Contrastive Regularized MSE (CRMSE) corrects distribution-level failures of MSE-trained ML interatomic potentials on ethanol and aspirin from MD17 by penalizing spurious minima sampled from the model's own Langevin chains.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
physics.chem-ph 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Contrastive Regularization of Machine Learning Potentials
Contrastive Regularized MSE (CRMSE) corrects distribution-level failures of MSE-trained ML interatomic potentials on ethanol and aspirin from MD17 by penalizing spurious minima sampled from the model's own Langevin chains.