Title :
Convergence properties of backpropagation algorithm
Author :
Prasanna, G.N.S.
Author_Institution :
MIT, Cambridge, MA, USA
Abstract :
Summary form only given, as follows. The author presents results regarding the convergence properties of the backpropagation algorithm for XOR circuits. Specifically, the behavior has been investigated in the presence of noise. As a performance metric, the author has determined the solution density as several parameters are varied. Some issues relating to asynchronous implementation as well as behavior in the presence of ambiguity are also briefly explored.<>
Keywords :
learning systems; neural nets; XOR circuits; asynchronous implementation; backpropagation algorithm; convergence properties; solution density; Learning systems; Neural networks;
Conference_Titel :
Neural Networks, 1989. IJCNN., International Joint Conference on
Conference_Location :
Washington, DC, USA
DOI :
10.1109/IJCNN.1989.118527
