Lady Gaga’s Bad Romance. Beyonce’s Crazy in Love. Adele’s Rolling in the Deep. They are all massive global hit songs, but what are the ingredients that make them so successful?
A team of researchers in the US have used a machine learning AI to predict hit songs with 97 per cent accuracy.
They did it by fitting a small group of volunteers with sensors designed to monitor their brainwaves as they listened to a set of 24 pop songs.
The approach is known as ‘neuroforecasting’ and aims to record the neural activity of the participants as they experience a sound, act or feeling.
Once they had the data, the team then used machine learning techniques and computational models to line up the participants’ neurophysical responses with how they rated the songs.
The researchersthen used a machine learning algorithm to figure out what was going on.
“By applying machine learning to neurophysiologic data, we could almost perfectly identify hit songs,” said research lead Paul Zak, professor at Claremont Graduate University.
“That the neural activity of 33 people can predict if millions of others listened to new songs is quite amazing. Nothing close to this accuracy has ever been shown before.
“The brain signals we’ve collected reflect activity of a brain network associated with mood and energy levels.”
Further, the model was able to predict the success of a tune with 82 per cent accuracy after the participants listened for just one minute.
The researchers suggest that the technique may help streaming services to predict the genres and styles that listeners are more interested in.
“This means that streaming services can readily identify new songs that are likely to be hits for people’s playlists more efficiently, making the streaming services’ jobs easier and delighting listeners,” Zak said.
“If in the future wearable neuroscience technologies, like the ones we used for this study, become commonplace, the right entertainment could be sent to audiences based on their neurophysiology.
“Instead of being offered hundreds of choices, they might be given just two or three, making it easier and faster for them to choose music that they will enjoy.”
The study was not without limitations. It only included a small number of songs and styles and the sample size, ethnic diversity and span of ages were relatively small.
However, the team believe that the same technique could be applied in the study of other kinds of art forms.
“Our key contribution is the methodology. It is likely that this approach can be used to predict hits for many other kinds of entertainment too, including movies and TV shows,” Zak said.
Read more about AI: