1 Jan 2016–31 Dec 2020
Computational data analysis
Current data analysis methods are ill prepared for the challenges underlying personalised medicine. Responses to treatments need to be predicted based on very small samples, even from only one, with a very large number of variables. Moreover, the data come from multiple, only partially relevant sources, access may be restricted due to reasons of privacy, and the complex models required for good predictivity would not scale up to whole collections of datasets.
Samuel Kaski develops computational methods for interactive multi-source data analysis and machine learning. They will be applied first to personalised medicine, but they will also be applicable to the ever widely available collections of datasets in science and data-based services.
Kaski’s research effectively combines advanced methods of machine learning and scalable visual analytics, creating a whole new approach to studies on personalised medicine. Kaski is a pioneer in many areas of machine learning technology and applications, and he has contributed to several significant scientific breakthroughs in this multidisciplinary field.