| Follow Us:
Spring blooms in Pennsylvania
Spring blooms in Pennsylvania - Aimee Dilger | Show Photo

Features

Q&A: Adrien Treuille, Carnegie Mellon University

About a decade ago, Adrien Treuille was an international studies major uninspired by his courses. Six weekends ago, the Carnegie Mellon University assistant professor of computer science, who could easily be misidentified as an undergrad, was sitting on a panel with Google CEO Eric Schmidt and Rensselaer Polytechnic Institute President Shirley Ann Jackson chatting about the Golden Triangle technologies of IT, biotech and nanotech.

Such a big leap took a big push into a new world in which Treuille is immersed in some of the most innovative thinking around.

When Treuille's economics professor at Georgetown University told the disenchanted junior, who was thinking about dropping out of school, that he should study science, something clicked for the New York City native. With numbers, Treuille liked that one couldn't just bully his way to a solution. The life-changing nudge led to his splash into computer science, earning his undergrad from the program at Georgetown in 2001 before taking his talents to one of the nation's top graduate programs at the University of Washington, where Treuille worked with one of computer graphics' legends, Zoran Popovic (a CMU alum).

Treuille is fast approaching legend status himself. Specializing in ways to make realistic computer simulations run in real-time on everyday computers, Treuille's most immediate success came in late 2006 when his published research was licensed to create Draft Track, the Emmy-nominated technology used on ESPN for broadcasts of NASCAR races that depicts the impact of air flow around cars.

Another signal that Treuile had made the right choice was his co-creation of Foldit, an online game that is helping biochemists unravel the mysteries of how proteins fold and could allow people without scientific background the opportunity to help cure diseases. Turns out solving puzzles for science is a popular idea--it was the first video game to be published in the scientific journal Nature.  Last year, Treuille made Technology Review's prestigious list of the world's top 35 innovators under the age of 35.

On June 22, Treuille, Schmidt and a host of the nation's brightest minds led a workshop organized by the President's Innovation and Technology Advisory Committee, and Treuille talked to Keystone Edge about the experience, as well as what led to his greatest successes.

How did you wind up working with ESPN and NASCAR?
A company called Sportvision approached us. They do the first-down line in football and the blue trail on the hockey pucks. They pretty much own the market for doing graphics for sports TV. I hadn't heard of them, but apparently they pay close attention to graphics research. We published a paper and within a month received an e-mail that they were thinking about different kinds of special effects and graphics, saw our technology and was excited about it, and wanted to send someone to us so we can teach them how it works and they'd give us a licensing agreement.

It was research-level code, so I thought there was no way they could pull this off. I had confidence in our technique but it's really advanced stuff and not easily packaged, like on a CD or something. So they a sent a Princeton-educated physics guy and he totally got it. I was amazed. We had three really long and hard days of me explaining the math and the code and how everything fit together and he left with a CD. I talked to him a number of times since then as they were developing the product, and then it was on TV. It was one of Men's Health's top coolest technologies of the year and very visible, and all within six months of publishing my research paper. When you do research, especially if it's not commercialized, it's not always clear if it's robust or will work every time. Sportvision was very nimble and smart, taking our ideas and having faith in us and proved it on a massive scale. They got the system down so it was just a button a broadcaster could push and it would show up on the screen.

How has Foldit grown?
We've built this new computational substrate for solving protein-folding problems. We sort of grew this community and cultivated it and trained players and by now, the players have routinely beat the best computational methods at coaching Foldit. This is a living being, a community kept alive and tended to like a plant in the sense that we have to make sure the users are happy and the scores are updated and the forum posts are responded to. But in exchange, (the players) give us this wonderful thing. When we post puzzles they compete very fiercely and perform top computational studies.

We've now got the first non-expert designed flu antigen that was designed by one of our game players. We made a flu antigen into a puzzle. We're sort of introducing new and bigger problems. Proteins are where the action is in biology and biochemistry. The flu antigen is not in clinical trials yet but it's being seriously tested in the lab as the very first initial screen whether the proteins bind the way we think they do. The most exciting thing about it is it's a resource you can participate in for free, just by going to a website and downloading a game. The level of public interest and enthusiasm for participating in science like this has been overwhelming.

What was the Golden Triangles panel like?
It was a huge honor. When I got there, many of these people I'd read about for years and here they are sitting two chairs down from me. It was really fun watching them in close quarters. Erik Schmidt is an incredibly bright guy, to put it mildly. It's not an accident he's CEO of Google. I felt that way about everyone there. It was sort of a mixture of young professors working at the intersection of computer science and biochemistry and a lot of CEOs of big companies who are working with research in health care, computer science, or IT. The mandate was to talk about various issues that are important to the scientific community, like research funding, and a report would be created based on the discussion for President Obama.

The conversation became mostly about that "valley of death" between our fertile academic environment and successful commercialization. Venture capital is only part of the solution. Even though VC is a risk-taking approach to investing, many academic inventions may require 10 years of testing, which exceeds the risk threshold of most venture capitalists. There are a number of other reasons for this valley of death. Academia has a tremendous wealth of scientific expertise in its ranks but has less business expertise and understanding of markets. There must be ways to bring these fields together better. The U.S. is probably sitting on a gold mine of commercial applications of its own scientific research that are not being sufficiently exploited today.

What solutions were discussed?
A really good example benefiting me is a fledgling experimental program at Carnegie Mellon called Project Olympus. It's different from a technology transfer office, which most universities have. A tech transfer office is great for neat academic technologies that can be readily commercialized by a company, but what's more difficult is having full spinouts coming out of universities. Usually that can produce more interesting and disruptive technologies, but it's very difficult to do. There's this problem of entrepreneurship and marketing expertise in science. Project Olympus is essentially a team of entrepreneurs embedded in the school who have the right to take on ambitious commercialization. They can be approached by any group of undergrads or professors and they become their mentors on these projects.

I have a preliminary plan to commercialize some of my research and it's going well but it would have really been difficult without Project Olympus. It's been really fun and completely eye-opening. There's no other resource like that that I know of. It's a simple idea really.

Joe Petrucci is managing editor of Keystone Edge. Send feedback here.

To receive Keystone Edge free every week, click here



Signup for Email Alerts
Share this page
0
Email
Print
Signup for Email Alerts