``Incorporating linguistic theories of phonological variation into speech recognition models,''
Royal Society Phil. Trans., to appear.
This paper describes the use of distinctive linguistic features to
represent acoustic variability of words for speech recognition.
Focusing on conventional HMM technology, we review implicit use of
linguistic features as questions in decision tree design for both
coarticulation and pronunciation modeling and describe possibilities
for more explicit use. The importance of conditioning on
(hierarchical) syllable and prosodic structure is discussed, and the
problem of modeling relative timing of feature-dependent acoustic cues
is raised as a key limitation of current models.
Return to SSLI Lab Publications