<ul><li>Multiscale problems in physics often require computationally expensive high-resolution simulations.</li><li>Data-driven surrogate models have been used as a faster alternative, but struggle with incorporating microscale physical constraints.</li><li>Equilibrium Neural Operator (EquiNO) is proposed as a complementary physics-informed surrogate that can predict microscale physics.</li><li>The EquiNO framework integrates the finite element method with operator learning and achieves significant speedup compared to traditional methods.</li></ul>

EquiNO: A Physics-Informed Neural Operator for Multiscale Simulations

Discover more