• July 11, 2002
  • By Sara Basson Program Director, Speech Transcription Strategy - IBM Research
  • Features

Speech Recognition and Accessible Education

Disabled populations are key drivers in many technological innovations. Many of these innovations are now so integral to mainstream markets that we can forget their historical roots. A fascinating list of a number of these innovations is chronicled by Steve Jacobs, of The Center for An Accessible Society. A couple of highlights are extracted below: · In 1808, Pellegrino Turri built the first typewriter, so that his blind friend could write intelligibly.
· Early speech synthesis work at Bell Labs, such as the “voice coder”, were originally envisioned as machines to help deaf people speak more intelligibly. The list continues, including such mainstream staples as captioned films, e-mail as a basic component of the Internet and scanners. In 1952, Bell Labs announced the first speech recognizer, which achieved 97% accuracy on digits one through 10, with speaker training. The race to “solve” the speech recognition challenge was on. Vocabulary size expanded over time, and user training requirements decreased. In 1998, continuous speech recognition products emerged, with IBM’s ViaVoiceTM and Dragon’s NaturallySpeakingTM. The vocabularies were large and expandable, the price was affordable, and users could speak in a normal voice as long as they spoke clearly. Speech recognition was now available to the mainstream retail marketplace. This was a dream come true for users with disabilities, waiting for just such a breakthrough. Users could now dictate documents into their computers, even if they had motor difficulties interfering with typing, or dyslexia interfering with writing. Disabled populations are once again pushing the envelope and prodding technologists to address their pressing need to receive accessible education. Imagine a regular classroom, for a deaf student. In order to succeed, the student will need intermediary access, such as a sign interpreter or a stenographer. These are expensive propositions, and not always available. Envision a student with motor or learning disabilities. To succeed in the classroom, these students must rely on “note-takers”, paid or volunteer. Again, the student (or school) may need to incur high costs, there will be uncertain quality, and a note-taker might not even be available. These were the problems facing St. Mary’s University in Halifax, Nova Scotia. Dr. David Leitch, director at St. Mary’s Atlantic Centre of Research, Access and Support for Students with Disabilities, opted courageously to do something about it by spawning the Liberated Learning Project (LLP). With a grant they received from the McConnell Foundation in Canada, and support from IBM, they embarked on a project to bring speech recognition to the college classroom. In the Liberated Learning environment, professors lecture via wireless microphones, and IBM ViaVoiceTM software converts speech to text. IBM Research developed software (IBM NetScribe) to manage the text output display and post-lecture editing requirements. The lecture appears on a large screen at the front of the classroom. Deaf and hard-of-hearing students can view the lecture in real time. All the students can obtain the text of the lecture online after the class and can download it, print it, convert it to Braille or listen to it via a voice synthesizer. The project team has noticed that international and non-disabled students also use the instantaneous display of the lecture as a reference check for their own notes and understanding. Furthermore, every student wants access to the speech recognition generated lecture notes. The project has expanded into an international initiative, including Saint Mary's University, IBM, Aliant Telecommunications, Deafax International, UK; University College of Cape Breton; the Alexander Graham Bell Institute; Ryerson University, Toronto; and the University of the Sunshine Coast, Queensland, Australia. The LLP continues to garner the interest of universities trying to improve access for students with disabilities. Keith Bain, the LLP project manager, hopes this will eventually receive widespread acceptance as a model for universities to better accommodate students in the classroom. The reach can extend beyond the university to lower school education, and to seminars and meetings in the workplace. Once again, the unique needs of disabled populations turn out to be not so unique after all, and technology advances for everyone. If you would like to learn more about participating in the Liberated Learning Project, go to http://www.liberatedlearning.com. To find out more about IBM’s accessibility initiatives, go to http://www.ibm.com/able. Dr. Sara Basson is manager of accessible research productization at IBM, and a member (and former president) of the Board of Directors of AVIOS. She can be reached at sbasson@us.ibm.com.
SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues