Our Past Projects
EVALUATIVE RESEARCH
GIFT Atlantic High School Pilot Project
The GIFT High School pilot project in Halifax, NS, Canada
was made possible by the generosity of the GIFT (Grocery
Industry Foundation Together) Atlantic Foundation. The goal
of the project was to explore how Liberated Learning speech
recognition technology could improve access for students with
disabilities in high school classrooms. The project gave the
Liberated Learning team the opportunity to explore the
transferability of speech recognition technology to the high school
classroom and to research the impact on high school students and teachers.
Work began in early 2006 to evaluate technology needs and train teachers
at the Halifax West High School. Grade 11 classes in English as a Second
Language, Oceanography and Social Studies were the pilot subject areas.
Following an exhaustive analysis of data from this first trial, work
began on a second phase of pilot testing. Partnering with the South
Shore Regional and Annapolis Valley Regional School Board, testing
began in the spring of 2007 at Park View Education Centre, Horton
High School, and Avon View High School. A province wide forum to
communicate project results and discuss future directions was conducted in January 2008. A comprehensive report titled A Study of the Transfer of Speech Recognition Technology from University Classrooms to High School Classrooms is available in our Resource section
Captioning and Transcription in Math Science
Lead: University of Kentucky
This research examines the use of computer-driven,
real time captioning and transcription to enhance student
learning in math and science. One key feature of this research
is the relationship between captioning display modality and
student performance. Captioning in the classroom research
has historically used whole-room projected display and/or
a laptop computer at each individual student's desk to deliver
captioning. However this project will use heads-up, micro
display glasses that allow users to focus on communication
cues, such as teacher body language or gesturing. This modality
also helps students attend to any visual material presented
in the classroom while simultaneously receiving real time
captioning. (2005-06)
- Primary impacts
The broader impact of captioning in the classroom will be
an increase in comprehension and retention of educational
content for all students. This increase in comprehension
and retention may ultimately result in higher levels of
classroom participation (knowledgeable students are more
confident students) and better performance in terms of grades.
More specifically, introducing captioning and transcription
targets five primary learning areas:
- Captioning will reduce working memory demand by simultaneously
displaying multiple lines of text allowing students
to "catch-up" on verbal information as they
look away to follow the teacher, take notes, or focus
on visual classroom material.
- Reading via captioning provides an additional learning
stimulus that provides supplementary reinforcement of
classroom material.
- Captioning has been shown to improve the oral reading
rates and word recognition of both disabled and non-disabled
students.
- Reading fluency in science and math is critical as
the teaching focus in both areas becomes increasingly
"real-world-problem" centred.
- Captioning helps students distracted by non-centralized
elements of a lecture to concentrate on the central
theme of the material being presented. Captioning is
an additional aid to student comprehension in vocabulary-heavy
science and math classes.
- Transcript Uses
The transcripts available at the end of class;
- allow students to easily extract and contextualize
vocabulary words from a given lecture
- provide supplements to notes taken during class
- provide a study aid for tests (especially in the review
of step-wise procedures),
- allow students to review verbatim lecture material
- allow students to review class discussion
RESEARCH AND DEVELOPMENT PROJECTS
ViaScribe Multi-speaker Feasibility Study
Lead: Saint Mary's University, IBM Research, University of
Southampton
The goal of this research project is to evaluate the feasibility
of using ViaScribe with multiple speakers contributing to
the same session. This study will explore promising alternative
routes for implementing multiple speaker capability in ViaScribe.
Two approaches to multiple speaker use will be evaluated:
an approach in which each speaker uses an individual computer
and an approach in which each speaker uses a different audio
input to one computer running multiple instances of ViaScribe.
This project will specify workable and optimal configurations
for each approach. The impact of several configuration variables
on error rates will be evaluated. (2005-06)
Multilingual Display for English Second Language
Learning
Lead: University of the Sunshine Coast
The global economy has resulted in English being used as
the main medium of communication for business and the internet
and this has led to international students going abroad to
learn English. Research is required to determine whether Japanese
or Chinese and English can be synchronized and displayed simultaneously
with other multimedia on one screen and if students find this
a useful method of improving their English language skills.
(2005-06)
ViaScribe for English for Academic Purposes
Lead: Massey University, University of the Sunshine Coast
The purpose of the study will be to investigate if, and to
what extent a metacognitive strategy training course supported
by speech recognition can help English for Academic Purposes
(EAP) students to recognize elements of lecture discourse
thus enhancing their listening comprehension and note taking
skills.
The specific objective of this project is to investigate
whether speech recognition in the form of digitized text can
more effectively support the teaching of strategies, lecture
discourse structure, and note taking ability than a traditional
EAP textbook course. This would be measured in listening test
performance scores, note taking quantity and quality, and
number of listening strategies used. (2005-06)
CORPORATE AND PUBLIC ACCESS
Since ongoing applied research has demonstrated the viability
of using the Liberated Learning approach in university settings,
the consortium has initiated a number of initiatives to introduce
enhanced accessibility into other environments.
Baddeck Liberated Learning Showcase
Alexander Graham Bell developed new teaching techniques that
assisted students with hearing disabilities in the late 1800's.
The Alexander Graham Bell Institute at Cape
Breton University, in Nova Scotia, Canada adopted the
Liberated Learning concept at The
Alexander Graham Bell National Historic site, administered
by Parks Canada, in Baddeck, Nova Scotia from 2004 to 2006.
The Baddeck Liberated Learning Showcase brought the newest
computer technology and the most recent software and technical
knowledge to help audiences with disabilities learn about
history and advancements in technology. This Showcase centrepiece
provided an interactive display of speech recognition technologies,
giving the research team the opportunity to demonstrate different
techniques for applying speech recognition in a public setting.
The Showcase offered a real life use of the client-server
and personalized display research and development projects.
The tour guide's commentary was captured by IBM ViaScribe
and relayed to the server. Individual audience members then
connected to the server using a handheld device. Each person
was able to personalize the display that the device received
from the server. The user selected font size and colour as
well as background to customize the screen for individual
viewing preferences. Each audience member could then read
the text of the tour guide's speech as it was delivered, gaining
full access to the information even at the back of the crowd.
COMMUNITY ACCESS
Universal Liberated Learning Phase I
Under the auspices of the Social Development Partnerships Program and through the Canadian Government’s Office of Disability Issues, a team comprised of Saint Mary's University, Trent University, and IBM Research completed a one-year project focused on engaging National Disability Organizations (NDOs) in Liberated Learning activities.
The most important project outcome was the signing of Memorandums of Collaboration (MOC) with three leading organizations:
- The Learning Disabilities Association of Canada (LDAC)
- The Canadian Hard of Hearing Association (CHHA)
- The Canadian Hearing Society (CHS)
The project team worked closely with LDAC, CHHA, and CHS to better align Liberated Learning with the actual needs of the Canadian disability community. The project not only raised awareness of Liberated Learning in the disability community, but also identified knowledge and technology gaps through direct interaction and analysis. A number of intriguing “application storyboards” emerged, which described real life usage scenarios for advanced speech recognition.
Universal Liberated Learning: Phase II
Saint Mary's University, Trent University, and IBM Research collaborated with leading National Disability Organizations - the Learning Disabilities Association of Canada, the Canadian Hard of Hearing Association and the Neil Squire Society -to research how emerging speech recognition technologies can increase information accessibility. ULL Phase II was funded through a contribution agreement with the Social Development Partnerships Program, Office of Disability Issues, Government of Canada and Saint Mary's University.
Building incrementally on a successful Phase I Development Venture, this project implemented three pilot tests of the Liberated Learning concept in educational, workplace, and virtual environments.
Education Pilot - This pilot was coordinated by the Learning Disabilities Association of Canada through their provincial office in Nova Scotia (LDANS) and implemented at Landmark East School in Wolfville, NS. Established in 1979, Landmark East is Canada’s international boarding and day school for students with learning disabilities. In one of Landmark East’s classrooms, students diagnosed with various learning disabilities were given access to SR generated multi-media transcripts to help them prepare for exams and complete assignments. Both the teacher and students provided feedback on how the SR generated transcripts impacted preparation for exams and completion of assignments. The project team analysed how the transcripts were used and how they enhanced the learning process.
Workplace Pilot - Canadian Hard of Hearing Association investigated how speech recognition can facilitate real time communication for employees who are hard of hearing. The pilot saught to understand the inherent speaker-training requirements and how various levels of SR accuracy impact the communication access needs of hard of hearing consumers. The analysis of these two components took place in a traditional office setting including one on one conversations and group/meeting settings.
Virtual Access Pilot - This pilot focused on enhancing the virtual learning environment of Neil Squire’s Distance Learning classroom, Wimba (http://www.wimba.com). The majority of these classes provide career and computer skills enhancement for people with a wide range of disabilities. During the pilot, course instructors used the real time text generated by ViaScribe to Close Caption the class’s webcasts. In the second part of the pilot, instructors will use the transcribed multimedia outputs to allow current students to review classroom presentations, allow students that are deaf or hard of hearing to review transcripts of class, and allow access for students that could not attend. Furthermore, Neil Squire will create a series of accessible lessons that can be accessed by past or future students as needed.
Universal Liberated Learning Phase III
Within the scope of phase III, the project team utilized a results-based approach focusing on two simultaneous activity streams.
First, the team scaled usage of speech recognition in educational, workplace, and virtual pilot applications, building incrementally from Phase II activities. In special classrooms for youth with severe Learning Disabilities, instructors used speech recognition as an alternative to conventional note taking. In a workplace setting, speech recognition will facilitate real-time communication between employees with Hearing Disabilities and hearing colleagues. A Virtual training program serving isolated clients with Physical Disabilities used speech recognition to improve access to online course content. All pilots gathered measurable evidence about technology benefits and limitations. Lessons learned were integrated into a transferable model that was widely shared both internally through our rich national network and externally, to likeminded organizations grappling with information accessibility issues.
A second stream simultaneously investigated technological, contextual, and geographic innovations. LDAC explored the literary and learning benefits for adult Aboriginal students diagnosed with a Learning Disability through a unique work transition program. By analyzing distinctive geographic and infrastructure challenges facing Canada’s Northern service providers, a implementation model was developed to support dissemination of new technologies and techniques. Trent University also developed a prototype “speech recognition in a box” solution for portable real time captioning.
|
