AG Tavan - Projects

LMU München
Fakultät für Physik



(c) 2002 BMO

Selforganized pattern classification and its application to molecular structures
(Data based Modelling and Neuroinformatics)

In a long-standing effort we have developed a series of biologically motivated neural network models that self-organize into useful tools for hierarchical pattern classification, function approximation, formation of hypotheses, control and so forth. The models have been formulated in such a way that they find clear interpretations in terms of classical statistical and probabilistic concepts allowing to identify what's new about them.

Currently, a main application is the development of tools for the identification and classification of molecular structures from simulation data.

Previous coworkers:  Heiko Carstens
Jan Busch
Sebastian Albrecht
Jeanne Rubner
Bernhard Rabus
Johann Kühnel
Niels Kunstmann
Martin Kloppenburg
Thomas Hirschberger
Claus Hillermeier
Dominik Dersch
Matthias Schmitz
Verena Schultheis
Paul Tavan

Publications:  "Extracting Markov models of peptide conformational dynamics from simulation data"
Verena Schultheis, Thomas Hirschberger, Heiko Carstens, and Paul Tavan
J. Chem. Theory Comput. 1, 515 - 526 (2005).

"Generalized radial basis function networks for classification and novelty detection: Self-organization of optimal Bayesian decision"
S. Albrecht, J. Busch, M. Kloppenburg, F. Metze und P. Tavan
Neural Networks 13: 1075-1093 (2000)

"Multivariate GRBF-Netzwerke und Systeme lokaler Experten: Neue Konzepte und ihre Anwendung bei komplexen Regelungsaufgaben"
Sebastian Albrecht
Doktorarbeit, Institut für Medizinische Optik der LMU, Fakultät für Physik, LMU München

"Deterministic annealing for density estimation by multivariate normal mixtures"
M. Kloppenburg, P. Tavan
Phys. Rev E 55: 2089-2092, 1997

"Asymptotic level density in topological feature maps"
D.R. Dersch, P. Tavan
IEEE Transactions on Neural Networks 6: 230-236, 199

"Toplogical feature maps with self-organized lateral connections: A population-coded one-layer model of associative memory"
C. Hillermeier, N. Kunstmann, B. Rabus, P. Tavan
Biol. Cybern. 72: 103-117, 1994

"An associative memory that can form hypotheses: A phase-coded neural network"
N. Kunstmann, C. Hillermeier, B. Rabus, P. Tavan
Biol. Cybern. 72: 118-132, 1994

"Load balanced vector quantization"
D. R. Dersch, P. Tavan
Proceedings of the ICANN'94, Sorrento M. Moreno and P.G. Morasso, eds., Springer (London 1994), 1067-1070

"Self-organization of associative memory and pattern classification: Recurrent signal processing on topological feature maps"
P. Tavan, H. Grubmüller, H. Kühnel
Biol. Cybern. 64: 95-105, 1990

"A self-organizing network for complete feature extraction"
J. Rubner, K. Schulten, P. Tavan
Parallel processing in neural systems and computers R. Eckmiller, G. Hartmann and G. Hauske, Eds., Elsevier, North-Holland, pp. 365-368, 1990

mailto: webmaster
Letzte Änderung: 2024-06-17 19:32