Closed-Form Expressions of Some Stochastic Adapting Equations for Nonlinear Adaptive Activation Function Neurons
Author:
- 1 December 2003
- journal article
- Published by MIT Press in Neural Computation
- Vol. 15 (12) , 2909-2929
Abstract
In recent work, we introduced nonlinear adaptive activation function (FAN) artificial neuron models, which learn their activation functions in an unsupervised way by information-theoretic adapting rules. We also applied networks of these neurons to some blind signal processing problems, such as independent component analysis and blind deconvolution. The aim of this letter is to study some fundamental aspects of FAN units' learning by investigating the properties of the associated learning differential equation systems.Keywords
This publication has 19 references indexed in Scilit:
- PROBABILITY DENSITY FUNCTION LEARNING BY UNSUPERVISED NEURONSInternational Journal of Neural Systems, 2001
- A contribution to (neuromorphic) blind deconvolution by flexible approximated Bayesian estimationSignal Processing, 2001
- Redundancy reduction revisitedNetwork: Computation in Neural Systems, 2001
- Notes on uniform approximation of time-varying systems on finite time intervalsIEEE Transactions on Circuits and Systems I: Regular Papers, 1998
- An asymptotic property of model selection criteriaIEEE Transactions on Information Theory, 1998
- An Information-Maximization Approach to Blind Separation and Blind DeconvolutionNeural Computation, 1995
- Information Processing in Dendritic TreesNeural Computation, 1994
- Local Synaptic Learning Rules Suffice to Maximize Mutual Information in a Linear NetworkNeural Computation, 1992
- System modeling using an adaptive delay filterIEEE Transactions on Circuits and Systems, 1987
- Oscillation and Chaos in Physiological Control SystemsScience, 1977