Diagnostic Benchmarks for Invariant Learning Dynamics: Empirical Validation of the Eidos Architecture
Fuente:
arXiv
Guardado en:
| Autor principal: | |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866910021978161152 |
|---|---|
| author | Anderson, Datorien L. |
| author_facet | Anderson, Datorien L. |
| contents | We present the PolyShapes-Ideal (PSI) dataset, a suite of diagnostic benchmarks designed to isolate topological invariance -- the ability to maintain structural identity across affine transformations -- from the textural correlations that dominate standard vision benchmarks. Through three diagnostic probes (polygon classification under noise, zero-shot font transfer from MNIST, and geometric collapse mapping under progressive deformation), we demonstrate that the Eidos architecture achieves >99% accuracy on PSI and 81.67% zero-shot transfer across 30 unseen typefaces without pre-training. These results validate the "Form-First" hypothesis: generalization in structurally constrained architectures is a property of geometric integrity, not statistical scale. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_13322 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Diagnostic Benchmarks for Invariant Learning Dynamics: Empirical Validation of the Eidos Architecture Anderson, Datorien L. Computer Vision and Pattern Recognition Machine Learning We present the PolyShapes-Ideal (PSI) dataset, a suite of diagnostic benchmarks designed to isolate topological invariance -- the ability to maintain structural identity across affine transformations -- from the textural correlations that dominate standard vision benchmarks. Through three diagnostic probes (polygon classification under noise, zero-shot font transfer from MNIST, and geometric collapse mapping under progressive deformation), we demonstrate that the Eidos architecture achieves >99% accuracy on PSI and 81.67% zero-shot transfer across 30 unseen typefaces without pre-training. These results validate the "Form-First" hypothesis: generalization in structurally constrained architectures is a property of geometric integrity, not statistical scale. |
| title | Diagnostic Benchmarks for Invariant Learning Dynamics: Empirical Validation of the Eidos Architecture |
| topic | Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2602.13322 |