"DE - Spolkov\u00E1 republika N\u011Bmecko" . "http://download.springer.com/static/pdf/237/chp%253A10.1007%252F978-3-642-32790-2_58.pdf?auth66=1352100308_6ca3feab5ab1f038650cd3b971c1f53b&ext=.pdf" . "Machlica, Luk\u00E1\u0161" . . "RIV/49777513:23520/12:43915990" . "Robust Adaptation Techniques Dealing with Small Amount of Data" . . "3"^^ . "Lecture Notes in Computer Science" . . . . . "3"^^ . . "The worst problem the adaptation is dealing with is the lack of adaptation data. This work focuses on the feature Maximum Likelihood Linear Regression (fMLLR) adaptation where the number of free parameters to be estimated significantly decreases in comparison with other adaptation methods. However, the number of free parameters of fMLLR transform is still too high to be estimated properly when dealing with extremely small data sets. We described and compared various methods used to avoid this problem, namely the initialization of the fMLLR transform and a linear combination of basis matrices varying in the choice of the basis estimation (eigen decomposition, factor analysis, independent component analysis and maximum likelihood estimation). Initialization methods compensate the absence of the test speaker's data utilizing other suitable data. Methods using linear combination of basis matrices reduce the number of estimated fMLLR parameters to a smaller number of weights to be estimated. Experiments are aimed to compare results of proposed basis and initialization methods." . . "7499" . "neuveden" . "165837" . "The worst problem the adaptation is dealing with is the lack of adaptation data. This work focuses on the feature Maximum Likelihood Linear Regression (fMLLR) adaptation where the number of free parameters to be estimated significantly decreases in comparison with other adaptation methods. However, the number of free parameters of fMLLR transform is still too high to be estimated properly when dealing with extremely small data sets. We described and compared various methods used to avoid this problem, namely the initialization of the fMLLR transform and a linear combination of basis matrices varying in the choice of the basis estimation (eigen decomposition, factor analysis, independent component analysis and maximum likelihood estimation). Initialization methods compensate the absence of the test speaker's data utilizing other suitable data. Methods using linear combination of basis matrices reduce the number of estimated fMLLR parameters to a smaller number of weights to be estimated. Experiments are aimed to compare results of proposed basis and initialization methods."@en . "P(DF12P01OVV022)" . "[7F5B05249D7E]" . . . . . . . "Robust Adaptation Techniques Dealing with Small Amount of Data" . . "8"^^ . . . . "RIV/49777513:23520/12:43915990!RIV13-MK0-23520___" . . . "Robust Adaptation Techniques Dealing with Small Amount of Data"@en . "Robust Adaptation Techniques Dealing with Small Amount of Data"@en . "0302-9743" . "Zaj\u00EDc, Zbyn\u011Bk" . "23520" . "ASR, adaptation, fMLLR, robustness, initialization, basis"@en . . "M\u00FCller, Lud\u011Bk" .