Mathematician without too many plans

Now, more than two years on, van de Geer recalls the unexpected news of her impending retirement with some amusement. The incident sits well with the Dutch researcher, who is not one for planning too much in advance. “I did not have a career plan,” she says from her office in the main building, which has been cleared of everything apart from a bookshelf, a desk, and a meeting table and chairs. Although she was designated professor emeritus of mathematics at the end of July, van de Geer will continue teaching at ETH next semester.

Van de Geer, who has now turned 65, had never planned to go into research or to be a professor. Nor to live in Switzerland one day. Even her choice of research topic while studying mathematics in her hometown of Leiden was coincidental: “One reason I became interested in statistics was that the professors in that subject were so friendly, and I was fascinated by their work and their findings.”

Searching for a structure

Van de Geer’s research focuses on statistics for high-dimensional problems, including probability and machine learning. She is always looking for solutions to reduce complexity. “We live in a complicated world – but in all things, there’s an underlying structure. It’s this structure we want to find. Statistics is a way of extracting meaningful information from the abundance of data,” says the scientist.

In the past, one developed a model and then analysed the associated data, van de Geer explains. “A classical paradigm of statistics is that you shouldn’t adjust the model after you’ve looked at the data. Now we’re moving away from that paradigm and deciding, for example, which variables to include in the model and which not, based on data.”

Van de Geer realised some 20 years ago that dealing with huge amounts of data was going to be crucial in the future. In 2002, she co-organised a workshop on the analysis of high-dimensional data. One theme was the evaluation of microarrays in biomedical research and medicine. These chips can be used to measure gene expression simultaneously across thousands of genes, which enables accurate diagnoses. “Instead of the doctor taking individual measurements – determining blood pressure, evaluating a urine sample or listening to heart sounds, for example – with microarrays you get 20,000 measurements in one fell swoop,” says the professor emeritus. “The challenge now is to make a statistical analysis of such microarray data from a relatively small sample of people, so as to assess the risk of a particular disease in an individual.”

Ten-fold surge in students

Van de Geer conducts research in a field that is now at the forefront: more and more powerful computers mean that more and more data can be stored, faster and faster. Anyone wanting to interpret and understand these data has to be able to record and evaluate them statistically.

The rising interest in statistics is reflected in the number of students: in 2005, when van de Geer started at ETH Zurich, a dozen people came to her lectures. Today, she says, at least 120 students attend, which is also due to the new interdisciplinary degree programme in data science.  

This surge in students is something that both pleases and worries the professor. It is particularly problematic when it comes to Master’s theses, she says. “With so many students, one-to-one supervision is almost impossible. On top of that, correcting exams other than those based on multiple-choice questions is becoming more and more time-consuming”.

For van de Geer, personal contact is paramount. She is delighted when students thank her for a lecture or relate years later that what they learned helped them in their careers. “Hearing something like that makes me feel very proud,” she says.

Knighted by the king

Van de Geer is a member of several scientific academies, such as the European Academy of Sciences, the German National Academy of Sciences Leopoldina, and the American National Academy of Sciences. In 2010, she was invited to give a lecture at the International Congress of Mathematicians – a special honour. From 2007 to 2015, she was a member of the Research Council of the Swiss National Science Foundation. In 2015, she was made a Knight of the Order of Orange-Nassau by the King of the Netherlands for her achievements. From 2015 to 2017, she was President of the Bernoulli Society for Mathematical Statistics and Probability.

Before coming to Zurich as ETH’s first woman mathematics professor in 2005, van de Geer was a full professor at Leiden University. She ended up in Switzerland primarily because her application for research funding in the Netherlands was rejected, as she recounts in her farewell lecture. With some amusement, she reveals the reason the National Research Council of the Netherlands gave for turning her down: “We have doubts as to whether Ms. van de Geer will be able to inspire doctoral students and postdocs,” their letter reads.

It seemed a ropey justification, particularly as the same research council considered her work scientifically sound. But at that moment, ETH professor Peter Bühlmann got in touch to suggest she apply for a post in Zurich. “Applying to ETH that was one of the best decisions I’ve ever made,” says van de Geer.

A woman in a man’s domain

Being one of only a few women in her discipline at first did not bother her. “As a student, I was happy to have so many men of the same age around me,” says the scientist with a twinkle in her eye. But in the course of her career, either her presence or her research work seemed to bother a couple of them: "When I was studying, for example, certain professors tried to intimidate me or imply that my place was elsewhere.”

Even after graduating, she was occasionally treated differently from her male colleagues. In Leiden, before she was appointed professor, certain members of staff asked her to make coffee, type texts or do their office tasks. “Then once I was made professor, some of the men found me a threat, and put me to the test. I felt I had to prove I was some kind of genius!”

She remembers how, as she embarked on her professorship in Leiden, her boss urged her to write a proposal for research funds – together with a colleague from a completely different field. "We saw little point in it, but nevertheless tried to link our two research areas. Of course, the proposal was rejected because there was no way of making a viable connection.” This, she believes, would probably not have happened to a man.

Van de Geer has not experienced anything like this in Zurich. “ETH is a great place to work – you can get on freely with your research and submit the proposals that you choose and that are relevant,” she says. She is especially grateful to the many kind people in Zurich who welcomed her and made her feel at home.

Statistics and the state of flow

The professor emeritus lives in Illnau-Effretikon with her son, who is studying computer science at ETH Zurich. “But he hates machine learning – and that’s more or less a modern version of statistics,” she says with a chuckle. In her free time, she likes using her hands: “I enjoy cleaning and gardening, and repairing things that are broken”. As well as drawing, she listens to and plays classical music; she has both a harpsichord and a piano at home.

Even after all these years, van de Geer is still captivated by her work. “So much of everyday life is based on simulations and empirical findings. I enjoy extrapolating evidence and it’s very rewarding to come up with a theoretical basis for why some algorithms work very well and others not at all.”

As a scientist, she finds working in mathematical statistics is a creative process that sometimes brings her to a flow state. It’s like solving a riddle or a puzzle, she explains. “You think on it and keep trying things out – and then suddenly the pieces start to fit together”. That’s a great moment, one when you completely forget the world around you. “You don’t want to be distracted by anything; you just want to get on and finish the puzzle.”

Lending a hand

This coming semester, the professor emeritus of mathematics will continue to give lectures as well as supervise and assess ongoing Master’s theses at ETH. And she will keep up her commitment various international committees until at least 2025. And after that? “Well, I don’t know quite yet; I’m up for anything! There’s more to life than the work I’ve been doing, and I’m looking forward to this new phase – even if I’m not sure what’s around the corner!”

She would certainly consider doing voluntary work, such as cleaning for elderly people. Improving her piano playing also comes to mind. Or simply quietly working out solutions on a blank piece of paper to trigger a flow state. Just as long as she does not have to plan too much!