A research team from the University of Copenhagen and the University of Helsinki presented a system
To train the model, the researchers setelectrodes on the heads of the study participants and showed them images of different faces, demonstrating how machine learning can use brain activity to determine which faces the subjects find the most attractive.
“Comparing the brain activity of other people, wefound that we can predict which faces each participant will find attractive before they see them. This way we can provide users with reliable recommendations - just like streaming services suggest new movies or TV shows based on the user's viewing history, ”explained senior study author Tuukka Ruotsalo of the Department of Computer Science at the University of Copenhagen.
Brain-computer interface cleared for first time by FDA for stroke rehabilitation
Scientists added that their model will be able to applycompanies that work with personalized recommendations and tailor-made content. However, existing collaborative filtering methods based on ratings, clicks, and content sharing are not always a reliable method for identifying user preferences.
“Due to social norms or other factorsusers may not disclose their real preferences through their online behavior. Therefore, explicit behavior can be biased. The brain signals we investigated are more related to immediate impressions than to elaborate behavior, ”said study co-author Michel Spape.
Read more:
Physicists explain why it is impossible to find dark matter: it is in another dimension
NASA scientists take pictures of clouds on Mars
Uranus has received the status of the strangest planet in the solar system. Why?