Machine Learning Challenges to Revolutionise Hearing Devices

One in six people in the UK has a hearing impairment, and this number will increase as the population ages. Yet only 40% of people who could benefit from hearing aids have them, and most people who have the devices don’t use them often enough. A major reason for this low uptake and use is the perception that hearing aids perform poorly.

We are organising a series of machine learning challenges to enhance hearing-aid signal processing and to better predict how people perceive speech-in-noise.

Enhancement challenge

The first Enhancement Challenge (CEC1) is about enhancing the processing of hearing aid signals to improve the intelligibility of speech in noise. The scenario is a living room, one target talker and one interfering noise source. The challenge is now closed. The challenge workshop was held on the 16th and 17th September 2021. You can find out more about the workshop, and find links to papers, slides and videos of presentations, here.

Prediction Challenge

The first Prediction Challenge (CPC1) launches in Autumn 2021. Please see our pre-announcement. This challenge is about predicting the speech intelligibility score when listeners listened to speech-in-noise processed by hearing aids, and told us what words they heard.