Matches in SemOpenAlex for { <https://semopenalex.org/work/W4302023003> ?p ?o ?g. }
Showing items 1 to 91 of
91
with 100 items per page.
- W4302023003 endingPage "044016" @default.
- W4302023003 startingPage "044016" @default.
- W4302023003 abstract "Abstract Spiking neural networks (SNNs) underlie low-power, fault-tolerant information processing in the brain and could constitute a power-efficient alternative to conventional deep neural networks when implemented on suitable neuromorphic hardware accelerators. However, instantiating SNNs that solve complex computational tasks in-silico remains a significant challenge. Surrogate gradient (SG) techniques have emerged as a standard solution for training SNNs end-to-end. Still, their success depends on synaptic weight initialization, similar to conventional artificial neural networks (ANNs). Yet, unlike in the case of ANNs, it remains elusive what constitutes a good initial state for an SNN. Here, we develop a general initialization strategy for SNNs inspired by the fluctuation-driven regime commonly observed in the brain. Specifically, we derive practical solutions for data-dependent weight initialization that ensure fluctuation-driven firing in the widely used leaky integrate-and-fire neurons. We empirically show that SNNs initialized following our strategy exhibit superior learning performance when trained with SGs. These findings generalize across several datasets and SNN architectures, including fully connected, deep convolutional, recurrent, and more biologically plausible SNNs obeying Dale’s law. Thus fluctuation-driven initialization provides a practical, versatile, and easy-to-implement strategy for improving SNN training performance on diverse tasks in neuromorphic engineering and computational neuroscience." @default.
- W4302023003 created "2022-10-06" @default.
- W4302023003 creator A5001875800 @default.
- W4302023003 creator A5026476666 @default.
- W4302023003 creator A5089883669 @default.
- W4302023003 date "2022-12-01" @default.
- W4302023003 modified "2023-10-01" @default.
- W4302023003 title "Fluctuation-driven initialization for spiking neural network training" @default.
- W4302023003 cites W1582051163 @default.
- W4302023003 cites W1970109917 @default.
- W4302023003 cites W2008008156 @default.
- W4302023003 cites W2015701831 @default.
- W4302023003 cites W2020676607 @default.
- W4302023003 cites W2032776279 @default.
- W4302023003 cites W2042076346 @default.
- W4302023003 cites W2064675550 @default.
- W4302023003 cites W2073929966 @default.
- W4302023003 cites W2079540311 @default.
- W4302023003 cites W2113345172 @default.
- W4302023003 cites W2123745704 @default.
- W4302023003 cites W2156640153 @default.
- W4302023003 cites W2160239507 @default.
- W4302023003 cites W2333833826 @default.
- W4302023003 cites W2513853720 @default.
- W4302023003 cites W2569813014 @default.
- W4302023003 cites W2578308616 @default.
- W4302023003 cites W2897580718 @default.
- W4302023003 cites W2951065015 @default.
- W4302023003 cites W2951784885 @default.
- W4302023003 cites W2962804204 @default.
- W4302023003 cites W2963335874 @default.
- W4302023003 cites W2978368159 @default.
- W4302023003 cites W2980513133 @default.
- W4302023003 cites W2984844508 @default.
- W4302023003 cites W3016391357 @default.
- W4302023003 cites W3102087395 @default.
- W4302023003 cites W3104409553 @default.
- W4302023003 cites W3124478039 @default.
- W4302023003 cites W3147709328 @default.
- W4302023003 cites W3206817059 @default.
- W4302023003 cites W4221157818 @default.
- W4302023003 doi "https://doi.org/10.1088/2634-4386/ac97bb" @default.
- W4302023003 hasPublicationYear "2022" @default.
- W4302023003 type Work @default.
- W4302023003 citedByCount "2" @default.
- W4302023003 countsByYear W43020230032023 @default.
- W4302023003 crossrefType "journal-article" @default.
- W4302023003 hasAuthorship W4302023003A5001875800 @default.
- W4302023003 hasAuthorship W4302023003A5026476666 @default.
- W4302023003 hasAuthorship W4302023003A5089883669 @default.
- W4302023003 hasBestOaLocation W43020230031 @default.
- W4302023003 hasConcept C114466953 @default.
- W4302023003 hasConcept C11731999 @default.
- W4302023003 hasConcept C119857082 @default.
- W4302023003 hasConcept C151927369 @default.
- W4302023003 hasConcept C154945302 @default.
- W4302023003 hasConcept C199360897 @default.
- W4302023003 hasConcept C41008148 @default.
- W4302023003 hasConcept C50644808 @default.
- W4302023003 hasConceptScore W4302023003C114466953 @default.
- W4302023003 hasConceptScore W4302023003C11731999 @default.
- W4302023003 hasConceptScore W4302023003C119857082 @default.
- W4302023003 hasConceptScore W4302023003C151927369 @default.
- W4302023003 hasConceptScore W4302023003C154945302 @default.
- W4302023003 hasConceptScore W4302023003C199360897 @default.
- W4302023003 hasConceptScore W4302023003C41008148 @default.
- W4302023003 hasConceptScore W4302023003C50644808 @default.
- W4302023003 hasFunder F4320320924 @default.
- W4302023003 hasFunder F4320322884 @default.
- W4302023003 hasIssue "4" @default.
- W4302023003 hasLocation W43020230031 @default.
- W4302023003 hasLocation W43020230032 @default.
- W4302023003 hasLocation W43020230033 @default.
- W4302023003 hasOpenAccess W4302023003 @default.
- W4302023003 hasPrimaryLocation W43020230031 @default.
- W4302023003 hasRelatedWork W2756276189 @default.
- W4302023003 hasRelatedWork W2948197126 @default.
- W4302023003 hasRelatedWork W3033176951 @default.
- W4302023003 hasRelatedWork W3117246098 @default.
- W4302023003 hasRelatedWork W3166663443 @default.
- W4302023003 hasRelatedWork W3211502403 @default.
- W4302023003 hasRelatedWork W4285069098 @default.
- W4302023003 hasRelatedWork W4287126156 @default.
- W4302023003 hasRelatedWork W4302023003 @default.
- W4302023003 hasRelatedWork W4283331723 @default.
- W4302023003 hasVolume "2" @default.
- W4302023003 isParatext "false" @default.
- W4302023003 isRetracted "false" @default.
- W4302023003 workType "article" @default.