Behavioral and Brain Sciences



Open Peer Commentary
Edelman: Representation of similarities

How to combine interpolation with feedback?


Guenther Palm a1
a1 Department of Neural Information Processing, University of Ulm, D-89069 Ulm, Germany palm@neuro.informatik.uni-ulm.de

Abstract

The Chorus representation is a sparse, similarity-preserving representation achieved by a feedforward neural network. Hence it is probably better suited for interpolation than for categorization. This commentary raises the question of how to combine categorization with interpolation, whether feedforward networks can be reasonable models for parts of the cerebral cortex, and whether people can perform more than one interpolation at a time.