Gill Pratt Looks At The Road to Artificial Intelligence

Artificial Intelligence  

Once the realm of science fiction, smart machines are rapidly becoming part of our world—and these technologies offer amazing potential to improve the way we live. Imagine intelligent, autonomous vehicles that reduce crashes and alleviate congestion in crowded cities.

Imagine robots that can help your aged grandma move around safely or instructors that can assist special-needs children in classrooms. Gill Pratt, former head of the Robotics Challenge at DARPA, now heads up the $1 billion Silicon Valley-based, Toyota Research Institute where he and his team are pushing the boundaries of human knowledge in autonomous vehicles and robotics.

Related articles

• ATLAS Is Now Unplugged 
• Will Baxter And Other Robots Take Your Job? 
• iRobot Co-Founder Urges Roboticists to Focus on Practicality and Costs, Not Gimmicks 

In this discussion, recorded at the Aspen Ideas Festival, Pratt and New York Times science writer John Markoff explore the breakthrough technologies on the horizon and the unprecedented issues we will face in this brave new world.

ATLAS, one of the key competitors at the DRC

Despite the development of neural networks and other artificial intelligence systems, what Pratt says has been the biggest instrument of change in the last five year is the mobile phone. "What does the cell phone have to do with neural networks? Cell phones allow all of us to take lots of pictures...As a result millions of pictures end up every day in the cloud."

One of the results of all these pictures has been the development of ImageNet, a huge database that has been cataloged by crowd sourcing. This database has presented the decades-old technology of deep convolutional neural networks a system that matches inputs with outputs. For instance a picture of a cow, is matched to the word 'cow.'

"What has been happening in the last four to five years has been tremendous progress on that idea of matching input to output."

"What has been happening in the last four to five years has been tremendous progress on that idea of matching input to output," states Pratt. It turns out that the systems based on this approach are already about as good or better than humans at this image recognition/object classification task.

Pratt further explores this idea in terms of what it means for our understanding of human cognition. 

Pratt was a program manager in the Defense Sciences Office at the US Defense Advanced Research Projects Agency (DARPA) from 2010 to 2015, a professor and associate dean of faculty affairs and research at Olin College, and an associate professor and director of MIT’s Leg Lab. Pratt’s primary interest is in robotics and intelligent systems, particularly in interfaces that enhance human/machine collaboration and mechanisms for enhanced mobility and manipulation, among others. He holds several patents in series elastic actuation and adaptive control. 

SOURCE  The Aspen Institute

By 33rd Square

Embed