On gradient adaptation with unit-norm constraints
- 1 June 2000
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Signal Processing
- Vol. 48 (6) , 1843-1847
- https://doi.org/10.1109/78.845952
Abstract
In this correspondence, we describe gradient-based adaptive algorithms within parameter spaces that are specified by ||w||=1, where ||/spl middot/|| is any vector norm. We provide several algorithm forms and relate them to true gradient procedures via their geometric structures. We also give algorithms that mitigate an inherent numerical instability for L/sub 2/-norm-constrained optimization tasks. Simulations showing the performance of the techniques for independent component analysis are provided.Keywords
This publication has 13 references indexed in Scilit:
- Adaptive data orthogonalizationPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2005
- KuicNet algorithms for blind deconvolutionPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- A self-stabilized minor subspace ruleIEEE Signal Processing Letters, 1998
- A unified algorithm for principal and minor components extractionNeural Networks, 1998
- Performance analysis of adaptive eigenanalysis algorithmsIEEE Transactions on Signal Processing, 1998
- Independent component analysis by general nonlinear Hebbian-like learning rulesSignal Processing, 1998
- Neural networks for blind decorrelation of signalsIEEE Transactions on Signal Processing, 1997
- Independent component analysis, A new concept?Signal Processing, 1994
- Simplified neuron model as a principal component analyzerJournal of Mathematical Biology, 1982
- Least squares type algorithm for adaptive implementation of Pisarenko's harmonic retrieval methodIEEE Transactions on Acoustics, Speech, and Signal Processing, 1982