Art is one of the few things which truly separate us from artificial intelligence. Even though there are robots which can write music, they don’t consider the emotions that it elicits, they merely weave notes into patterns determined by the physical parameters of the notes. All that may change, thanks to the work of Japanese researchers.

This is a Brain Music EEG headset. The machine learning algorithm is hooked up to the listener’s brain, learning and adapting as it goes. Image credits: Osaka University.

Although it wouldn’t seem so at a first glance, music and mathematics have a lot in common. There’s an almost mathematical beauty to musical patterns, which musicologists have remarked for centuries. Therefore, it would seem, Artificial Intelligence (AI) should have a decent time writing songs — but that doesn’t really happen.

While AI can write OK songs, there’s one aspect which it can’t incorporate: feelings. Really good songs mess with our feelings, they send us to despair or bring great joy to our hearts, and that’s something machines can’t do. At least, not yet.

An international research team led by Osaka University together with Tokyo Metropolitan University, imec in Belgium, and Crimson Technology has released a new machine-learning device that analyzes people’s brain waves as they listen to songs. Their machine learns how listeners feel and adapts to it, giving them more of what they want to hear.

“Most machine songs depend on an automatic composition system,” says Masayuki Numao, professor at Osaka University. “They are preprogrammed with songs but can only make similar songs.”

Numao figured that if he can somehow tap into the listener’s emotional state and see what works for him/her, the AI could deliver more of the same. Basically, he found a way to artificially insert feelings into AI songs. Users listened to music while wearing wireless headphones that contained brain wave sensors. These sensors detected EEG readings, which the robot uses to make music interactively reacting to the listener. The music was then created by the machine through Musical Instrument Digital Interface (MIDI) technology on the spot and played in a rich tone using a synthesizer.

“We preprogrammed the robot with songs, but added the brain waves of the listener to make new music.” Numao found that users were more engaged with the music when the system could detect their brain patterns.

So far, no audio samples have been released, but the machine was featured at the 3rd Wearable Expo in Tokyo, Japan, receiving positive reviews from volunteers.

Aside from being a potential breakthrough in AI, this technology could also have real-life applications. Having customizable music that adapts to your feelings could for instance boost your productivity when you’re working or motivate you when you’re working out. As we all know, finding the music that adapts to your feelings can sometimes be pretty daunting.

Enjoyed this article? Join 40,000+ subscribers to the ZME Science newsletter. Subscribe now!