AI & ML New Capability

Introduces Any-Subgroup Equivariant Networks (ASEN), a single model that can adapt to multiple different symmetry groups via input modulation.

March 23, 2026

Original Paper

Any-Subgroup Equivariant Networks via Symmetry Breaking

Abhinav Goel, Derek Lim, Hannah Lawrence, Stefanie Jegelka, Ningyuan Huang

arXiv · 2603.19486

The Takeaway

Currently, equivariant models are hard-coded for specific symmetries (like rotation or translation). This framework allows a single base model to handle diverse geometric data types and symmetries dynamically, moving closer to universal geometric foundation models.

From the abstract

The inclusion of symmetries as an inductive bias, known as equivariance, often improves generalization on geometric data (e.g. grids, sets, and graphs). However, equivariant architectures are usually highly constrained, designed for symmetries chosen a priori, and not applicable to datasets with other symmetries. This precludes the development of flexible, multi-modal foundation models capable of processing diverse data equivariantly. In this work, we build a single model -- the Any-Subgroup Equ