The Universal Law of Generalization Holds for Naturalistic Stimuli
This work addresses the gap in evidence for a foundational psychological law in real-world contexts, offering insights for cognitive science and AI perception, though it is incremental in applying existing methods to new data.
The study tackled the problem of testing Shepard's universal law of generalization with naturalistic stimuli, providing direct evidence by analyzing large datasets of human similarity and generalization judgments across natural images, involving over 2400 participants and hundreds of thousands of data points.
Shepard's universal law of generalization is a remarkable hypothesis about how intelligent organisms should perceive similarity. In its broadest form, the universal law states that the level of perceived similarity between a pair of stimuli should decay as a concave function of their distance when embedded in an appropriate psychological space. While extensively studied, evidence in support of the universal law has relied on low-dimensional stimuli and small stimulus sets that are very different from their real-world counterparts. This is largely because pairwise comparisons -- as required for similarity judgments -- scale quadratically in the number of stimuli. We provide direct evidence for the universal law in a naturalistic high-dimensional regime by analyzing an existing dataset of 214,200 human similarity judgments and a newly collected dataset of 390,819 human generalization judgments (N=2406 US participants) across three sets of natural images.