Title :
Hamiltonian neural nets as a universal signal processor
Author :
Sienko, Wieslaw ; Citko, Wieslaw M. ; Wilamowski, Bogdan M.
Author_Institution :
Gdynia Maritime Acad., Tech. Univ. of Koszalin, Gdynia, Poland
Abstract :
This paper presents how to find an architecture for very large scale lossless neural nets, which can be used as Haar-Walsh spectrum analyzers. This analysis relies on the orthogonality of weight matrices W, where W could be Hurwitz-Radon matrices. The unique feature of these nets is the possibility to treat them either as algorithms or as Hamiltonian physical objects (Haar-Walsh Signal Processors).
Keywords :
Haar transforms; Walsh functions; neural net architecture; neural nets; signal processing; spectral analysers; Haar-Walsh Signal Processors; Haar-Walsh spectrum analyzers; Hamiltonian neural nets; Hamiltonian physical objects; Hurwitz-Radon matrices; lossless neural nets; orthogonal filter; orthogonality; universal signal processor; weight matrices; Artificial neural networks; Biological systems; Equations; Large-scale systems; Neural networks; Neurons; Signal processing; Signal processing algorithms; Silicon; Spectral analysis;
Conference_Titel :
IECON 02 [Industrial Electronics Society, IEEE 2002 28th Annual Conference of the]
Print_ISBN :
0-7803-7474-6
DOI :
10.1109/IECON.2002.1182910
