Towards structural systematicity in distributed, statically bound visual representations

Abstract

The problem of representing the spatial structure of images, which arises in visual object processing, is commonly described using terminology borrowed from propositional theories of cognition, notably, the concept of compositionality. The classical propositional stance mandates representations composed of symbols, which stand for atomic or composite entities and enter into arbitrarily nested relationships. We argue that the main desiderata of a representational system -productivity and systematicity -can (indeed, for a number of reasons, should) be achieved without recourse to the classical, proposition-like compositionality. We show how this can be done, by describing a systematic and productive model of the representation of visual structure, which relies on static rather than dynamic binding and uses coarsely coded rather than atomic shape primitives.

Young Children's Self-Generated Object Views and Object Recognition

“…Theories (as well as studies with nonhuman primates) suggest that the construction of sparse shape representations is critically dependent on category learning (Kourtzi & Connor, 2011; Doumas & Hummel, 2010; Edelman & Intrator, 2003). Thus, it could be that both children’s dynamic viewing preferences and their ability to represent common objects in terms of sparse geometric shape depend on category learning, which is indexed by vocabulary size.…”

Symbolic play connects to language through visual object recognition

“…Perry, Samuelson, Malloy & Schiffer, 2010; Smith, Jones, Landau, Gershkoff‐Stowe & Samuelson, 2002). Further, computational models of object recognition succeed on the assumption that sparse representations of 3‐dimensional object shape are a consequence of category learning (Edelman & Intrator, 2003; Doumas & Hummel, 2010). Thus, abstracting the underlying geometric structure of an object’s shape – an ability important to mature object recognition and, by the present results, to object substitutions in play – may depend on learning names for a range of basic‐level category instances (see Perry et al ., 2010).…”

Economic Choice as an Untangling of Options into Actions

“…Indeed, some have argued that eliminating such information is a necessary precondition for accomplishing the invariant representations associated with later regions in the visual hierarchy (e.g. Edelman and Intrator, 2003; Serre et al, 2007). …”