A. Hyttinen, F. Eberhardt, and P. O. Hoyer, Learning Linear Cyclic Causal Models with Latent Variables,
Accepted to a journal, currently in revision, 2011/2012.
Full theory of learning linear cyclic models with latent variables using only 2nd order statistics from several experimental datasets.
(Article) (Presentation) (Code)

A. Hyttinen, F. Eberhardt, and P. O. Hoyer, Noisy-OR Models with Latent Confounding,
Conference on Uncertainty in Artificial Intelligence, Barcelona, Spain, 2011.
Generally causal models with latent variables are not identifiable from passive observational data or from experimental data in which only a few variables are subject to interventions at a time. The article shows that if the local CPDs of the model are restricted to follow the noisy-OR parameterization, we can identify the causal model for example from experiments intervening on a single variable at a time and passive observational data. Learning methods include a constraint based method and maximimizing the likelihood with the EM-algorithm.
(Article) (Poster) (Code)

A. Hyttinen, F. Eberhardt, and P. O. Hoyer, Causal discovery for linear cyclic models with latent variables,
Proceedings of the Fifth European Workshop on Probabilistic Graphical Models, Helsinki, Finland, 2010.
A method for learning linear cyclic networks with latent variables is revised to get a clearer picture of the underdetermination when full identifiability is not yet reached, identifiability conditions are updated and completeness with regard to search space and assumption is proved. Additionally the common assumption of faithfulness is incorporated into the procedure.
(Article) (Presentation) (Poster) (Code)

P. O. Hoyer and A. Hyttinen, Bayesian Discovery of Linear Acyclic Causal Models,
Proceedings of the 25th Conference on Uncertainty in Artificial Intelligence, Montreal, Canada, 2009.
Introduces a Bayesian score-based LiNGAM method, which also properly handles data from mixed Gaussian/non-Gaussian linear acyclic models.
(Article) (Poster) (Code)

I have been a Ph.D. student since June 2008. The tentative title of my thesis is "Causal Discovery: Theory and Practice". My research interests include causality, machine learning, and graphical models. Previously I have gotten a M.Sc. in Information Technology from Tampere University of Technology, studied theoretical physics at the University of Helsinki, and worked as a research scientist at the Finnish Geodetic Institute.