Approximation, sampling, and compression in high dimensional problems

Created: 2019-06-18 08:28
Institution: Isaac Newton Institute for Mathematical Sciences
Description: In a number of problems, both in theory and applications, one faces a situation when the ambient dimension is extremely high. Such problems often include approximating, sampling, or compressing functions on high-dimensional domains. Classical methods fail to be effective in this case due to the effect known as `curse of dimensionality'; hence new tools and algorithms need to be devised. Compressed sensing, which has gained great popularity in this century, is one example of a circle of ideas which make high-dimensional problems feasible. Methods which allow one to overcome the curse of dimensionality come from a mixture of mathematical fields: approximation, probability, functional and harmonic analysis, linear algebra, combinatorics, geometry, etc. In addition to pure mathematical interest, this field has great importance in numerous applications, in particular in data science and signal processing. Despite decades of research, many important questions in this area are still open. This workshop will bring together researchers in pure and applied mathematics, who attack high-dimensional problems.
 

Media items

This collection contains 14 media items.

  •  

Media items

A sequence of well-conditioned polynomials

   3 views

Ortega-Cerdà, J
Tuesday 18th June 2019 - 13:30 to 14:20

Collection: Approximation, sampling, and compression in high dimensional problems

Institution: Isaac Newton Institute for Mathematical Sciences

Created: Wed 19 Jun 2019


Discrete translates in function spaces

   12 views

Olevskii, A
Tuesday 18th June 2019 - 09:50 to 10:40

Collection: Approximation, sampling, and compression in high dimensional problems

Institution: Isaac Newton Institute for Mathematical Sciences

Created: Wed 19 Jun 2019