Current Main Research focus
One of the current main focuses of the group is on the development of new tools to uncover and model high order patterns of data.
Highlight: Check our paper on the use of Minimally Complex spin Models (MCM) to identify and model groups of highly correlated variables in binary data: here.
The procedure performs exact Bayesian model selection and takes into account all possible high order patterns of data (3-body correlations, 4-body correlations, etc.) in the detection of these « communities » of variables. The use of Minimally Complex Models provides robust predictions on dependencies between variables.
-
Codes: You can find our codes that performs this « community detection » for binary data on Github:
— MinCompSpin performs an exhaustive search and works best for systems with a small number of variables (<= 15).
— MinCompSpin_Greedy performs a greedy search and works best for systems with a large number of variables (up to 128 variables).
Current group members
-
PhD students: Ebo Peerbooms, Merijn Moody (co-supervised with Jo Ellis-Monaghan and Patrick Forré)
Research Assistant: Aaron De Clercq
Master students:
-
— Physics: Sietse Verbeek, Matteo Sokratis Lioumis
— Mathematics: Khaled Tamimy
— Computational Science: Jonas Argelo
-
Master students graduated:
-
— Theoretical Physics: Jair Lenssen, Karel Geraedts
— Computational Science: Sam Kamphof, Marije Dekker, Karim Semin, Maria Iosif, Aaron De Clercq, Paul Hosek, Lotte Wolfenter, Mylène van der Maas, Karel Geraedts
— Theoretical physics and Neuroscience: Martijn Klop (Utrecht University)
Open positions
-
Don’t hesitate to contact me by email to ask about possible open positions or Master projects.