KineticNet: deep learning a transferable kinetic energy functional for orbital-free density functional theory
Orbital-free density functional theory (OF-DFT) holds promise to compute ground state molecular properties at minimal cost. However, it has been held back by our inability to compute the kinetic energy as a functional of electron density alone. Here, we set out to learn the kinetic energy functional...
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Dokumenttyp: | Article (Journal) |
| Sprache: | Englisch |
| Veröffentlicht: |
13 October 2023
|
| In: |
The journal of chemical physics
Year: 2023, Jahrgang: 159, Heft: 14, Pages: [1], 1-13 |
| ISSN: | 1089-7690 |
| DOI: | 10.1063/5.0158275 |
| Online-Zugang: | Verlag, lizenzpflichtig, Volltext: https://doi.org/10.1063/5.0158275 Verlag, lizenzpflichtig, Volltext: https://pubs.aip.org/aip/jcp/article/159/14/144113/2916356 
|
| Verfasserangaben: | R. Remme, T. Kaczun, M. Scheurer, A. Dreuw and F.A. Hamprecht |
MARC
| LEADER | 00000caa a2200000 c 4500 | ||
|---|---|---|---|
| 001 | 1876922923 | ||
| 003 | DE-627 | ||
| 005 | 20240307031945.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 231222s2023 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.1063/5.0158275 |2 doi | |
| 035 | |a (DE-627)1876922923 | ||
| 035 | |a (DE-599)KXP1876922923 | ||
| 035 | |a (OCoLC)1425208043 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rda | ||
| 041 | |a eng | ||
| 084 | |a 29 |2 sdnb | ||
| 100 | 1 | |a Remme, Roman |d 1996- |e VerfasserIn |0 (DE-588)1226781462 |0 (DE-627)174780827X |4 aut | |
| 245 | 1 | 0 | |a KineticNet |b deep learning a transferable kinetic energy functional for orbital-free density functional theory |c R. Remme, T. Kaczun, M. Scheurer, A. Dreuw and F.A. Hamprecht |
| 264 | 1 | |c 13 October 2023 | |
| 300 | |a 14 | ||
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a Computermedien |b c |2 rdamedia | ||
| 338 | |a Online-Ressource |b cr |2 rdacarrier | ||
| 500 | |a Gesehen am 22.12.2023 | ||
| 520 | |a Orbital-free density functional theory (OF-DFT) holds promise to compute ground state molecular properties at minimal cost. However, it has been held back by our inability to compute the kinetic energy as a functional of electron density alone. Here, we set out to learn the kinetic energy functional from ground truth provided by the more expensive Kohn-Sham density functional theory. Such learning is confronted with two key challenges: Giving the model sufficient expressivity and spatial context while limiting the memory footprint to afford computations on a GPU and creating a sufficiently broad distribution of training data to enable iterative density optimization even when starting from a poor initial guess. In response, we introduce KineticNet, an equivariant deep neural network architecture based on point convolutions adapted to the prediction of quantities on molecular quadrature grids. Important contributions include convolution filters with sufficient spatial resolution in the vicinity of nuclear cusp, an atom-centric sparse but expressive architecture that relays information across multiple bond lengths, and a new strategy to generate varied training data by finding ground state densities in the face of perturbations by a random external potential. KineticNet achieves, for the first time, chemical accuracy of the learned functionals across input densities and geometries of tiny molecules. For two-electron systems, we additionally demonstrate OF-DFT density optimization with chemical accuracy. | ||
| 700 | 1 | |a Kaczun, Tobias |e VerfasserIn |0 (DE-588)1314532952 |0 (DE-627)1876923156 |4 aut | |
| 700 | 1 | |a Scheurer, Maximilian |d 1994- |e VerfasserIn |0 (DE-588)1157021441 |0 (DE-627)1020013451 |0 (DE-576)502574127 |4 aut | |
| 700 | 1 | |a Dreuw, Andreas |d 1972- |e VerfasserIn |0 (DE-588)1060214598 |0 (DE-627)799305626 |0 (DE-576)416304974 |4 aut | |
| 700 | 1 | |a Hamprecht, Fred |e VerfasserIn |0 (DE-588)1020505605 |0 (DE-627)691240280 |0 (DE-576)360605516 |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t The journal of chemical physics |d Melville, NY : American Institute of Physics, 1933 |g 159(2023), 14, Artikel-ID 144113, Seite [1], 1-13 |h Online-Ressource |w (DE-627)268760675 |w (DE-600)1473050-9 |w (DE-576)077610261 |x 1089-7690 |7 nnas |a KineticNet deep learning a transferable kinetic energy functional for orbital-free density functional theory |
| 773 | 1 | 8 | |g volume:159 |g year:2023 |g number:14 |g elocationid:144113 |g pages:[1], 1-13 |g extent:14 |a KineticNet deep learning a transferable kinetic energy functional for orbital-free density functional theory |
| 856 | 4 | 0 | |u https://doi.org/10.1063/5.0158275 |x Verlag |x Resolving-System |z lizenzpflichtig |3 Volltext |
| 856 | 4 | 0 | |u https://pubs.aip.org/aip/jcp/article/159/14/144113/2916356 |x Verlag |z lizenzpflichtig |3 Volltext |
| 951 | |a AR | ||
| 992 | |a 20231222 | ||
| 993 | |a Article | ||
| 994 | |a 2023 | ||
| 998 | |g 1020505605 |a Hamprecht, Fred |m 1020505605:Hamprecht, Fred |d 700000 |d 708070 |e 700000PH1020505605 |e 708070PH1020505605 |k 0/700000/ |k 1/700000/708070/ |p 5 |y j | ||
| 998 | |g 1060214598 |a Dreuw, Andreas |m 1060214598:Dreuw, Andreas |d 700000 |d 708000 |e 700000PD1060214598 |e 708000PD1060214598 |k 0/700000/ |k 1/700000/708000/ |p 4 | ||
| 998 | |g 1157021441 |a Scheurer, Maximilian |m 1157021441:Scheurer, Maximilian |d 110000 |e 110000PS1157021441 |k 0/110000/ |p 3 | ||
| 998 | |g 1314532952 |a Kaczun, Tobias |m 1314532952:Kaczun, Tobias |d 500000 |d 501249 |e 500000PK1314532952 |e 501249PK1314532952 |k 0/500000/ |k 1/500000/501249/ |p 2 | ||
| 998 | |g 1226781462 |a Remme, Roman |m 1226781462:Remme, Roman |d 700000 |d 708000 |e 700000PR1226781462 |e 708000PR1226781462 |k 0/700000/ |k 1/700000/708000/ |p 1 |x j | ||
| 999 | |a KXP-PPN1876922923 |e 4443175148 | ||
| BIB | |a Y | ||
| SER | |a journal | ||
| JSO | |a {"name":{"displayForm":["R. Remme, T. Kaczun, M. Scheurer, A. Dreuw and F.A. Hamprecht"]},"language":["eng"],"note":["Gesehen am 22.12.2023"],"recId":"1876922923","person":[{"display":"Remme, Roman","given":"Roman","role":"aut","family":"Remme"},{"family":"Kaczun","role":"aut","display":"Kaczun, Tobias","given":"Tobias"},{"display":"Scheurer, Maximilian","given":"Maximilian","family":"Scheurer","role":"aut"},{"given":"Andreas","display":"Dreuw, Andreas","family":"Dreuw","role":"aut"},{"role":"aut","family":"Hamprecht","display":"Hamprecht, Fred","given":"Fred"}],"physDesc":[{"extent":"14 S."}],"id":{"eki":["1876922923"],"doi":["10.1063/5.0158275"]},"title":[{"title_sort":"KineticNet","subtitle":"deep learning a transferable kinetic energy functional for orbital-free density functional theory","title":"KineticNet"}],"type":{"bibl":"article-journal","media":"Online-Ressource"},"origin":[{"dateIssuedDisp":"13 October 2023","dateIssuedKey":"2023"}],"relHost":[{"name":{"displayForm":["American Institute of Physics"]},"recId":"268760675","id":{"zdb":["1473050-9"],"issn":["1089-7690"],"eki":["268760675"]},"physDesc":[{"extent":"Online-Ressource"}],"title":[{"subtitle":"bridges a gap between journals of physics and journals of chemistry","title":"The journal of chemical physics","title_sort":"journal of chemical physics"}],"origin":[{"publisherPlace":"Melville, NY","dateIssuedKey":"1933","dateIssuedDisp":"1933-","publisher":"American Institute of Physics"}],"part":{"year":"2023","volume":"159","text":"159(2023), 14, Artikel-ID 144113, Seite [1], 1-13","extent":"14","pages":"[1], 1-13","issue":"14"},"language":["eng"],"pubHistory":["1.1933 -"],"note":["Gesehen am 16.06.2023"],"type":{"media":"Online-Ressource","bibl":"periodical"},"disp":"KineticNet deep learning a transferable kinetic energy functional for orbital-free density functional theoryThe journal of chemical physics"}]} | ||
| SRT | |a REMMEROMANKINETICNET1320 | ||