Incremental Processing for Flexible and Adaptive Speech Output?

old_uid14800
titleIncremental Processing for Flexible and Adaptive Speech Output?
start_date2014/12/11
schedule13h
onlineno
location_infoAmpère
summaryHuman speakers and listeners process spoken language incrementally, piece-by-piece, and just-in-time, with the different processes involved working concurrently, on different time scales, and with varying degrees of specificity and flexibility. Typical language processing systems, in contrast, use a non-incremental pipeline architecture which can be an obstacle in interactive use-cases, where the future context is only partially known. I present the IU approach towards incremental processing in general and, focusing on incremental speech output in particular, I outline ideas how it could be used in very-low-delay applications, such as simultaneous interpreting, or speech/gesture coordination.
responsiblesHueber