Professor Elisabeth Paté-Cornell

The Stanford Challenge

Elisabeth Paté-Cornell specializes in understanding risk—going behind the scenes to perceive threats and trying to prevent them from materializing. "If we're doing our job well, nobody hears about it," says Paté-Cornell, Stanford's Burt and Deedee McMurtry Professor of Engineering and chair of the Department of Management Science and Engineering. It's a kind of analysis that has applications far from engineering itself, from detecting terrorists' plans to fighting infectious diseases to predicting human error in a variety of situations.

Paté-Cornell, who co-chairs the International Initiative, has done her job so effectively that her research foresaw a major breakdown in the U.S. space program. Asked by NASA to assess risks to the space shuttle in the early 1990s, most researchers would have focused on the O-rings that had failed in the 1986 Challenger crash. Instead, Paté-Cornell chose to study the vulnerability of the shuttle's thermal protection tiles.

"It's essential to think beyond the recurrence of the past event," she says. She worked with Stanford graduate student Paul S. Fischbeck (now a professor at Carnegie Mellon University) to examine tile performance in previous shuttle missions. Paté-Cornell and Fischbeck identified which of the shuttle's 25,000 tiles were most prone to failure and computed the risk of losing tiles and an entire orbiter vehicle due to poor bonding or the impact of a piece of detached insulation from the external fuel tank—precisely what led to the loss of the Columbia a decade later. (Though she had shared her findings with NASA, the agency was unable to correct some design and maintenance flaws in time to prevent the accident.)

Paté-Cornell's work on the shuttle illustrates both the sensitivity and the volatility of her subject matter. In any given case, she reviews a wide range of potentially catastrophic, ever-changing threats and seeks to understand the risks they pose. Because of the many variables she takes into account, her research is fundamentally interdisciplinary, combining engineering and mathematics with human factors, from organizational policies to workers' fatigue and stress. Stanford—and especially the merger of three departments to form the Department of Management Science and Engineering, which Paté-Cornell spearheaded in 1999—has given her leeway to pursue each of these angles in a wide breadth of projects. In addition to her NASA work, she has used this combination of engineering methods and management analysis to assess rates of error among hospital anesthesiologists and failures of offshore oil platforms and to investigate a variety of defense and security issues. Paté-Cornell expects that the new International Initiative will increase interest in and support for this kind of cross-disciplinary, problem-solving work.

Uncovering so many layers of a problem can mean developing a special relationship with her research subjects. At NASA, Paté-Cornell met with technicians under the shuttle orbiters and also in a local bar. As the workers opened up about their experiences, she gained some surprising insights about why and how corners were sometimes cut in the maintenance of the tiles.

From a technical point of view, Paté-Cornell's approach makes inventive use of systems analysis and also Bayesian probability in analyzing risks. First developed by Thomas Bayes in the 18th century, the Bayesian model allows a researcher to assign rates of likelihood to random events.

These risk-assessment tools can be applied to issues of national security with encouraging results. After the attacks of September 11, 2001, Paté-Cornell turned her attention to detecting and preventing terrorism. Because of limited budgets for counterterrorism, governments must determine which threats are most serious and deserve the most resources. "We are trying to set priorities given what we know," says Paté-Cornell, who was invited to serve on the President's Foreign Intelligence Advisory Board. The biggest challenge is that terrorism—like all other unpredictable threats—does not necessarily repeat itself. To understand today's risks, it is not enough to study the past. Risk managers must extrapolate into the future.

To do so, Paté-Cornell has looked at the ways intelligence officers digest the massive amounts of information they receive each day. She showed them how to compute approximate probabilities of future events based on what we know and on the likelihood of false positives and false negatives in the information gathered. Her model, explored in several journal articles, has applications for intelligence operations not only in the United States but around the world.

But Paté-Cornell is not content to stop there. She plans to broaden this model's impact with a new project on "early warnings"—an effort to create alarm systems for major threats of all kinds, from terrorism to tsunamis to epidemic diseases. She hopes to see several Stanford faculty members team up to research how to implement such systems on a global scale. "Stanford's International Initiative is an ideal framework for this type of work," says Paté-Cornell. Because of the university's strengths in engineering, security, geophysics, medicine, and other fields, Stanford is in an excellent position to pioneer new solutions, she explains.

Whatever this polymath scholar turns her attention to next, it's predictable that she will make a difference to those trying to avert tragedy. "I like to get my arms around a complex project and make sense of it, even when people are saying, 'You can't do it—it's too complicated,'" Paté-Cornell says.