Feature, Alignment, and Supervision in Category Learning: A Comparative Approach With Children and Neural Networks
Abstract
Understanding how humans and machines learn from sparse data is central to cognitive science and machine learning. Using a matched task design, we compare children and convolutional neural networks (CNNs) in a few-shot semi-supervised category learning task. Both learners received mixtures of labeled and unlabeled exemplars while supervision (1/3/6 labels), target feature (size, shape, pattern), and perceptual alignment (high/low) are systematically varied. We find that children generalize rapidly from minimal labels but show strong feature-specific biases and sensitivity to alignment. CNNs show a different interaction profile: added supervision improves performance, but both alignment and feature structure moderate the impact additional supervision has on learning. These findings show that comparisons between humans and machines should be controlled and sensitive to task structure. Comparing accuracy alone may hinder a more nuanced understanding of the conditions that differentially support success in each system.