Cognitive Neural Architecture Search Reveals Hierarchical Entailment

TL;DR

This paper uses neural architecture search to find convolutional models that align with brain hierarchies, revealing that hierarchical structure is fundamental in primate visual processing and offering a new approach for computational neuroscience.

Contribution

It demonstrates that neural architecture search can discover models with brain-like hierarchies, challenging previous assumptions and reducing reliance on manual network design.

Findings

Models with random weights surpass pretrained models in brain alignment.

Optimized architectures become competitive classifiers after supervised training.

Hierarchical structure is confirmed as fundamental in primate visual processing.

Abstract

Recent research has suggested that the brain is more shallow than previously thought, challenging the traditionally assumed hierarchical structure of the ventral visual pathway. Here, we demonstrate that optimizing convolutional network architectures for brain-alignment via evolutionary neural architecture search results in models with clear representational hierarchies. Despite having random weights, the identified models achieve brain-alignment scores surpassing even those of pretrained classification models - as measured by both regression and representational similarity analysis. Furthermore, through traditional supervised training, architectures optimized for alignment with late ventral regions become competitive classification models. These findings suggest that hierarchical structure is a fundamental mechanism of primate visual processing. Finally, this work demonstrates the potential of neural architecture search as a framework for computational cognitive neuroscience research that could reduce the field's reliance on manually designed convolutional networks.