MindSet: Vision. A toolbox for testing DNNs on key psychological experiments
Lists
Only you see this
Abstract
Multiple benchmarks have been developed to assess the alignment between deep neural networks (DNNs) and human vision. In almost all cases these benchmarks are observational in the sense they are composed of behavioural and brain responses to naturalistic images that have not been manipulated to test hypotheses regarding how DNNs or humans perceive and identify objects. Here we introduce the toolbox MindSet: Vision, consisting of a collection of image datasets and related scripts designed to test DNNs on 30 psychological findings. In all experimental conditions, the stimuli are systematically manipulated to test specific hypotheses regarding human visual perception and object recognition. In addition to providing pre-generated datasets of images, we provide code to regenerate these datasets, offering many configurable parameters which greatly extend the dataset versatility for different research contexts, and code to facilitate the testing of DNNs on these image datasets using three different methods (similarity judgments, out-of-distribution classification, and decoder method), accessible via GitHub. We test ResNet-152 on each of these methods as an example of how the toolbox can be used.
BibTeX
@misc{biscione2024mindset,
author = {Valerio Biscione and Dong Yin and Gaurav Malhotra and Marin Dujmovic and Milton L. Montero and Guillermo Puebla and Federico Adolfi and Rachel F Heaton and John E. Hummel and Benjamin D. Evans and Karim G. Habashy and Jeffrey Bowers},
title = {{MindSet: Vision. A toolbox for testing DNNs on key psychological experiments}},
year = {2024},
eprint = {2404.05290},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2404.05290},
}