Time: 215pm-330pm, Friday, April 29, 2011
Place: Room 4102, CUNY Graduate Center, 365 Fifth Ave (34str&35str).
Speaker: Andrew Rosenberg (CUNY)
Title: Speech is More than Noisy Text: Automatic Prosodic Analysis Techniques and Application

Abstract:

Spoken material can cause headaches for text-based natural language processing techniques.  
Automatic speech recognition errors befuddle bag-of-word and n-gram models.  Disfluencies, 
and informal syntax can cause serious problems for parsers.  Common solutions to these 
problems seek to make traditional techniques more robust to spoken material.  But speech 
is more than a noisy sequence of words.  Through intonation, or prosody, speech conveys 
important communicative information which is largely untapped in spoken language 
processing.  This talk will describe previous and current approaches to the automatic 
analysis of intonation, and ultimately outline a path towards improved spoken language 
processing through prosodic analysis.

Speaker Bio:

Andrew Rosenberg is an Assistant Professor at Queens College (CUNY).  His research concerns 
spoken language processing and machine learning. He received his Ph.D. from Columbia 
University in 2009.