Research Work
My primary interests are in statistical machine learning, particularly applied to domains of Automatic Speech Recognition (ASR) and Natural Language Processing (NLP). Although ASR systems perform reasonably well in a number of applications, their performance degrades rapidly when conditions are no longer ideal. I am interested in examining techniques by which ASR systems may be made robust in the face of background noise and variations in accent.
My other interests include Linguistics and Cognitive Science.
Publications
- R. Prabhavalkar, P. Jyothi, W. Hartmann, J. Morris, E. Fosler-Lussier. "Investigations into the Crandem Approach to Word Recognition", NAACL-HLT short
paper
session (to appear),
Los Angeles, USA, 2010.
- R. Prabhavalkar, E. Fosler-Lussier. "Backpropagation Training for Multilayer Conditional Random Field based Phone Recognition". Proceedings of
ICASSP, Dallas,
USA, 2010. (abstract | bibtex | slides )
- R. Prabhavalkar, Z. Jin, E. Fosler-Lussier. "Monaural Segregation of Voiced Speech using Discriminative Random Fields". Proceedings of Interspeech, Brighton, UK, 2009. (abstract | bibtex | poster | slides)