How was the experiment done?
For this study, 12 faces were studied. Researchers have created an artificial neural network that has identified 12 different signs, such as assessing a person’s facial width and lip or eye height. The data obtained from this assessment are used to classify a person based on five personality traits: honesty, neuroticism, extrasion, acceptability and openness.
3% accuracy of artificial intelligence
Artificial intelligence had an 8% accuracy when compared to a quiz filled out by volunteers. Researchers say that at this point, there is a good chance that it will be 90 percent true and that humans have less consistency in facial recognition.
More accurate in knowing the personality of women than men
Scientists, using the so-called ‘Big Five’ traits, use a well-established method for classifying personalities. This system was found to be more accurate on women than men and was the best to recognize honesty.
A study of 21,000 selfies of 12,000 volunteers
A total of 15,000 volunteers uploaded a total of 21,000 selfies and they were divided into two groups. One was used to train the AI system and the other group was used to test the network.
Helpful in finding a good employee
Researchers claim that the system can help the boss identify the applicant for the most suitable job. Knowing the personality from the face the employer will choose the employee. Researchers from two Moscow universities, the HSE University (Economics of Higher Education) and the Open University Fair Humanities and Economics, say that when an artificial neural network is asked to rate a person’s personality without getting any other information, on average.
Dating websites best match selection
The technology can also improve the selection of ‘best matches’ for dating websites, customer service or online tutoring. Co-author Dr. Evgeny Osin said that a lot of information about a person’s inner work is hidden in the face. He said: “The help of artificial intelligence in choosing a spouse can help individuals achieve more satisfactory interaction results.”
Dr. “Our results show that real-life photographs taken in uncontrolled conditions can be used to predict personality traits using sophisticated computer vision algorithms. The advantage of our method is that it is relatively simple – it relies on 3D scanners or 3D facial landmark maps.” No and can be easily implemented using a desktop computer.