Search papers, labs, and topics across Lattice.
1
0
13
Equivariant neural networks can bypass the brutal scaling of many-body $GW$ perturbation theory, achieving few-meV quasiparticle accuracy and cross-scale transferability at linear cost.