Kou is shaking up the world of statistics

5 min read

From molecules to the stock market, statistics professor asks, ‘What are the chances?’

Harvard statistics professor Samuel Kou, now 34, grew up in Lanzhou, a city in China’s mountainous northwest near the border with Inner Mongolia. The altitude there is higher than Denver’s storied mile, and earthquakes rumble through town several times a year.

Lanzhou, though large, is remote enough that the denizens of Beijing or Shanghai will ask people from Kou’s province, Gansu, if they ride camels to school.

Kou (it’s pronounced “Cole”) laughs at the idea. But it is a sign of how far across the world he has come, and how fast. Twenty years ago he was a boy riding a bicycle to a provincial Chinese secondary school, where he excelled at physics and math — and occasionally rode past buildings where occupants rushed outside as the earth shook.

Today, Kou is a young scholar shaking up the world of statistics. He is inventing new ways — for instance — to infer the probability of biochemical reactions in nanoscale molecules, and to model the behavior of American stock markets.

Working with Harvard chemist X. Sunney Xie, Kou has come up with statistical models that predict random phenomena in a single molecule, a tiny scale that so far has baffled classical theorists.

With his brother Steven Kou, a professor of industrial engineering and operations research at Columbia University, he has puzzled out statistical predictors of market behavior. In an insightful 2004 paper, Kou and his brother offered a counterintuitive view of high-tech stocks: The more volatile such stocks are, the more predictably they behave in an orderly way.

Thinking of Kou’s interdisciplinary versatility, Statistics Department Chair Xiao-Li Meng likes to quote the legendary John Tukey, who coined the terms “bits” and “software.” Tukey said, “The best thing about being a statistician is that you get to play in everyone’s backyard.”

In turn, said Meng, “Sam signifies a new generation of statisticians who are at the very forefront of the scientific arena because of their deep statistical insights, firsthand scientific knowledge, and powerful analytical skills.”

But Kou brings more, he added, “not only a beautiful mind for research, but also a beautiful heart for teaching.”

In many ways, Kou is still the modest, intense, and hardworking young man riding that bicycle in Lanzhou, where by high school he played center on the basketball team and took up bridge and reading military history as his chief diversions from problem sets and equations. (Today, after hours, he is deep into reading about the Napoleonic era — still captivated by war, battle, and what he calls “a boy’s fascination.”)

His father is a retired civil engineer who specialized in making buildings earthquake-proof, and his mother is a retired high school chemistry teacher. So math and science had a place of honor in Kou’s boyhood. By his first year of high school, Kou was enrolled in a fast-track “Olympic class” of 32 young science scholars recruited from across his province of 20 million. They were treated to guest lecturers, university-level professors, and advanced classes in math, physics, and English.

Those advantages came with a price. Class started at 7:30 a.m., broke for a long European-style lunch, and ended at 6:30 p.m. Homework took Kou from 8 to 11 p.m.

By the end of high school, Kou was one of 30 students on China’s national training team for the International Physics Olympics, and one of 15 on the national training team for the International Olympiad in Informatics. By the time he arrived at the prestigious Peking University, three directions called to him: math, physics, and computer science. (In China, freshmen declare a major right away.)

But Kou entered college already armed with a favored route, inspired by a visit to his high school by Columbia University mathematician and applied statistician Yuan-Shih Chow. “He very much convinced me that math was the way to go,” said Kou, especially since statistics could be used to understand social phenomena.

Halfway through Peking University, Kou had another inspiring experience: two courses, taken concurrently, pointed to two different intellectual futures. Kou did very well in “Abstract Algebra,” but its remote, cerebral cast did not appeal to his heart. The course “Introduction to Probability,” on the other hand, with its intimations of action upon the real world, “directly appealed to me,” said Kou. “Suddenly, I [fell] in love with it.”

Math, to him, he said, was best used to solve “real-world problems.”

When he applied to graduate schools in 1997, only American universities were in Kou’s sights. He chose Stanford University over Harvard, since at the time the California center had a bigger statistics faculty.

The only barrier to U.S. study, at first, was language, said Kou, who took three years to gain confidence in his English. It was tough when studying applied statistics, he said — and even tougher when someone told a joke.

Beyond language, American schools offered a steep learning curve in classroom dynamics. In China, said Kou, “you’re given all this knowledge systematically,” but never encouraged to think critically or ask hard questions — even of the professor. “That’s a very different academic environment.”

The Chinese academic system results in students testing very well, but for academic research Kou prefers the American way — stimulating, free, and fearless. “You’re not punished for asking stupid or irrelevant questions,” he said. “You’re more encouraged to think on your own.”