TRANSFORM-INVARIANT RECOGNITION BY ASSOCIATION IN A RECURRENT NETWORK

Objects can be recognized independently of the view they present, of t heir position on the retina, or their scale. It has been suggested tha t one basic mechanism that makes this possible is a memory effect, or a trace, that allows associations to be made between consecutive views of one object. In this work, we explore the possibility that this mem ory trace is provided by the sustained activity of neurons in layers o f the visual pathway produced by an extensive recurrent connectivity. We describe a model that contains this high recurrent connectivity and synaptic efficacies built with contributions from associations betwee n pairs of views that is simple enough to be treated analytically. The main result is that there is a change of behavior as the strength of the association between views of the same object, relative to the asso ciation within each view of an object, increases. When its value is sm all, sustained activity in the network is produced by the views themse lves. As it increases above a threshold value, the network always reac hes a particular state (which represents the object) independent of th e particular view that was seen as a stimulus. In this regime, the net work can still store an extensive number of objects, each defined by a finite (although it can be large) number of views.