Controlling the Magnification Factor of Self-Organizing Feature Maps

The magnification exponents occurring in adaptive map formation algorithms like Kohonen's self-organizing feature map deviate for the information theoretically optimal value = 1 as well as from the values that optimize, e.g., the mean square distortion error ( = 1/3 for one-dimensional maps). At the same time, models for categorical perception such as the "perceptual magnet" effect, which are based on topographic maps, require negative magnification exponents < 0. We present an extension of the self-organizing feature map algorithm, which utilizes adaptive local learning step sizes to actually control the magnification properties of the map. By change of a single parameter, maps with optimal information transfer, with various minimal reconstruction errors, or with an inverted magnification can be generated. Analytic results on this new algorithm are complemented by numerical simulations.

[1]  B. Boycott,et al.  Cortical magnification factor and the ganglion cell density of the primate retina , 1989, Nature.

[2]  C. Gielen,et al.  Neural computation and self-organizing maps, an introduction , 1993 .

[3]  Prahlad Gupta,et al.  Investigating Phonological Representations: A Modeling Agenda , 1994 .

[4]  Teuvo Kohonen,et al.  Self-Organizing Maps , 2010 .

[5]  Christopher L. Scofield,et al.  Neural networks and speech processing , 1991, The Kluwer international series in engineering and computer science.

[6]  Klaus Schulten,et al.  Kohonens Self-Organizing Maps for Modeling the Formation of the Auditory Cortex of a Bat , 1988 .

[7]  Paul L. Zador,et al.  Asymptotic quantization error of continuous signals and the quantization dimension , 1982, IEEE Trans. Inf. Theory.

[8]  Thomas Villmann,et al.  Growing a hypercubical output space in a self-organizing feature map , 1997, IEEE Trans. Neural Networks.

[9]  Lipo Wang,et al.  On competitive learning , 1997, IEEE Trans. Neural Networks.

[10]  F. O. Schmitt,et al.  The Organization of the Cerebral Cortex. , 1982 .

[11]  Y.K. Kim,et al.  Adaptive learning method in self-organizing map for edge preserving vector quantization , 1995, IEEE Trans. Neural Networks.

[12]  Teuvo Kohonen,et al.  The self-organizing map , 1990, Neurocomputing.

[13]  K. Stevens,et al.  Linguistic experience alters phonetic perception in infants by 6 months of age. , 1992, Science.

[14]  Nobuo Suga,et al.  Cortical computational maps for auditory imaging , 1990, Neural Networks.

[15]  D. Whitteridge,et al.  The representation of the visual field on the cerebral cortex in monkeys , 1961, The Journal of physiology.

[16]  H. Ritter,et al.  A principle for the formation of the spatial structure of cortical feature maps. , 1990, Proceedings of the National Academy of Sciences of the United States of America.

[17]  Helge Ritter.  Asymptotic level density for a class of vector quantization processes , 1991, IEEE Trans. Neural Networks.

[18]  P. Kuhl.  Human adults and human infants show a “perceptual magnet effect” for the prototypes of speech categories, monkeys do not , 1991, Perception & psychophysics.

[19]  Dominik R. Dersch,et al.  Asymptotic level density in topological feature maps , 1995, IEEE Trans. Neural Networks.

[20]  J. Kaas.  The functional organization of somatosensory cortex in primates. , 1993, Annals of anatomy = Anatomischer Anzeiger : official organ of the Anatomische Gesellschaft.

[21]  R. Palmer,et al.  Introduction to the theory of neural computation , 1994, The advanced book program.

[22]  Juan K. Lin,et al.  Faithful Representation of Separable Distributions , 1997, Neural Computation.

[23]  Carlo H. Séquin,et al.  Optimal adaptive k-means algorithm with dynamic adjustment of learning rate , 1995, IEEE Trans. Neural Networks.

[24]  G A van Velzen,et al.  Instabilities in Kohonen's self-organizing feature map , 1994 .

[25]  Richard Durbin,et al.  An analogue approach to the travelling salesman problem using an elastic net method , 1987, Nature.

[26]  Jürgen P. Kropp,et al.  Phenomenological pattern recognition in the dynamical structures of tidal sediments from the German Wadden Sea , 1997 .

[27]  Duane DeSieno,et al.  Adding a conscience to competitive learning , 1988, IEEE 1988 International Conference on Neural Networks.

[28]  Ralf Der,et al.  The "Perceptual Magnet" Effect: A Model Based on Self-Organizing Feature Maps , 1995 .

[29]  Stanley C. Ahalt,et al.  Competitive learning algorithms for vector quantization , 1990, Neural Networks.