MMS
Meeting in Mathematical Statistics 2016

Advances in nonparametric and high-dimensional Statistics

December 12-16





Home


Welcome to the 2016 edition of the Meeting in Mathematical Statistics. It is our pleasure to announce that this year the workshop will take place in Frejus (south of France) in the villa Clythia. The goal of the workshop is to bring together leading and upcoming researchers working on mathematical statistics.



Tutorials

  • Tsybakov, Alexandre (ENSAE ParisTech - CREST, France)

    Optimality in variable selection

    Abstract:

    Variable selection is a classical and extensively studied problem. It had received a boost in the sparsity era with the appearance of Lasso type methods. Its most recent reincarnations are related to clustering, community detection, croudsourcing. However, in the theory not much is known beyond the consistency of some methods. These lectures will focus on a decision theoretic perspective on variable selection. The main goal will be to construct minimax optimal and adaptive procedures.

  • Juditsky, Anatoly (University Grenoble-Alpes, France)

    Estimation and testing by convex optimization

    Abstract:

    We discuss a "computation-friendly" approach to estimation and hypothesis testing via Convex Programming. Most of the existing results aim to quantify in a closed analytic form estimation rates or rates of separation between sets of distributions allowing for reliable decision in precisely stated observation models. In contrast to this descriptive (and highly instructive) traditional framework, the approach we promote here can be qualified as "operational" - the proposed estimating and testing routines and their risks are yielded by an efficient computation. All we know in advance is that, under specific favorable circumstances, the risk of such decision, whether high or low, is provably near-optimal. As a compensation for the lack of "explanatory power," this approach is applicable to a much wider family of observation schemes and estimation or hypotheses testing problems than those where "closed form descriptive analysis" is possible.

    Joint work with A. Nemirovski (Georgia Institute of Technology)

    Lecture notes for the extended version of this tutorial are available here.



Organizing Committee



Sponsors


Sponsors