Multilayer perceptrons can compute all relevant functions
The limitations described by Minsky & Papert don't apply to multilayer networks.
Minsky & Papert ignored such networks in part because there were no useful learning procedures for training them. Today, however there are several effective algorithms for training multilayer networks.

David Rumelhart & James McClelland, 1986b.
Immediately related elementsHow this works
-
Artificial Intelligence Â»Artificial Intelligence
▲
Can computers think? [1] Â»Can computers think? [1]
▲
Yes: connectionist networks can think [5a] Â»Yes: connectionist networks can think [5a]
▲
One-layer perceptrons can’t compute certain functions Â»One-layer perceptrons can’t compute certain functions
■
Multilayer perceptrons can compute all relevant functions
◄
James McClelland Â»James McClelland
◄
David Rumelhart Â»David Rumelhart
+Commentaires (0)
+Citations (0)
+About