Spectral reconstruction with deep neural networks

We explore artificial neural networks as a tool for the reconstruction of spectral functions from imaginary time Green’s functions, a classic ill-conditioned inverse problem. Our ansatz is based on a supervised learning framework in which prior knowledge is encoded in the training data and the inver...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Kades, Lukas (VerfasserIn) , Pawlowski, Jan M. (VerfasserIn) , Rothkopf, Alexander (VerfasserIn) , Scherzer, Manuel (VerfasserIn) , Urban, Julian M. (VerfasserIn) , Wetzel, Sebastian (VerfasserIn) , Wink, Nicolas (VerfasserIn) , Ziegler, Felix P. G. (VerfasserIn)
Dokumenttyp: Article (Journal)
Sprache:Englisch
Veröffentlicht: 9 November 2020
In: Physical review
Year: 2020, Jahrgang: 102, Heft: 9, Pages: 1-19
ISSN:2470-0029
DOI:10.1103/PhysRevD.102.096001
Online-Zugang:Verlag, lizenzpflichtig, Volltext: https://doi.org/10.1103/PhysRevD.102.096001
Verlag, lizenzpflichtig, Volltext: https://link.aps.org/doi/10.1103/PhysRevD.102.096001
Volltext
Verfasserangaben:Lukas Kades, Jan M. Pawlowski, Alexander Rothkopf, Manuel Scherzer, Julian M. Urban, Sebastian J. Wetzel, Nicolas Wink, and Felix P.G. Ziegler

MARC

LEADER 00000caa a2200000 c 4500
001 1759905704
003 DE-627
005 20220819222622.0
007 cr uuu---uuuuu
008 210607s2020 xx |||||o 00| ||eng c
024 7 |a 10.1103/PhysRevD.102.096001  |2 doi 
035 |a (DE-627)1759905704 
035 |a (DE-599)KXP1759905704 
035 |a (OCoLC)1341415139 
040 |a DE-627  |b ger  |c DE-627  |e rda 
041 |a eng 
084 |a 29  |2 sdnb 
100 1 |a Kades, Lukas  |d 1991-  |e VerfasserIn  |0 (DE-588)117814528X  |0 (DE-627)1049231686  |0 (DE-576)517696266  |4 aut 
245 1 0 |a Spectral reconstruction with deep neural networks  |c Lukas Kades, Jan M. Pawlowski, Alexander Rothkopf, Manuel Scherzer, Julian M. Urban, Sebastian J. Wetzel, Nicolas Wink, and Felix P.G. Ziegler 
264 1 |c 9 November 2020 
300 |a 19 
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 07.06.2021 
520 |a We explore artificial neural networks as a tool for the reconstruction of spectral functions from imaginary time Green’s functions, a classic ill-conditioned inverse problem. Our ansatz is based on a supervised learning framework in which prior knowledge is encoded in the training data and the inverse transformation manifold is explicitly parametrized through a neural network. We systematically investigate this novel reconstruction approach, providing a detailed analysis of its performance on physically motivated mock data, and compare it to established methods of Bayesian inference. The reconstruction accuracy is found to be at least comparable and potentially superior in particular at larger noise levels. We argue that the use of labeled training data in a supervised setting and the freedom in defining an optimization objective are inherent advantages of the present approach and may lead to significant improvements over state-of-the-art methods in the future. Potential directions for further research are discussed in detail. 
700 1 |a Pawlowski, Jan M.  |d 1965-  |e VerfasserIn  |0 (DE-588)1047077388  |0 (DE-627)777525925  |0 (DE-576)400331381  |4 aut 
700 1 |a Rothkopf, Alexander  |d 1982-  |e VerfasserIn  |0 (DE-588)1147581789  |0 (DE-627)1006325409  |0 (DE-576)495767360  |4 aut 
700 1 |a Scherzer, Manuel  |e VerfasserIn  |0 (DE-588)1156540917  |0 (DE-627)1019338962  |0 (DE-576)502235950  |4 aut 
700 1 |a Urban, Julian M.  |d 1994-  |e VerfasserIn  |0 (DE-588)1173444157  |0 (DE-627)1043277242  |0 (DE-576)515261769  |4 aut 
700 1 |a Wetzel, Sebastian  |d 1987-  |e VerfasserIn  |0 (DE-588)1085180069  |0 (DE-627)848860608  |0 (DE-576)456916016  |4 aut 
700 1 |a Wink, Nicolas  |d 1994-  |e VerfasserIn  |0 (DE-588)1153918382  |0 (DE-627)1015518893  |0 (DE-576)500627967  |4 aut 
700 1 |a Ziegler, Felix P. G.  |d 1990-  |e VerfasserIn  |0 (DE-588)1147584567  |0 (DE-627)1006327932  |0 (DE-576)495768162  |4 aut 
773 0 8 |i Enthalten in  |t Physical review  |d Ridge, NY : American Physical Society, 2016  |g 102(2020), 9, Artikel-ID 096001, Seite 1-19  |h Online-Ressource  |w (DE-627)846313510  |w (DE-600)2844732-3  |w (DE-576)454495811  |x 2470-0029  |7 nnas  |a Spectral reconstruction with deep neural networks 
773 1 8 |g volume:102  |g year:2020  |g number:9  |g elocationid:096001  |g pages:1-19  |g extent:19  |a Spectral reconstruction with deep neural networks 
856 4 0 |u https://doi.org/10.1103/PhysRevD.102.096001  |x Verlag  |x Resolving-System  |z lizenzpflichtig  |3 Volltext 
856 4 0 |u https://link.aps.org/doi/10.1103/PhysRevD.102.096001  |x Verlag  |z lizenzpflichtig  |3 Volltext 
951 |a AR 
992 |a 20210607 
993 |a Article 
994 |a 2020 
998 |g 1153918382  |a Wink, Nicolas  |m 1153918382:Wink, Nicolas  |d 130000  |d 130300  |e 130000PW1153918382  |e 130300PW1153918382  |k 0/130000/  |k 1/130000/130300/  |p 7 
998 |g 1173444157  |a Urban, Julian M.  |m 1173444157:Urban, Julian M.  |d 130000  |d 700000  |d 728500  |e 130000PU1173444157  |e 700000PU1173444157  |e 728500PU1173444157  |k 0/130000/  |k 0/700000/  |k 1/700000/728500/  |p 5 
998 |g 1047077388  |a Pawlowski, Jan M.  |m 1047077388:Pawlowski, Jan M.  |d 130000  |d 130300  |d 700000  |d 728500  |e 130000PP1047077388  |e 130300PP1047077388  |e 700000PP1047077388  |e 728500PP1047077388  |k 0/130000/  |k 1/130000/130300/  |k 0/700000/  |k 1/700000/728500/  |p 2 
998 |g 117814528X  |a Kades, Lukas  |m 117814528X:Kades, Lukas  |d 130000  |d 130300  |d 700000  |d 728500  |e 130000PK117814528X  |e 130300PK117814528X  |e 700000PK117814528X  |e 728500PK117814528X  |k 0/130000/  |k 1/130000/130300/  |k 0/700000/  |k 1/700000/728500/  |p 1  |x j 
999 |a KXP-PPN1759905704  |e 3935090617 
BIB |a Y 
SER |a journal 
JSO |a {"id":{"eki":["1759905704"],"doi":["10.1103/PhysRevD.102.096001"]},"language":["eng"],"name":{"displayForm":["Lukas Kades, Jan M. Pawlowski, Alexander Rothkopf, Manuel Scherzer, Julian M. Urban, Sebastian J. Wetzel, Nicolas Wink, and Felix P.G. Ziegler"]},"type":{"media":"Online-Ressource","bibl":"article-journal"},"recId":"1759905704","physDesc":[{"extent":"19 S."}],"person":[{"family":"Kades","given":"Lukas","role":"aut","display":"Kades, Lukas"},{"role":"aut","display":"Pawlowski, Jan M.","given":"Jan M.","family":"Pawlowski"},{"given":"Alexander","family":"Rothkopf","role":"aut","display":"Rothkopf, Alexander"},{"given":"Manuel","family":"Scherzer","role":"aut","display":"Scherzer, Manuel"},{"role":"aut","display":"Urban, Julian M.","given":"Julian M.","family":"Urban"},{"family":"Wetzel","given":"Sebastian","role":"aut","display":"Wetzel, Sebastian"},{"family":"Wink","given":"Nicolas","role":"aut","display":"Wink, Nicolas"},{"family":"Ziegler","given":"Felix P. G.","display":"Ziegler, Felix P. G.","role":"aut"}],"title":[{"title_sort":"Spectral reconstruction with deep neural networks","title":"Spectral reconstruction with deep neural networks"}],"relHost":[{"pubHistory":["3rd series, volume 93, number 1 (January 2016)-"],"id":{"eki":["846313510"],"issn":["2470-0029"],"zdb":["2844732-3"]},"recId":"846313510","name":{"displayForm":["published by American Physical Society"]},"physDesc":[{"extent":"Online-Ressource"}],"origin":[{"dateIssuedKey":"2016","publisherPlace":"Ridge, NY","publisher":"American Physical Society","dateIssuedDisp":"2016-"}],"note":["Gesehen am 14.03.2023"],"disp":"Spectral reconstruction with deep neural networksPhysical review","titleAlt":[{"title":"Particles, fields, gravitation, and cosmology"}],"language":["eng"],"type":{"media":"Online-Ressource","bibl":"periodical"},"part":{"volume":"102","year":"2020","extent":"19","pages":"1-19","text":"102(2020), 9, Artikel-ID 096001, Seite 1-19","issue":"9"},"corporate":[{"display":"American Physical Society","role":"isb"}],"title":[{"title_sort":"Physical review","title":"Physical review"}]}],"origin":[{"dateIssuedDisp":"9 November 2020","dateIssuedKey":"2020"}],"note":["Gesehen am 07.06.2021"]} 
SRT |a KADESLUKASSPECTRALRE9202