Maxine Eskenazi

Timeline

1. 1973–1994

   Doctorate and LIMSI-CNRS research

   She earns a BA in Modern Languages from Carnegie-Mellon University in 1973, a DEA in linguistics from the University of Paris VII in 1981, and a Doctorat de Troisième Cycle in computer science from the University of Paris XI in 1984, with a thesis on the acoustic characterization of French vowels for automatic recognition advised by Jean-Sylvain Liénard. From 1980 to 1994, she works at LIMSI-CNRS in Orsay, France, first as a graduate student and then as a tenured researcher. She joins the International Speech Communication Association at its formation in the late 1980s and runs its first newsletter.

2. 1999–2006

   NativeAccent and Carnegie Speech

   In 1999, she publishes a prototype for foreign-language pronunciation tutoring with automatic speech processing in Language Learning & Technology. She becomes president and chief technical officer of Carnegie Speech, a CMU spinoff, in 2001. In 2006, LTI director Jaime Carbonell names her as the inventor of the techniques behind the company's NativeAccent tutor, which exaggerates the gap between a learner's sound and the target sound so the learner can hear it.

3. 2005–2014

   Let's Go bus information

   With Antoine Raux, Brian Langner, Dan Bohus and Alan W. Black, she describes in 2005 how the Let's Go Public spoken dialog system gives Pittsburgh callers bus schedule information during off-peak times. Let's Go goes live to real users in March 2005 and answers the phone daily for the Port Authority of Allegheny County until 2014. Its data and system are used in over 22 theses and over 250 non-CMU publications.

4. 2006–2009

   The REAP vocabulary tutor

   In 2006, with Michael Heilman, Kevyn Collins-Thompson and Jamie Callan, she reports classroom results for REAP, an intelligent tutoring system that selects authentic web readings for each student's vocabulary practice. REAP gains acceptance at the University of Pittsburgh's English Language Institute. By 2009, she is an associate teaching professor in the Language Technologies Institute of Carnegie Mellon's School of Computer Science.

5. 2013–2018

   Crowdsourcing and zero-shot dialog

   In 2013, she co-edits the Wiley book Crowdsourcing for Speech Processing, and her team's system wins the 2013 Dialog State Tracking Challenge. With Tiancheng Zhao, she wins a SIGDIAL best paper award in 2018 for zero-shot dialog generation, a method that lets a neural dialog system handle a new domain from only a domain description. She also publishes Advanced Social Interaction with Agents with Laurence Devillers and Joseph Mariani through Springer in 2018.

6. 2021

   An ISCA fellowship

   In April 2021, ISCA names her a fellow for contributions to phonetics, language learning, spoken dialog systems and language resources, and for service to the community. As a principal systems scientist in Carnegie Mellon's Language Technologies Institute, she builds on GetGoing, a slow-speaking bus information system for seniors, to guide people through large buildings such as UPMC Presbyterian Hospital.

7. 2022

   The DialPort tools and retirement

   In 2022, with Jessica Huynh, Shikib Mehri and Cathy Jiao, she describes the DialPort tools, NSF-funded services for the dialog research community that include the DialPort Portal and DialCrowd. She retires from Carnegie Mellon in December 2022.

8. 2023–2025

   ISCA fellow selection committee

   She serves on the ISCA Fellow Selection Committee for 2023/24 and 2024/25. In 2024, she coauthors a study with Cathy Jiao and Aaron Steinfeld in which 198 people, 80 of them with self-reported disabilities, listen to spoken navigation instructions; a slower speech rate helps them recall numbers.