A self-refining training loop, where a neural network samples molecular conformations from its own predicted energy and trains on them, reduces the need for large labeled DFT datasets in amortized density functional theory.
W., Musaelian, A., Owen, C
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Self-Refining Training for Amortized Density Functional Theory
A self-refining training loop, where a neural network samples molecular conformations from its own predicted energy and trains on them, reduces the need for large labeled DFT datasets in amortized density functional theory.