Catalogo Articoli (Spogli Riviste)

OPAC HELP

Titolo:
Irreversible thermodynamics of neural networks: Calortropy production in logic operations
Autore:
Barragan, D; Eu, BC;
Indirizzi:
McGill Univ, Dept Chem, Montreal, PQ H3A 2K6, Canada McGill Univ MontrealPQ Canada H3A 2K6 Chem, Montreal, PQ H3A 2K6, Canada McGill Univ, Ctr Phys Mat, Montreal, PQ H3A 2K6, Canada McGill Univ Montreal PQ Canada H3A 2K6 Mat, Montreal, PQ H3A 2K6, Canada Asia Pacific Ctr Theoret Phys, Seoul, South Korea Asia Pacific Ctr TheoretPhys Seoul South Korea hys, Seoul, South Korea
Titolo Testata:
JOURNAL OF PHYSICAL CHEMISTRY B
fascicolo: 29, volume: 105, anno: 2001,
pagine: 7104 - 7114
SICI:
1520-6106(20010726)105:29<7104:ITONNC>2.0.ZU;2-7
Fonte:
ISI
Lingua:
ENG
Soggetto:
CHEMICAL IMPLEMENTATION; SYSTEMS; OSCILLATOR; MACHINES; MODEL; GATES;
Tipo documento:
Article
Natura:
Periodico
Settore Disciplinare:
Physical, Chemical & Earth Sciences
Citazioni:
29
Recensione:
Indirizzi per estratti:
Indirizzo: Eu, BC McGill Univ, Dept Chem, 801 Sherbrooke St W, Montreal, PQ H3A 2K6, Canada McGill Univ 801 Sherbrooke St W Montreal PQ Canada H3A 2K6 , Canada
Citazione:
D. Barragan e B.C. Eu, "Irreversible thermodynamics of neural networks: Calortropy production in logic operations", J PHYS CH B, 105(29), 2001, pp. 7104-7114

Abstract

Irreversible thermodynamics of chemical neural networks is formulated, andenergy and matter dissipation accompanying logic operations is investigated in this paper. This formalism therefore puts the dynamics of neural networks within the framework of the laws of thermodynamics and thereby providesa mathematical basis to study neural dynamics as irreversible processes subjected to the thermodynamic principles. By using the minimal bromate oscillator, we apply the formalism of irreversible thermodynamics to study numerically the modes of energy dissipation related to logic operations. The calortropy of the system serves as an integral surface of the evolution equations for the concentrations of chemical species in the neurons of the network. It is shown that logic operations in the neural networks can be regardedas evolution of the states of neurons in the network toward the steady-state values of the calortropy production characteristic of the logic operations.

ASDD Area Sistemi Dipartimentali e Documentali, Università di Bologna, Catalogo delle riviste ed altri periodici
Documento generato il 08/04/20 alle ore 09:12:21