SLIM: Learning an Operator for Material-Efficient 3D Lattice Refinement
A neural lattice-refinement model that replaces per-instance topology optimization with a single forward prediction, removing redundant material while preserving homogenized stiffness.
arXiv preprint, 2026.