000083153 001__ 83153
000083153 005__ 20180317093239.0
000083153 037__ $$aREP_WORK
000083153 245__ $$aSequence Classification with Input-Output Hidden Markov Models
000083153 269__ $$a2004
000083153 260__ $$bIDIAP$$c2004
000083153 336__ $$aReports
000083153 520__ $$aWe present a training and testing method for Input-Output Hidden Markov Model that is particularly suited for classification of sequences in which class information accumulates over time. We discuss two such cases: the discrimination of mental tasks from sequences of EEG features, common in Brain Computer Interface research, and phoneme classification from sequences of acoustic features for speech recognition. The objective function is modified so that training focuses on the improvement of classification accuracy. For both tasks the algorithm performs significantly better than the alternative solution proposed in the literature, specifically designed for other types of sequences.
000083153 6531_ $$alearning
000083153 700__ $$aChiappa, Silvia
000083153 700__ $$0243961$$aBengio, Samy$$g140142
000083153 8564_ $$uhttp://publications.idiap.ch/downloads/reports/2004/rr04-13.pdf$$zURL
000083153 909CO $$ooai:infoscience.tind.io:83153$$preport$$pSTI
000083153 909C0 $$0252189$$pLIDIAP$$xU10381
000083153 937__ $$aEPFL-REPORT-83153
000083153 970__ $$asilviac04sequence/LIDIAP
000083153 973__ $$aEPFL$$sPUBLISHED
000083153 980__ $$aREPORT