Fast and energy-efficient neuromorphic deep learning with first-spike times
For a biological agent operating under environmental pressure, energy consumption and reaction times are of critical importance. Similarly, engineered systems are optimized for short time-to-solution and low energy-to-solution characteristics. At the level of neuronal implementation, this implies ac...
Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article (Journal) |
| Language: | English |
| Published: |
17 September 2021
|
| In: |
Nature machine intelligence
Year: 2021, Volume: 3, Issue: 9, Pages: 823-835 |
| ISSN: | 2522-5839 |
| DOI: | 10.1038/s42256-021-00388-x |
| Online Access: | Verlag, lizenzpflichtig, Volltext: https://doi.org/10.1038/s42256-021-00388-x Verlag, lizenzpflichtig, Volltext: https://www.nature.com/articles/s42256-021-00388-x 
|
| Author Notes: | J. Göltz, L. Kriener, A. Baumbach, S. Billaudelle, O. Breitwieser, B. Cramer, D. Dold, A.F. Kungl, W. Senn, J. Schemmel, K. Meier and M.A. Petrovici |
MARC
| LEADER | 00000caa a2200000 c 4500 | ||
|---|---|---|---|
| 001 | 1774586355 | ||
| 003 | DE-627 | ||
| 005 | 20220820061543.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 211020s2021 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.1038/s42256-021-00388-x |2 doi | |
| 035 | |a (DE-627)1774586355 | ||
| 035 | |a (DE-599)KXP1774586355 | ||
| 035 | |a (OCoLC)1341421991 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rda | ||
| 041 | |a eng | ||
| 084 | |a 29 |2 sdnb | ||
| 100 | 1 | |a Göltz, Julian |e VerfasserIn |0 (DE-588)1197708723 |0 (DE-627)1679440802 |4 aut | |
| 245 | 1 | 0 | |a Fast and energy-efficient neuromorphic deep learning with first-spike times |c J. Göltz, L. Kriener, A. Baumbach, S. Billaudelle, O. Breitwieser, B. Cramer, D. Dold, A.F. Kungl, W. Senn, J. Schemmel, K. Meier and M.A. Petrovici |
| 264 | 1 | |c 17 September 2021 | |
| 300 | |a 13 | ||
| 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 20.10.2021 | ||
| 520 | |a For a biological agent operating under environmental pressure, energy consumption and reaction times are of critical importance. Similarly, engineered systems are optimized for short time-to-solution and low energy-to-solution characteristics. At the level of neuronal implementation, this implies achieving the desired results with as few and as early spikes as possible. With time-to-first-spike coding, both of these goals are inherently emerging features of learning. Here, we describe a rigorous derivation of a learning rule for such first-spike times in networks of leaky integrate-and-fire neurons, relying solely on input and output spike times, and show how this mechanism can implement error backpropagation in hierarchical spiking networks. Furthermore, we emulate our framework on the BrainScaleS-2 neuromorphic system and demonstrate its capability of harnessing the system’s speed and energy characteristics. Finally, we examine how our approach generalizes to other neuromorphic platforms by studying how its performance is affected by typical distortive effects induced by neuromorphic substrates. | ||
| 700 | 1 | |a Kriener, Laura |e VerfasserIn |0 (DE-588)113836181X |0 (DE-627)895871386 |0 (DE-576)492500061 |4 aut | |
| 700 | 1 | |a Baumbach, Andreas |d 1992- |e VerfasserIn |0 (DE-588)1169784216 |0 (DE-627)1035890704 |0 (DE-576)512249334 |4 aut | |
| 700 | 1 | |a Billaudelle, Sebastian |d 1991- |e VerfasserIn |0 (DE-588)118726153X |0 (DE-627)1666378356 |4 aut | |
| 700 | 1 | |a Breitwieser, Oliver |d 1987- |e VerfasserIn |0 (DE-588)1163617652 |0 (DE-627)1027933408 |0 (DE-576)508090997 |4 aut | |
| 700 | 1 | |a Cramer, Benjamin |d 1993- |e VerfasserIn |0 (DE-588)1125028068 |0 (DE-627)879345616 |0 (DE-576)483278572 |4 aut | |
| 700 | 1 | |a Dold, Dominik |d 1990- |e VerfasserIn |0 (DE-588)1125027908 |0 (DE-627)879345179 |0 (DE-576)483278009 |4 aut | |
| 700 | 1 | |a Kungl, Ákos Ferenc |d 1991- |e VerfasserIn |0 (DE-588)1187260193 |0 (DE-627)1666373842 |4 aut | |
| 700 | 1 | |a Senn, W. |e VerfasserIn |4 aut | |
| 700 | 1 | |a Schemmel, Johannes |e VerfasserIn |0 (DE-588)1025834607 |0 (DE-627)72488291X |0 (DE-576)370821440 |4 aut | |
| 700 | 1 | |a Meier, Karlheinz |d 1955-2018 |e VerfasserIn |0 (DE-588)1025835115 |0 (DE-627)724884114 |0 (DE-576)370822269 |4 aut | |
| 700 | 1 | |a Petrovici, Mihai A. |e VerfasserIn |0 (DE-588)1072021005 |0 (DE-627)826788823 |0 (DE-576)433488700 |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t Nature machine intelligence |d [London] : Springer Nature Publishing, 2019 |g 3(2021), 9 vom: Sept., Seite 823-835 |h Online-Ressource |w (DE-627)1025147669 |w (DE-600)2933875-X |w (DE-576)506804771 |x 2522-5839 |7 nnas |a Fast and energy-efficient neuromorphic deep learning with first-spike times |
| 773 | 1 | 8 | |g volume:3 |g year:2021 |g number:9 |g month:09 |g pages:823-835 |g extent:13 |a Fast and energy-efficient neuromorphic deep learning with first-spike times |
| 856 | 4 | 0 | |u https://doi.org/10.1038/s42256-021-00388-x |x Verlag |x Resolving-System |z lizenzpflichtig |3 Volltext |
| 856 | 4 | 0 | |u https://www.nature.com/articles/s42256-021-00388-x |x Verlag |z lizenzpflichtig |3 Volltext |
| 951 | |a AR | ||
| 992 | |a 20211020 | ||
| 993 | |a Article | ||
| 994 | |a 2021 | ||
| 998 | |g 1072021005 |a Petrovici, Mihai A. |m 1072021005:Petrovici, Mihai A. |d 130000 |d 130700 |e 130000PP1072021005 |e 130700PP1072021005 |k 0/130000/ |k 1/130000/130700/ |p 12 |y j | ||
| 998 | |g 1025835115 |a Meier, Karlheinz |m 1025835115:Meier, Karlheinz |d 910000 |d 950000 |d 950950 |e 910000PM1025835115 |e 950000PM1025835115 |e 950950PM1025835115 |k 0/910000/ |k 1/910000/950000/ |k 2/910000/950000/950950/ |p 11 | ||
| 998 | |g 1025834607 |a Schemmel, Johannes |m 1025834607:Schemmel, Johannes |d 130000 |d 130700 |d 130000 |e 130000PS1025834607 |e 130700PS1025834607 |e 130000PS1025834607 |k 0/130000/ |k 1/130000/130700/ |k 0/130000/ |p 10 | ||
| 998 | |g 1187260193 |a Kungl, Ákos Ferenc |m 1187260193:Kungl, Ákos Ferenc |d 910000 |d 950000 |d 950950 |e 910000PK1187260193 |e 950000PK1187260193 |e 950950PK1187260193 |k 0/910000/ |k 1/910000/950000/ |k 2/910000/950000/950950/ |p 8 | ||
| 998 | |g 1125027908 |a Dold, Dominik |m 1125027908:Dold, Dominik |d 910000 |d 950000 |d 950950 |e 910000PD1125027908 |e 950000PD1125027908 |e 950950PD1125027908 |k 0/910000/ |k 1/910000/950000/ |k 2/910000/950000/950950/ |p 7 | ||
| 998 | |g 1125028068 |a Cramer, Benjamin |m 1125028068:Cramer, Benjamin |d 130000 |d 130700 |e 130000PC1125028068 |e 130700PC1125028068 |k 0/130000/ |k 1/130000/130700/ |p 6 | ||
| 998 | |g 1163617652 |a Breitwieser, Oliver |m 1163617652:Breitwieser, Oliver |d 910000 |d 950000 |d 950950 |e 910000PB1163617652 |e 950000PB1163617652 |e 950950PB1163617652 |k 0/910000/ |k 1/910000/950000/ |k 2/910000/950000/950950/ |p 5 | ||
| 998 | |g 118726153X |a Billaudelle, Sebastian |m 118726153X:Billaudelle, Sebastian |d 130000 |d 130700 |e 130000PB118726153X |e 130700PB118726153X |k 0/130000/ |k 1/130000/130700/ |p 4 | ||
| 998 | |g 1169784216 |a Baumbach, Andreas |m 1169784216:Baumbach, Andreas |d 130000 |d 130700 |e 130000PB1169784216 |e 130700PB1169784216 |k 0/130000/ |k 1/130000/130700/ |p 3 | ||
| 998 | |g 1197708723 |a Göltz, Julian |m 1197708723:Göltz, Julian |d 130000 |d 130700 |e 130000PG1197708723 |e 130700PG1197708723 |k 0/130000/ |k 1/130000/130700/ |p 1 |x j | ||
| 999 | |a KXP-PPN1774586355 |e 3991588250 | ||
| BIB | |a Y | ||
| SER | |a journal | ||
| JSO | |a {"relHost":[{"physDesc":[{"extent":"Online-Ressource"}],"origin":[{"dateIssuedDisp":"[2019]-","publisher":"Springer Nature Publishing","publisherPlace":"[London]"}],"language":["eng"],"disp":"Fast and energy-efficient neuromorphic deep learning with first-spike timesNature machine intelligence","part":{"text":"3(2021), 9 vom: Sept., Seite 823-835","extent":"13","pages":"823-835","volume":"3","year":"2021","issue":"9"},"pubHistory":["Volume 1, no. 1 (January 2019)-"],"type":{"bibl":"periodical","media":"Online-Ressource"},"id":{"eki":["1025147669"],"issn":["2522-5839"],"zdb":["2933875-X"]},"title":[{"title":"Nature machine intelligence","title_sort":"Nature machine intelligence"}],"note":["Gesehen am 30.04.25"],"recId":"1025147669"}],"language":["eng"],"origin":[{"dateIssuedDisp":"17 September 2021","dateIssuedKey":"2021"}],"physDesc":[{"extent":"13 S."}],"type":{"media":"Online-Ressource","bibl":"article-journal"},"name":{"displayForm":["J. Göltz, L. Kriener, A. Baumbach, S. Billaudelle, O. Breitwieser, B. Cramer, D. Dold, A.F. Kungl, W. Senn, J. Schemmel, K. Meier and M.A. Petrovici"]},"id":{"doi":["10.1038/s42256-021-00388-x"],"eki":["1774586355"]},"title":[{"title_sort":"Fast and energy-efficient neuromorphic deep learning with first-spike times","title":"Fast and energy-efficient neuromorphic deep learning with first-spike times"}],"person":[{"display":"Göltz, Julian","given":"Julian","role":"aut","family":"Göltz"},{"display":"Kriener, Laura","given":"Laura","role":"aut","family":"Kriener"},{"family":"Baumbach","display":"Baumbach, Andreas","given":"Andreas","role":"aut"},{"family":"Billaudelle","display":"Billaudelle, Sebastian","given":"Sebastian","role":"aut"},{"role":"aut","given":"Oliver","display":"Breitwieser, Oliver","family":"Breitwieser"},{"given":"Benjamin","role":"aut","display":"Cramer, Benjamin","family":"Cramer"},{"family":"Dold","display":"Dold, Dominik","role":"aut","given":"Dominik"},{"role":"aut","given":"Ákos Ferenc","display":"Kungl, Ákos Ferenc","family":"Kungl"},{"given":"W.","role":"aut","display":"Senn, W.","family":"Senn"},{"family":"Schemmel","given":"Johannes","role":"aut","display":"Schemmel, Johannes"},{"family":"Meier","given":"Karlheinz","role":"aut","display":"Meier, Karlheinz"},{"display":"Petrovici, Mihai A.","given":"Mihai A.","role":"aut","family":"Petrovici"}],"recId":"1774586355","note":["Gesehen am 20.10.2021"]} | ||
| SRT | |a GOELTZJULIFASTANDENE1720 | ||