T:A:L:K:S

close this window
title:
Hierarchical and Black-Box Approximation of Tensors
name:
Grasedyck
first name:
Lars
location/conference:
SPP-JT09
WWW-link:
http://personal-homepages.mis.mpg.de/lgr
PREPRINT-link:
http://www.dfg-spp1324.de/download/preprints/preprint020.pdf
PRESENTATION-link:
http://dfg-spp1324.de/download/jt09/talks/grasedyck.pdf
abstract:
The talk covers the advances in black-box tensor approximation as well as the new reliable tensor representation and arithmetics in the hierarchical rank model. The results for the black-box approximation of tensors are in the classical tensor rank model where a priori estimates of truncation errors are missing. In the new hierarchical model we present a priori
error bounds in terms of the best approximation. The hierarchical rank k model contains all tensors of rank k (even border rank k) and is thus a much richer class.