A Rational Model of Dimension-reduced Human Categorization
This addresses the challenge of modeling human-like categorization efficiency for cognitive science and AI, though it appears incremental as it builds on existing mPPCA methods.
The paper tackled the problem of how humans categorize with few samples despite many features by proposing a dimension-reduced representation using a mixture of probabilistic principal component analyzers (mPPCA), which effectively predicted human categorization on the CIFAR-10H dataset with only a single principal component per category.
Humans can categorize with only a few samples despite the numerous features. To mimic this ability, we propose a novel dimension-reduced category representation using a mixture of probabilistic principal component analyzers (mPPCA). Tests on the ${\tt CIFAR-10H}$ dataset demonstrate that mPPCA with only a single principal component for each category effectively predicts human categorization of natural images. We further impose a hierarchical prior on mPPCA to account for new category generalization. mPPCA captures human behavior in our experiments on images with simple size-color combinations. We also provide sufficient and necessary conditions when reducing dimensions in categorization is rational.