PROFILE Profile of Christos Papadimitriou Farooq Ahmed Science Writer When Christos Papadimitriou graduated in closely guarded but that was the only 1972 from the National Technical University thing that really got my attention.” of Athens in his native Greece, he arrived at a breaking point. Like all Greek men he was Complexity required to join the armed forces, and he After graduating from college and serving enlisted while the country struggled un- in the military, Papadimitriou left Greece der military rule. for northern New Jersey. When he arrived “I went in fully expecting to serve and then at Princeton in 1973, the computer science work quietly as an engineer somewhere,” group at the university had five people. Papadimitriou recalls. “But during military However, for Papadimitriou the setting service, the ruthlessness of the regime struck provided a stimulating intellectual environ- me in a very powerful way. I knew that I ment and an opportunity to influence an could not stay.” He applied to graduate emerging field. “ ” “ school and joined Princeton University’s For the first time in my life, he admits, I — fledgling computer science group. Demo- found something that I really enjoyed and I cratic reforms in 1974 ousted Greece’s mili- was good at it! Suddenly, I decided to take ” tary junta the same year that Papadimitriou my graduate studies very seriously. Papadi- earned a Master’s degree. He would com- mitriou worked with Kenneth Steiglitz and Christos Papadimitriou. Image courtesy of plete his doctorate two years later in elec- began his lifelong investigations into com- Eirene-Chloe Markenscoff Papadimitriou. trical engineering and computer science. putational complexity. “ ” In the past four decades, Papadimitriou’s It all started with Alan Turing, Papadi- influence on the field of computer science has mitriou says. In 1936 Turing, a British between several cities such that each city been far-reaching. His articles launched new logician who established the conceptual is only visited once, is of the NP-complete research areas, his textbooks have become framework for computer science, proved (3) type. His work on complexity, often standards, and his works of fiction—two nov- that some problems could not be solved with collaborator Mihalis Yannakakis, now els and a graphic novel—have introduced the by computers: that no algorithm could be at Columbia University, transformed and history of logic and computer science to written to find solutions, even if the questions “ expanded the subject (4). broad audiences. In 1996 Papadimitriou were well-defined (1). Computer science joined the University of California, Berkeley, was a field born with a dire understanding Academic Freedom ” where he holds the C. Lester Hogan Chair of its limitations, adds Papadimitriou. When he received his doctorate from ’ ’ of Electrical Engineering and Computer Sci- Building on Turing sandothers work, in Princeton, computer science was still a small ’ ence. He was elected to the National Acad- 1971 the University of Toronto sStephen enough field that Papadimitriou felt like he “ emy of Sciences in 2009. Cook devised a way to distinguish between knew all that there was to know at the problems that could be solved quickly, in time.” He left as a theoretician with a de- Leaving Athens polynomial time, and others for which the termination to advance the field and ensure Born to educators in a small town outside of time to uncover a solution increased ex- that it became one that would solve “sig- Delphi in the Greek mountains, Papadimitriou ponentially (2). In the field’s parlance, the nificant and meaningful” problems. moved to Athens with his family when he latter class of problems is now referred to Over the next two decades, Papadimitriou was five years old. The “mountain accent” he as “NP-complete” or “NP-hard.” taught at six different institutions: Harvard carried into the financial and political capi- Papadimitriou describes the difference University, Massachusetts Institute of Tech- tal of Greece earned him few friends. “That between these types of problems as similar nology, Stanford University, the National grief led me to studying as a way of escaping to trying to find a needle in a haystack. “If Technical University of Athens, the Uni- my social problems.” the needle’s ferromagnetic then you can use versity of California at San Diego, and At the National Technical University of amagnet,andyou’re done. But what if that the University of California at Berkeley. Athens, Papadimitriou studied mechani- needle is plastic? You have to look under Papadimitriou’s return in 1981 to Athens, cal and electrical engineering. However, every blade of hay.” where he taught for at least one semester he wasn’t keen on the subject. Instead, “I decided to dedicate my career to try and for the next seven years, provided, he says, Papadimitriou stole into the university’s understand the difference between the nee- a vindication. “It was a bold, perhaps risky computer laboratory and taught himself dles,” Papadimitriou says. “That’scomplexity how to program. “There was an IBM 1620 theory.” Papadimitriou proved that the his- This is a Profile of a recently elected member of the National Academy computer, probably a million times less toric traveling salesman problem, in which of Sciences to accompany the member’s Inaugural Article on page powerful than today’s cellphones. It was one has to calculate the shortest route 15881. 15858–15860 | PNAS | November 11, 2014 | vol. 111 | no. 45 www.pnas.org/cgi/doi/10.1073/pnas.1405579111 Downloaded by guest on September 30, 2021 PROFILE decision for me to return. But I was excited and colleagues found that a two-dimensional to contribute to the development of com- protein-folding model, the hydrophobic- puter science in Greece.” polar model, was of the NP-complete type It was at Berkeley’s northern California (10). Papadimitriou’s work on databases in campus, however, that Papadimitriou found the 1970s and 1980s culminated in a 1986 an intellectual home. “There is a very ag- monograph, The Theory of Database Con- gressive form of academic freedom at Ber- currency Control (11). keley. I first visited as a Miller Fellow in 1978, and it gave me a feeling of extreme freedom: The World Wide Web If it feels good intellectually, then you must After returning to the University of do it! And do it with passion.” California, Berkeley in January 1996, From the mid-1970s to the mid-1990s, Papadimitriou reached another inflection as computer science matured, Papadimitriou point in his career. In the early 1990s, the taught nearly every course, from program- Internet and the World Wide Web had ming languages to compilers, operating reached a critical moment. A plethora of systems, and databases. His 1982 textbook, small voices were virtually publishing con- Elements of the Theory of Computation (5), tent, and major companies were profiting written with Harvard University’s Harry without a brick-and-mortar presence. Lewis, explained the foundations of emerg- “The Internet,” explains Papadimitriou, ing fields and influenced how courses on the “was something completely new and dif- subject were taught. With his graduate ferent. Like mathematics, its existence is adviser Steiglitz, Papadimitriou published based on minimalistic principles, but you a graduate textbook on algorithms and cannot point to a designer, a committee, a complexity called Combinatorial Optimi- user, or a company that made it happen. It Christos as featured in LogiComix. Image grew out of the interaction of many.” zation in 1982 (6). A more recent text- courtesy of Apostolos Doxiadis, Christos book on algorithms has become a standard This line of reasoning led to a realization: Papadimitriou, Alecos Papadatos, and Annie for undergraduates (7). “Computer science was no longer about the di Donna, 2009, Logicomix Print Ltd. and “Frankly,” Papadimitriou explains, “teach- computer. The creation of the Internet was ing is the only way I know to understand transforming computer science into a nat- Bloomsbury Publishing Plc. something. But these days,” he adds with a ural science. But it was also transforming laugh, “the courses have become so com- computer science into a social science! Be- Knuth Prize, awarded jointly by the Associ- plicated that I doubt I could pass them!” cause you cannot understand the Internet ation for Computing Machinery and the In his research, Papadimitriou used mathe- without understanding the incentives of the Institute of Electrical and Electronics ” matical tools to uncover complexity, wher- entities that use and continue to create it. Engineers. ever it exists. He defined new classes of Papadimitriou focused on understand- complexity, which have led to breakthroughs ing the Internet as a neuroscientist ap- Anarchy and new ways of understanding computa- proaches the brain or an astrophysicist Papadimitriou’s work has helped create and tional problems (8, 9). Extending his work views the universe. For his contributions, engage new scientific disciplines. Investiga- beyond computer science, Papadimitriou in 2002 Papadimitriou received the Donald E. tions into decision making and mathematical models of conflict resolution drove contri- butions to algorithmic game theory, a field at the intersection of economics and com- puter science. In 1999 Papadimitriou and Elias Kout- soupias, now at the University of Oxford, defined the price of anarchy, a mathematical way of measuring how performance in a system degrades as a result of the selfishness of its users (12). “Imagine,” Papadimitriou explains, “that the Internet was not run by thousands of providers but instead by one benevolent dictator who wanted to maxi- mize social welfare. How much better would it be than what we have today? That’s what we call the price of anarchy.” Often, the answer is, “Not that much.” His recent work in game theory concerns the Nash equilibrium, a fundamental concept Papadimitriou and his team discuss how to tell the story of British mathematician Betrand in economics for which Princeton’sJohn Russell’s epic search for the truth.