Semi-supervised training of Gaussian mixture models by conditional entropy minimization

Keyphrases

• Conditional Entropy 100%
• Gaussian Mixture Model 100%
• Unlabeled Data 100%
• Training Methods 100%
• Semi-supervised Training 100%
• Entropy Minimization 100%
• Maximum mutual Information 50%
• Mutual Information Criterion 50%
• Classification Performance 50%
• Convergence Rate 50%
• Preconditioned Conjugate Gradient Method 50%
• Parameter Update 50%
• Phonetic Classification 50%
• Discriminative Training 50%
• Training Criterion 50%
• TIMIT Corpus 50%

Computer Science

• Conditional Entropy 100%
• Gaussian Mixture Model 100%
• Unlabeled Data 100%
• Mutual Information 50%
• Information Criterion 50%
• Conjugate Gradient Method 50%
• Convergence Rate 50%