Graphical Models in Science
Probabilistic graphical models are powerful tools for the description of data generating processes and dependence structures of large numbers of variables, which can be seen as nodes of a vast network. Graphical models have found application in disparate fields, including neuroscience, genomics, astronomy, physics, and forensic science. In neuroscience, graphical models can be used to infer neuronal functional connectivity, which is the statistical dependence of neurons’ activities. In genomics, graphical models can be used to identify associations between tens of thousands of genes. In astronomy, graphical models can help us make sense of thousands of cosmic measurements that are collected every day using cutting-edge space technologies. In forensic science, graphical models allow for the concurrent examination of evidence of various nature in intricate crime investigations.
In this project, the student will master the use of the most advanced statistical methods for the estimation of graphical models in high-dimensional settings and will focus on one application in neuroscience, genomics, astronomy, physics, or forensic science, and/or possibly other fields of interest.