Michael Gutmann
search externally:   Google Scholar,   Springer,   CiteSeer,   Microsoft Academic Search,   Scirus ,   DBlife


My research is about structure in natural sensory input and its neural representation.

Data representation methods such as Principal Component Analysis (PCA), Independent Component Analysis (ICA), and sparse coding are/were used to study this topic from the point of view of computational neuroscience.

I work on:

(1) extensions of such data representation methods to further elicit structure in natural stimuli.

The encoding of the sensory input in such methods relates, however, to rather abstract neuron models. In order to reduce this gap in the modeling of neural representation, I also work on

(2) data representation with more detailed, spiking, neuron models that originate from dynamical systems theory.


flag Unsupervised Learning by Discriminating Data from Artificial Noise
as author at  Generative / Discriminative Interface,