This is a printer-friendly version of an article from Zip06.com.

Daniel Spielman: Prized Possessions

Daniel Spielman has just done something that he is getting increasingly good at: He has won another prestigious prize. This time, Dan, who is Sterling Professor of Computer Science at Yale and professor of statistics and data science, won the Held Prize, given by the National Academy of Sciences for research in the analysis of computer algorithms.

Dan, who lives in Guilford, has already won the Fulkerson Prize, given for a paper in a specialized area of mathematics, and he has twice won the Godel Prize for outstanding papers in the area theoretical computer science and in 2013 he won the best-known prize of all, a MacArthur Fellowship, popularly known as a genius grant.

“It’s a little awkward; being identified as a genius is never a good conversation starter,” he says. “But it gives you credibility.”

The prize of which he is the proudest is far less well known: the Rolf Nevanlinna Prize, given by the International Mathematical Union and which he won in 2009.

“It was from my colleagues,” he explains.

The prize, given every four years, is now called the Abacus Medal.

The research that earned him the Held prize was done with two colleagues, Adam Marcus, then a postdoctoral student at Yale, and Nikhil Srivastava, a Yale graduate student at the same time. The team found an answer to the Kadison-Singer problem, which had defied solution for more than 50 years.

The Kadison-Singer problem asks if unique information about a whole network could be deduced in some situations from studying a discrete part of it. Though the prevailing opinion before the work of Dan and his colleagues had been no, Dan’s team proved that yes was correct response.

The solution proposed by the Yale team was unconventional, using information gleaned from other work on a specialized series of graphs. Spielman has noted that other mathematicians who had worked on the problem found their approach “very strange.”

The National Academy of Science labeled the team’s work “groundbreaking” and “one of the most important developments in mathematics in the last decade” because it provided new tools to look a series of problems in pure mathematics, quantum physics, and computer science.

At first, given their work on graphs, the team though solving Kadison-Singer would be relatively straightforward. It didn’t turn out that way. After four years, they had reached a point where they considered giving up. In the fifth year of work, however, they came up with their solution.

Dan attributes his knack of doing the kind of research that garners awards to his first job, an associate professor at MIT. A colleague told him that at MIT the most important thing was a particular version of the academic dictum to publish or perish; it was not how much he published but the nature of what he published.

“The advice I got was to do one or two really excellent papers,” he says. “Don’t publish unless it’s a big deal.”

Dan decided to work on what he saw as “ambitious problems.” Still, he recognized then there was an obvious pitfall in this.

“It’s dangerous because all your eggs are in the same basket,” he says.

In fact, he says his work on the Kadison-Singer problem was the biggest gamble he has ever taken in his professional career.

When Dan, who grew up in Philadelphia, was still a youngster, he went to a summer camp at the Center for Talented Youth at Johns Hopkins University—”Nerd camp,” he calls it.

By high school, he had developed an expertise in computers. His parents had gotten him a Commodore PET computer secondhand. He had never taken any computer classes at school, but the machine’s prior owner had included a lot of literature on computers. Dan read it all.

By the summer of 9th grade, he had taught himself enough about computers to get a job working for a company involved key card security. It was fairly difficult, he admits, to talk the business into hiring someone who had just finished 9th grade.

“It took some convincing,” he says, “but the job was pretty cool.”

The next summer, he got a job in a financial services and brokerage firm.

Dan started using computers to solve challenging puzzles as a result of an article he read in Scientific American about a complicated three-dimensional puzzle called Bill’s Baffling Burr.

He ordered the puzzle thinking the challenge would be to disassemble it; instead, it came disassembled and the first challenge was to putting it together. As an alternative to the trial and error of fitting pieces together, Dan looked the solution up on the computer.

For college, Dan chose Yale, precisely because it gave him the chance to focus on subjects that were not math- and computer-related.

“I was interested in a lot of different things,” he says.

He did his graduate work at MIT and taught there before coming back to Yale.

Around his home in Guilford, Dan likes to do construction projects. He built loft beds for his two children, now 17 and 19, when they were small and has constructed both a shed and a greenhouse on his property. He thinks this summer might be the time for him to start work on another greenhouse.

He describes his work as rough carpentry.

“I don’t do fine carpentry,” he says.

But he does do fine mathematics.