Experience
I am currently an Assistant Professor in the Electronic and Computer Engineering Department of the Technical University of Crete, where I teach undergraduate and graduate courses in digital communications, speech processing and speech recognition and lead a research project on speaker and dialect adaptation of speech recognition systems in collaboration with SRI International and a European Telecommunications Company.
I am also a Senior Research Engineer at the Speech Technology and Research Laboratory of the Stanford Research Institute (SRI International) in Menlo Park, California. There, in the period between January 1992 and February 1995, I developed a speech-recognition based system for the automatic grading of English pronunciation of non-native speakers, I investigated methods to improve the accuracy of large-vocabulary speech recognizers by modeling the temporal dependencies of the speech features and performed research in the area of speech to speech translation.
At SRI, I was a principal investigator for ARPA research contracts and and lead the speech recognition part of a commercial project that developed a system for spoken English to Swedish translation in the domain of air travel planning (ATIS).
I also performed research in the areas of speaker adaptation and acoustic modeling. I invented a new acoustic modeling technique, genones, that significantly reduced the error rate of SRI's DECIPHER speech recognition system. This modeling technique is scalable and can achieve the best trade-off between accuracy and amount of computation for a particular application. Genones are used by the recognition engine of Nuance Communications, a spin-off of SRI that provides high-performance software to financial services, telecommunications, and other markets.
Education Projects
Affiliations Experience Publications Personal info
