0509948 20240111141025.520191024235959.9 000492301000002 85077750421 10.1109/LSP.2019.2943060 5 s. P cav_un_epca*02532121070-99082611 (2019)1653-1657Institute of Electrical and Electronics Engineers PARAFAC convolutive neural networks tensor cav_un_auth*0101212 Tichavský Petr UTIA-B Stochastická informatika Department of Stochastic Informatics SI SI Department of Stochastic Informatics K Ústav teorie informace a automatizace AV ČR, v. v. i. cav_un_auth*0382249 Phan A. H. RU cav_un_auth*0382250 Cichocki A. RU http://library.utia.cas.cz/separaty/2019/SI/tichavsky-0509948.pdf 732 kB https://ieeexplore.ieee.org/document/8846103 cav_un_auth*0345929 GA17-00902S GA ČR Canonical polyadic (CP) tensor decomposition is an important task in many applications. Many times, the true tensor rank is not known, or noise is present, and in such situations, different existing CP decomposition algorithms provide very different results. In this paper, we introduce a notion of sensitivity of CP decomposition and suggest to use it as a side criterion (besides the fitting error) to evaluate different CP decomposition results. Next, we propose a novel variant of a Krylov-Levenberg-Marquardt CP decomposition algorithm which may serve for CP decomposition with a constraint on the sensitivity. In simulations, we decompose order-4 tensors that come from convolutional neural networks. We show that it is useful to combine the CP decomposition algorithms with an error-preserving correction. WOS BB 20000 20200 20201 2020 3 4 A hod sml 4ah 4as 20231122144345.3 RVO:67985556 http://hdl.handle.net/11104/0301141 1 Department of Stochastic Informatics UTIA-B 20201 ENGINEERING, ELECTRICAL & ELECTRONIC S Copyright receipt 20190919 3+4 Proceedings Paper Engineering Biomedical|Radiology Nuclear Medicine Medical Imaging C ENGINEERING.ELECTRICAL&ELECTRONIC 3.414 0.482 5.4 0.02646 1.106 11929 167 1.145 25.9 4.55 5579 Q1 2.866 371 Q2 81.438 67.9 Q1 Q4 1.08 Q1 67.857 2019 Soubory v repozitáři: tichavsky-0509948.pdf, tichavsky-0509948-CopyrightReceipt.pdf 85077750421 SCOPUS PUBMED 000492301000002 WOS 732 kB cav_un_epca*0253212 IEEE Signal Processing Letters 1070-9908 1558-2361 Roč. 26 č. 11 2019 1653 1657 Institute of Electrical and Electronics Engineers