Qualitative Spatial Reasoning for Geometric Analogies

In this paper we provide evidence supporting the hypothesis that qualitative spatial reasoning and representations provide a bridge between the perceptual and the conceptual. We demonstrate that qualitative visual structure combined with analogical processing can produce human-like results on a classic geometric analogies task. Importantly, the bulk of the computations are not particular to the task but are general purpose: we sketch the problems using our existing sketch understanding system, sKEA, which computes visual structure that is used by our existing analogical matcher, SME. The results are significant both as evidence in our ongoing research into general-purpose spatial representation, and in illustrating the utility of second-order analogical matches.