden Markov models (HMMs) is presented. Expand
  • 2,575
  • 299
  • PDF
Minimum Phone Error and I-smoothing for improved discriminative training
TLDR
In this paper we introduce the Minimum Phone Error (MPE) and Minimum Word Error (MWE) criteria for the discriminative training of HMM systems. Expand
  • 787
  • 97
  • PDF
Mean and variance adaptation within the MLLR framework
TLDR
A new technique for adapting both the means and the variances of a set of continuous density HMMs within the MLLR framework has been described. Expand
  • 477
  • 52
  • PDF
Large scale discriminative training of hidden Markov models for speech recognition
TLDR
This paper describes, and evaluates on a large scale, the lattice based framework for discriminative training of large vocabulary speech recognition systems based on Gaussian mixture hidden Markov models based on the maximum mutual information estimation criterion which has been used to train HMM systems for conversational telephone speech transcription using up to 265 hours of training data. Expand
  • 348
  • 34
Posterior probability decoding, confidence estimation and system combination
TLDR
In this paper the estimation of word posterior probabilities is discussed and their application in the CU-HTK system used in the Hub5 Conversational Telephone Speech evaluation is described. Expand
  • 268
  • 28
  • PDF
Flexible speaker adaptation using maximum likelihood linear regression
Substituted acetylenes wherein the amino and acetylene groups are directly connected to aromatic ring carbon atoms and wherein the substituted acetylene has at least 3 carbon atoms and a hydroxylExpand
  • 178
  • 19
MMIE training of large vocabulary recognition systems
TLDR
This paper describes a framework for optimising the structure and parameters of a continuous density HMM-based large vocabulary recognition system using the Maximum Mutual Information Estimation (MMIE) criterion. Expand
  • 208
  • 18
Speaker adaptation of HMMs using linear regression
  • 110
  • 18
Speaker adaptation for continuous density HMMs: a review
TLDR
This paper reviews some popular speaker adaptation schemes that can be applied to continuous density hidden Markov models. Expand
  • 172
  • 16
  • PDF
Recurrent neural network language model training with noise contrastive estimation for speech recognition
TLDR
In recent years recurrent neural network language models (RNNLMs) have been successfully applied to a range of tasks including speech recognition. Expand
  • 70
  • 15
  • PDF