A remarkable robot by OpenAI, able to solve the cube of a Rubik on its own, has shown how far robotics have gone-but at the same time, experts say, how far we still have to go.
The OpenAI program used a computer simulation to train the robot hand to solve the puzzle. This robot hand went through procedures that would take around 10,000 years for a single human to complete.
Once taught, a cube that had been slightly modified was introduced and solved by the robot to help the machine tell how it was held during the process.
The research team said the execution period differed, but it usually took about four minutes to complete the job.
Before this, machine-learning and robotics were used to solve the cube of a Rubik. Notably, an MIT engineered machine managed to solve a cube in just 0.38 seconds in March 2018.
What is essential with the initiative of OpenAI is the use of a multi-purpose device. In this case, a human-had-like model, rather than a computer strictly designed to handle the cube of a Rubik but nothing else.
“The ability to solve the Rubik’s cube in the real world, on a robot hand, is extremely difficult,” said Matthias Plappert, team leader for robotics at OpenAI.
“You need to very precisely control your fingers, you need to do it for a very long time without the kind of messing up in between. and a lot of different things can happen in the process.”
Mr Plappert supported the team’s method of progressively injecting complications into the process which simulated difficulties that would force the robot to adapt in order to complete the cube.
In a demonstration video released by OpenAI, the researchers demonstrated that increasing the robots complexity-such as nudging the cube with a giraffe stuffed toy and shielding the cube with a black sheet-did not necessarily stop the cube from being completed (although the hand did not have a 100% record, researchers said).
Humans have instincts in managing disturbances and problems, especially when dealing with carrying and controlling objects. But for robotics this was identified as a major challenge.
A problem which needs to be addressed if advanced robotics are to become ubiquitous in every home and business.
“We use human hands for all kinds of things,” said Peter Welinder, also a team leader at OpenAI.
“We use them to solve Rubik’s cubes, but we also use them for cooking meals. This is one of the reasons we picked the robotic hand. because it promises much more general-purpose robots.
“The big obstacle has been manipulation. So we took the Rubik’s Cube as an example of really how far can we push manipulation.”
Yet, argues UC Berkeley’s Prof. Ken Goldberg, “OpenAI’s work should not be inflated, given what he described as an outstanding “showmanship” move.”
“The average human isn’t particularly good at solving Rubik’s cubes,” he told the.
“So when they see a robot doing it, they say, ‘well, this is better than a human’. But that’s a little deceptive – because games are not reality.”
He said “the ADR methodology of OpenAI is a real improvement, but more work will be required to treat objects that are more complicated and volatile than a cube of Rubik.”
“Will we get to the point where a robot could pick up a deck of cards and shuffle them like a Las Vegas croupier? Or anyone who’s reasonably good at doing that? That could be 10-20 years off.”
“We’re far from being able to replace kitchen workers who chop up vegetables or even pick up and you know, do dish washing. All those are very complex tasks.”
OpenAI was formed in October 2015, co-funded by CEO Elon Musk of Tesla, start-up consultant Sam Altman and all other investors (although Mr.Musk is no longer involved with OpenAI).
Its goal was to continue artificial intelligence research with a focus on a technique known as “reinforcement learning”-teaching AI with repetitive tasks and “rewards” for correct results. OpenAI is considered a key rival to the Deepmind, a product of Google.