Betting on the Best Statistical Modeling

How do methods hold up when data is blurred for privacy? Lasse Vuursteen wants to know.

Lasse Vuursteen

The following is a summary of a story that originally appears on the Trinity College of Arts & Sciences website.

As a new assistant professor in Duke’s Department of Statistical Science, Lasse Vuursteen brings an unusual origin story to the field. His path began as a teenager playing poker semi‑professionally in the Netherlands, devouring books on strategy. Game theory led him to math, econometrics,and statistics. He still plays for fun, but now he focuses on the math that helps researchers make sense of data.

A big part of his work is figuring out when a statistical method is truly the best choice. Some problems have straightforward answers, but many don’t. “We don’t have theory that tells us what works or what is optimal,” he said. Even small changes in a problem can turn a once‑ideal method into the wrong tool.

He also examines how statistical methods hold up when they run into real‑world limitations. Vuursteen looks at what happens when data must be blurred or restricted for privacy, or when massive volumes of information strain communication systems. “A practitioner would like to know: what is the utility of this data if I impose privacy and how much privacy can I impose?” says Vurrsteen.

His tools stretch into transfer learning as well, helping determine when results from one dataset can responsibly apply to another.

And for Vuursteen, the appeal of all this work is simple: “I use math to understand the laws of statistics. … “Sometimes you can find very interesting dynamics where for certain settings [a method] is optimal, but if you tweak the setting a little bit, it is no longer the optimal thing to do,” he said.

Read more about Vuursteen and other new Trinity colleagues on Trinity’s website.