Machine Learning for Personalized Medicine

Impact of Machine Learning on Personalised Health: past, present and future

by Marta Milo

Last decade has seen a massive increase of data production in science. Particularly in the biomedical field, data has grown exponentially thanks to the development of technologies like sequencing and high-throughput proteomics. The information that this data contains is only partially uncovered to this date, but the impact that it has on human progression and well being is already very clear.

Despite the ability to process large amount of data and to quantify fine details of biological processes, the costs and the time to perform such experiments remain in some cases still very prohibitive. For this reasons the use of a variety of statistical, probabilistic and optimization techniques methods, like machine learning techniques, that allows to “learn” from the available data, to detect hidden patterns from large, noisy and complex datasets, is particularly suitable for application in medicine.

In this talk I will present examples of using machine learning techniques for a variety datasets from medical and biological problems and what are the advantages and disadvantages of this approach. I will also give examples when these techniques enabled to discover informative knowledge from a large complex system in the presence of small number of samples. Finally I will discuss how we use Machine Learning today for analysis of sequencing data and how we can use it for future more complex datasets generated integrating data from different sources.

>back to "Talks and Speakers"