Work

G-Maximization: an Unsupervised Learning Procedure \nfor Discovering Regularities

Year: 1986
Type: article
Abstract: Hill climbing is used to maximize an information theoretic measure of the difference \nbetween the actual behavior of a unit and the behavior that would be predicted by a \nstatistician who knew the first order statistics of the inputs but believed them to be \nindependent. This causes the unit to detect higher order correlations among its inputs. \nInitial simulations are presented, and seem encouraging. We describe an extension of the \nbasic idea which makes it resemble competitive learning and which causes members of a \npopulation of these units to differentiate, each extracting different structure from the input.
Language: English
Open Access status: green
OpenAlex ID: W2591802459
FWCI: 1.345
Cites:
Cited by: 26
Related to: 10