Learning without relearning physical symmetry: an SO(3)-equivariant convolutional network (CNN) on the sphere
arXiv:2610.23821v1 Announce Type: new Abstract: Many of the most important datasets in the physical sciences live natively on the sphere, and these data often carry no preferred position or orientation. Using methods that do not reflect this symmetry leads to far higher simulation requirements, whi…
Read the full story at arXiv stat.ML ↗
Timeline · 1 report
- 2026-10-19 04:00 · arXiv stat.ML
Learning without relearning physical symmetry: an SO(3)-equivariant convolutional network (CNN) on the sphere