Your Face Mirrors Your Deepest Beliefs-Predicting Personality and Morals through Facial Emotion Recognition
A current paper on arXiv.org shows that Al can be used for the undertaking of facial emotion recognition. It can generate characteristics that can, in turn, forecast people’s individuality and moral values. Determining individuality qualities in others is a undertaking the place individuals typically struggle.
First of all, feelings are tracked via facial emotion recognition though the viewer is seeing emotionally triggering online video snippets. The scientists demonstrate that the emotional reaction in the deal with of the viewer will in fact forecast the individuality and moral values of the viewer via correlations, regression, and device finding out.
The investigation confirms that feelings in reaction to positive and unfavorable ordeals are at the center of our moral values. The paper proposes a novel and more trustworthy way to measure individuality qualities, attitudes to possibility, and moral values.
Can we actually “read the mind in the eyes”? Additionally, can AI guide us in this undertaking? This paper responses these two concerns by introducing a device finding out program that predicts individuality qualities of people on the basis of their deal with. It does so by tracking the emotional reaction of the individual’s deal with via facial emotion recognition (FER) though seeing a series of 15 shorter films of different genres. To calibrate the program, we invited 85 people to enjoy the films, though their emotional responses have been analyzed via their facial expression. At the identical time, these people also took 4 effectively-validated surveys of individuality qualities and moral values: the revised NEO FFI individuality inventory, the Haidt moral foundations test, the Schwartz individual value program, and the domain-specific possibility-using scale (DOSPERT). We identified that individuality qualities and moral values of an unique can be predicted via their emotional reaction to the films as revealed in their deal with, with an precision of up to 86% working with gradient-boosted trees. We also identified that different individuality qualities are much better predicted by different films, in other terms, there is no single online video that will present accurate predictions for all individuality qualities, but it is the reaction to the combine of different films that allows for accurate prediction.
Analysis paper: Gloor, P. A., “Your Deal with Mirrors Your Deepest Beliefs-Predicting Persona and Morals via Facial Emotion Recognition”, 2021. Website link: https://arxiv.org/abs/2112.12455