# Universal dimensions of visual representation

arXiv: https://arxiv.org/abs/2408.12804 · HTML: https://mlsanity.com/paper/2408.12804/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- Builds on 12 of its 37 references that are in the map.
- First posted on arXiv 23 Aug 2024.

## Builds on (most cited first, 12 of 12)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [ImageNet Large Scale Visual Recognition Challenge](https://mlsanity.com/paper/1409.0575/) (2014)
- [Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification](https://mlsanity.com/paper/1502.01852/) (2015)
- [MPNet: Masked and Permuted Pre-training for Language Understanding](https://mlsanity.com/paper/2004.09297/) (2020)
- [The Platonic Representation Hypothesis](https://mlsanity.com/paper/2405.07987/) (2024)
- [Understanding Neural Networks Through Deep Visualization](https://mlsanity.com/paper/1506.06579/) (2015)
- [Recurrence is required to capture the representational dynamics of the human visual system](https://mlsanity.com/paper/1903.05946/) (2019)
- [The neuroconnectionist research programme](https://mlsanity.com/paper/2209.03718/) (2022)
- [Model Zoos: A Dataset of Diverse Populations of Neural Network Models](https://mlsanity.com/paper/2209.14764/) (2022)
- [Visual representations in the human brain are aligned with large language models](https://mlsanity.com/paper/2209.11737/) (2022)

## Built on by (most cited first, 12 of 12)

- [Brain Mapping with Dense Features: Grounding Cortical Semantic Selectivity in Natural Images With Vision Transformers](https://mlsanity.com/paper/2410.05266/) (2024)
- [Language models align with brain regions that represent concepts across modalities](https://mlsanity.com/paper/2508.11536/) (2025)
- [Superposition disentanglement of neural representations reveals hidden alignment](https://mlsanity.com/paper/2510.03186/) (2025)
- [Representational Alignment Across Model Layers and Brain Regions with Multi-Level Optimal Transport](https://mlsanity.com/paper/2510.01706/) (2025)
- [Shared representations in brains and models reveal a two-route cortical organization during scene perception](https://mlsanity.com/paper/2507.13941/) (2025)
- [Characterizing Universal Object Representations Across Vision Models](https://mlsanity.com/paper/2605.13675/) (2026)
- [Artificial Phantasia: Emergent Mental Imagery in Large Language Models](https://mlsanity.com/paper/2509.23108/) (2025)
- [Extremely coarse learning objectives induce human-aligned representations in AI vision models](https://mlsanity.com/paper/2605.05556/) (2026)
- [Platonic Representations in the Human Brain: Unsupervised Recovery of Universal Geometry](https://mlsanity.com/paper/2605.20496/) (2026)
- [Toward a Machine Bertin: Why Visualization Needs Design Principles for Machine Cognition](https://mlsanity.com/paper/2602.01527/) (2026)
- [The Umwelt Representation Hypothesis: Rethinking Universality](https://mlsanity.com/paper/2604.17960/) (2026)
- [The Cross-Architecture Substrate: A Domain-Transcendent, Calibration-Surviving Geometric Invariant of Modern Vision Encoders](https://mlsanity.com/paper/2606.07882/) (2026)

## Abstract (arXiv)

Do neural network models of vision learn brain-aligned representations because they share architectural constraints and task objectives with biological vision or because they learn universal features of natural image processing? We characterized the universality of hundreds of thousands of representational dimensions from visual neural networks with varied construction. We found that networks with varied architectures and task objectives learn to represent natural images using a shared set of latent dimensions, despite appearing highly distinct at a surface level. Next, by comparing these networks with human brain representations measured with fMRI, we found that the most brain-aligned representations in neural networks are those that are universal and independent of a network's specific characteristics. Remarkably, each network can be reduced to fewer than ten of its most universal dimensions with little impact on its representational similarity to the human brain. These results suggest that the underlying similarities between artificial and biological vision are primarily governed by a core set of universal image representations that are convergently learned by diverse systems.