Structural Inference Day

University of Hamburg
April 18, 2011


In recent years new methods to gain efficiency by using structures in complex infinite dimensional models have attracted much attention. To mention just two examples, the recovery of sparse signals or matrices and the adaptive estimation of functions of unknown regularity are particular active fields of research.

On April 18 a workshop devoted to such problems of structural inference will take place at the Department of Mathematics of the University of Hamburg. The two main speakers are Vladimir Koltchinskii (Georgia Institute of Technology, Atlanta), a leading expert in sparse recovery problems and adaptation, and Richard Nickl (University of Cambridge), who is renowned for his contribution of empirical process theory and its application to adaptive estimation. Below you find a preliminary program, which will be regularly updated.

The participation is for free and everyone is heartily invited. However, to facilitate the organization of this one-day workshop, please send an e-mail with the subject Structural Inference Day to holger.drees@uni-hamburg.de if you plan to participate.


Program

Time Speaker and title Room
10:30 Reception
11:00 Vladimir Koltchinskii (Georgia Institute of Technology, Atlanta): Error bounds in low rank matrix recoveryH4
12:00 Coffee break
12:20 Ulrike von Luxburg (University of Hamburg): Convergence of Hitting Times and Commute Distances on Random Geometric Graphs H2
12:40 Angelika Rohde (University of Hamburg): Estimation of High-Dimensional Low-Rank MatricesH2
13:00 Lunch break
14:40 Richard Nickl (University of Cambridge): Wavelet density estimators on compact manifolds. From concentration to adaptationH4
15:40 Natalie Neumeyer (University of Hamburg): The residual-based empirical process in nonparametric multiple regressionH4
16:00 Coffee break
16:20 Thorsten Dickhaus (Humboldt University Berlin): Multiple testing under structural assumptionsH1
16:40 Vladimir Spokoiny (Humboldt University and WIAS, Berlin): Structure adaptive estimation by alternating H1
17:00 Alexander Meister (University of Rostock): Asymptotic equivalence for nonparametric regression with non-regular errorsH1
17:30 Closing


Here you will find a description (in German) how to come to Department of Mathematics by public transport.

On an interactive map of the campus of the University of Hamburg the so-called Geomatikum (Bundesstr. 55, 20146 Hamburg) where the workshop will take place is indicated.

If you have further questions concerning the Structural Inference Day please do not hesitate to contact Holger Drees (holger.drees@uni-hamburg.de).