Sciweavers

IJCAI
1997

On the Efficient Classification of Data Structures by Neural Networks

14 years 27 days ago
On the Efficient Classification of Data Structures by Neural Networks
Marco Gori Dipartimento di Ingegneria deU'Informazione Universita di Siena Via Roma 56 53100 Siena, Italy Alessandro Sperduti Dipartimento di Informatica Universita di Pisa Corso Italia, 40 56125 Pisa, Italy In the last few years it has been shown that recurrent neural networks are adequate for processing general data structures like trees and graphs, which opens the doors to a number of new interesting applications previously unexplored. In this paper, we analyze the efficiency of learning the membership of DO AGs (Directed Ordered Acyclic Graphs) in terms of local minima of the error surface by relying on the principle that their absence is a guarantee of efficient learning. We give sufficient conditions under which the error surface is local minima free. Specifically, we define a topological index associated with a collection of DOAGs that makes it possible to design the architecture so as to avoid local minima.
Paolo Frasconi, Marco Gori, Alessandro Sperduti
Added 01 Nov 2010
Updated 01 Nov 2010
Type Conference
Year 1997
Where IJCAI
Authors Paolo Frasconi, Marco Gori, Alessandro Sperduti
Comments (0)