A new Data Science group and PhD programme at SISSA
The Department of Physics at SISSA opened in 2020 a new research line and a new path of doctoral training in Data Science, funded by a generous grant of the Italian government as part of the Dipartimenti di Eccellenza initiative. Since its foundation in 1978, SISSA's main mission has been to train PhD students to the highest international standards. To maintain its competitive level, the School continuously adapts to the evolution of science, which always offers new challenges, often across the borders of traditional disciplines.
Data Science is nowadays fundamental for the study of many physical systems. Analysis tools developed in this field were fundamental in many recent discoveries within particle physics, astronomy, materials sciences and biophysics. For example, in particle physics and astronomy, scientists grapple with very large data sets, high-dimensional parameter spaces and often low signa-to-noise. Applying machine learning to sift through the data for possible instances of a particle or process allows researchers to pinpoint an interesting event within a stack of trillions. In astronomy, data science and machine learning are often used to identify processes or interesting features (new galaxies, supernovas, possible black holes, dark matter signatures...), and to accelerate manyfolds otherwise untractable computational problems.
To analyze data sets of such complexity it is necessary to develop new algorithmic strategies aimed at extracting from the data the relevant features that can be later used by the scientists to build models and, above all, to make or test predictions. The development and the understanding of these strategies, which are often inspired by statistical physics, form the core of the "Machine Learning for the Natural Sciences" approach that SISSA is developing.
Please visit the new site of the Theoretical and Scientific Data Science group for further information about our research, training, public engagement and consulting activities.