The researchers in the MIT Media Lab have recently developed a new personalized machine-learning model, which takes computers a step ahead for the interpretation of human emotions as naturally as done by the humans. In the high improvement of the affective computing, the computers and robots are expected to analyze the facial expressions and then interpret and understand the emotions and then further respond accordingly. The major applications include, monitoring the health of an individual and their well-being, helping in diagnosing signs of a certain disease, gauging the student interest in classrooms, and development of the helpful robot companions.
On the other hand, the challenge for the developers is that, the people around the world express their emotions in a quite different way, depending on several factors. Some of the general differences can be observed among genders, age groups, and cultures. However, some of the other differences are even more grained, such as how much the individual has slept, the time of the day, the level of familiarity with the person, and others.
The team of researchers have found out that human brains instinctively catch the deviations; however, the machines struggle. The deep-learning techniques were developed in the last few years in order to help catch the subtleties, but still they are not quite accurate or as adaptable in the different population as they could be.
The researchers in the Media Lab have further developed a new personalized machine-learning model, which is predicted to outperform the traditional system. It is also expected that this new machine will capture the small facial expression variations in order to understand the mood while training on thousands of the images of faces.