We once again have new evidence of the universality of facial expressions (and emotions), but this time from a new approach. In a new study was examined the occurrence of 16 expressions in 6 million videos from 144 countries using machine learning.
Potential applications of this study include helping people who have trouble reading emotions, such as children and adults with autism, recognize the faces that humans make to convey certain feelings.
Neural networks
At least since the time of Aristotle, scholars have tried to understand how and why the face reveals our feelings, from joy to sadness. The debate on universality is fundamental to understanding the nature, causes and functions of emotions.
The study, in two experiments using deep neural networks, examined the extent to which 16 types of facial expressions were systematically produced in thousands of contexts in 6 million videos from 144 countries.
In each region, certain facial configurations were observed relatively more frequently in certain contexts. The associations were subtle (i.e., the magnitude of associations between facial expression and context tended to be weak), but, surprisingly, the expression-context association pattern observed in videos from one region of the world was similar to those in from other regions of the world.
For example, across the various regions sampled, people in the videos performed facial-muscular movements labeled 'awe' more frequently in contexts that involved fireworks, a parent, toys, a pet, and dancing than in contexts that did not include these. items. such as those related to music, art, police and team sports.
It was thus discovered that each type of facial expression had different associations with a set of contexts that were conserved in a 70% in 12 regions of the world.
16 facial expressions that one tends to associate with amusement, anger, amazement, concentration, confusion, contempt, satisfaction, desire, disappointment, doubt, elation, interest, pain, sadness, surprise and triumph.
According to these associations, regions varied in the frequency with which different facial expressions occurred depending on which contexts were most salient. The results reveal fine patterns in human facial expressions that are conserved throughout the modern world.
–
The news
Thanks to AI we now know that there are 16 universal facial expressions
was originally published in
Xataka Science
by
Sergio Parra
.