Relevance vector machines with empirical likelihood-ratio kernels for PLDA speaker verification.
Wei RaoMan-Wai MakPublished in: ISCSLP (2014)
Keyphrases
- likelihood ratio
- speaker verification
- relevance vector machines
- hypothesis testing
- noisy environments
- learning machines
- audio visual
- kernel function
- emotion recognition
- multilayer perceptron
- feature space
- language identification
- support vector
- confidence intervals
- speech recognition
- multimedia
- match scores
- dynamic programming
- linear combination
- pattern recognition
- genetic algorithm
- machine learning methods
- policy search
- computer vision