Topographic Independent Component Analysis
- 1 July 2001
- journal article
- Published by MIT Press in Neural Computation
- Vol. 13 (7) , 1527-1558
- https://doi.org/10.1162/089976601750264992
Abstract
In ordinary independent component analysis, the components are assumed to be completely independent, and they do not necessarily have any meaningful order relationships. In practice, however, the estimated “independent” components are often not at all independent. We propose that this residual dependence structure could be used to define a topo-graphic order for the components. In particular, a distance between two components could be defined using their higher-order correlations, and this distance could be used to create a topographic representation. Thus, we obtain a linear decomposition into approximately independent components, where the dependence of two components is approximated by the proximity of the components in the topographic representation.Keywords
This publication has 32 references indexed in Scilit:
- GTM: The Generative Topographic MappingNeural Computation, 1998
- A Unifying Objective Function for Topographic MappingsNeural Computation, 1997
- Equivariant adaptive source separationIEEE Transactions on Signal Processing, 1996
- Robust neural networks with on-line learning for blind identification and blind separation of sourcesIEEE Transactions on Circuits and Systems I: Regular Papers, 1996
- An Information-Maximization Approach to Blind Separation and Blind DeconvolutionNeural Computation, 1995
- Adaptive blind separation of independent sources: A deflation approachSignal Processing, 1995
- Putative strategies of scene segmentation in monkey visual cortexNeural Networks, 1994
- Independent component analysis, A new concept?Signal Processing, 1994
- Directionally selective complex cells and the computation of motion energy in cat visual cortexVision Research, 1992
- A dimension reduction framework for understanding cortical mapsNature, 1990