Pregled bibliografske jedinice broj: 366061
Parallel Execution of Square Approximation Learning Algorithm for MLP Neural Networks
Parallel Execution of Square Approximation Learning Algorithm for MLP Neural Networks // 16th Mediterranean Conference on Control and Automation
Ajaccio: Institute of Electrical and Electronics Engineers (IEEE), 2008. str. 1816-1821 (predavanje, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)
CROSBI ID: 366061 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Parallel Execution of Square Approximation Learning Algorithm for MLP Neural Networks
Autori
Antunović, Mladen ; Filko, Damir ; Hocenski, Željko
Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni
Izvornik
16th Mediterranean Conference on Control and Automation
/ - Ajaccio : Institute of Electrical and Electronics Engineers (IEEE), 2008, 1816-1821
ISBN
978-1-4244-2505-1
Skup
16th Mediterranean Conference on Control and Automation
Mjesto i datum
Ajaccio, Francuska, 25.06.2008. - 27.06.2008
Vrsta sudjelovanja
Predavanje
Vrsta recenzije
Međunarodna recenzija
Ključne riječi
neural network; learning algorithm; gradient method; parallel execution; time efficiency; Hessian
Sažetak
This work gives improvement of gradient learning algorithms for adjusting neural network weights. Suggested improvement results in alternative method that converge in less iteration and is inherently parallel, convenient for implementation on computer grid. Experimental results show time savings in multiple thread execution for a wide range of MLP neural network parameters, such as size of input/output data matrix, number of neurons and layers.
Izvorni jezik
Engleski
Znanstvena područja
Elektrotehnika, Računarstvo
POVEZANOST RADA
Projekti:
165-0361621-2000 - Distribuirano računalno upravljanje u transportu i industrijskim pogonima (Hocenski, Željko, MZO ) ( CroRIS)
Ustanove:
Fakultet elektrotehnike, računarstva i informacijskih tehnologija Osijek
