We also find that previously-parameterized classical potentials model two separate anion polarization states that drastically influence resulting lithium solvation and transference.
A schematic of AutoBADDIE's workflow.
We introduce a low-data ML framework to predict classical interatomic potentials, showing strong agreement with experimental findings and DFT polymer backbone energy barriers. We use orders of magnitude less DFT training. All thanks to chemistry-informed symmetry and charge regularization terms.
Our work on "End-To-End Learning of Classical Interatomic Potentials for Benchmarking Anion Polarization Effects in Lithium Polymer Electrolytes" is out now in Chemistry of Materials! pubs.acs.org/doi/10.1021/...
