Clifford-Steerable CNNs

Clifford-Steerable CNNs (ICML 2024) — equivariant CNNs on multivector fields for relativistic and fluid PDE data, continuing UvA's Clifford Group Equivariant Networks line and sitting upstream of Erwin. UvA-led (Zhdanov, Ruhe, Forré) with JKU coauthors; extended in the equivariant PDE-surrogate literature. The Microsoft-branded sibling "Clifford Neural Layers for PDE Modeling" is excluded and attributed to Microsoft.