Speaker recognition using multiple kernel learning based on conditional entropy minimization.
Tetsuji OgawaHideitsu HinoNima ReyhaniNoboru MurataTetsunori KobayashiPublished in: ICASSP (2011)
Keyphrases
- multiple kernel learning
- entropy minimization
- linear combination
- kernel learning
- kernel methods
- multiple kernel
- lp norm
- multi task
- learning problems
- learning models
- binary classification
- feature selection
- kernel function
- multiple features
- support vectors
- support vector machine
- prediction accuracy
- data mining
- semi supervised learning
- semi supervised
- prior knowledge
- training data