Meet our new PIs: Mikko Karttunen

Karttunen's position is part of the international recruitment of principal investigators by the ELLIS Institute and joint professorships with partner universities funded by the Foundation PS established by Peter Sarlin. Through the recruitments, the institute aims to strengthen top research in artificial intelligence and machine learning in Europe.

Computational methods to help medicine and biology

Karttunen's research area combines materials science, biology and medical technology with high-performance computing simulations, machine learning and artificial intelligence. At Western University in Canada, he has worked as Professor of Physics and Chemistry and as Canada Research Chair (Tier 1) in Computational Materials and Biomaterials Research.

"Part of my research is purely theoretical and computational methods. The other part almost always includes an applied or experimental partner. As an example of applied materials science, we worked with the National Research Council (NRC) in Canada in material optimisation with the aim of optimising certain properties of ceramic materials in terms of durability and thermal conductivity. We used machine learning to create a model that predicted how the ceramic material should be produced. The NRC produced the model and carried out the experimental work. The simulation-machine learning-experimental pipeline worked very well," Karttunen says.

A significant part of Karttunen's research relates to medicine and biology, especially biological membranes. The research topics include how drugs affect and penetrate the membrane, how drugs can be transported, and how proteins attach to membranes.

"Membranes are environments that have been part of my research in different forms for nearly 25 years. Some are rooted in basic research, while others are linked to experimental research. Typically, we have tried to involve a partner who conducts the experimental side, as my work is computational and theoretical."

Karttunen's research is highly interdisciplinary. Computational methods are useful across a wide range of disciplines.

"Both physics and chemistry are involved. At the atomic level, the boundaries between physics and chemistry become semantic. We have specific molecules whose chemical structure we model. Chemical reactions, however, are different from what we normally study."

Hard and soft substances at the core of research

Karttunen has seen the academic world from many angles and continents. He graduated as Master of Science in Technology from Tampere University of Technology in 1993 and headed to Canada as a doctoral researcher by chance.

"I knew I wanted to study abroad. My initial plan was to go to the United States, and I had already been accepted there. During my last year in Tampere, Professor Tapio Ala-Nissilä returned from his trips. We started to talk about living abroad, and Ala-Nissilä knew a very good research group in Montreal. After meeting their principal investigator in Finland, I applied there for my doctoral research."

Karttunen defended his doctoral dissertation at the McGill University's Department of Physics in Canada in 1999. Since then, Karttunen's research has taken him all over the world: he has worked, for example, at the University of Fudan in China, the University of Hokkaido in Japan, the University of Santiago de Chile in Chile, the Max Planck Institute in Germany and elsewhere in Europe.Internationality has opened new perspectives on the academic world.

"You learn new approaches and to communicate with different kinds of people. Interaction is the key. If you stay in one place, you will hit a wall eventually. It's good to see the world. It’s important to see different approaches," Karttunen reflects.

In his career, Karttunen has been involved in interesting projects where multi- and interdisciplinarity have played a key role. His work has involved classical physics and superconductors as well as biological and material physics.

"Since the dissertation days, I've carried material physics and biological physics with me, hard and soft substances."

Out of the Finnish ELLIS partners, Karttunen has worked with Aalto University for a long time, and the cooperation will continue. With UEF, Karttunen has cooperated in pharmaceutical research and co-published articles with Professor of Biopharmaceutics Arto Urtti. 

"It's interesting to move to a new, more networked environment. I hope that what I bring to the table will give others ideas and opportunities for cooperation."

AI development requires an interdisciplinary approach

Along with the professorship in computational engineering at the University of Eastern Finland, Karttunen joins the ELLIS Institute and the PS Fellow researcher network, which focuses on and invests heavily in AI and machine learning research.

AI applications have taken massive leaps forward in recent years. The leading research in the field is currently conducted in the United States and China.The ambitious task of the ELLIS Institute is to bring European AI and machine learning research to the global forefront.

"Networking is extremely important in Finland's conditions. The ELLIS Institute and Foundation PS network is a wonderful thing. There is also networking towards Europe, which is absolutely essential on a global scale."

Karttunen is familiar with the development of the field: he became acquainted with machine learning already as a student at the Tampere University of Technology in the early 1990s.

"Our group, led by Professor Kimmo Kaski, was very interdisciplinary, and the research topics ranged from theoretical physics to medical machine learning. In terms of interdisciplinarity, we were certainly ahead of our time, and it highlighted the advantages of the approach: while the phenomena may be different, the methods are interlinked," Karttunen says.

Since then, much has happened behind the scenes of artificial intelligence and machine learning research, both in the development of methods and in computational capacity. In particular, Karttunen highlights Geoffrey Hinton's pioneering work in method development.

In recent years, artificial intelligence has become a reality for ordinary consumers. Interest in the use of AI has been unprecedented, and it is expanding at a huge rate in everyday lives. At the same time, there is a growing concern about what the ever-expanding use and development of AI will lead to.

"I share Geoffrey Hinton's view that there are areas of uncertainty in AI development. We shouldn't follow everything blindly. AI is a bigger change than the arrival of the internet back in the day. It involves many questions, such as who controls data, how it can be collected, and how we approach autonomous actors. The development of AI must be interdisciplinary, meaning that social scientists should also be included. Ethical aspects and regulation are also essentially linked to AI," Karttunen concludes.