HLT4LL: Maxine Eskenazi

Dr Eskenazi is Principal Systems Scientist in the Language Technologies Institute (LTI) at Carnegie Mellon University where she is the director of the Dialog Research Center. She received her B.A. in French and Education from Carnegie Mellon University. Her Doctorat de Troiseme Cycle in Computer Science is from the Universite de Paris 11. She was a Chargee de Recherche at LIMSI-CNRS in Orsay France before becoming a Systems Scientist at Carnegie Mellon. Her interests cover language learning systems using language technologies, crowdsourcing for speech processing and spoken dialog systems. She has published extensively in these areas and in the area of speech processing in general. She is on the scientific committees of many organizations and is Founder and was first Chair of the ISCA special interest group, SLaTE (Speech and Language Technologies for Education). Dr. Eskenazi has been a panelist for NSF and ChistEra. She serves at present as head of admissions for the LTI MLT and PhD programs.

Speech Technology and Language Learning
Language learning systems are just beginning to use speech technology. It is a promising marriage that, if done correctly, holds much potential. With a point of view inspired by findings in cognitive psychology, we will first discuss what we believe speech processing has to offer. We will then look at the present state of speech technology in language learning. First we will examine the tension between two types of uses of this technology: learning drills and immersion exercises. We will offer some solutions . Then we will explore the present limitations of the use of automatic speech recognition and discuss what can be done to create reliable software. Finally, we will attempt to peer into the future, discussing the trends and characteristics of the systems to come.