Human Shape Representations Are Not an Emergent Property of Learning to Classify Objects

Humans are particularly sensitive to relationships between parts of objects. It remains unclear why this is. One hypothesis is that relational features are highly diagnostic of object categories and emerge as a result of learning to classify objects. We tested this by analyzing the internal represen...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Journal of experimental psychology. General 2023-12, Vol.152 (12), p.3380-3402
Hauptverfasser: Malhotra, Gaurav, Dujmović, Marin, Hummel, John, Bowers, Jeffrey S.
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Humans are particularly sensitive to relationships between parts of objects. It remains unclear why this is. One hypothesis is that relational features are highly diagnostic of object categories and emerge as a result of learning to classify objects. We tested this by analyzing the internal representations of supervised convolutional neural networks (CNNs) trained to classify large sets of objects. We found that CNNs do not show the same sensitivity to relational changes as previously observed for human participants. Furthermore, when we precisely controlled the deformations to objects, human behavior was best predicted by the number of relational changes while CNNs were equally sensitive to all changes. Even changing the statistics of the learning environment by making relations uniquely diagnostic did not make networks more sensitive to relations in general. Our results show that learning to classify objects is not sufficient for the emergence of human shape representations. Instead, these results suggest that humans are selectively sensitive to relational changes because they build representations of distal objects from their retinal images and interpret relational changes as changes to these distal objects. This inferential process makes human shape representations qualitatively different from those of artificial neural networks optimized to perform image classification. Public Significance StatementThis study shows significant differences in how the human visual system and recent artificial intelligence (AI) models represent objects. This difference between the two systems likely stems from their different goals-while AI models are built to classify objects, humans must additionally reason and interact with them. Our results suggest that the human visual system represents objects in a manner that enables us to perform these additional tasks.
ISSN:0096-3445
1939-2222
DOI:10.1037/xge0001440