TY  - RPRT
TI  - How the Optimizer Shapes Learned Solutions in Equivariant Neural Networks
AU  - Teodor-Mihai Stupariu
AU  - Andrei Manolache
PY  - 2026
UR  - https://arxiv.org/abs/2605.27662
ID  - 2605.27662
ER  -