Contemporary frontiers in high-dimensional statistical data analysis

Workshop theme

The motivation for the research programme itself derived from the realization that, with the spectacular evolution of computing facilities and the proliferation of applications in which the number of experimental units is comparatively small but the underlying dimension is massive, it has become necessary to fit complex models for which the effective number of parameters is very large. Areas of application include image analysis, microarray analysis, finance, document classification, astronomy and atmospheric science. Methodological advances have been made, but with them comes the need for further development and appropriate theoretical underpinning. The workshop will review the state-of-the-art of this rapidly developing field. Timed as it is at the beginning of the six-months research programme, the workshop will be broad in scope, with the aim of laying the groundwork for research interactions during the rest of the programme and beyond. Particular topics likely to be covered include the following: strategies for explicit and implicit dimension-reduction; classification methods for complex datasets, including machine-learning approaches; asymptotic theory for increasing dimension; graphical and other visualisation methods for complex datasets; and presentation of topical case studies, probably drawn from the areas of application listed above.

The workshop will therefore cover fundamental areas of modern statistical theory and methodology required for the analysis of important large-scale practical problems. The nature of the topics is such that it should be of interest to those working in machine learning research and others in computer science, as well as to mainstream statisticians.

Applications for this workshop are now closed

Timetable

Monday 7th January 2008
08:30 to 09:55	Registration	No Room Required
09:55 to 10:00	David Wallace - Welcome	Room 1
10:00 to 11:00	Miriam Donoho San José State University. Breakdown point of model selection when the number of variables exceeds the number of observations	Room 1
11:00 to 11:30	Coffee	No Room Required
11:30 to 12:30	Sara Anna van de Geer ETH Zürich The deterministic lasso	Room 1
12:30 to 13:30	Lunch at Wolfson Court	No Room Required
14:00 to 15:00	Edward Wegman George Mason University Methods for visualizing high dimensional data	Room 1
15:00 to 15:30	Tea	No Room Required
15:30 to 16:30	Alastair Young Imperial College London Bootstrap and parametric inference: successes and challenges	Room 1
16:30 to 17:30	Short advertisements by poster presenters "Gong Show"	Room 1
17:30 to 18:30	Welcome Wine Reception and Poster session	No Room Required
18:45 to 19:30	Dinner at Wolfson Court (Residents only)	No Room Required

Tuesday 8th January 2008
09:00 to 10:00	Practical and information-theoretic limitations in high-dimensional inference	Room 1
10:00 to 11:00	Richard Samworth University of Cambridge Some thoughts on nonparametric classification: nearest neighbours, bagging and max likelihood estimation of shape-constrained densities	Room 1
11:00 to 11:30	Coffee	No Room Required
11:30 to 12:30	Ralph Dennis Cook University of Minnesota Model-based sufficient dimension reduction for regression	Room 1
12:30 to 13:30	Lunch at Wolfson Court	No Room Required
14:00 to 15:00	Michael Jordan University of California, Berkeley Kernel-based contrast functions for sufficient dimension reduction	Room 1
15:00 to 15:30	Tea	No Room Required
15:30 to 16:30	Jianqing Fan Princeton University Challenge of dimensionality in model selection and classification	Room 1
16:30 to 17:30	Peter Bickel University of California, Berkeley Regularised estimation of high dimensional covariance matrices	Room 1
18:45 to 19:30	Dinner at Wolfson Court (Residents only)	No Room Required

Wednesday 9th January 2008
09:00 to 10:00	Fionn Murtagh Birkbeck, University of London The ultrametric topology perspective on analysis of massive, very high dimensional data stores	Room 1
10:00 to 11:00	Lutz Duembgen Universität Bern P-values for computer-intensive classifiers	Room 1
11:00 to 11:30	Coffee	No Room Required
11:30 to 12:30	Werner Stuetzle University of Washington Nonparametric cluster analysis: estimating the cluster tree of a density	Room 1
12:30 to 13:30	Lunch at Wolfson Court	No Room Required
14:00 to 15:00	Mike West Duke University Sparsity modelling in large-scale dynamic models for portfolio analysis	Room 1
15:00 to 15:30	Tea	No Room Required
15:30 to 16:30	Computationally tractable statistical estimation when there are more variables than observations	Room 1
16:30 to 17:30	Boaz Nadler University of California, Irvine and Weizmann Institute of Science Learning in high dimensions, noise, sparsity and treelets	Room 1
19:30 to 23:00	Conference Dinner - Emmanuel College (The Old Library)	No Room Required

Thursday 10th January 2008
09:00 to 10:00	Aad Willem van der Vaart Vrije Universiteit Amsterdam Estimating a response parameter in missing data models with high-dimensional covariates	Room 1
10:00 to 11:00	Jon August Wellner University of Washington Persistence: alternative proofs of some results of Greenshtein and Ritov	Room 1
11:00 to 11:30	Coffee	No Room Required
11:30 to 12:30	Ralph Dennis Cook University of Minnesota Looking at models in high-dimensional data spaces	Room 1
12:30 to 13:30	Lunch at Wolfson Court	No Room Required
14:00 to 15:00	Jared Tanner University of Edinburgh The surprising structure of Gaussian point clouds and its implications for signal processing	Room 1
15:00 to 15:30	Tea	No Room Required
15:30 to 16:30	Ann B Lee Carnegie Mellon University Finding low-dimensional structure in high-dimensional data	Room 1
16:30 to 17:30	Partha Niyogi University of Chicago A geometric perspective on learning theory and algorithms	Room 1
18:45 to 19:30	Dinner at Wolfson Court (Residents only)	No Room Required

Friday 11th January 2008
09:00 to 10:00	High-dimensional variable selection and graphs: sparsity, faithfulness and stability	Room 1
10:00 to 11:00	Enno Mammen Universität Mannheim Time series regression with semiparametric factor dynamics	Room 1
11:00 to 11:30	Coffee	No Room Required
11:30 to 12:30	Keming Yu Brunel University Using side information for prediction	Room 1
12:30 to 13:30	Lunch at Wolfson Court	No Room Required
14:00 to 15:00	David Hoyle University of Manchester A physicist's approach to high-dimensional inference	Room 1
15:00 to 15:30	Tea	No Room Required
15:30 to 16:30	Bertrand Salem Clarke University of British Columbia Models, model lists, model spaces and predictive optimality	Room 1
16:30 to 17:30	Discussion of INI programme directions	Room 1
18:45 to 19:30	Dinner at Wolfson Court (Residents only)	No Room Required

Contemporary frontiers in high-dimensional statistical data analysis

Speakers

Peter Bickel University of California, Berkeley
Bertrand Salem Clarke University of British Columbia
Ralph Dennis Cook University of Minnesota
Miriam Donoho San José State University.
Lutz Duembgen Universität Bern
Jianqing Fan Princeton University
David Hoyle University of Manchester
Michael Jordan University of California, Berkeley
Ann B Lee Carnegie Mellon University
Enno Mammen Universität Mannheim
Fionn Murtagh Birkbeck, University of London
Boaz Nadler University of California, Irvine and Weizmann Institute of Science
Partha Niyogi University of Chicago
Richard Samworth University of Cambridge
Werner Stuetzle University of Washington
Jared Tanner University of Edinburgh
Sara Anna van de Geer ETH Zürich
Aad Willem van der Vaart Vrije Universiteit Amsterdam
Edward Wegman George Mason University
Jon August Wellner University of Washington
Mike West Duke University
Alastair Young Imperial College London
Keming Yu Brunel University

Show All

Contemporary frontiers in high-dimensional statistical data analysis

Workshop theme

Applications for this workshop are now closed

Timetable

Speakers

Supported By

Register for updates