Adaptive estimation of functionals under sparsity

Add to your list(s) Download to your calendar using vCal

• Alexandre Tsybakov (CREST: Centre de Recherche en Économie et Statistique)
• Thursday 18 January 2018, 09:45-10:30
• Seminar Room 1, Newton Institute.

If you have a question about this talk, please contact INI IT.

STSW01 - Theoretical and algorithmic underpinnings of Big Data

Adaptive estimation of functionals in sparse mean model and in sparse regression exhibits some interesting effects. This talk focuses on estimation of the L_2-norm of the target vector and of the variance of the noise. In the first problem, the ignorance of the noise level causes changes in the rates. Moreover, the form of the noise distribution also infuences the optimal rate. For example, the rates of estimating the variance differ depending on whether the noise is Gaussian or sub-Gaussian without a precise knowledge of the distribution. Finally, for the sparse mean model, the sub-Gaussian rate cannot be attained adaptively to the noise level on classes of noise distributions with polynomial tails, independently on how fast is the polynomial decay. Joint work with O.Collier and L.Comminges.

This talk is part of the Isaac Newton Institute Seminar Series series.

This talk is included in these lists:

• All CMS events
• Featured lists
• INI info aggregator
• Isaac Newton Institute Seminar Series
• School of Physical Sciences
• Seminar Room 1, Newton Institute
• bld31

Note that ex-directory lists are not shown.