50 citations · 103 across the 10 of their papers we have counts for
10 papers
Statistical physics of independent component analysis
R. Urbanczik
Statistical physics is used to investigate independent component analysis with polynomial contrast functions. While the replica method fails, an adapted cavity approach yields vali…
Learning Curves for Mutual Information Maximization
Robert Urbanczik
An unsupervised learning procedure based on maximizing the mutual information between the outputs of two networks receiving different but statistically dependent inputs is analyzed…
Statistical Physics of Feedforward Neural Networks
Robert Urbanczik
The article is a lightly edited version of my habilitation thesis at the University Wuerzburg. My aim is to give a self contained, if concise, introduction to the formal methods us…
Learning multilayer perceptrons efficiently
C. Bunzmann, M. Biehl, R. Urbanczik
A learning algorithm for multilayer perceptrons is presented which is based on finding the principal components of a correlation matrix computed from the example inputs and their t…
Online Learning with Ensembles
R. Urbanczik
Supervised online learning with an ensemble of students randomized by the choice of initial conditions is analyzed. For the case of the perceptron learning rule, asymptotically the…
Noisy regression and classification with continuous multilayer networks
Martin Ahr, Michael Biehl, Robert Urbanczik
We investigate zero temperature Gibbs learning for two classes of unrealizable rules which play an important role in practical applications of multilayer neural networks with diffe…