2000 character limit reached
Beyond MD17: the reactive xxMD dataset
Published 22 Aug 2023 in cs.LG, cs.AI, physics.chem-ph, and quant-ph | (2308.11155v3)
Abstract: System specific neural force fields (NFFs) have gained popularity in computational chemistry. One of the most popular datasets as a bencharmk to develop NFFs models is the MD17 dataset and its subsequent extension. These datasets comprise geometries from the equilibrium region of the ground electronic state potential energy surface, sampled from direct adiabatic dynamics. However, many chemical reactions involve significant molecular geometrical deformations, for example, bond breaking. Therefore, MD17 is inadequate to represent a chemical reaction. To address this limitation in MD17, we introduce a new dataset, called Extended Excited-state Molecular Dynamics (xxMD) dataset. The xxMD dataset involves geometries sampled from direct non-adiabatic dynamics, and the energies are computed at both multireference wavefunction theory and density functional theory. We show that the xxMD dataset involves diverse geometries which represent chemical reactions. Assessment of NFF models on xxMD dataset reveals significantly higher predictive errors than those reported for MD17 and its variants. This work underscores the challenges faced in crafting a generalizable NFF model with extrapolation capability.
- Chmiela, S. et al. Machine learning of accurate energy-conserving molecular force fields. \JournalTitleScience advances 3, e1603015 (2017).
- On the role of gradients for machine learning of molecular energies and forces. \JournalTitleMachine Learning: Science and Technology 1, 045018 (2020).
- Trajectory surface hopping approach to nonadiabatic molecular collisions: The reaction of h+ with d2. \JournalTitleThe Journal of chemical physics 55, 562–572 (1971).
- Trajectory-surface-hopping study of Na(3p P2superscriptP2{}^{2}\text{P}start_FLOATSUPERSCRIPT 2 end_FLOATSUPERSCRIPT P) + H2subscriptH2\text{H}_{2}H start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT -> Na(3s S2superscriptS2{}^{2}\text{S}start_FLOATSUPERSCRIPT 2 end_FLOATSUPERSCRIPT S) + H2subscriptH2\text{H}_{2}H start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT(v’, j’, θ𝜃\thetaitalic_θ). \JournalTitleThe Journal of chemical physics 79, 1334–1342 (1983).
- Herman, M. F. Nonadiabatic semiclassical scattering. i. analysis of generalized surface hopping procedures. \JournalTitleThe Journal of chemical physics 81, 754–763 (1984).
- Tully, J. C. Molecular dynamics with electronic transitions. \JournalTitleThe Journal of Chemical Physics 93, 1061–1071 (1990).
- Yarkony, D. R. Diabaolical conical intersections. \JournalTitleReviews of Modern Physics 68, 985 (1996).
- Levine, B. G. et al. Conical intersections at the nanoscale: Molecular ideas for materials. \JournalTitleAnnual Review of Physical Chemistry 70, 21 (2019).
- Ws22 database, wigner sampling and geometry interpolation for configurationally diverse molecular datasets. \JournalTitleScientific Data 10, 95 (2023).
- Batzner, S. et al. E (3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials. \JournalTitleNature communications 13, 2453 (2022).
- Mace: Higher order equivariant message passing neural networks for fast and accurate force fields. \JournalTitleAdvances in Neural Information Processing Systems 35, 11423–11436 (2022).
- Batatia, I. et al. The design space of e (3)-equivariant atom-centered interatomic potentials. \JournalTitlearXiv preprint arXiv:2205.06643 (2022).
- Equivariant message passing for the prediction of tensorial properties and molecular spectra. In International Conference on Machine Learning, 9377–9388 (PMLR, 2021).
- Liu, Y. et al. Spherical message passing for 3d graph networks. \JournalTitlearXiv preprint arXiv:2102.05013 (2021).
- A complete active space scf method (casscf) using a density matrix formulated super-ci approach. \JournalTitleChemical Physics 48, 157 (1980).
- The m06 suite of density functionals for main group thermochemistry, thermochemical kinetics, noncovalent interactions, excited states, and transition elements: two new functionals and systematic testing of four m06-class functionals and 12 other functionals. \JournalTitleTheoretical chemistry accounts 120, 215–241 (2008).
- Barbatti, M. Nonadiabatic dyanmics with trajectory surface hopping method. \JournalTitleWIREs Computational Molecular Science 1, 620 (2011).
- Conical intersections and double excitations in time-dependent density functional theory. \JournalTitleMolecular Physics 104, 1039–1051 (2006).
- Dual-functional tamm-dancoff approximation: A convenient density functional method that correctly describes s1/s0 conical intersections. \JournalTitleJournal of Physical Chemistry Letters 8, 2107–2112 (2017).
- Shu, Y. et al. Dynamics algorithms with only potential energies and gradients: Curvature-driven coherent switching with decay of mixing and curvature-driven trajectory surface hopping. \JournalTitleJournal of Chemical Theory and Computation 18, 1320 (2022).
- Fewest switches surface hopping with baeck-an couplings. \JournalTitleOpen Research Europe 1, 49 (2021).
- Zhang, L. et al. Nonadiabatic dynamics of 1,3-cyclohexadiene by curvature-driven coherent switching with decay of mixing. \JournalTitleJournal of Chemical Theory and Computation 18, 7073 (2022).
- Direct nonadiabatic dynamics of ammonia with curvature-driven coherent switching with decay of mixing and with fewest switches with time uncertainty: An illustration of population leaking in trajectory surface hopping due to frustrated hops. \JournalTitleJournal of Chemical Theory and Computation 19, 1672 (2023).
- Zhao, X. et al. Nonadiabatic coupling in trajectory surface hopping: Accurate time derivative coupling by the curvature-driven approximation. \JournalTitleJournal of Chemical Theory and Computation 19, 6577 (2023).
- Diabatization by machine intelligence. \JournalTitleJournal of Chemical Theory and Computation 16, 6456–6464 (2020).
- Permutationally restrained diabatization by machine intelligence. \JournalTitleJournal of Chemical Theory and Computation 17, 1106–1116 (2021).
- Diabatic states of molecules. \JournalTitleJournal of Physical Chemistry A 126, 992 (2022).
- Schütt, K. et al. Schnet: A continuous-filter convolutional neural network for modeling quantum interactions. \JournalTitleAdvances in neural information processing systems 30 (2017).
- Fast and uncertainty-aware directional message passing for non-equilibrium molecules. \JournalTitlearXiv preprint arXiv:2011.14115 (2020).
- Musaelian, A. et al. Learning local equivariant representations for large-scale atomistic dynamics. \JournalTitleNature Communications 14, 579 (2023).
- reactive xxMD dataset. https://doi.org/10.5281/zenodo.10393858 (2023).
- Larsen, A. H. et al. The atomic simulation environment—a python library for working with atoms. \JournalTitleJournal of Physics: Condensed Matter 29, 273002 (2017).
- Thirty years of density functional theory in computational chemistry: an overview and extensive assessment of 200 density functionals. \JournalTitleMolecular Physics 115, 2315–2372 (2017).
- Reaxff: A reactive force field for hydrocarbons. \JournalTitleJournal of Physical Chemistry A 105, 9396–9409 (2001).
- Parametrically managed activation function for fitting a neural network potential with physical behavior enforced by a low-dimensional potential. \JournalTitleJournal of Physical Chemistry A 127, 5287 (2023).
- Parametrization and validation of intramolecular force fields derived from dft calculations. \JournalTitleJournal of Chemical Theory and Computation 3, 1803 (2007).
- Vanduyfhuys, L. et al. Quickff: A program for a quick and easy derivation of force fields for metal-organic frameworks from ab initio input. \JournalTitleJournal of Computational Chemistry 36, 1015 (2016).
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.