SA-NNBF creates fully antisymmetric, spin-symmetric neural wavefunctions via sum-of-products spin eigenfunctions and tensor compression, enabling VMC calculations that outperform standard NNBF and SA-DMRG on systems with over 100 electrons including FeMoco.
(b) Spin correlation function⟨ ˆS1 · ˆSn⟩as a function ofnestimated with the optimized NQS’s (h= 128), in comparison with the DMRG result as reference
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
physics.chem-ph 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Spin-adapted neural network backflow for strongly correlated electrons
SA-NNBF creates fully antisymmetric, spin-symmetric neural wavefunctions via sum-of-products spin eigenfunctions and tensor compression, enabling VMC calculations that outperform standard NNBF and SA-DMRG on systems with over 100 electrons including FeMoco.