Shulou(Shulou.com)11/24 Report--

CTOnews.com, June 21 (Xinhua)-- artificial intelligence (AI) can accurately identify popular songs, which could threaten the work of music producers and talent show judges, according to a new study.

Scientists at Claremont Graduate University in California used neural networks, an AI technique that mimics the human brain, to analyze the neurophysiological responses of listeners to 24 songs. They found that by measuring the brain activity of a few volunteers, they could predict whether millions of people would like the songs. Their method is called "neural prediction" and its accuracy is as high as 97%.

Paul Zack, a professor at the university and senior author of the study, said at a media conference: "by applying machine learning to neurophysiological data, we can almost perfectly identify popular songs. The neural activity of 33 individuals can predict whether millions of people will listen to new songs, which is very surprising. I've never had such accuracy before."

CTOnews.com noted that this AI system can not only replace the judges of TV talent shows, but also improve the efficiency of streaming services. Tens of thousands of songs are released every day, and apps like Spotify, Tidal and Deezer struggle to choose which songs to add to the playlist. Previous attempts to identify songs that appeal to the public have only a 50% success rate.

In this experiment, participants were wearing scanners, and while listening to a group of 24 songs, they were also asked to provide their preferences and basic demographic data, and the scanners measured their neurophysiological responses. "the brain signals we collected reflect the activity of a brain network associated with emotional and energy levels," Zack said. "

The team found that even analyzing the neural response of the song in the first minute could achieve 82% accuracy.

"this means that streaming services can more effectively identify new songs that may become hits, making the work of streaming services easier and more satisfying for listeners," Zack explained. If wearable neuroscience techniques like the one we used for this study become common in the future, appropriate entertainment content can be sent to listeners depending on their neurophysiological state. Instead of having to choose from hundreds of choices, they only give them two or three choices to make it easier and faster for them to choose the music they like. "

The researchers point out that the technology can be used not only to identify popular songs, but also for other types of entertainment content, such as movies and TV shows. The study was published in the journal Frontiers of artificial Intelligence.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.