Aarhus University Seal / Aarhus Universitets segl

Professor appointment: Kasper Green Larsen

Kasper Green Larsen is 35 years old and a newly appointed professor at the Department of Computer Science - that makes him one of the youngest professors at Aarhus University. But despite his young age, Kasper Green Larsen has an impressive array of groundbreaking research results and awards behind him and is passionate about making his mark on a field that is more needed than ever.

When Kasper started studying Computer Science at Aarhus University in 2008, he had an idea that the study should be used for practical programming in the labor market, but research and theory ended up being his passion and today he is internationally recognized for his research results.

-    I’d done a bit of coding before my studies, and I also had a student job with programming, but I was very quickly caught up in the satisfaction you get from being able to work something out theoretically. When you can rule two lines under a correct answer.

Kasper Green Larsen became good at drawing lines and started his research career as a PhD student at the basic research center MADALGO. As a mere 26-year-old, he made a breakthrough when he was able to mathematically calculate how fast and efficient a search in a database can become - even if computers get faster and faster with time. Kasper's research proved that there is a ceiling for how fast one can search, and that that ceiling is not to be broken. Thus, large companies - such as Google - could simply stop improving databases, provided they had reached the mathematical limit, and thus can spend their energy on something else.

The research was carried out in the research group Algorithms, Data Structures and Foundations of Machine Learning - the very same research group he currently manages. Kasper's research assumes both practical and theoretical form, and the results form the basis for research in many other branches of computer science. In his research, Kasper in particular focus on how to optimize algorithms, and not least on how to find the ultimate best algorithm.

- I am very fascinated by developing methods and algorithms to analyze and process data faster than has previously been possible. And maybe even more of finding out if it can be done faster and more efficiently at all.  There's an awful lot of mathematics in it, and I've always had a flair for math, but where traditional mathematics can be a little abstract, computer science is relatively concrete. We can use mathematics to design algorithms and thereby get computers to solve problems for us, faster and more efficiently. We can also describe exactly what a computer is capable of, and thereby find definitive formulas that tell us whether we have the best algorithm now, or whether there’s something more to be gained in the future," says Kasper Green Larsen.

And it is exactly the mathematics Kasper used when he, in collaboration with his colleague Jelani Nelson from Harvard University, solved a data compression problem that computer scientists have been trying to improve since 1984. With huge amounts of data, it may be necessary for analyzes to compress the data, but it must be done in a way which retains the properties of the raw data, while ensuring that one can still search quickly and efficiently in the compressed data set. Through his research, Kasper proved that the 1984 method (the so-called Johnson-Lindenstrauss transformation) is the most optimal method of compressing data.

- We now have a complete mathematical understanding of how much high-dimensional data can be compressed. In addition to the fact that computer scientists no longer have to spend time researching methods for compressing data, it also means that we can put all our efforts into making the compression process faster. The amounts of data will only increase in the future, so faster algorithms are essential, Kasper Green Larsen told Danish manazine Ingeniøren in 2017.

And even though faster data analysis can benefit everyone, Kasper is very much aware that much of the more theoretical research he does will not necessarily have great significance in ordinary people's everyday lives, but still, much of the basic research being carried out now will end up being used by companies in the future. 

-    The vaccines we’ve developed during the corona epidemic were developed on the basis of research conducted at universities, but none of us knew that this research would benefit us now – or that it would be in that form," says Kasper, and he continues:  - So who knows? It may well be that a lot of what we’re doing now will be used for something that we had not expected at all. 

Right now, his focus is on a topic with wide-ranging interest: what the Department of Computer Science calls foundations of machine learning, and many others call artificial intelligence. For Kasper Green Larsen, it’s a real treat to know that we simply can't explain why neural networks and image recognition, for example, work quite as well as they do. The usual formulas simply cannot give an answer.

-    Practitioners in companies are delighted that everything works, but what theorists like me find really exciting is finding out why," says Kasper Green Larsen.

For him, it is thrilling that gigantic networks, with perhaps a billion different adjustable parameters, can see patterns and recognize images with perhaps only a million input images or data elements – for example images of people or DNA strings. 

"I want to understand it better and thereby inspire others to make it even smarter. However, the fact is that many of these calculations and networks require enormous resources. Google can cope because they have massive capacity and can get 10,000 computers to calculate something. The power consumption alone makes it impossible for smaller companies or small research units to get involved, but if we can find models that require less data and yet can make good predictions or recognition, then we will have come a long way. Everything we do is available to the public. It will all benefit everyone," says Kasper Green Larsen.

A researcher, but also a communicator

Although Kasper is a busy researcher, he also finds time to spread the message about what computer science is and why it is possibly one of the most important areas of research.

-    Many people around the university consider their own research as the most important, but if I look at the big picture, from a societal perspective, the 200 students I teach every year are probably just as important to society. That’s important to remember," says Kasper Green Larsen.

Therefore, the newly appointed professor is very aware that his communication skills must be utilized. Also, for those who have not even considered a field of study yet - e.g. by lecturing to high school classes.

- I would like to give many just a little understanding of computer science and the importance of the subject in our everyday lives. There are so many people who have no idea of what computer science is, and here I would like to use my abilities, says Kasper.

Therefore, he has been involved in developing a program under the Science Club (Videnskabsklubben), which is aimed at students in 5th and 6th grade, where you get acquainted with computer science and the mathematics behind it. Simply to raise awareness about the field and perhaps create an interest that will persist. A quick search on the Jobindex portal reveals that there’s a lot to be done. Almost 2,100 information technology vacancies, in Denmark, are testimony to the huge need for manpower with precisely these skills.

- It is important that we get rid of the erroneous image we have had in the past. It would be sad to miss talented students because they have a wrong idea of what computer science is or the exciting opportunities the education offers, says Kasper Green Larsen.

The interest in the dissemination of a complex subject is great and is also unfolded in the younger department of Royal Danish Academy of Sciences and Letters, The Danish Young Academy, where Kasper Green Larsen has been a member for the past five years and thus been able to connect with lecturers and researchers from other departments at Danish universities. A path suggested by Kasper's former mentor, Lars Arge, who died just 53 years old just over a year ago.

- He was a great inspiration for me and the man who made me interested in research, says Kasper Green Larsen.

And what luck that he chose research, because two lines can already be drawn under the fact that Kasper Green Larsen has had an impressive significance for computer science. Congratulations on your new title Kasper - we look forward to following you and your research for many years to come.

Facts

Kasper Green Larsen

35 years

Lives in Viborg

2 children at 10 and 12

Education

Bachelor in Computer Science, Aarhus University, 2005-2008.

Ph.D. in Computer Science, Aarhus University, 2008-2013.

Tenured at Department of Computer Science since 2018.

Appointed Professor at Aarhus University, Feb. 1, 2022.

Research group leader at Department of Computer Science, Aarhus University.

Kasper Green Larsen

Professor Department of Computer Science