Neuromorphic hardware learns to learn
Hyperparameters and learning algorithms for neuromorphic hardware are usually chosen by hand to suit a particular task. In contrast, networks of neurons in the brain were optimized through extensive evolutionary and developmental processes to work well on a range of computing and learning tasks. Occ...
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Dokumenttyp: | Article (Journal) |
| Sprache: | Englisch |
| Veröffentlicht: |
21 May 2019
|
| In: |
Frontiers in neuroscience
Year: 2019, Jahrgang: 13 |
| ISSN: | 1662-453X |
| DOI: | 10.3389/fnins.2019.00483 |
| Online-Zugang: | Verlag, Volltext: https://doi.org/10.3389/fnins.2019.00483 Verlag: https://www.frontiersin.org/articles/10.3389/fnins.2019.00483/full 
|
| Verfasserangaben: | Thomas Bohnstingl, Franz Scherr, Christian Pehle, Karlheinz Meier and Wolfgang Maass |
MARC
| LEADER | 00000caa a2200000 c 4500 | ||
|---|---|---|---|
| 001 | 1685460569 | ||
| 003 | DE-627 | ||
| 005 | 20220817184514.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 191213s2019 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.3389/fnins.2019.00483 |2 doi | |
| 035 | |a (DE-627)1685460569 | ||
| 035 | |a (DE-599)KXP1685460569 | ||
| 035 | |a (OCoLC)1341282172 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rda | ||
| 041 | |a eng | ||
| 084 | |a 33 |2 sdnb | ||
| 100 | 1 | |a Ortner, Thomas |e VerfasserIn |0 (DE-588)120157689X |0 (DE-627)1685466478 |4 aut | |
| 245 | 1 | 0 | |a Neuromorphic hardware learns to learn |c Thomas Bohnstingl, Franz Scherr, Christian Pehle, Karlheinz Meier and Wolfgang Maass |
| 264 | 1 | |c 21 May 2019 | |
| 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 13.12.2019 | ||
| 520 | |a Hyperparameters and learning algorithms for neuromorphic hardware are usually chosen by hand to suit a particular task. In contrast, networks of neurons in the brain were optimized through extensive evolutionary and developmental processes to work well on a range of computing and learning tasks. Occasionally this process has been emulated through genetic algorithms, but these require themselves hand-design of their details and tend to provide a limited range of improvements. We employ instead other powerful gradient-free optimization tools, such as cross-entropy methods and evolutionary strategies, in order to port the function of biological optimization processes to neuromorphic hardware. As an example, we show these optimization algorithms enable neuromorphic agents to learn very efficiently from rewards. In particular, meta-plasticity, i.e., the optimization of the learning rule which they use, substantially enhances reward-based learning capability of the hardware. In addition, we demonstrate for the first time Learning-to-Learn benefits from such hardware, in particular, the capability to extract abstract knowledge from prior learning experiences that speeds up the learning of new but related tasks. Learning-to-Learn is especially suited for accelerated neuromorphic hardware, since it makes it feasible to carry out the required very large number of network computations. | ||
| 650 | 4 | |a HICANN-DLS | |
| 650 | 4 | |a learning-to-learn | |
| 650 | 4 | |a Markov decision processes (MDPs) | |
| 650 | 4 | |a Meta-plasticity | |
| 650 | 4 | |a mixed-signal hardware | |
| 650 | 4 | |a Multi-armed bandits (MAB) | |
| 650 | 4 | |a neuromorphic hardware | |
| 650 | 4 | |a Spiking Neural Networks (SNN) | |
| 650 | 4 | |a Transfer Learning | |
| 700 | 1 | |a Scherr, Franz |e VerfasserIn |4 aut | |
| 700 | 1 | |a Pehle, Christian |d 1988- |e VerfasserIn |0 (DE-588)1093238550 |0 (DE-627)853202338 |0 (DE-576)46265527X |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 Maass, Wolfgang |e VerfasserIn |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t Frontiers in neuroscience |d Lausanne : Frontiers Research Foundation, 2007 |g 13(2019) Artikel-Nummer 483, 14 Seiten |h Online-Ressource |w (DE-627)55908109X |w (DE-600)2411902-7 |w (DE-576)281378959 |x 1662-453X |7 nnas |a Neuromorphic hardware learns to learn |
| 773 | 1 | 8 | |g volume:13 |g year:2019 |g extent:14 |a Neuromorphic hardware learns to learn |
| 856 | 4 | 0 | |u https://doi.org/10.3389/fnins.2019.00483 |x Verlag |x Resolving-System |3 Volltext |
| 856 | 4 | 0 | |u https://www.frontiersin.org/articles/10.3389/fnins.2019.00483/full |x Verlag |
| 951 | |a AR | ||
| 992 | |a 20191213 | ||
| 993 | |a Article | ||
| 994 | |a 2019 | ||
| 998 | |g 1025835115 |a Meier, Karlheinz |m 1025835115:Meier, Karlheinz |d 130000 |d 130700 |e 130000PM1025835115 |e 130700PM1025835115 |k 0/130000/ |k 1/130000/130700/ |p 4 | ||
| 998 | |g 1093238550 |a Pehle, Christian |m 1093238550:Pehle, Christian |d 130000 |d 130700 |e 130000PP1093238550 |e 130700PP1093238550 |k 0/130000/ |k 1/130000/130700/ |p 3 | ||
| 999 | |a KXP-PPN1685460569 |e 3564555374 | ||
| BIB | |a Y | ||
| SER | |a journal | ||
| JSO | |a {"origin":[{"dateIssuedKey":"2019","dateIssuedDisp":"21 May 2019"}],"relHost":[{"recId":"55908109X","note":["Gesehen am 04.06.20"],"language":["eng"],"title":[{"title":"Frontiers in neuroscience","title_sort":"Frontiers in neuroscience"}],"physDesc":[{"extent":"Online-Ressource"}],"name":{"displayForm":["Frontiers Research Foundation"]},"disp":"Neuromorphic hardware learns to learnFrontiers in neuroscience","pubHistory":["1.2007 -"],"part":{"year":"2019","volume":"13","extent":"14","text":"13(2019) Artikel-Nummer 483, 14 Seiten"},"id":{"zdb":["2411902-7"],"issn":["1662-453X"],"eki":["55908109X"]},"origin":[{"dateIssuedDisp":"2007-","publisher":"Frontiers Research Foundation","publisherPlace":"Lausanne","dateIssuedKey":"2007"}],"type":{"media":"Online-Ressource","bibl":"periodical"}}],"type":{"media":"Online-Ressource","bibl":"article-journal"},"id":{"eki":["1685460569"],"doi":["10.3389/fnins.2019.00483"]},"physDesc":[{"extent":"14 S."}],"person":[{"role":"aut","roleDisplay":"VerfasserIn","given":"Thomas","family":"Ortner","display":"Ortner, Thomas"},{"family":"Scherr","given":"Franz","display":"Scherr, Franz","role":"aut","roleDisplay":"VerfasserIn"},{"family":"Pehle","given":"Christian","display":"Pehle, Christian","roleDisplay":"VerfasserIn","role":"aut"},{"role":"aut","roleDisplay":"VerfasserIn","given":"Karlheinz","family":"Meier","display":"Meier, Karlheinz"},{"family":"Maass","given":"Wolfgang","display":"Maass, Wolfgang","roleDisplay":"VerfasserIn","role":"aut"}],"name":{"displayForm":["Thomas Bohnstingl, Franz Scherr, Christian Pehle, Karlheinz Meier and Wolfgang Maass"]},"language":["eng"],"title":[{"title":"Neuromorphic hardware learns to learn","title_sort":"Neuromorphic hardware learns to learn"}],"recId":"1685460569","note":["Gesehen am 13.12.2019"]} | ||
| SRT | |a ORTNERTHOMNEUROMORPH2120 | ||