Article

Model metamers illuminate divergences between biological and artificial neural networks

2022-05-20

Abstract excerpt

Deep neural network models of sensory systems are often proposed to learn representational transformations with invariances like those in the brain. To reveal these invariances we generated “model metamers” – stimuli whose activations within a model stage are matched to those of a natural stimulus. Metamers for state-of-the-art supervised and unsupervised neural network models of vision and audition were often com...

Topics

Open a Topic to create a Post that cites this publication.

Identifiers and source

Literature Corpus work: c6382ebe-d7c1-5e3e-8d87-183577a484d9
DOI: 10.1101/2022.05.19.492678

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

Model metamers illuminate divergences between biological and artificial neural networksDOI 10.1101/2022.05.19.492678Read

Investigating Invariances in Auditory Event Categorization with Model MetamersSemantic proximity Read

Neural prediction decorrelation reveals that adversarial robustness substantially improves DNN prediction accuracy across the entire human auditory cortexSemantic proximity Read

Comparing neural models using their perceptual discriminability predictionsSemantic proximity Read

Probing Human Categorization of Naturalistic Images by Combining Deep Neural Networks with Cognitive ModelsSemantic proximity Read

Human Sensitivity to Perturbations Constrained by a Model of the Natural Image ManifoldSemantic proximity Read

Aligning Model and Macaque Inferior Temporal Cortex Representations Improves Model-to-Human Behavioral Alignment and Adversarial RobustnessSemantic proximity Read

The role of capacity constraints in Convolutional Neural Networks for learning random versus natural dataSemantic proximity Read

Gradual sensory maturation promotes abstract representation learningSemantic proximity Read

Simulating a Primary Visual Cortex at the Front of CNNs Improves Robustness to Image PerturbationsSemantic proximity Read

Robust deep learning object recognition models rely on low frequency information in natural imagesSemantic proximity Read

Direct-fit to nature: an evolutionary perspective on biological (and artificial) neural networksSemantic proximity Read

Biological convolutions improve DNN robustness to noise and generalisationSemantic proximity Read