Interactive machine learning from multiple biodata sources

Project Details

Description

Precision medicine, which is tailored to individual patients, brings a difficult challenge to data analysis methods: The goal is to learn to predict effectiveness of treatments using data with extremely small, even equal to one patient, sample size but an extremely large number of variables. Additional data exist but their relevance is unknown, access may be restricted due to privacy, and the complex models required for good predictivity are computationally heavy. Prof. Kaski develops computational, so-called machine learning methods for analysing multiple data sources, taking into account tacit knowledge of domain experts with interactive modelling. Along with the increasing data-intensivity, similar problems have become very general in other fields of science and services, starting from information retrieval.
Short titleKaski Samuel AP-kulut II
StatusFinished
Effective start/end date01/01/201931/08/2021

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.