Title :
Should activation functions be affinely recursive?
Author_Institution :
Dept. of Math., Georgetown Univ., Washington, DC, USA
Abstract :
Some activation functions can be represented as linear combinations of affine transformations of themselves. This phenomenon, investigated by Kurkova and Kainen (1996), is reviewed here, and some examples are considered
Keywords :
multilayer perceptrons; recursive functions; transfer functions; affine transformations; affinely recursive activation functions; one-hidden-layer neural net; Equations; Mathematics; Neural networks; Neurons; Vectors;
Conference_Titel :
Neural Networks, 2000. IJCNN 2000, Proceedings of the IEEE-INNS-ENNS International Joint Conference on
Conference_Location :
Como
Print_ISBN :
0-7695-0619-4
DOI :
10.1109/IJCNN.2000.857850
