Robot that can ‘imagine itself’ built by scientists
Image credit: Kittipong Jirasukhanont | Dreamstime.com
Scientists from Columbia University have created a robot that can learn without prior programming via ‘deep learning’ – an advancement said to be the first step towards machine self-awareness.
The device consists of a jointed artificial arm and grasping ‘hand’, like those used in numerous production plants, with the scientists from the New York-based university giving it the ability to ‘imagine itself’ using a process of self-simulation.
Professor Hod Lipson, director of the Creative Machines Lab, said: “If we want robots to become independent, to adapt quickly to scenarios unforeseen by their creators, then it’s essential that they learn to simulate themselves. While our robot’s ability to imagine itself is still crude compared to humans, we believe that this ability is on the path to machine self-awareness.”
Initially, the robot experienced 24 hours of behaving like a “babbling infant”, moving randomly while attempting various tasks. Within about a day of intensive deep learning, the robot built up an internal picture of its structure and abilities, with the machine being able to grasp objects from specific locations and drop them with 100 per cent accuracy.
Even when relying entirely on its internal self-model – the machine’s ‘imagination’ – the robot was able to complete the pick-and-place task with a 44 per cent success rate.
PhD student Robert Kwiatkowski, a member of the team, said: “That’s like trying to pick up a glass of water with your eyes closed, a process difficult even for humans.”
Other tasks included writing text on a board using a marker.
However, to test whether the robot could detect damage to itself, the scientists replaced part of its body with a deformed version. As a result, the machine was able to recognise the change and work around it with little loss of performance.
According to Lipson, self-aware robots may shed new light on the age-old mystery of consciousness: “Philosophers, psychologists, and cognitive scientists have been pondering the nature of self-awareness for millennia but have made relatively little progress. We still cloak our lack of understanding with subjective terms like ‘canvas of reality’, but robots now force us to translate these vague notions into concrete algorithms and mechanisms.”
Self-aware robots and computers running amok or threatening humans have been a rich source of material for sci-fi novels and films, and the scientists say they are aware of the potential dangers involved in giving robots the gift of self-awareness.
Writing in the journal Science Robotics, the researchers warn: “Self-awareness will lead to more resilient and adaptive systems, but also implies some loss of control. It’s a powerful technology, but it should be handled with care.”
Last October, Pepper the humanoid robot, created by SoftBank Robotics, was the first robot to make an appearance at a UK parliamentary meeting, speaking before the Education Select Committee about the future of artificial intelligence in the classroom.