One debate for reasons why the robots will never fully be like humans is because they lack human-like social skills.
However, scientists are working with new ways of giving robots social skills to be able to interact with people. In this type of research, there is evidence of progress in two further studies.
The Massachusetts Institute of Technology, MIT researchers, carried out one test. The researchers built a machine learning system for self-driving vehicles designed to understand the social features of other drivers.
The team learned driving situations to understand how other drivers are likely to act on other roads. Because not all human drivers behave the same way, the data was for teaching the driverless car on how to avoid dangerous situations.
The team states that the technology uses equipment acquired from the social psychology field. In this test, scientists developed a system that tried to decide a person’s way of driving is selfish or selfless. In road tests, vehicles that are self-driven equipped with the system improved their prediction ability by up to 25 percent in predicting what other drivers would do.
In one experiment, they observed the self-driven car making a turn on the left-hand side. The study discovered that the system could result in the vehicle to wait before the corner if it predicted the selfish act of the oncoming drivers since it might be unsafe. However, when the oncoming car predicted to be selfless, the self-driven car would turn without delaying since it saw no risk of hazardous driving.
Wilko Schwarting, who is the lead writer of the article, says more about the research, told MIT News that all the robots operating around or working with people need to effectively be able to understand their intentions to be able to understand their behavior.
Wilko stated that humans’ acts of being collaborative or competitive usually spillover on their behavior as drivers. He added that the MIT studies were seeking to learn the possibilities of a system trained to measure and predict those behaviors.
They built the system to learn the right behaviors for different driving situations. For instance, the researchers noted that even the most careful driver should understand that fast and decisive action is usually required to avoid danger.
The MIT crew considers expanding its model of research to comprise other things that a self-driven vehicle needs to deal with. They include predictions about humans trotting around traffic, motorcycles and other things seen in driving environments.
This post was originally published on Food and Beverage Herald