Title :
Logical networks capable of computing weighted regular languages
Author :
Ludermir, Teresa B.
Author_Institution :
Dept. of Comput., King´´s Coll. London, UK
Abstract :
The author introduces a logical node. The logical node is a connectionist model implementable as a random access memory. The node proposed is an extension of the probabilistic logic node where the node outputs I/O depending on the values stored in the memory position addressed by the input plus a probability. The advantage of this node is that the networks constructed of such nodes are capable of recognizing all weighted regular languages, that is, it is possible to recognize more than finite state languages with such networks
Keywords :
formal languages; neural nets; probabilistic logic; connectionist model; finite state languages; logical neural networks; probabilistic logic node; random access memory; weighted regular languages; Automata; Computer networks; Educational institutions; Hardware; Neural networks; Neurons; Pattern recognition; Probabilistic logic; Random access memory; Read-write memory;
Conference_Titel :
Neural Networks, 1991. 1991 IEEE International Joint Conference on
Print_ISBN :
0-7803-0227-3
DOI :
10.1109/IJCNN.1991.170662
