DAC’s keynotes are the highest profile events at the conference, a chance to highlight the innovation happening now and coming soon in the years ahead. Stone’s speech will likely do both. That’s because his area of expertise, AI, is rapidly crossing over from research labs into the realm of application. Consider how Gartner’s forecasts for 2016 and beyond made news with the prediction that more than 3M workers around the world will be supervised by a robo-boss by 2018. By 2020, autonomous software agents outside of human control will generate 5% of all economic transactions. AI is very much here and now.
|Highlights from the 2012 RoboCup final in Mexico City, won by Peter Stone's Austin Villa robot soccer team.|
Stone, founder and director of the Learning Agents Research Group (LARG) within the AI Laboratory in the UT’s Department of Computer Science, has published extensively on autonomous software, among other topics. A recent example is this paper from AAAI 2016 showing how autonomous electricity trading might help bring about the smart-grid. The paper details an algorithm written by Stone and a colleague that took first place in a 2014 competition simulating a large-scale retail electricity market. Eventually such autonomous trading could help adapt customer demand to supply conditions, leading to a more sustainable, efficient energy supply.
While disembodied algorithms may have more immediate impact, robots are more apt to capture the imagination — especially when they’re doing things like playing the world’s most popular sport. Stone’s teams have an impressive track record in RoboCup, the annual international robotics competition featuring various challenges and games, most notably soccer.
RoboCup doesn’t shirk from ambition in its official goal: “By the middle of the 21st century, a team of fully autonomous humanoid robot soccer players shall win a soccer game, complying with the official rules of FIFA, against the winner of the most recent World Cup.”
This seems to stretch credibility if you consider the balky movements of the robots on Stone’s 2012 team, which won the final. However Stone made an interesting point in a TEDxYouth talk the following year. By mid-century RoboCup will have been around for about 50 years, roughly the same amount of time that elapsed between the Wright Brothers and the Apollo Mission, and between the first computer and Deep Blue’s defeat of defeat of world chess champion Garry Kasparov.
|Stone at TEDxYouth event in Austin in 2013.|
“We’ve learned in AI that predictions are so apt to be wrong that it’s best to pick a year after you’ve retired if you’re going to make a prediction,” he said on stage, to a crowd that laughed and applauded throughout his talk.
One prediction I feel good about making: Stone will steal the show Thursday on DAC’s main stage.