I'm a Ph.D. student at IDSIA, working in the IPG (Imprecise Probability Group).

My research topic is Bayesian network structure learning. You can take a look to my work below or in my scholar.


Improved Local Search in Bayesian Networks Structure Learning
M. Scanagatta, G. Corani, M. Zaffalon
AMBN 2017 (Advanced Methodologies for Bayesian Networks), 45-56

Learning Treewidth-Bounded Bayesian Networks with Thousands of Variables
M. Scanagatta, G. Corani, C.P. de Campos, M. Zaffalon
NIPS 2016 (Advances in Neural Information Processing Systems 29), 1462-1470

Learning extended tree augmented naive structures
C.P. de Campos, G. Corani, M. Scanagatta, M. Cuccu, M. Zaffalon
IJAR 2016 (International Journal of Approximate Reasoning 68), 153-163

Air pollution prediction via multi-label classification
G. Corani, M. Scanagatta
EMS 2016 (Environmental Modelling & Software 80), 259-264

Early classification of time series by hidden Markov models with set-valued parameters
A. Antonucci, M. Scanagatta, D.D. Mauá, C.P. de Campos
Proceedings of the NIPS Time Series Workshop 2015

Learning Bayesian networks with thousands of variables
M. Scanagatta, C.P. de Campos, G. Corani, M. Zaffalon
NIPS 2015 (Advances in Neural Information Processing Systems 28), 1864-1872

Min-BDeu and max-BDeu scores for learning Bayesian networks
M. Scanagatta, C.P. De Campos, M. Zaffalon
PGM 2014 (European Workshop on Probabilistic Graphical Models 7), 426-441


I developed a web-service for learning the structure of a Bayesian network from data: BLIP.

Thanks to the state-of-the-art methods developed during my studies it can process datasets with thousands of variables in few hours, either in the general case or with a bound on the treewidth!

Contact me: mauro [at] idsia [dot] ch