Geometry-aware similarity metrics for neural representations on Riemannian and statistical manifolds
Fuente:
arXiv
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912999398178816 |
|---|---|
| author | Cayco-Gajic, N Alex Pellegrino, Arthur |
| author_facet | Cayco-Gajic, N Alex Pellegrino, Arthur |
| contents | Similarity measures are widely used to interpret the representational geometries used by neural networks to solve tasks. Yet, because existing methods compare the extrinsic geometry of representations in state space, rather than their intrinsic geometry, they may fail to capture subtle yet crucial distinctions between fundamentally different neural network solutions. Here, we introduce metric similarity analysis (MSA), a novel method which leverages tools from Riemannian geometry to compare the intrinsic geometry of neural representations under the manifold hypothesis. We show that MSA can be used to i) disentangle features of neural computations in deep networks with different learning regimes, ii) compare nonlinear dynamics, and iii) investigate diffusion models. Hence, we introduce a mathematically grounded and broadly applicable framework to understand the mechanisms behind neural computations by comparing their intrinsic geometries. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_28764 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Geometry-aware similarity metrics for neural representations on Riemannian and statistical manifolds Cayco-Gajic, N Alex Pellegrino, Arthur Machine Learning Artificial Intelligence Differential Geometry Neurons and Cognition Similarity measures are widely used to interpret the representational geometries used by neural networks to solve tasks. Yet, because existing methods compare the extrinsic geometry of representations in state space, rather than their intrinsic geometry, they may fail to capture subtle yet crucial distinctions between fundamentally different neural network solutions. Here, we introduce metric similarity analysis (MSA), a novel method which leverages tools from Riemannian geometry to compare the intrinsic geometry of neural representations under the manifold hypothesis. We show that MSA can be used to i) disentangle features of neural computations in deep networks with different learning regimes, ii) compare nonlinear dynamics, and iii) investigate diffusion models. Hence, we introduce a mathematically grounded and broadly applicable framework to understand the mechanisms behind neural computations by comparing their intrinsic geometries. |
| title | Geometry-aware similarity metrics for neural representations on Riemannian and statistical manifolds |
| topic | Machine Learning Artificial Intelligence Differential Geometry Neurons and Cognition |
| url | https://arxiv.org/abs/2603.28764 |