Advancements in artificial intelligence has been progressing at a very fast pace for the past few decades, and it has been very good at reading humans' emotional reactions. Now, "deep learning" system is said to have the capacity to not just read, but really understand human emotions and even react to them appropriately.
Deep Learning's 90 Percent Accuracy in Facial Recognition
Deep learning is a type of AI that builds a "neural network" almost like how a human brain does, Newsweek reported. Researchers have combined this approach with "face rotation model" giving the AI the capability to recognize criminals just by looking at their facial features. This system was used on Chinese ID Photos and it resulted to an incredible accuracy rate of almost 90 percent.
According to researchers, there are three major features that helps the artificial neural networks identify the criminals. The features are the nose's tip and the corners of the mouth, the curves of the upper lip, and the eyes' inner circles.
This approach has become an important part in AI technology that Microsoft has recently announced its acquisition of deep learning startup Maluuba, Venturebeat reported. Microsoft is yet to reveal its plans with the company, but Microsoft executive vice president Harry Shum they are planning to democratize AI. This means they want to make it available to consumers, developers, and businesses.
Other Approaches that Help AI Understand Human Emotions
Affective computing is another approach for AI to recognize human emotions. It can recognize the various levels of frustrations and engagements among people. It has helped researchers recognize whether or not students are having a hard time with certain lessons.
A tech company called BeyondVerbal has also developed an AI that can detect emotions through the sound of the voice. The approach resulted to an 80 percent accuracy rate. It can help autistic individuals identify the emotions of people in the future.
There are even reports that Sony is a robot that is capable of building relationships with humans. The company hasn't revealed how they will achieve this amazing technology though.
