A differentiable phase-equilibrium loss function enables gradient-based optimization of thermodynamic model parameters, demonstrated on binary, ternary, and EAM potential systems.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cond-mat.mtrl-sci 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Machine Learning Compatible CALPHAD-type Optimization from Phase Equilibria by Auto-differentiation
A differentiable phase-equilibrium loss function enables gradient-based optimization of thermodynamic model parameters, demonstrated on binary, ternary, and EAM potential systems.