A new study describes how researchers have developed a robot that can predict how another robot will behave, a first step in the development of what is known as “theory of mind”.
The study was published in Scientific Reports on January 11th. The research took place at Columbia University’s Creative Machines Lab.
A robot theory-of-mind test
The researchers believe that the experiments described in the current study are the first Steps are to endow robots with what cognitive scientists call this “theory of mind.” Around the age of three, children begin to understand that other people have different goals, needs, and perspectives. As children develop theory of mind, this can lead to playful activities such as hide-and-seek. But it also allows for more sophisticated behaviors like lying.
Researchers consider theory of mind to be a key differentiator of human and primate cognition. It is essential for complex social interactions such as cooperation, competition, empathy, and deceit.
Empathy as a way of anticipating the behavior of others
When people live together for a long time, they quickly learn to predict the short-term actions of those in their immediate vicinity. This ability to anticipate the actions of others makes it easier to live and work together.
But robots are conspicuously lacking in this ability, which is why this study was conducted. It is part of a broader effort to give robots the ability to understand and anticipate the behavior and intentions of others, solely through visual observations.
To conduct the study, the team first built a small robot and put him in a playpen. They programmed the robot to look for and approach green circles that it could see. But sometimes those green circles got blocked from the robot’s view.
A second robot observed this behavior for two hours and began to anticipate the other robot’s intent. This “observing robot” was eventually able to predict its partner’s destination and path 98 times out of 100, despite never being explicitly told about the other robot’s visual impairment.
A primitive form of empathy
“Our results are beginning to show how robots can see the world from another robot’s perspective,” said lead author Boyuan Chen. Chean also said this ability is “perhaps a primitive form of empathy.”
This ability will make robots more useful. But if robots can predict how humans think, they may also learn to manipulate those thoughts.
“We know that robots will not remain passive machines that follow instructions for long,” said the head of the lab , Hod Lipson. “As with other forms of advanced AI, we hope policymakers can help keep this type of technology in check so we can all benefit,” he said.
Other breaking news from science and psychology:
- A new theory suggests that Jupiter acts like a pinball machine, sending comets like Chicxulub into orbits that bring them very close to the Sun.
- How can you stop your cat from chasing wild animals? A new study shows that cats hunt less when their owners give them plenty of meat and playtime.
- A full moon and sleep patterns: A new study finds that in the days leading up to a full moon, people go to bed later and sleep less.
Study: “Visual Behavior Modeling for Robotic Theory of Mind”Authors: Boyuan Chen, Carl Vondrick and Hod Lipson Published in: Scientific Reports.Published date: 11. January 2021DOI: 10.1038/s41598-020-77918-xPhoto: by Andy Kelly via Unsplash
.
