Machine Learning Challenges for Hearing Aids (Clarity-2022)

29th June 2022, Virtual Workshop

The workshop is now over. Links to the papers are now available on the programme page and the full results of the Clarity Prediction Challenge are published on the results page.

To allow the development of better hearing aids, we need better ways to automaticallly evaluate the speech intelligibility of audio signals. In particular, we need prediction models that can take the audio produced by a hearing aid and the listener's characteristics (e.g. audiogram) and estimate the speech intelligibility score that the listener would achieve in a listening test.

This virtual workshop will report the outcomes of the Clarity Prediction Challenge, the first machine learning challenge to consider the task of hearing-impaired speech-in-noise intelligibility prediction.

The challenge was based around a novel dataset of hearing-impaired listeners responses to hearing-aid algorithm processed speech-in-noise. The challenge was launched in November 2021 and ran until March 2022 and attracted the submission of 16 different systems from 9 separate teams.

This virtual workshop will feature

• announcement of the challenge results
• presentations from participating teams
• a keynote talk
• award of prizes for best system and best student contribution
• a discussion of future directions for hearing aid machine learning challenges.

Programme details now available.

Note, registration is free but prior registration is required.

Clarity-2022 Workshop

Organisers

• Michael Akeroyd
  University of Nottingham
• Will Bailley
  University of Sheffield
• Jon Barker
  University of Sheffield
• Trevor Cox
  University of Salford
• John Culling
  University of Cardiff
• Lara Harris
  University of Salford
• Graham Naylor
  University of Nottingham
• Zuza Podwinska
  University of Salford
• Zehai Tu
  University of Sheffield
• More info

Dates

• Registration opens: 15th May
• More info