Probability Techniques in Analysis and Approximation Theory

Low-rank approximation analysis

26 Nov 2024, 10:00

45m

Saint-Petersburg University

Speaker

Eugene Tyrtyshnikov (Marchuk Institute of Numerical Mathematics of RAS, Moscow)

Description

Tensor decompositions become a very popular tool for modelling data in many application problems. However, a better understanding of why they are so efficient is still a hot issue with a machinery based on some relevant probability models for data. We discuss some open questions and new developments
of cross-approximation approach to optimization problems with the tensor-train model.

References:

1. $E. Tyrtyshnikov$,
   Tensor decompositions and rank increment conjecture, Russian Journal of Numerical Analysis and Mathematical Modelling, 25 (4), 239--246 (2020).

2. $D. Zheltkov, E. Tyrtyshnikov$,
   Global optimization based on TT-decomposition, Russian Journal of Numerical Analysis and Mathematical Modelling, 25 (4), 247--261 (2020).