Workshop 2011: High-dimensional problems in statistics

Poster of the Workshop

This workshop is organized by the members of the Seminar for Statistics (SfS) S. van de Geer, P. Bühlmann, M. Maathuis and  H.R. Künsch.

Date: September 19-23, 2011
Place: Institute for Mathematical Research, ETH Zürich

For any further information about this workshop please contact: http://www.fim.math.ethz.ch/conferences/2011/High_dimensional

Modern statistical theory concerns the estimation of objects in complex parameter spaces, for example a space of regression functions with a huge number of variables, or a collection of convex sets in image analysis, etc. A key point is the way one describes smoothness. For example, smoothness may be sparsity, e.g. in the number of coefficients in a wavelet expansion, or the dimension of a manifold. An important topic in this workshop is the adaptation to unknown smoothness, using penalty based methods which are computationally feasible for high-dimensional problems.
There will be many connections with analysis and approximation theory. There are also quite a few further apparent relations with other branches of mathematics. For example, concentration inequalities from probability theory are nowadays a main statistical tool. As another example, statistics uses and extends various  techniques from optimization theory (e.g., convex optimization, exponential weighting, interior point methods). Moreover, from the algorithmic point of view, statistical problems have clear relations with e.g. compressing and learning algorithms in computer science.

The workshop has as sub-theme "Graphical modeling and causal inference", with important connections to the theory of sparse (random) graphs, discrete optimization including randomized algorithms, and sparse approximation.

These Invited Speakers have already accepted:

Bartlett, Peter, Department of Statistics, University of California, Berkeley, USA

Bickel, Peter, Department of Statistics, University of California, Berkeley, USA

Bunea, Florentina, Department of Statistics, Florida State University, Tallahassee, USA

Candes, Emmanuel, Department of Statistics, Stanford University, USA

Cohen, Albert, Laboratoire Jacques-Louis Lions, Université Marie Curie, Paris, France

Koltchinskii, Vladimir, School of Mathematics, Georgia Inst. of Technology, Atlanta,USA

Mallik, Bani K., Department of Statistics, Texas A&M University, USA

Meinshausen, Nicolai, Department of Statistics, University of Oxford, UK

Mizera, Ivan, Dept. of Mathematical and Statistical Sciences, University of Alberta, Canada

Murphy, Susan, University of Michigan, Ann Arbor, USA

Nesterov, Yurii, Département d'ingénierie mathématique, Université catholique de Louvain, Belgium

Ritov, Ya'acov, Department of Statistics, The Hebrew University of Jerusalem, Israel

Robins, James, Department of Biostatistics, Harvard, Boston USA

Rohde, Angelika, Departement Mathematik, Universität Hamburg, Germany

Schneider, Ulrike, Institute for Mathematical Stochastics, Göttingen, Germany

Schölkopf, Bernhard, Max Planck Institute for Biological Cybernetics, Tübingen, Germany

Tropp, Joel, California Institute of Technology, Pasadena, USA

Tsybakov, Alexandre, Laboratoire de Statistique, CREST, Malakoff Cedex, France

Wainwright, Martin, Department of Statistics, University of California, Berkely, USA

Wegkamp, Marten, Department of Statistics, Florida State University, Tallahassee, USA

Zhang, Cun-Hui, Department of Statistics, Rutgers University, New Jersey, USA

------------------------------------------------------------

Thematic semester:
High Dimensional Approximation, Learning Theory and Stochastic Partial Differential Equations (Fall 2011)

The thematic semester aims at gathering leading mathematicians in order to fertilize and stimulate new mathematical research.

The format of the thematic semester is centered around two embedded workshops:

Stochastic Partial Differential Equations
Analysis, Numerics, Geometry and Modeling (12-16 September 2011)

High-dimensional problems in statistics (19-23 September 2011)

and around long-term visitors who will work with D-MATH faculty members, Post doctoral and doctoral students during their stay.

Important themes in quantitative modeling in engineering and in the sciences during the decade ahead will be mathematical approaches to quantification of uncertainty, knowledge extraction and mathematical modeling from massive datastreams. In view of an increasing availability of large volumes of data, e.g., in financial industry, in biological systems engineering or in internet traffic, this requires application and development of new mathematical and computational tools. Additional difficulty and challenge arises from the fact that data might be delivered at low quality or data might be related to strongly time-dependent systems.
Methods from stochastics, statistics and numerics will be considered simultaneously, to come up with innovative solutions.