When he was an undergraduate at Georgetown University, Adrien Treuille was considering dropping out of college because he hated his international relations major so much.
That's when an economics professor gave him some unusual advice: "The problem is," the professor told him, "you need to be doing math, and I was like, that's absurd. I had no interest in math, and no apparent talent for it."
But he ended up following the advice anyway, segued into computer science, and last month, was named by Technology Review magazine as one of the TR35 -- 35 stellar innovators under the age of 35.
Dr. Treuille, which he pronounces "Troy," is now a professor at Carnegie Mellon University, and his specialty is creating visual simulations of complex phenomena that can be run on a desktop computer.
Adrien Treuille
Position: Assistant professor in computer science and robotics, Carnegie Mellon University.
Age: 30
Residence: Lawrenceville
Education: Bachelor's degree, Georgetown University, 2001; master's and Ph.D. in computer science, University of Washington, 2006 and 2008.
Previous positions: Postdoctoral research fellow, University of Washington, 2007-08; visiting scholar, Institut National de Recherche en Informatique et en Automatique, Grenoble, France, 2004.
Professional honors: TR35, named one of top 35 innovators under 35 by Technology Review, 2009; dean's innovation award, Carnegie Mellon, 2009.
One example is animation that shows the turbulent movement of air over speeding race cars in real time. Another is simulating how crowds of people move when they mix together, an algorithm already being developed for computer games.
But the project he seems most excited about is Foldit, an online computer game in which contestants try to fold proteins into the most efficient shapes possible.
The game already has attracted 90,000 players from around the world, and holds out the hope that one day, average citizens with a passion for computer games may be able to develop useful new enzymes and drugs.
Proteins are the workhorses of biology, and they function by assembling themselves into certain shapes. If they fold incorrectly, they can cause such disorders as the memory loss of Alzheimer's, the paralysis of Lou Gehrig's disease or the tremors of Parkinson's.
In Foldit, players are given the simulated image of a protein and are asked to change its shape by moving parts of it with a computer cursor.
Nature likes proteins to assume a configuration that uses the least energy possible, and the game is able to calculate that using a scientifically validated formula, which it converts into a score: the more efficient the player's protein shape, the higher her score.
At first, even Dr. Treuille was skeptical that many people would get excited by such a game, but then he and his co-creator, Seth Cooper, tested it on 15 biochemistry graduate students, and "to our astonishment, they went totally crazy. One woman worked on the game for 12 hours straight, and they were trying to brutally knock each other off the top of the standings.
"That's when we realized we had really tapped into some sort of basic competitive instinct," he said in a video interview with Technology Review.
Now that Foldit is being played all over the globe, he has discovered something else. The game seems to have drawn people from all walks of life who are legitimately interested in advancing scientific research.
"This is completely different than shooting things, and it's really important for the world," he said.
The regular competitors include everyone from an Australian farm laborer to high schoolers who tell fellow players that they have to stop for a while because their moms say it's time to do their homework. "The players are not scientists for the most part, so I actually feel we've enabled them to become scientists," Dr. Treuille said. "By putting simulations on their screens, we've given them access to a new part of the world."
For now, scientists already know the shapes of most of the proteins that Foldit players are manipulating. But someday, scientists may be able to post the shapes of biological receptors and see if Foldit players can design brand-new proteins that will fit into them like a lock and key.
While Foldit has been an unexpected success, his race car simulation has undoubtedly reached more people.
Millions of NASCAR fans have now seen Draft Track, the commercial version of a simulation Dr. Treuille created while he was a Ph.D. student at the University of Washington. It shows technicolor images of the air turbulence created by a moving race car.
While the TV version hasn't been universally applauded, the simulation does prove an important principle that underpins much of Dr. Treuille's work -- that it is possible to take complex physical phenomena that were once confined to supercomputers and streamline them so they can be shown on a desktop computer and other smaller devices.
Simulations on supercomputers have enormous power, he said, but much of that power is used to depict outcomes that are unlikely to occur in real life. His algorithms are designed to pluck out only the most probable results, so that in the race car example, they "allow us to take the most salient features of the fluid flow around a car and then simulate only those."
Another streamlining technique he has developed is to show how people move in a crowd, a technology that has been licensed by game maker Electronic Arts. Instead of treating each animated person as an individual, it plots them as points in a connected wave, and can realistically show how people move into the gaps that are created when they encounter obstacles or each other in a crowd.
He discovered recently that his crowd simulation had been used by graphics chip maker AMD for an advanced test game called "Froblins," which shows armored, toad-like creatures moving around en masse in a mountainous landscape. A sample of the game can be viewed on YouTube.
One of Dr. Treuille's next big projects will be to expand Foldit into new realms of biology. He has one prototype now in which players will compete to see who can fold RNA into the best possible knot, a project supported by scientists at Stanford University.
Some scientists believe that RNA -- ribonucleic acid -- was the first living molecule, and one day, Dr. Treuille said, he would like to create an interactive game in which players could synthesize their own RNA molecules and let them interact with each other in a computer version of the prehistoric "primordial soup."
For now, Foldit and its future cousins are just games.
But one day, he told Technology Review, "what we want to do is to build a platform where scientists can contribute their problems and humans can help them explore the possible solutions by playing these games.
"So in some sense we're democratizing science and opening it up to the public and saying, 'You help solve the pressing scientific questions of the 21st century.'"





Comments
Welcome to Post-Gazette subscriber-only commenting community. Please review our community guidelines.