Introduces representability-aware neural networks for predicting and variationally optimizing 2-RDMs, achieving high accuracy and competitive energies versus exact diagonalization in a fractional Chern insulator model of twisted bilayer MoTe2.
Specifically, we note that the energies predicted by the interpolation-trained NNs (App
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cond-mat.str-el 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Representability-Aware Neural Networks for Reduced Density Matrices: Application to Fractional Chern Insulators
Introduces representability-aware neural networks for predicting and variationally optimizing 2-RDMs, achieving high accuracy and competitive energies versus exact diagonalization in a fractional Chern insulator model of twisted bilayer MoTe2.