August 16, 2007Her job: Lead member of technical staff, AT&T Labs - Research
Describe what you do in your current work situation?
I'm a lead technical researcher in AT&T Labs--Research, based in Florham Park, NJ. I work on text-to-speech (TTS) synthesis, one of several speech technologies represented in the IP and Voice Services Research Lab. TTS technology converts standard text input into audible speech output. It can be used in many ways, such as speaking for people who cannot speak (you may have seen TTS used this way by physicist Stephen Hawking), screen reading for blind or dyslexic individuals, email reading over a cell phone, reading name and address information over the phone, hands and eyes free navigation while you're driving (or piloting a plane), entertainment (games and movies) and educational applications such as teaching reading, writing, or second language learning. You can check out our web page and interactive TTS demos at http://www.research.att.com/projects/tts/ -- it's a very popular website and lots of fun to play with! I also lead a text-to-speech technology American National Standards working group. We're currently working on defining standard methods of evaluating the intelligibility of TTS output.
Why did you choose engineering?
I did not receive my undergraduate or doctoral degrees in engineering, but following my interests led me more and more to that field, beginning in graduate school. I did my graduate work at the Center for Cognitive Sciences of the University of Minnesota, and I was also a visiting affiliate of the Haskins Laboratories, New Haven, CT. A few years after receiving my doctorate, I was awarded a five-year Research Career Development Award through NIH (National Institutes of Health), and I chose to spend much of my time in the Speech Communication group of the Research Lab of Electronics in the Electrical Engineering Dept. at MIT. It was a very simulating environment for me. After my award ended, I was thrilled to accept a job in a speech technology department within AT&T Bell Laboratories.
Where did you go to school and what degree(s) do you have?
University of Minnesota, Minneapolis: B.A. & Ph.D. * Postdoctoral work at Research Lab of Electronics, MIT
What kinds of activities have typically been part of your work?
My academic research focused on the human perception of speech sounds: I used computer programs to synthesize speech and similar sounds from detailed acoustic parameters that I specified. The audio output was recorded and used as stimuli in perceptual experiments with listeners. I also made acoustic measurements of natural recorded speech. With data about the acoustic properties of speech and the way people identified synthetic speech sounds and discriminated between them, I built and tested computer models that tried to duplicate how various speech sounds are recognized by humans. My work in speech technology at AT&T is on improving the quality of text-to-speech synthesis. For TTS, standard text is the input rather than acoustic parameters. This requires work on text processing (automatically identifying and appropriately handling abbreviations and other symbols that occur in text), syntax (automatically identifying nouns, verbs, and other parts of speech), word pronunciation, and prosody (the rhythm and melody of speech). Our TTS system uses a large database of recorded and carefully labeled small units of human speech as building blocks to synthesize any desired utterance. It uses an algorithm to select automatically the most appropriate speech units for a particular utterance. Much of my work has focused on ensuring a high quality accurately labeled speech database, on improving intelligibility and prosody, and on experiments to evaluate the most fruitful algorithms and approaches to synthesis. Part of what I do as a researcher also involves statistical analysis of experimental data, writing research papers, giving talks about my work, publishing articles in scientific journals, and writing books or book chapters.
What do you like best about being an engineer?
Speech technology is especially fun because it is highly interdisciplinary, and combines several of my major interests. Besides the intrinsic enjoyment I find in technical work and problem-solving, it's very rewarding to have contributed to a technology that people in the real world use to help them communicate.
Which of your career accomplishments are you proudest of?
I'm proud to be or have been part of the several excellent research organizations I mentioned above. In my previous academic research career, I'm most proud of some classic work that I did on auditory modeling of speech perception. I'm very proud of the unprecedentedly natural voice quality of the TTS system we're developing at AT&T Labs Research, and of being able to make substantial contributions to that effort. In particular, I'm proud of the work I've done to improve the quality of synthesis of the female voice (which, for various reasons, was a greater challenge than male voice synthesis). Our TTS team created what I believe was the first truly high quality female synthetic voice.
What challenges have you met and conquered in your pursuit of an engineering career?
When I was a student, I personally didn't know a single woman in engineering. Although I was an excellent science and math student, no one had ever suggested to me the possibility of studying engineering, and I didn't really understand at that time what a wide variety of interesting work engineering could encompass. Fortunately, the research and technical area that most interested me, speech communication, is multi-disciplinary, so I could still make a contribution to the technology without an engineering degree.
Please tell us a little about your family.
I'm the proud mother of three grown children -- one son and two daughters. My family is very important to me.
What are your short-term (1-2 years) and long-term (10+ years) goals?
My short-term goal is to continue to improve the naturalness of TTS so that for standard applications (such as man-machine dialogues, navigation, or email reading) it sounds essentially equivalent to recorded human speech. My longer-term goal is to expand the range of expressiveness of TTS, and to see our TTS system used in a wider variety of real-world applications.
What (or who) had/has the greatest influence on your life choices?
The greatest influence in making that choice was my experiences and colleagues while I was a research affiliate at MIT. Wherever my expanding interests took me, there always seemed to be a world expert in that field just down the hall. It was such an exciting and absorbing time for me that some nights I could hardly sleep, and I'd have dreams about new research ideas!
What advice would you give to a young woman considering a career in engineering?
1. Look into a variety of different areas in engineering (and in other disciplines too) and pursue what you find most fascinating.
2. An excellent, stimulating academic or work environment can inspire you to achieve more than you thought possible.
Describe something about your life outside of work: your hobbies, or perhaps a favorite book.
I enjoy spending time with my family and friends, travel, art, creative projects of various kinds, reading (current favorite: Ian Rankin's Rebus series), nature and being outdoors, especially hiking. I like to cook and I'm an avid gardener.