A robust biologically plausible implementation of ICA-like learning

Felipe Gerhard, Cristina Savin, Jochen Triesch

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We present a model that can perform ICA-like learning by simple, local, biologically plausible rules. By combining synaptic learning with homeostatic regulation of neuron properties and adaptive lateral inhibition, the neural network can robustly learn Gabor-like receptive fields from natural images. With spatially localized inhibitory connections, a topographic map can be achieved. Additionally, the network can solve the Földiák bars problem, a classical nonlinear ICA task.

Original languageEnglish (US)
Title of host publicationESANN 2009 Proceedings, 17th European Symposium on Artificial Neural Networks - Advances in Computational Intelligence and Learning
Pages147-152
Number of pages6
StatePublished - 2009
Event17th European Symposium on Artificial Neural Networks - Advances in Computational Intelligence and Learning, ESANN 2009 - Bruges, Belgium
Duration: Apr 22 2009Apr 24 2009

Publication series

NameESANN 2009 Proceedings, 17th European Symposium on Artificial Neural Networks - Advances in Computational Intelligence and Learning

Other

Other17th European Symposium on Artificial Neural Networks - Advances in Computational Intelligence and Learning, ESANN 2009
CountryBelgium
CityBruges
Period4/22/094/24/09

ASJC Scopus subject areas

  • Artificial Intelligence
  • Information Systems

Fingerprint Dive into the research topics of 'A robust biologically plausible implementation of ICA-like learning'. Together they form a unique fingerprint.

Cite this