Proceedings 3rd Twente Workshop on Language Technology, 27-38, 1992 GENERALIZATION PERFORMANCE OF BACKPROPAGATION LEARNING ON A SYLLABIFICATION TASK

We investigated the generalization capabilities of backpropagation learning in feed forward and re current feed forward connectionist networks on the assignment of syllable boundaries to ortho graphic representations in Dutch hyphenation This is a di cult task because phonological and morphological constraints interact leading to am biguity in the input patterns We compared the results to di erent symbolic pattern matching ap proaches and to an exemplar based generaliza tion scheme related to a k nearest neighbour approach but using a similarity metric weighed by the relative information entropy of positions in the training patterns Our results indicate that the generalization performance of backprop agation learning for this task is not better than that of the best symbolic pattern matching ap proaches and of exemplar based generalization