Study claims success, inspires app using face biometrics to match dating pairs
It turns out that all humans might just be books that can be accurately judged – by AI – for neuroticism and romantic compatibility.
A study published in the Journal IEEE Access reportedly found that one’s personality can be “reliably predicted” at least 70 percent of the time using facial recognition software.
Five Chinese researchers say their algorithm accurately classified static images of people according using the so-called big five personality traits: neuroticism, extroversion, openness, agreeableness and conscientiousness. Accuracy was judged against subjects’ self-reporting.
The results are based on a dataset of 13,347 pairs of face images and personality traits created by the researchers. A deep neural network of 10,667 sample pairs pulled from the dataset to test the leftover biometric samples to validate against self-reported traits.
Neuroticism and extroversion apparently are most obvious on a face, at least for the algorithm. They were predicted with 90 percent accuracy, according to the study’s authors.
Developers of dating software are on the same scent. A pair of sisters saw the personality trait study and started SciMatch based on its findings.
SciMatch claims that emotional relationships can be stably built on facial templates. Common interests are fine for predicting compatibility, according to the company, but getting those templates analyzed and categorized is necessary.
Skepticism might be due in this case. The company describes its methods as “utilizing Ruby, an AI-powered matchmaking robot and personal virtual assistant.” Take a selfie and find a lifetime mate.
For the most part, most facial recognition apps are being used for more tactical dating purposes, like the identity verification from Yoti for matchmaking app Muzz.
Comments