In this project we are working on the development of truly ``secure'' speech processing technologies. Project investigators include Bhiksha Raj, Manas Pathak, Jose' Portelo, and Isabel Trancoso of INESC (Portugal). This project has the following grants:
- NSF grant no. 1017256
- FCT-Portual: PTDC/EIA-CCO/122542/2010
- Cisco grant, 2010
Deriving semantic information from multimedia recordings
In this project we are developing "audio grammars" to derive semantic information from youtube style audio. Project investigators include Bhiksha Raj, Rita Singh, Richard Stern and Sourish Chaudhuri. This project has been funded by the following grant:
- IARPA "Aladdin" program, through Draper labs.
- Adobe grant, 2010.
SenseBank: A semantic Coherence Dataset
In this project we atttempt to collect a semantic-coherence data set from the mechanical turk. This was funded through a supplemental NSF grant.
In this project we atttempt to develop "compositional models" that can derive the underlying latent structure in audio recordings, for various purposes such as signal separation and enhancement, speech recognition, and audio analysis. Investigators include Bhiksha Raj, Rita Singh, Sourish Chaudhuri and various external collaborators and interns. The funding sources include:
- Google grant, 2009
- Google grant, 2010