Structure Estimation, Graphical Modelling and Causal Inference in High Dimensions

• Applicant: Professor Dr. Peter Bühlmann
• Subject Area: Mathematics
• Term: from 2008 to 2015
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 40095828

This project develops new methodology, theory and algorithms for inferring discrete structures (i.e. graphs) and corresponding parameters from noisy data. The graph represents a model structure or whether an association or a causal effect between variables is effective. We focus on high-dimensional problems where the number of variables or nodes in the graph is much larger than sample size (the number of observations) but assuming some underlying sparse structure. New regularization techniques are required for efficiently estimating high-dimensional graphs and corresponding parameters. Many of the statistical problems are directly connected to questions from molecular and systems biology where our collaborators (plant-biotechnology, molecular systems biology, biochemistry, cell biology, pathology) are able to do biological validation of quantitative models and algorithms.

• DFG Programme: Research Units
• Subproject of: FOR 916: Statistical Regularisation and Qualitative Constraints - Inference, Algorithms, Asymptotics and Applications
• International Connection: Switzerland