Current students

The focus of the research is on Machine Learning for Climate Science. Specifically, the main objective is to design new methods to be applied on large spatio-temporal dataset of climate variables. The research is based on two particular fields of interest.
The first one is related to Detection, which is the identification of the variables (e.g., temperature) that are the main drivers of an Extreme Event, such as Droughts or Tropical Cyclones. From a Machine Learning perspective this can be articulated on a Feature Selection step (possibly after a suitable dimensionality reduction procedure) that identifies the variables that share significant information with the Extreme Event and a subsequent Supervised Learning step aimed to construct a data-driven indicator of the Extreme Event.
The second one is related to Causation, whose purpose is to identify the causal relationship between variables. This is a step forward with respect to Detection, since the interest is not only in selecting the variables able to identify an Extreme Event but also to find which are the variables that can be the its cause.
The research is aimed to give a contribution to the EU-funded CLINT (Climate Intelligence) project with the design and application of new algorithms for Detection and Causation.