Abstract
Humans are particularly sensitive to changes in the 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 analysing 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 behaviour was best predicted by the amount 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.
Competing Interest Statement
The authors have declared no competing interest.
Footnotes
This research was supported by the European Research Council Grant Generalization in Mind and Machine, ID number 741134.
A pre-print of this article has been made available on bioRxiv: doi:https://doi.org/10.1101/2021.12.14.472546 and a version of code for running the simulations reported in the manuscript as well as participant data from Experiment 4 is available at: https://github.com/gammagit/distal
We have revised the Abstract and Title for clarity. We have also restructured the experiments and simulations so the Methods of each Experiment / Simulations is reported with the experiment.
