A neural network to determine children’s age of by their eye movement  

Artificial intelligence’ main advantage over human intelligence is its accuracy and attention to detail. Scientists from the University of Minnesota have once again confirmed the fact by creating software capable to determine a child age based on his or her eye movement.

The basic assumption was that people of different ages were attracted to different things. So, it is possible to determine a person’s age by eye movement.

The experiment involved 19 toddlers who were 18 months old and 22 children aged two and a half. Each child had a special device on his or her head that tracked gaze direction. The subjects were shown various objects. The toddlers paid attention to people’s faces while the 2.5-year-olds tended to look more at touchable things.

All data were carefully recorded and used for machine learning. Subsequently, the neural network learned to determine a child’s age based solely on his gaze direction trajectory. The answers were 83% accurate. This unusual computer application only highlights the fact that there is no limit to technological development and humanity has many new frontiers to strive for.