Main area of research
Sharmistha Sarkar Gray is a Senior Research Scientist in Acoustic Modelling Algorithm group of Dragon Research. She joined Nuance at November 2012 and has been responsible for acoustic model building in various languages for TV and car domain. She was also involved in building the child specific acoustic models and create child speech data base and coordinate transcriptions.
Sharmistha has a Ph.D. from University of Colorado, Boulder on automatic speech recognition and language analysis for accented speech of US English. Prior to joining Nuance, she has worked at Avaya Inc., and Lena Research Foundation. She specializes in child speech processing and was involved in the study of automatic tracking of language growth and early detection of autism and language delay for infants and toddlers. She has also coauthored more than 25 journal and conference papers and 7 international patents, including a paper on automatic autism detection which was ranked as the best paper in “2010 Top Ten List of Achievements in Autism Research” by Autism Speaks.
Sharmistha is also a proud mother of two kids (a 4 year old girl and a 5 year old boy) who keep her busy and always entertained with first hand child speech analysis right at home.
Adult-targeted automatic speech recognition (ASR) has made significant advancements in recent years and can produce speech-to-text output with very low word-error-rate, for multiple languages, and …