Artificial Intelligence in Hearing Aids

Artificial Intelligence is the buzz word of the year. But what does it mean that a hearing aid uses AI?

The number one concern for individuals with hearing loss is hearing over background noise. Whether the noise is the car, dishes clanking, water running, or other conversations in a crowd, improving the signal to noise ratio is the goal of hearing aid technology.

For many years, hearing aids have used techniques like directional microphones to reduce sounds from behind the user. They have also tried to identify which frequency bands contain speech and which do not by looking for the common modulation pattern of speech. They can do a pretty good job of canceling out a steady noise like a fan running once it has stayed stable for 30 seconds.

But what about a situation that is more dynamic? What if you turned your head and are not looking at the person you are trying to hear? It can be easy for the other sounds to overwhelm the person you are trying to understand.

Now several hearing aid manufacturers have incorporated a deep neural network (DNN) into their hearing aids. They have trained their hearing aid chip on millions of sounds. What this means is that the hearing aid can do a better job of picking out a voice from the noise because the hearing aid can think - “that is a vacuum cleaner, that is a bird chirping, that is a man’s voice talking” You still need to hear the environmental sounds, but the voices can really 'pop out’ from the background when the hearing aid knows what is speech and what is not. These modern hearing aids can also decipher how many people are talking at once by the characteristics of the different talkers voices.