The next generation of SoundSense Learn™ – how using two AI engines helps us break the personalisation barrier
With our first generation of SoundSense Learn™, we set out to give as much personalisation power to the user as possible, without requiring them to delve into the mechanics of their hearing aid settings. To achieve this, we leveraged the immense processing power of artificial intelligence to create a personalisation experience that guided the user to a better sound via a series of A/B comparisons.
This was a huge success with hundreds of thousands new listening profiles being created as a result. However, we felt that there is a limit to how much personalisation can be offered before the benefits outweigh the burden on the user. We also know that not all users have the desire to be so engaged with their hearing aids and want their sound to be personalised automatically.
A new way to an instantly better sound
Our solution to this was to introduce another AI to deliver a second path to personalisation. This method analyses the SoundSense Learn™ evaluations of users all over the world to create a series of automatic recommendations that can be instantly accessed by users. Widex Lead Audiologist Oliver Townend explains:
“When we looked at this data and analysed not only the programs created, but all of the comparisons made along the way, we could build probabilistic models that could accurately predict – within a given moment, within a given listening intention – two ideal settings that will help most users most of the time. Not chosen by us but powered by thousands of users’ data that came before them”.
Users can find this new option (called ‘Made for You’) alongside the traditional SoundSense Learn™ (under ’Create your own’) in the My Sound section of the WIDEX MOMENT™ app.
To hear a new Made for You recommendation, the user simply selects the most relevant listening situation from a list and the new sound profile is activated. Instant access to a better listening experience – made for users, by users.
The art of taking the complex and making it simple
It’s easy to underestimate what it takes to create a feature like SoundSense Learn™. AI is a term that we hear everywhere these days, but we should remember that when it’s applied in the right way, it can help us in ways that have never before been possible. “This is truly empowering for users, because it allows them to help themselves in the moment”, says Oliver, “This could significantly reduce unnecessary clinical appointments where they're bringing back listening problems from the real world into the clinical setting, where there is really no relationship between the place they’re having the problem and where they're trying to have that problem fixed”.
The dual AI used in SoundSense™ Learn must parse huge amounts of data in real time in order to present user with a simple, intuitive and immediately beneficial experience. This is all possible not only because of the expertise, skill and experience of our innovation department, but also because, at WIDEX, we remain committed to one clear and simple vision – to bring the best and most natural sound to all users. That means giving the opportunity of personalisation both to those users that want to be more hands-on with their hearing aids, and to those that just want a recommendation tailored to them.