Non linear neurons in the low noise limit : a factorial code maximizes information transferJean

By Jean-Pierre Nadal. 1994

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

The "independent components" of natural scenes are edge filters. (1997, unknown): “…is the same (up to scaling and shifting) as the cumulative density functions (c.d.f.s) of the underlying independent components, it can be shown (Nadal & Parga, 1994)* that such a nonlinear "infomax" procedure also minimizes the mutual information between the ui, exactly what is required for…”
Independent component analysis: algorithms and applications (2000, methodology): “Another related contrast function was derived from a neural network viewpoint in Bell and Sejnowski (1995) and Nadal and Parga (1994).”
A Unifying Information-Theoretic Framework for Independent Component Analysis (2000, background): “3 Information Maximization Nadal and Parga (1994) showed that in the low-noise case, the maximum of the mutual information between the inputs x and outputs y of a neural processor implied that the output distributions were factorial.”
Continuous latent variable models for dimensionality reduction and sequential data reconstruction (2001, background): “31), or approximate versions of it, can be justified from several points of view: Information maximisation In the low noise case, the infomax principle states that if the mutual information between the inputs t and the outputs y of a processor is maximum, then the output distribution p(y) is factori”
Independent component analysis: recent advances (2013, result): “Furthermore, a neural network approach called infomax was proposed by Bell and Sejnowski (1995); Nadal and Parga (1994) and was shown to be equivalent to likelihood by Cardoso (1997).”