Cognitive Neural Architecture Search Reveals Hierarchical Entailment

2025-02-16

AI Summary

This study challenges the strict hierarchical processing model of the visual cortex and proposes a novel evolutionary neural architecture search (ENAS) paradigm optimized for brain signal alignment. Method: Leveraging fMRI response regression modeling and representational similarity analysis (RSA), the approach directly optimizes neural architectures to fit primate visual cortical responses in an unsupervised manner. Contributions/Results: (1) The method automatically discovers lightweight, shallow convolutional architectures that significantly outperform pretrained models (e.g., ResNet) on fMRI alignment metrics; (2) even with random weight initialization, these architectures achieve superior brain–machine alignment compared to supervised pretrained models; (3) the learned architectural priors intrinsically encode cognitive hierarchical structure. After standard supervised fine-tuning, the discovered architectures attain competitive ImageNet classification accuracy. These findings establish ENAS as a core computational tool for cognitive neuroscience, bridging neural architecture design and biological vision principles.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Deep Neural Architectures and Foundation Models
• Search and Optimization: Evolutionary Computation

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Web Mining and Content Analysis: Models for Web evolution

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.

Problem

• Optimize convolutional network architectures for brain-alignment
• Challenge traditional hierarchical structure of visual pathway
• Reduce reliance on manually designed convolutional networks

Innovation

• Evolutionary neural architecture search
• Brain-alignment optimization
• Hierarchical representational models