Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2605.29109 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911726093467648 |
|---|---|
| author | Smith, Wyatt A. Fernández-Ramírez, César Greensite, Jeff Szczepaniak, Adam P. |
| author_facet | Smith, Wyatt A. Fernández-Ramírez, César Greensite, Jeff Szczepaniak, Adam P. |
| contents | As a first step towards machine identification of confining objects in thermalized lattice gauge configurations, we present our 2dVoId model for center vortex identification on pure SU(2) lattices in $D = 2$ dimensions. We create a training set by inserting thin Z2 vortices at various locations on a zero action lattice, and then distort those configurations by applying random SU(2) gauge transformations, noise, and by thickening the vortices via cooling. For moderate vortex visibility, our model is able to reliably identify the location of center vortices. We additionally demonstrate scalability through tiling strategies, which will enable generalization to higher dimensions while reducing training costs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_29109 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | First steps towards gauge-independent vortex identification through machine learning Smith, Wyatt A. Fernández-Ramírez, César Greensite, Jeff Szczepaniak, Adam P. High Energy Physics - Lattice Nuclear Theory As a first step towards machine identification of confining objects in thermalized lattice gauge configurations, we present our 2dVoId model for center vortex identification on pure SU(2) lattices in $D = 2$ dimensions. We create a training set by inserting thin Z2 vortices at various locations on a zero action lattice, and then distort those configurations by applying random SU(2) gauge transformations, noise, and by thickening the vortices via cooling. For moderate vortex visibility, our model is able to reliably identify the location of center vortices. We additionally demonstrate scalability through tiling strategies, which will enable generalization to higher dimensions while reducing training costs. |
| title | First steps towards gauge-independent vortex identification through machine learning |
| topic | High Energy Physics - Lattice Nuclear Theory |
| url | https://arxiv.org/abs/2605.29109 |