Hey, nice to meet you! I am Francesco Montagna:
A (very happy to be!) researcher in the fields of causality and AI for science, working at the Institute of Science and Technology Austria and the Chan Zuckerberg Initiative, under the supervision of Francesco Locatello and Theofanis Karletsos.
To overclaim a lot, I am also a cook and a writer (read: I like to do both, but wouldn't be ready to defend that I am good at either). Soon to come, a website on these topics!
I earned my PhD at the University of Genoa in the ELLIS PhD program supervised by Lorenzo Rosasco and Francesco Locatello. During my PhD, I was an intern at Amazon AWS in Tubingen, Germany, and a visiting researcher at the CLeaR group at Carnegie Mellon University, hosted by Kun Zhang. I work on theoretical and algorithmic causal discovery and causal representation learning, and their application to understanding and simulating cell behaviours.
Selected work
On the identifiability of causal graphs with multiple environments
TLDR: Given a structural causal model with arbitrary bijective mechanisms, three environments suffice to identify graphs of arbitrary size.Causal discovery with score matching:
TLDR: I developed a branch of causal discovery connecting the causal graph inference with score matching estimation techniques
TLDR: I show the limitations of supervised learning-based causal discovery with identifiability theory
Software
causally Python library for data sampling with structural causal model under realistic assumptions
I implemented several causal discovery algorithms for the PyWhy project dodiscover
Get in touch at <name> dot <surname>997 at gmail dot com
Get in touch for supervision on causal discovery if you are a PhD or a strong master student looking for a research internship or master thesis project.