Hypersolid: Emergent Vision Representations via Short-Range Repulsion
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866911405728333824 |
|---|---|
| author | Rodríguez-Betancourt, Esteban Casasola-Murillo, Edgar |
| author_facet | Rodríguez-Betancourt, Esteban Casasola-Murillo, Edgar |
| contents | A recurring challenge in self-supervised learning is preventing representation collapse. Existing solutions typically rely on global regularization, such as maximizing distances, decorrelating dimensions or enforcing certain distributions. We instead reinterpret representation learning as a discrete packing problem, where preserving information simplifies to maintaining injectivity. We operationalize this in Hypersolid, a method using short-range hard-ball repulsion to prevent local collisions. This constraint results in a high-separation geometric regime that preserves augmentation diversity, excelling on fine-grained and low-resolution classification tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_21255 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Hypersolid: Emergent Vision Representations via Short-Range Repulsion Rodríguez-Betancourt, Esteban Casasola-Murillo, Edgar Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning A recurring challenge in self-supervised learning is preventing representation collapse. Existing solutions typically rely on global regularization, such as maximizing distances, decorrelating dimensions or enforcing certain distributions. We instead reinterpret representation learning as a discrete packing problem, where preserving information simplifies to maintaining injectivity. We operationalize this in Hypersolid, a method using short-range hard-ball repulsion to prevent local collisions. This constraint results in a high-separation geometric regime that preserves augmentation diversity, excelling on fine-grained and low-resolution classification tasks. |
| title | Hypersolid: Emergent Vision Representations via Short-Range Repulsion |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2601.21255 |