Physical Bongard Problems

Abstract

In this paper, we introduce Physical Bongard Problems (PBPs) as a novel and potentially rich approach to study the impact the constraints of a physical world have on mechanisms of concept learning and scene categorization. Each PBP consists of a set of 2D physical scenes which are positive or negative examples of a concept that must be identified. We discuss the properties that make PBPs challenging, analyze computational and representational requirements for a computational solver, and describe a first implementation of such a system. It can solve a subset of non-trivial PBPs using a version space approach for achieving its scene categorizations. The key element is a physics engine that is used both for the construction of information-rich physical features and for the prediction of how a given situation might evolve.

Does Vocabulary Help Structure the Mind?

“…Rather than using the original Bongard problems, we used a variant of Bongard problems which depict simple physical events and relationships requiring participants to reason about similarities and differences between these events, and often requiring participants to mentally simulate how an event will unfold in time (Weitnauer & Ritter, 2012). 9 For example, in problem 9 (Figure 5A) what makes the scenes on the left different from the scenes on the right is that the two objects will move in the same direction in the scenes on the right, while they will move in opposite directions in the scenes on the left.…”

“…To view the full set of the physical Bongard problems created byWeitnauer and Ritter (2012) along with their solutions, see https://github.com/eweitnauer/PBPs/tree/master/pngs/all-with-sol. We selected 11 of these physical Bongard problems for our experiments, using the first 12 example images for each problem (see Fig.5A).…”

“…We selected 11 of these physical Bongard problems for our experiments, using the first 12 example images for each problem (see Fig.5A). The numbers in Figures5 and 6correspond to the numbering used byWeitnauer & Ritter (2012).10 Many of these scenes were short animations that demonstrated the unfolding of the event over time, e.g., https://github.com/mzettersten/vocab-mind-2020/blob/master/pbp/stimuli/namingTask/pbp09_naming.gif.…”

Does vocabulary help structure the mind?

“…Rather than using the original Bongard problems, we used a variant of Bongard problems which depict simple physical events and relationships requiring participants to reason about similarities and differences between these events, and often requiring participants to mentally simulate how an event will unfold in time (Weitnauer & Ritter, 2012). For example, in problem 9 ( Figure 5A) what makes the scenes on the left different from the scenes on the right is that the two objects will move in the same direction in the scenes on the right, while they will move in opposite directions in the scenes on the left.…”

A Situated Dialogue System for Learning Structural Concepts in Blocks World

“…With these constraints in mind, we use a physical Bongard problem (Bongard et al, 1970;Weitnauer and Ritter, 2012) task for evaluating our system in a situated Blocks World environment. The user is provided with a set of visual examples, some positive and some negative, and must determine the constraints on the class of structure that allows the positive examples (and perhaps others) while avoiding any negative examples (Figure 1).…”