Maximum a posteriori adaptation of subspace Gaussian mixture models for cross-lingual speech recognition.
Liang LuArnab GhoshalSteve RenalsPublished in: ICASSP (2012)
Keyphrases
- cross lingual
- maximum a posteriori
- speech recognition
- gaussian mixture model
- em algorithm
- expectation maximization
- maximum likelihood
- language modeling
- language model
- mixture model
- speaker identification
- speaker recognition
- probabilistic model
- feature space
- machine translation
- markov random field
- hidden markov models
- image reconstruction
- bayesian framework
- pattern recognition
- speech signal
- automatic speech recognition
- n gram
- dimensionality reduction
- gaussian distribution
- high dimensional
- unsupervised learning
- query expansion
- probability density function
- document clustering
- generative model
- low dimensional
- retrieval model
- feature vectors
- noisy environments
- news articles
- information retrieval
- mel frequency cepstral coefficients
- bayesian networks
- transfer learning
- text classification
- principal component analysis
- information extraction
- image retrieval
- reinforcement learning
- feature extraction
- learning algorithm