Since the dawn of computers there has always been a not-so irrational fear that they would eventually supersede and replace humans. While that hasn’t happened yet, it’s possible that the first step has already happened; computers can now recognize human emotions with 87% accuracy based on keystrokes, text analysis, and depending on the emotion. While this isn’t the same as being able to tell due to expression, body language, or tone, this is the first time that computers have been accomplish this task with such a high percentage of success. This is the first step towards creating computers that can both accurately read as well as respond to emotional input from the humans using them.
The program reads emotion by combining both keystroke dynamics and analyzing the text patterns from the sentences that the subjects wrote. The researches created a database of words commonly associated with the 7 emotions that they were testing for: joy, fear, sadness, guilt, disgust, anger, and shame. Of all the emotions tested, the computers had the easiest time analyzing joy and anger, with a success rate of 87% and 81% respectively. Computers that can read emotion are planned to be used in areas ranging from education and gaming to user authentication and video/imaging processing. Any activity that relies on a human’s emotional state might soon be done on a computer that can read how you feel depending on what you type. The future is coming close every day.
If you’d like to read more, the link is here.
from Yaser Khalifa and Evolutionary Algorithms http://ift.tt/1t0tEFK
from WordPress http://ift.tt/1qy6SyB
