Large-margin minimum classification error training: A theoretical risk minimization perspective
Computer Speech and Language
The application of hidden Markov models in speech recognition
Foundations and Trends in Signal Processing
Invited paper: Automatic speech recognition: History, methods and challenges
Pattern Recognition
Training data selection for improving discriminative training of acoustic models
Pattern Recognition Letters
Discriminative training of HMMs for automatic speech recognition: A survey
Computer Speech and Language
Minimum classification error learning for sequential data in the wavelet domain
Pattern Recognition
Penalized logistic regression with HMM log-likelihood regressors for speech recognition
IEEE Transactions on Audio, Speech, and Language Processing
Expert Systems with Applications: An International Journal
Optimized discriminative transformations for speech features based on minimum classification error
Pattern Recognition Letters
Efficient training of discriminative language models by sample selection
Speech Communication
Computer Speech and Language
Computers in Biology and Medicine
Structural Bayesian Linear Regression for Hidden Markov Models
Journal of Signal Processing Systems
| Hi-index | 0.00 |
The minimum classification error (MCE) framework for discriminative training is a simple and general formalism for directly optimizing recognition accuracy in pattern recognition problems. The framework applies directly to the optimization of hidden Markov models (HMMs) used for speech recognition problems. However, few if any studies have reported results for the application of MCE training to large-vocabulary, continuous-speech recognition tasks. This article reports significant gains in recognition performance and model compactness as a result of discriminative training based on MCE training applied to HMMs, in the context of three challenging large-vocabulary (up to 100 k word) speech recognition tasks: the Corpus of Spontaneous Japanese lecture speech transcription task, a telephone-based name recognition task, and the MIT Jupiter telephone-based conversational weather information task. On these tasks, starting from maximum likelihood (ML) baselines, MCE training yielded relative reductions in word error ranging from 7% to 20%. Furthermore, this paper evaluates the use of different methods for optimizing the MCE criterion function, as well as the use of precomputed recognition lattices to speed up training. An overview of the MCE framework is given, with an emphasis on practical implementation issues