H2CO Dataset
Lists
Only you see this
Abstract
<p>The deposited data sets were used to compare three state-of-the art machine<br>\nlearning (ML) approaches to obtain representations of potential energy<br>\nsurfaces (PESs). The comparison is meant to be representative as it examines<br>\na purely kernel-based approach (reproducing kernel Hilbert space plus<br>\nforces (RKHS+F))[1], a purely neural network based approach (PhysNet)[2]<br>\nand includes the FCHL representation [3] within kernel ridge regression.<br>\nFormaldehyde, H2CO, is used as a benchmark system.</p>\n\n<p>H2CO is a small molecule for which PESs can be calculated<br>\nat different levels of theory and, thus, suitable for an in-depth theoretical<br>\nstudy. Also, very high-level calculations have already been presented (see e.g.<br>\nRef. [4]) and experimental reference data is available to compare with [5].</p>\n\n<p>Using reference data calculated at three different levels of quantum chemical<br>\ntheory (B3LYP/cc-pVDZ, MP2/aug-cc-pVTZ and CCSD(T)-F12/aug-cc-pVTZ-F12) ML<br>\nmodels are trained using the different ML methods. The performance of the<br>\nmodels is then examined by considering energy and force learning curves,<br>\nharmonic frequencies and IR spectra from finite-Temperature molecular dynamics<br>\n(MD) simulations.</p>\n\n<p>The data sets contain different geometries for the H2CO molecule generated using<br>\nthe normal mode sampling approach [6] performed at different temperatures. Four<br>\ndata sets are deposited:</p>\n\n<p><br>\ni) "h2co_B3LYP_cc-pVDZ_4001.npz": 4001 geometries of H2CO generated using normal mode<br>\n sampling and calculated using ORCA [7] (B3LYP/cc-pVDZ).<br>\nii) "h2co_mp2_avtz_4001.npz": 4001 geometries of H2CO generated using normal mode<br>\n sampling and calculated using MOLPRO 2019 [8] (MP2/aug-cc-pVTZ).<br>\niii) "h2co_ccsdt_avtz_4001.npz": 4001 geometries of H2CO generated using normal mode<br>\n sampling and calculated using MOLPRO 2019 [8] (CCSD(T)-F12/aug-cc-pVTZ-F12).<br>\niv) "h2co_ccsdt_avtz_2500_extrapol.npz": 2500 geometries of H2CO generated using normal mode<br>\n sampling and calculated using MOLPRO 2019 [8] (CCSD(T)-F12/aug-cc-pVTZ-F12). This sampling<br>\n was carried out at higher temperature (5000 K compared to 2000K) to test the extrapolation<br>\n ability of the ML methods.</p>\n\n<p>For more details, see <a href="http://arxiv.org/abs/2006.16752">http://arxiv.org/abs/2006.16752</a></p>\n\n<p>---------------------------------------------------------------------------------------<br>\nHOW TO CITE:</p>\n\n<p>When using this dataset, please cite the following paper:<br>\nKäser, S. and Koner, D. and Christensen, A. S. and von Lilienfeld, O. A. and Meuwly, M.<br>\n"ML Models of Vibrating H2CO: Comparing Reproducing Kernels, FCHL and PhysNet"<br>\narXiv:2006.16752</p>\n\n<p>and the digital object identifier (DOI):<br>\nKäser, S. and Koner, D. and Christensen, A. S. and von Lilienfeld, O. A. and Meuwly, M. (2020).<br>\nH2CO Dataset. Zenodo. http://doi.org/10.5281/zenodo.3923823</p>\n\n<p>---------------------------------------------------------------------------------------</p>\n\n<p>[1] Koner, D.; Meuwly, M. arXiv e-prints 2020, arXiv:2005.04667<br>\n[2] Unke, O. T.; Meuwly, M. J. Chem. Theory Comput. 2019, 15, 3678–3693<br>\n[3] Faber, F. A.; Christensen, A. S.; Huang, B.; von Lilienfeld, O. A. J. Chem. Phys. 2018, 148, 241717<br>\n[4] Zhang, X.; Zou, S.; Harding, L. B.; Bowman, J. M. J. Phys. Chem. A 2004, 108, 8980–8986<br>\n[5] Herndon, S. C.; Nelson Jr, D. D.; Li, Y.; Zahniser, M. S. J. Quant. Spectrosc. Radiat. Transf. 2005, 90, 207–216<br>\n[6] Smith, J. S.; Isayev, O.; Roitberg, A. E. Sci. Data 2017, 4, 170193<br>\n[7] Neese, F. The ORCA program system. WIREs Comput. Mol. Sci. 2012, 2, 73–78<br>\n[8] Werner, H.-J.; Knowles, P. J.; Knizia, G.; Manby, F. R.; Schütz, M.; et al. https://www.molpro.net</p>
BibTeX
@inproceedings{kaser2020h,
author = {Silvan Käser and Debasish Koner and Anders S. Christensen and O. Anatole von Lilienfeld and Markus Meuwly},
title = {{H2CO Dataset}},
booktitle = {Figshare},
year = {2020},
doi = {10.5281/zenodo.3923823},
}Loading related content...