FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3042725081> ?p ?o ?g. }

• W3042725081 abstract "Spiking neural networks (SNNs) are expected to be part of the future AI portfolio, with heavy investment from industry and government, e.g., IBM TrueNorth, Intel Loihi. While Artificial Neural Network (ANN) architectures have taken large strides, few works have targeted SNN hardware efficiency. Our analysis of SNN baselines shows that at modest spike rates, SNN implementations exhibit significantly lower efficiency than accelerators for ANNs. This is primarily because SNN dataflows must consider neuron potentials for several ticks, introducing a new data structure and a new dimension to the reuse pattern. We introduce a novel SNN architecture, SpinalFlow, that processes a compressed, time-stamped, sorted sequence of input spikes. It adopts an ordering of computations such that the outputs of a network layer are also compressed, time-stamped, and sorted. All relevant computations for a neuron are performed in consecutive steps to eliminate neuron potential storage overheads. Thus, with better data reuse, we advance the energy efficiency of SNN accelerators by an order of magnitude. Even though the temporal aspect in SNNs prevents the exploitation of some reuse patterns that are more easily exploited in ANNs, at 4-bit input resolution and 90% input sparsity, SpinalFlow reduces average energy by $1.8 times$, compared to a 4-bit Eyeriss baseline. These improvements are seen for a range of networks and sparsity/resolution levels; SpinalFlow consumes $5 times$ less energy and $5.4 times$ less time than an 8-bit version of Eyeriss. We thus show that, depending on the level of observed sparsity, SNN architectures can be competitive with ANN architectures in terms of latency and energy for inference, thus lowering the barrier for practical deployment in scenarios demanding real-time learning." @default.
• W3042725081 created "2020-07-23" @default.
• W3042725081 creator A5002568331 @default.
• W3042725081 creator A5043417372 @default.
• W3042725081 creator A5055536889 @default.
• W3042725081 creator A5086620434 @default.
• W3042725081 creator A5087056095 @default.
• W3042725081 date "2020-05-01" @default.
• W3042725081 modified "2023-09-30" @default.
• W3042725081 title "SpinalFlow: An Architecture and Dataflow Tailored for Spiking Neural Networks" @default.
• W3042725081 cites W1604973310 @default.
• W3042725081 cites W1682403713 @default.
• W3042725081 cites W1981509416 @default.
• W3042725081 cites W2004277475 @default.
• W3042725081 cites W2009839417 @default.
• W3042725081 cites W2012902942 @default.
• W3042725081 cites W2036963181 @default.
• W3042725081 cites W2082690044 @default.
• W3042725081 cites W2097446068 @default.
• W3042725081 cites W2116360511 @default.
• W3042725081 cites W2121458485 @default.
• W3042725081 cites W2130360162 @default.
• W3042725081 cites W2131763976 @default.
• W3042725081 cites W2237922334 @default.
• W3042725081 cites W2285660444 @default.
• W3042725081 cites W2314470091 @default.
• W3042725081 cites W2330398877 @default.
• W3042725081 cites W2442974303 @default.
• W3042725081 cites W2508602506 @default.
• W3042725081 cites W2518281301 @default.
• W3042725081 cites W2518511512 @default.
• W3042725081 cites W2560647685 @default.
• W3042725081 cites W2606722458 @default.
• W3042725081 cites W2735289987 @default.
• W3042725081 cites W2761132374 @default.
• W3042725081 cites W2783525259 @default.
• W3042725081 cites W2808550672 @default.
• W3042725081 cites W2883012749 @default.
• W3042725081 cites W2884103949 @default.
• W3042725081 cites W2892077605 @default.
• W3042725081 cites W2896895627 @default.
• W3042725081 cites W2964338223 @default.
• W3042725081 cites W4240168186 @default.
• W3042725081 cites W4251155475 @default.
• W3042725081 doi "https://doi.org/10.1109/isca45697.2020.00038" @default.
• W3042725081 hasPublicationYear "2020" @default.
• W3042725081 type Work @default.
• W3042725081 sameAs 3042725081 @default.
• W3042725081 citedByCount "30" @default.
• W3042725081 countsByYear W30427250812021 @default.
• W3042725081 countsByYear W30427250812022 @default.
• W3042725081 countsByYear W30427250812023 @default.
• W3042725081 crossrefType "proceedings-article" @default.
• W3042725081 hasAuthorship W3042725081A5002568331 @default.
• W3042725081 hasAuthorship W3042725081A5043417372 @default.
• W3042725081 hasAuthorship W3042725081A5055536889 @default.
• W3042725081 hasAuthorship W3042725081A5086620434 @default.
• W3042725081 hasAuthorship W3042725081A5087056095 @default.
• W3042725081 hasConcept C113775141 @default.
• W3042725081 hasConcept C11731999 @default.
• W3042725081 hasConcept C119599485 @default.
• W3042725081 hasConcept C127413603 @default.
• W3042725081 hasConcept C154945302 @default.
• W3042725081 hasConcept C173608175 @default.
• W3042725081 hasConcept C18903297 @default.
• W3042725081 hasConcept C206588197 @default.
• W3042725081 hasConcept C2524010 @default.
• W3042725081 hasConcept C2742236 @default.
• W3042725081 hasConcept C2777210771 @default.
• W3042725081 hasConcept C33923547 @default.
• W3042725081 hasConcept C41008148 @default.
• W3042725081 hasConcept C50644808 @default.
• W3042725081 hasConcept C86803240 @default.
• W3042725081 hasConcept C9390403 @default.
• W3042725081 hasConcept C96324660 @default.
• W3042725081 hasConceptScore W3042725081C113775141 @default.
• W3042725081 hasConceptScore W3042725081C11731999 @default.
• W3042725081 hasConceptScore W3042725081C119599485 @default.
• W3042725081 hasConceptScore W3042725081C127413603 @default.
• W3042725081 hasConceptScore W3042725081C154945302 @default.
• W3042725081 hasConceptScore W3042725081C173608175 @default.
• W3042725081 hasConceptScore W3042725081C18903297 @default.
• W3042725081 hasConceptScore W3042725081C206588197 @default.
• W3042725081 hasConceptScore W3042725081C2524010 @default.
• W3042725081 hasConceptScore W3042725081C2742236 @default.
• W3042725081 hasConceptScore W3042725081C2777210771 @default.
• W3042725081 hasConceptScore W3042725081C33923547 @default.
• W3042725081 hasConceptScore W3042725081C41008148 @default.
• W3042725081 hasConceptScore W3042725081C50644808 @default.
• W3042725081 hasConceptScore W3042725081C86803240 @default.
• W3042725081 hasConceptScore W3042725081C9390403 @default.
• W3042725081 hasConceptScore W3042725081C96324660 @default.
• W3042725081 hasLocation W30427250811 @default.
• W3042725081 hasOpenAccess W3042725081 @default.
• W3042725081 hasPrimaryLocation W30427250811 @default.
• W3042725081 hasRelatedWork W1031874 @default.
• W3042725081 hasRelatedWork W10515246 @default.
• W3042725081 hasRelatedWork W11854027 @default.
• W3042725081 hasRelatedWork W14874453 @default.
• W3042725081 hasRelatedWork W3714851 @default.

• next