Title :
Learnability in sequential RAM-based neural networks
Author :
De Souto, Marcílio C P ; Adeodato, Paulo J L
Author_Institution :
Dept. of Electr. Eng., Imperial Coll. of Sci., Technol. & Med., London, UK
Abstract :
It is well known that, in a broad sense, recurrent neural networks are equivalent to Turing machines. However, in general, such computational power has not been achieved by the current learning algorithms. In this paper, the learning capability of the existing algorithms for sequential RAM-based neural networks is analysed. These learning algorithms are proved to have limitations which prevent the networks from attaining their computability
Keywords :
computability; finite automata; learning (artificial intelligence); recurrent neural nets; computability; finite automata; learnability; learning algorithms; recurrent neural networks; sequential RAM-based neural networks; Artificial neural networks; Computer architecture; Computer networks; Educational institutions; Intelligent networks; Learning automata; Multilayer perceptrons; Neural networks; Postal services; Recurrent neural networks;
Conference_Titel :
Neural Networks, 1998. Proceedings. Vth Brazilian Symposium on
Conference_Location :
Belo Horizonte
Print_ISBN :
0-8186-8629-4
DOI :
10.1109/SBRN.1998.730988