Naslov
New Neural Network Architevture and Algorithm for Memory Sequence Learning and Modelling

Autori
Kišasondi, Tonimir ; Lovrenčić, Alen

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Conference Proceedings of the 16th International Conference on Information and Intelligent Systems / Aurer, Boris ; Bača, Miroslav - Varaždin : Fakultet organizacije i informatike Sveučilišta u Zagrebu, 2005, 415-422

Skup
16th International Conference on Information and Intelligent Systems (16 ; 2005)

Mjesto i datum
Varaždin, Hrvatska, 21.09.2005. - 23.09.2005

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
neuralnetworks; MLP training agorithms; neural network ADT; neural network graph
(neural networks; MLP training agorithms; neural network ADT; neural network graph)

Sažetak
In this paper we present a modified neural network architecture and an algorithm that enables neural networks to learn vectors in accordance to user designed sequences. This enables us to use the modified network algorithm to identify, generate or complete specified patterns that are learned in the training phase. The algorithm is based on the idea that neural networks in the human neurocortex represent a distributed memoryof sequences that are stored in invariant hierarchical form with associative access. The algorithm was tested on our custom built simulator that supports the usage of out ADT neural network with standard and our custom built training algorithms.

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo, Informacijske i komunikacijske znanosti

POVEZANOST RADA