Eric Elmoznino

Convolutional architectures are cortex-aligned de novo

The paper shows that convolutional neural networks with minimal training can produce cortex-aligned visual representations through specific dimensionality manipulations, suggesting architectural constraints are key.

High-performing neural network models of visual cortex benefit from high latent dimensionality

This study shows that high-dimensional latent spaces improve generalization in DNN models of visual cortex when predicting cortical responses to new stimuli.