Researcher's Robots Learn From Environment, Not Programming

June 2, 2010 By La Monica Everett-Haynes, University of Arizona
Ian Fasel said that in building self-teaching robots, researchers can learn much more about the naturally occurring behaviors in humans, such as why young children often are constantly exploring the world.

( -- Ian Fasel, a UA assistant research professor, recently received two grants to fund research and design projects toward creating highly intelligent robots.

Humans have trained robots to build vehicles, fly airplanes, automatically test blood pressure in hospital patients and even play table tennis.

But robots have no concept of self, nor do they truly understand what it is they are programmed to do, said Ian Fasel, an assistant research professor of computer science at the University of Arizona.

As a consequence, robots often are unable to problem solve - a design and implementation glitch that is driving Fasel's research.

He recently earned grant funding to advance his robotics work, which is focused on improving what robots are able to understand and learn, but without the aid of human programming.

"The prevailing technique for recognizing objects is for someone to collect a large database of examples - cars, pedestrians, faces," Fasel said. "It's a lot of work to find a lot of good examples and to prepare them in a learning algorithm."

The $250,000 grant - his first as a new faculty member - came from the Defense Advanced Research Projects Agency, or , to fund "CLIME: Concept Learning from Intrinsically Motivated sensory-motor Experience."

The CLIME project involves training robots to learn concepts from trial and error in the world, whereas a second project teaches robots that already have some human-programmed concepts to seek out new information as efficiently as possible.

In effect, Fasel and his team are engineering a system that can teach concepts to itself, instead of requiring a human to program them in, as it learns how to interpret and control its sensors and motors.

"A lot of our language (as human beings) is metaphorical," Fasel said. "The argument we're making is if the robots can understand sensory and motor concepts, they won't make dumb mistakes when we try to talk to them."

Fasel also is collaborating with a team from Texas on another project, "Active Learning for Sequential Sensing and Efficient Human Interaction in Collaborative Human-Robot Teams," which the U.S. Office of Naval Research is funding for three years at nearly $600,000.

Fasel is a subcontractor and co-principal investigator on the grant, which is part of a larger effort led by collaborators at the University of Texas at Austin and Texas Tech University.

"Further in the future, the real goal is for them to teach themselves commands," he said. "We really do want robots in which they learn commands to help humans interact with them."

Fasel and his colleagues also are training robots able to detect emotion in their human counterparts - all with little or no programming.

Such work in enhancing artificial intelligence in robots is an increasingly visible area in computer science.

Researcher's Robots Learn From Environment, Not Programming
UA researcher Ian Fasel is building robots that are learning to perform tasks with very little human programming.

In 2003, Sony created "QRIO," a bipedal robot that recognizes voices and faces, which was subsequently featured in a music video by the popular music artist, Beck.

Other humanoid robotic models have been designed using recognition technology to comprehend certain movements, gestures and sounds and also to act as companions.

If at this point you cannot stop thinking about science fiction films such as Star Wars, Artificial Intelligence and I-Robot - just try. While filmmakers have long displayed robotic characters and technologies seemingly out of this world, most are tremendously difficult and time consuming to try and replicate, Fasel said.

Also, much of the actual advancement in robotics has been driven by researchers feeding robots with algorithms that provide step-by-step explanations for processes and tasks and robots learning via computation.

Fasel's approach is to determine less exhausting and time-consuming ways to teach robots.

"In real-world applications, it would literally be some kind of unmanned aerial vehicle that needs to understand its flyover area," Fasel said. "Or, imagine if you had a household robot. It needs to know what it is, where it is and where objects belong."

Fasel also said understanding the learning patterns in robots may also help further knowledge about human nature.

"What we want to do is learn about their behaviors as well," Fasel said. "Another part is understanding apparent curiosity in infants in constantly exploring the world."

Explore further: Androids might soon become science fact

Related Stories

Androids might soon become science fact

May 9, 2006

New Zealand and European scientists say it's time to send R2-D2 back to science fiction land -- and get ready to greet androids that think and act like humans.

Robots attend nursery school

June 23, 2005

Two robots are attending a nursery school at the University of California San Diego so they can learn to play well with the group.

'The robots are coming'

January 29, 2009

Alexander Stoytchev and his three graduate students recently presented one of their robot's long and shiny arms to a visitor. Here, they said, swing it around.

Japanese Firms Start Testing Robots In Office Buildings

February 7, 2009

( -- It may not be too long before visitors are greeted by a robotic receptionist in Japanese Smart Office Buildings. Shimizu Corp and Yasukawa Electric Corp have opened the "Smart Showroom" that is part of the ...

Developing long-term relations with robots

April 13, 2008

Scientists at Queen Mary, University of London are leading an international project which is set to advance the relationship between robots and humans, as part of new European project called LIREC - Living with Robots and ...

Recommended for you

Balancing nuclear and renewable energy

April 25, 2018

Nuclear power plants typically run either at full capacity or not at all. Yet the plants have the technical ability to adjust to the changing demand for power and thus better accommodate sources of renewable energy such as ...

Researchers 3-D print electronics and cells directly on skin

April 25, 2018

In a groundbreaking new study, researchers at the University of Minnesota used a customized, low-cost 3D printer to print electronics on a real hand for the first time. The technology could be used by soldiers on the battlefield ...

Electrode shape improves neurostimulation for small targets

April 24, 2018

A cross-like shape helps the electrodes of implantable neurostimulation devices to deliver more charge to specific areas of the nervous system, possibly prolonging device life span, says research published in March in Scientific ...

China auto show highlights industry's electric ambitions

April 22, 2018

The biggest global auto show of the year showcases China's ambitions to become a leader in electric cars and the industry's multibillion-dollar scramble to roll out models that appeal to price-conscious but demanding Chinese ...


Adjust slider to filter visible comments by rank

Display comments: newest first

2 / 5 (1) Jun 02, 2010
Yeah, I've always imagined a lab or company that sells robots would have teams training the robot and teaching it by experience around the clock and then they would periodically release a copy of what the robot had learned to customers who will update their robot. Or perhaps all the robots that are sold learn from their respective homes and jobs and then they all share the information learned with each other. Kind of like a robot apps center.
not rated yet Jun 03, 2010
well gaining and sharing knowladge and skills... thats what i call inteligence.
1 / 5 (1) Jun 03, 2010

Not that I'd like true AI due to the obvious risks of a robot rebellion, but if you actually wanted to design the "ultimate" robotic civilization, you simply have them all linked by something like a radio network, and each time any member learns anything that is "all new" it simply uploads it to all the others immediately.

Ironically, the closest science fiction "robots" to what a real race of "AI" would look like ends up being the Borg from Star Trek: TNG. (The Voyager Borg and movie Borg are sort of plagued by non-sensical plot devices, so ignore those.)

And yeah, I agree, if it is "learning" by experience, then that is intelligence to some degree, even if it is only perhaps "animal level".

I think that this raises serious concerns in regards to human safety, because such robots would be dangerous to humans even by "accident", such as is the case with a particularly curious pet or a small child.

Robots should have tight parameters in a limited scope.
1 / 5 (1) Jun 03, 2010
Good stuff but can we please drop this idea that these robots learn "without programming". Everything about these and any other kinds of robots is circuits and programming from the ground up. This programming may be clever and flexible enough for the robot to "realize" on its own that, for example, some things move and some don't but the potential for such a "realization" is still a product of the robot's programming.

This is not to diminish the accomplishments of these or any other robotics researchers.
1 / 5 (1) Jun 04, 2010
@ralph wiggum, everything about you is all programing from your genes and the environment, you tell me how you are more than that.
1 / 5 (1) Jun 04, 2010
I am talking about these robots, not living organisms. Robots have no genes other than what their humans designers put in them. They have no awareness of the environment other than what their programming allows. All of these robot's abilities are still a function of its human-designed hardware and software. Not naturally evolved HW/SW, but explicitly human-designed HW/SW. Clear enough for you? Do you think that a newly minted Boeing 777 just learns to fly or do you think its ability to fly is a function of its design?

This point doesn't take anything away from the accomplishments of these researchers, it's just calling a spade a spade. Writing SW able to "adopt" and "learn" is what AI is all about, but it's still human-written SW, whether its an open-ended neural network or a precise step-by-step algorithm.
not rated yet Jun 07, 2010
The best robot I've seen so far :
It seems that we should learn from robots if we are to stay ahead of the game.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.