FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3044604993> ?p ?o ?g. }

• W3044604993 abstract "Machine learning on tiny IoT devices based on microcontroller units (MCU) is appealing but challenging: the memory of microcontrollers is 2-3 orders of magnitude less even than mobile phones. We propose MCUNet, a framework that jointly designs the efficient neural architecture (TinyNAS) and the lightweight inference engine (TinyEngine), enabling ImageNet-scale inference on microcontrollers. TinyNAS adopts a two-stage neural architecture search approach that first optimizes the search space to fit the resource constraints, then specializes the network architecture in the optimized search space. TinyNAS can automatically handle diverse constraints (i.e. device, latency, energy, memory) under low search costs. TinyNAS is co-designed with TinyEngine, a memory-efficient inference library to expand the design space and fit a larger model. TinyEngine adapts the memory scheduling according to the overall network topology rather than layer-wise optimization, reducing the memory usage by 2.7x, and accelerating the inference by 1.7-3.3x compared to TF-Lite Micro and CMSIS-NN. MCUNet is the first to achieves >70% ImageNet top1 accuracy on an off-the-shelf commercial microcontroller, using 3.6x less SRAM and 6.6x less Flash compared to quantized MobileNetV2 and ResNet-18. On visual&audio wake words tasks, MCUNet achieves state-of-the-art accuracy and runs 2.4-3.4x faster than MobileNetV2 and ProxylessNAS-based solutions with 2.2-2.6x smaller peak SRAM. Our study suggests that the era of always-on tiny machine learning on IoT devices has arrived." @default.
• W3044604993 created "2020-07-29" @default.
• W3044604993 creator A5012428737 @default.
• W3044604993 creator A5021398655 @default.
• W3044604993 creator A5037897290 @default.
• W3044604993 creator A5040877128 @default.
• W3044604993 creator A5065001783 @default.
• W3044604993 creator A5070926896 @default.
• W3044604993 date "2020-07-20" @default.
• W3044604993 modified "2023-09-27" @default.
• W3044604993 title "MCUNet: Tiny Deep Learning on IoT Devices" @default.
• W3044604993 cites W1724438581 @default.
• W3044604993 cites W2031489346 @default.
• W3044604993 cites W2108598243 @default.
• W3044604993 cites W2131524184 @default.
• W3044604993 cites W2194775991 @default.
• W3044604993 cites W2260663238 @default.
• W3044604993 cites W2300242332 @default.
• W3044604993 cites W2319920447 @default.
• W3044604993 cites W2419597278 @default.
• W3044604993 cites W2469490737 @default.
• W3044604993 cites W2518108298 @default.
• W3044604993 cites W2560017826 @default.
• W3044604993 cites W2570343428 @default.
• W3044604993 cites W2612445135 @default.
• W3044604993 cites W2748428003 @default.
• W3044604993 cites W2752037867 @default.
• W3044604993 cites W2784372305 @default.
• W3044604993 cites W2786771851 @default.
• W3044604993 cites W2796438033 @default.
• W3044604993 cites W2797583228 @default.
• W3044604993 cites W2810075754 @default.
• W3044604993 cites W2883780447 @default.
• W3044604993 cites W2885820039 @default.
• W3044604993 cites W2924888702 @default.
• W3044604993 cites W2932077855 @default.
• W3044604993 cites W2949251082 @default.
• W3044604993 cites W2949941638 @default.
• W3044604993 cites W2950027350 @default.
• W3044604993 cites W2953384591 @default.
• W3044604993 cites W2962988160 @default.
• W3044604993 cites W2963125010 @default.
• W3044604993 cites W2963363373 @default.
• W3044604993 cites W2963374479 @default.
• W3044604993 cites W2963674932 @default.
• W3044604993 cites W2963918968 @default.
• W3044604993 cites W2963960923 @default.
• W3044604993 cites W2964081807 @default.
• W3044604993 cites W2964259004 @default.
• W3044604993 cites W2964299589 @default.
• W3044604993 cites W2967733054 @default.
• W3044604993 cites W2970949810 @default.
• W3044604993 cites W2979711838 @default.
• W3044604993 cites W2981563141 @default.
• W3044604993 cites W2982083293 @default.
• W3044604993 cites W2982479999 @default.
• W3044604993 cites W2994749257 @default.
• W3044604993 cites W2996721521 @default.
• W3044604993 cites W3013186616 @default.
• W3044604993 cites W3016265891 @default.
• W3044604993 cites W3037865064 @default.
• W3044604993 cites W3118608800 @default.
• W3044604993 hasPublicationYear "2020" @default.
• W3044604993 type Work @default.
• W3044604993 sameAs 3044604993 @default.
• W3044604993 citedByCount "14" @default.
• W3044604993 countsByYear W30446049932020 @default.
• W3044604993 countsByYear W30446049932021 @default.
• W3044604993 countsByYear W30446049932022 @default.
• W3044604993 crossrefType "posted-content" @default.
• W3044604993 hasAuthorship W3044604993A5012428737 @default.
• W3044604993 hasAuthorship W3044604993A5021398655 @default.
• W3044604993 hasAuthorship W3044604993A5037897290 @default.
• W3044604993 hasAuthorship W3044604993A5040877128 @default.
• W3044604993 hasAuthorship W3044604993A5065001783 @default.
• W3044604993 hasAuthorship W3044604993A5070926896 @default.
• W3044604993 hasConcept C108583219 @default.
• W3044604993 hasConcept C113775141 @default.
• W3044604993 hasConcept C149635348 @default.
• W3044604993 hasConcept C154945302 @default.
• W3044604993 hasConcept C173018170 @default.
• W3044604993 hasConcept C2776214188 @default.
• W3044604993 hasConcept C41008148 @default.
• W3044604993 hasConcept C50644808 @default.
• W3044604993 hasConcept C68043766 @default.
• W3044604993 hasConcept C9390403 @default.
• W3044604993 hasConceptScore W3044604993C108583219 @default.
• W3044604993 hasConceptScore W3044604993C113775141 @default.
• W3044604993 hasConceptScore W3044604993C149635348 @default.
• W3044604993 hasConceptScore W3044604993C154945302 @default.
• W3044604993 hasConceptScore W3044604993C173018170 @default.
• W3044604993 hasConceptScore W3044604993C2776214188 @default.
• W3044604993 hasConceptScore W3044604993C41008148 @default.
• W3044604993 hasConceptScore W3044604993C50644808 @default.
• W3044604993 hasConceptScore W3044604993C68043766 @default.
• W3044604993 hasConceptScore W3044604993C9390403 @default.
• W3044604993 hasLocation W30446049931 @default.
• W3044604993 hasOpenAccess W3044604993 @default.
• W3044604993 hasPrimaryLocation W30446049931 @default.
• W3044604993 hasRelatedWork W1686810756 @default.

• next