A solution to the tag-assignment problem for neural networks

Abstract

Purely parallel neural networks can model object recognition in brief displays – the same conditions under which illusory conjunctions (the incorrect combination of features into perceived objects in a stimulus array) have been demonstrated empirically (Treisman 1986; Treisman & Gelade 1980). Correcting errors of illusory conjunction is the “tag-assignment” problem for a purely parallel processor: the problem of assigning a spatial tag to nonspatial features, feature combinations, and objects. This problem must be solved to model human object recognition over a longer time scale. Our model simulates both the parallel processes that may underlie illusory conjunctions and the serial processes that may solve the tag-assignment problem in normal perception. One component of the model extracts pooled features and another provides attentional tags that correct illusory conjunctions. Our approach addresses two questions: (i) How can objects be identified from simultaneously attended features in a parallel, distributed representation? (ii) How can the spatial selectional requirements of such an attentional process be met by a separation of pathways for spatial and nonspatial processing? Our analysis of these questions yields a neurally plausible simulation of tag assignment based on synchronizing feature processing activity in a spatial focus of attention.

Transsaccadic integration of bystander locations

“…Experiment 1 seemed to suggest that the source might serve as the anchor of a configurational coding of the layout of the bystanders: The postsaccadic absence of source and bystanders had separate detrimental effects on displacement detection. That the source would serve as an anchor would be consistent with its central position in the display and the fact that it was at the centre of foveal attention during presaccadic coding, allowing for precise allocentric coding of its position relative to the other items in the display (Ballard, Hayhoe, Pook, & Rao, 1997;Ballard, Hayhoe, Salgian, & Shinoda, 2000;Strong & Whitehead, 1989). However, in Experiment 2, a factorial manipulation of postsaccadic presence of source and bystanders revealed that the source effect was entirely dependent upon the presence of the bystanders.…”

“…Given that the source is in the focus of attention during the better part of the presaccadic fixation, this may not be very surprising. Specifically, one prominent theory on how we keep track of object position during scene exploration proposes that only when an object is selectively attended its position is coded under the form of a set of vectors pointing to other objects in the visual field (Strong & Whitehead, 1989). On this view, the foveated source would make a prime candidate for acting as the centre of an environmentcentred reference frame within which the locations of other objects are coded.…”

Learning to Segment Images Using Dynamic Feature Binding

“…Malsburg and Schneider 1986), Gray et al (1989), Eckhorn et al (1988), and Strong and Whitehead (1989), among others, have suggested a biologically plausible mechanism of labeling through temporal correlations among neural signals, either the relative timing of neuronal spikes or the synchronization of oscillatory activities in the nervous system. The key idea here is that each processing unit conveys not just an activation value-average firing frequency in neural terms-but also a second, independent value that represents the relative phase of firing.…”

The how and why of what went where in apparent motion: Modeling solutions to the motion correspondence problem.

“…73-77). Third, if a FINST refers to a geocentric location, then it can be used to coordinate perceptual-motor functions.In order to provide these three functions, the FINST mechanism must solve what Strong and Whitehead (1989) called the tag-assignment problem: A FINST must keep pointing to the same feature cluster as the cluster changes retinal position, even though the cluster may undergo changes in appearance because of its movement. Pylyshyn (1988Pylyshyn ( , 1989 assumed that tag assignment is a primitive property of his model, but he did not provide details of how it is realized in the FINST mechanism.…”