Getting to Know Your Fellow Researchers: Lori Lamel

What sparked your interest in speech and language processing?

This is an easy question. I sort of fell into it. I was a coop student in EECS and also a first year graduate student. My first teaching assistant assignment was with Victor Zue "Introduction to Electronics". We hit it off well and he asked me to be a TA for a new graduate course that he was offering in the spring based on the new book that had come out "Digital Processing of Speech Signals" by Rabiner and Schafer. It was challenging! I knew absolutely nothing about speech production (other than talking a lot!) - I was younger than most of the students in the course and well in the minority being a women in EECS at MIT. But working with Victor and Stephanie Seneff, as well as other colleagues in the Speech Group (led by Ken Stevens), was a wonderful experience and also sparked my interest in spectrogram reading. To this day I still trust my eyes more than my ears.

How do you think speech and language processing is changing the society for the new generation?

Speech is something that, for the most part, we all take for granted - at least in our native language.  So many people have the same high expectations for machines. I find some voice services are amazingly good, but I also am concerned that we will end up communicating less with other humans, and more with machines.  How many times have we seen young kids sending each other messages, rather than speaking together. I hope that the future will show that these technologies will facilitate interaction with others (technology enabled communication), and not just replace human-human communication.

What is your holy grail in speech and language processing? When will we achieve it?

Really understanding what people mean, and not just the words they say. I do not want to speculate on if we will ever achieve this in unconstrained contexts, with speakers of different languages.

Do you have any specific advice for students, junior faculty or others early in their careers?

Find a work environment where you can continue to learn. Find a supportive environment where you are at the same time comfortable but also challenged. Learn to accept that there will be some things that you are better at than others. Be open to collaborating with others with different strengths/competences and opinions. Always treat people with respect, even if they have different views and values.

Keep some time for yourself to do things only for you. At certain points in your career it can be hard to find the right balance between your professional and personal lives, but this is important. Sometimes you may need to just be willing to let some things go and should not feel guilty about that.

What development in the field has most surprised you? Was there a hard problem that turned out to be easy? An easy problem that proved surprisingly difficult?

One of the first problems I addressed during my Masters work at AT&T was endpoint detection. After all these years this is still a research problem, in particular in far field conditions. It is now more generally called speech or voice activity detection.

Oftentimes, the work that people get noticed for is not the same as the work which they find most exciting/rewarding/interesting. Which of your publications is your favorite? Why?

I have a hard time guessing what work is most associated with me. My most cited papers are about TIMIT. That was a fun project and I am glad to see how much the corpus and other corpora that have been derived from it have served the community.

I guess one of my favorite publications is based on work with Jean-Luc Gauvain that we called "Identifying non-linguistic speech features". We first presented this at the Human Language Technology workshop and at Eurospeech back in 1993, and later published an extended version "A phone-based approach to non-linguistic speech feature identification" in Computer Speech & Language. This paper took a unified view using phone-based acoustic likelihoods for speaker, language and gender identification.