Artificial Intelligence Has Learned To Identify The Songs Someone Is Listening To From Their Brain Readings

Saturday, January 30, 2021

The world of artificial intelligence has been moving fast for a number of years now. Are you up to date with the latest technology?

There are a lot of applications that AI and machine learning are applied to in our day to day lives. Anything from the music we listen to, our online searches, and transport needs are helped by artificial intelligence. However, the potential of AI goes so much further beyond what we currently have.

This point is highlighted by a 2018 announcement from a joint scientific research group from University Hospital Leipzig in Germany and the D’Or Institute for Research and Education in Brazil. A research project they worked on has managed to showcase that artificial intelligence can learn to identify the songs someone is listening to from their brain readings. A similar, more recent experiment has been conducted by Derek Lomas at Delft University of Technology in the Netherlands, which came to similar results.

As part of the experiment between the D'Or Institute for Research and Education in Brazil and University Hospital Leipzig in Germany, volunteers were played 40 pieces of music, ranging from jazz, rock, pop, and classical, as well as others.

While listening to the music, the volunteers were also having their brain scanned by an MRI machine. The neural fingerprint that was then captured was analysed by a computer to distinguish patterns in the brain that each piece of music created. There were a number of musical factors taken into consideration by the computer, these included tonality, dynamics, rhythm and timbre.

Once the AI had an opportunity to analyse the brain patterns, the researchers theorised that the computer would be able to reverse the order. Basically, now that it understood the patterns involved with each music genre in a person’s brain, it could analyse a person’s brain activity and say what music they were listening to. This technique is known as brain decoding.

The AI computer was tested by having two song options given to it, the AI had to choose which was the song that the person was listening to. It passed this test with an 85% score of correct answers. Following this trail, the computer was then given a further test, wherein 10 options were given and 1 had to be chosen. In this case, the computer was right 74% of the time.

What applications does this discovery have?

The implications of this can be far reaching and huge. As a new way of viewing and understanding brain activity, it can provide insights into the differences between a healthy brain and a dysfunctional one. Beyond this, it is a positive discovery for people who are unable to communicate properly. 

The hope is that this technology could one day develop into a way of giving voice to people by the AI analysing areas of a brain and projecting their voice for them. While this is not predicted to happen immediately, it is certainly an area of research that shows a lot of promise for the future.

In fact, there are already laboratories around the world that are able to transform brain signals into speech. Unfortunately, the accuracy of this technology is still not there yet, so application has remained within the lab. Within this environment it has been possible to create whole sentences that usually can be comprehended.

Currently, people who are unable to speak use technology that usually is based around them controlling a cursor with miniscule movements to select words and letters on a monitor. The most famous user of such a device is Stephen Hawking. These devices can be extremely slow, because a person has to type out words and letters one at a time. The average rate of this is ten words per minute, however, with mind mapping technology, this average can be raised up to 150 words a minute.

For some people it becomes all too easy to fall back on old tropes and stereotypes around artificial intelligence. Without understanding it, they can speak of an uneasiness around the technology and quote lines from films. While the technology is extremely powerful and has to be treated with a lot of respect, discoveries such as the one above show the huge potential for good that can be done with artificial intelligence.