feature article
Dr. Peter Stone.
Photo by: Matt Lankes.
f
A
I
i
s
t
o
“
T
h
e
g
o
a
l
o
a
t
h
a
v
e
n
’
t
d
o
t
h
i
n
g
s
t
h
b
e
e
n
p
o
s
s
i
b
l
e
b
e
f
o
r
e
.”
i
n

l
e
a
r
n
i
n
g

n a small practice field on the first floor of Taylor Hall, In the living world, learning is a trait unique to animals. An animal’s
robot dogs playing soccer scuttle around the field like ability to learn, or improve its behavior through experience, helps it
m
a
c
h
i
n
e
s
infants crawling across a playpen. They survey their survive in a changing environment. AI agents too need the ability to
field with Cyclops-like camera eyes and nudge balls around with learn and the flexibility to adapt if they are to survive in a world that
silver chins. And like little humans, these robots learn to walk, rec- constantly provides new challenges. “You don’t want an agent that’s
ognize color and hold a ball. They experiment and adapt, learning going to step into the same hole or drive into the same pothole every
new tricks every day. day,” says Stone.
The robodogs—members of the UT Austin Villa Robot Soccer Soccer fields don’t have potholes, per se, but the game offers plenty
Team—serve as a test bed for Dr. Peter Stone’s research in machine of challenges for robots. Field surfaces, lighting conditions, and colors
learning and artificial intelligence (AI). Stone, an assistant professor of teammates, soccer balls, and goals are unique to each field. And
Soccer-playing robots that learn
of computer sciences and Alfred P. Sloan Research Fellow, ultimately the game itself is dynamic and interactive. Programming robots to
seeks to create completely autonomous agents—independent AI ma- respond to every feasible variable in a soccer game by hand (called
pave the way to the future of
chines that learn and interact like living animals, challenging our hardcoding) would be next to impossible, which is why Stone loads
artificial intelligence perceptions of intelligence and consciousness. his soccer-bots’ brains with machine learning code.
“We always push the boundary of what it means to be human,” says “We’re always looking for ways for the robots to learn themselves,
Stone. “The goal of AI is to do things that haven’t been possible before.” rather than us hardcoding them,” says Stone. “Anything you hardcode
In addition to soccer-playing robodogs, Stone is developing AI- is brittle. The more the agents can learn to adapt, the more robust they
powered cars that drive themselves, self-repairing computers, and will be.”
software agents that manage supply chains. Regardless of the task Stone and his students didn’t build the Aibo robodogs—that
these AI agents perform, for them to be independent and success- credit goes to Sony—but they do program their brains from scratch.
ful in our complex world, they must all have the power to learn. The robodogs’ brains, which are tiny, removable memory sticks, run
6 s p r i n g 2 0 0 6 f o c u s o n s c i e n c e 7
Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24