An international team of scientists considered people’s preferences and generated an appearance that they consider to be good-looking. 80% of them passed the “attractiveness test”.
A team of researchers has developed artificial intelligence (AI) for mind reading. It reads a person’s personal preferences and can create a portrait with an appearance that will appeal to the user. The system uses a person’s preferences to create portraits of pretty people who don’t really exist.
This was stated by scientists from the University of Helsinki and Copenhagen. They published an article detailing the system by which the computer-brain interface is used to transfer data to the AI system. It then interprets the data and uses it to train the image generator.
After the user’s preferences were interpreted, the machine generated a new series of images, tuned to be more attractive to the user. After analysis, the researchers found that 80% of the personalized images generated by the machine passed the attractiveness test.
Researchers initially challenged a generative neural network (GAN) to create hundreds of artificial portraits. The images showed 30 volunteers who were asked to pay attention to faces they found attractive, while their brain responses were recorded using EEG.
The researchers then analyzed the EEG data using machine learning methods, connecting the individual data through a brain computer interface (BCI) to a generative neural network.
The researchers note that in the hands of corporations, this tool can become “intimidating” Some companies may knowingly generate images and set up more relevant targeted ads or political campaigns.