The key objectives are:
·
Automated transcription
of emotional and accented speech in multiple languages in the presence of spontaneous
speech phenomena such as whispering and uncued language switching.
·
Enhanced text and speech
processing technology such as segment boundary determination in recorded
interviews, nomination of new categories and named-entity tagging during text classification - Automated Generation of Metadata and
Cataloguing
IBM will focus its initial
efforts on processing testimonies in English.