An algorithm helps predict who will develop diabetes

Mikko Koivisto, professor of computer science, wishes to develop reliable automated reasoning.

What are your research topics?

I investigate algorithms, or how computationally challenging mathematical problems can be solved efficiently and accurately, usually with computers.

Where and how does the topic of your research have an impact?  

Well-functioning algorithms are needed for many things. For instance, they help in establishing a reliable assessment of the risk of individuals developing diabetes. To make such assessments, you need to have information on the genes and lifestyles of these individuals, as well as certain values measured from the body. In addition to assessing the risk of developing the disease, people can be offered guidance on reducing this risk.

Not all factors impacting the development of diseases are observed or even known, making it necessary to run through a number of hypotheses and assigning them varying probabilities when carrying out reliable assessments. This is often challenging from a computing perspective.

What is particularly inspiring in your field right now?  

At the moment, I am especially inspired by the new needs brought about by AI research. Causalities should be better identified and reliable automated reasoning developed.

 

Mikko Koivisto is a professor of computer science at the Faculty of Science.