Krylov-Levenberg-Marquardt Algorithm for Structured Tucker Tensor Decompositions

0541614 - ÚTIA 2022 RIV US eng J - Článek v odborném periodiku
Tichavský, Petr - Phan, A. H. - Cichocki, A.
Krylov-Levenberg-Marquardt Algorithm for Structured Tucker Tensor Decompositions.
IEEE Journal on Selected Topics in Signal Processing. Roč. 15, č. 3 (2021), s. 550-559. ISSN 1932-4553. E-ISSN 1941-0484
Grant ostatní: GA ČR(CZ) GA20-17720S
Institucionální podpora: RVO:67985556
Klíčová slova: canonical polyadic tensor decomposition * parallel factor analysis * tensor chain * sensitivity
Obor OECD: Electrical and electronic engineering
Impakt faktor: 7.695, rok: 2021
Způsob publikování: Omezený přístup
http://library.utia.cas.cz/separaty/2021/SI/tichavsky-0541614.pdf https://ieeexplore.ieee.org/document/9354901

Structured Tucker tensor decomposition models complete or incomplete multiway data sets (tensors), where the core tensor and the factor matrices can obey different constraints. The model includes block-term decomposition or canonical polyadic decomposition as special cases. We propose a very flexible optimization method for the structured Tucker decomposition problem, based on the second-order Levenberg-Marquardt optimization, using an approximation of the Hessian matrix by the Krylov subspace method. An algorithm with limited sensitivity of the decomposition is included. The proposed algorithm is shown to perform well in comparison to existing tensor decomposition methods.

Trvalý link: http://hdl.handle.net/11104/0319267