Results from some projects working with Speech Recognition

1
Results from some projects working with Speech
Recognition

• CALL Centre,
• University of Edinburgh, 2002

2
The Potential.

• For so long they dyslexic pupils have had to
  rely on others to scribe for them that they have
  begun to despise themselves and switch off.
  VoiceText Dragon Dictate allows these pupils to
  respond independently, gives them control, and
  gives them the freedom to express themselves as
  others do.
• Elaine Donald, Perth High School, 1998

3
Perth High School, Scotland

• Elaine Donald, Principal Teacher, Support for
  Learning
• Started using VoiceText (Dragon Dictate) in 1995
• 12 pupils from S1 to S5
• 11 with specific writing and spelling
  difficulties, all with reasonably good word
  recognition skills

4
Perth High Results

• More than 30 pupils using Speech Recognition by
  1998, sharing six laptops
• Benefits reported
• Increase in volume and quantity of work
• Great improvements in self-esteem
• Improvement in reading ability
• Greater motivation to write
• More sophisticated language and vocabulary
• Better punctuation

Click here (340Kb) to read the full report
5
Miles, Martin Owen Devon LEA

• In 1998, Martin Miles, Di Martin and Jim Owen ran
  an evaluation using IBM Simply Speaking
• 11 pupils aged between 14 and 16 were trained to
  use Simply Speaking and progress monitored over
  10 weeks
• 8 pupils were located in a specialist Dyslexia
  Support Centre in a secondary school
• 3 pupils were in a Support for Learning
  department in a different secondary school
• 8 pupils were at Key Stage 4 and 3 at Key Stage 3

6
Miles et al reported

• Improvement in recognition accuracy
• Improvements in hand-written work quantity,
  spelling, and quality
• Average reading age improved by 13.4 months over
  10 weeks
• Average spelling age improved by 6.1 months over
  10 weeks
• Improved self-esteem and motivation

Click here (1.8Mb) to read the report
7
BECTA Speech Recognition Project

• From April 1998 to March 2000 BECTA ran a project
  involving twelve centres and schools. The centres
  worked with student with a range of ages, skills
  and needs (e.g. physical difficulty, dyslexia,
  visual impairment)
• They used both Discrete and continuous
  programs
• Each centre addressed different research
  questions and produced a report about the work.

8
Some results from BECTA

• Pupils learned some spellings through frequent
  use
• Pupils began to recognise words they could not
  previously read
• Some pupils showed improvement in diction
• Slow, clear speech is important, although pupil
  with poor speech were still successful
• The programs are more accurate with long words,
  so pupils used more complex vocabulary than when
  writing
• Pupils learned to draft ideas and outline work
• It was time consuming for staff and pupils

9
More results from BECTA

• The programs are too complex for most students
  with V.I. to use independently
• H.I. students with poor speech had low accuracy
• Motivation to use SR was not a problem with most
  students even when accuracy was low
• Students were proud of the results of their
  dictation
• Peers and staff praised the quality of their work
• Many students needed help to complete the initial
  training

Click here (165Kb) to read the BECTA Project
Final Report (PDF)
Click below to go to the Becta web site
http//www.becta.org.uk/technology/speechrecog/ind
ex.html
10
BECTA project schools - click to read the report
from a school

• ACE Centre (88Kb) Advisory Trust, Oxford
• Bishop Thornton CE Primary School (281Kb) ,
  Harrogate
• Black Moss Special School (32Kb) , Skelmersdale
• Exhall Grange School (38Kb) / RNIB, Coventry
• Kingsbridge School (99Kb) and Community College,
  Kingsbridge
• Kingston upon Hull SENSS (18Kb) (Kingston
  Secondary School and Amy Johnson Secondary
  School)
• Kirk Hallam Community School (147Kb) , Kirk
  Hallam
• New Tunmarsh Centre (60Kb) , London
• Stanbridge Earl School (22Kb) - Accelerated
  Learning Centre, Romsey
• St Thomas of Canterbury RCP School (22Kb) , Bolton

11
The BECTA report says that success with speech
recognition requires the Three Ts

• Time
• To choose and set up the software
• For staff to learn the system
• Training
• Individual training is needed for students
• Technology
• Better computer better results
• Technical support is often needed

12
Speech Recognition in SchoolsReports and
articles

• Click to read the reports (Word or PDF format)
• Ready for Happy Talking?, (20Kb) Mick Donegan,
  ACE Centre, 2001
• Introducing Speech Recognition Technology for
  writing, to a young person with special
  educational needs, (60Kb) Mick Donegan, ACE
  Centre, 2000
• Speech Recognition Systems (165Kb), Allan Wilson,
  CALL Centre, 2000
• Usable Voice Recognition Software The Mark
  College Experience!, (14Kb) by Malcolm Litten,
  Mark College, 2000.
• Supporting Writing with Speech Recognition
  (168Kb), by Paul Nisbet, CALL Centre, 1999

13
Some useful web sites

• BECTa (British Educational Communications
  Technology Agency) http//www.BECTa.org.uk/inclusi
  on/speechrecog/index.cfm Information, reports
  and case studies from the BECTa Speech
  Recognition Project. Good place to start looking
  for information about speech recognition in
  schools.
• CALL Centre (UK) http//callcentre.education.ed.a
  c.uk Electronic versions of this Pack plus other
  resources on speech recognition, plus information
  on a range of curriculum-based topics related to
  augmentative communication and assistive
  technology.
• Iansyst Ltd. http//www.dyslexic.com Specialist
  supplier for speech recognition in education.
  This excellent site is regularly updated with
  comparative reviews and the latest pricing
  information.
• Speaking to Write (USA) http//www.edc.org/spk2wrt
  This is an American project exploring the use
  of voice recognition technology to support
  secondary age students with disabilities.
• Words Worldwide Ltd. http//www.keyspell.com
  Information on Keystone ScreenSpeaker and Roamer,
  and voice recognition systems.

14
End