Matches in SemOpenAlex for { <https://semopenalex.org/work/W4308479822> ?p ?o ?g. }
Showing items 1 to 63 of
63
with 100 items per page.
- W4308479822 endingPage "4526" @default.
- W4308479822 startingPage "4516" @default.
- W4308479822 abstract "Random forests (RFs) use a collection of decision trees (DTs) to perform the classification or regression. RFs are adopted in a wide variety of machine learning (ML) applications, and they are finding increasing use also in scenarios at the extreme edge of the Internet of Things (TinyML) where memory constraints are particularly tight. This article addresses the optimization of the computational and storage costs for running DTs on the microcontroller units (MCUs) typically deployed in TinyML scenarios. We introduce three alternative DT kernels optimized for memory- and compute-limited MCUs, providing insight into the key memory-latency tradeoffs on an open-source RISC-V platform. We identify key bottlenecks and demonstrate that SW optimizations enable up to significant memory footprint and latency decrease. Experimental results show that the optimized kernels achieve up to 4.5 <inline-formula xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink> <tex-math notation=LaTeX>$mu text{s}$ </tex-math></inline-formula> latency, <inline-formula xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink> <tex-math notation=LaTeX>$4.8times $ </tex-math></inline-formula> speedup, and 45% storage reduction against the widely-adopted naive DT design. We carry out a detailed performance and energy cost analysis of various optimized DT variants: the best approach requires just 8 instructions and 0.155 pJ per decision." @default.
- W4308479822 created "2022-11-12" @default.
- W4308479822 creator A5004842257 @default.
- W4308479822 creator A5017136532 @default.
- W4308479822 creator A5036407549 @default.
- W4308479822 date "2022-11-01" @default.
- W4308479822 modified "2023-10-18" @default.
- W4308479822 title "Optimizing Random Forest-Based Inference on RISC-V MCUs at the Extreme Edge" @default.
- W4308479822 doi "https://doi.org/10.1109/tcad.2022.3199903" @default.
- W4308479822 hasPublicationYear "2022" @default.
- W4308479822 type Work @default.
- W4308479822 citedByCount "2" @default.
- W4308479822 countsByYear W43084798222023 @default.
- W4308479822 crossrefType "journal-article" @default.
- W4308479822 hasAuthorship W4308479822A5004842257 @default.
- W4308479822 hasAuthorship W4308479822A5017136532 @default.
- W4308479822 hasAuthorship W4308479822A5036407549 @default.
- W4308479822 hasConcept C111919701 @default.
- W4308479822 hasConcept C154945302 @default.
- W4308479822 hasConcept C162307627 @default.
- W4308479822 hasConcept C173608175 @default.
- W4308479822 hasConcept C199360897 @default.
- W4308479822 hasConcept C26517878 @default.
- W4308479822 hasConcept C2776214188 @default.
- W4308479822 hasConcept C41008148 @default.
- W4308479822 hasConcept C48145219 @default.
- W4308479822 hasConcept C68339613 @default.
- W4308479822 hasConcept C74912251 @default.
- W4308479822 hasConcept C76155785 @default.
- W4308479822 hasConcept C82876162 @default.
- W4308479822 hasConceptScore W4308479822C111919701 @default.
- W4308479822 hasConceptScore W4308479822C154945302 @default.
- W4308479822 hasConceptScore W4308479822C162307627 @default.
- W4308479822 hasConceptScore W4308479822C173608175 @default.
- W4308479822 hasConceptScore W4308479822C199360897 @default.
- W4308479822 hasConceptScore W4308479822C26517878 @default.
- W4308479822 hasConceptScore W4308479822C2776214188 @default.
- W4308479822 hasConceptScore W4308479822C41008148 @default.
- W4308479822 hasConceptScore W4308479822C48145219 @default.
- W4308479822 hasConceptScore W4308479822C68339613 @default.
- W4308479822 hasConceptScore W4308479822C74912251 @default.
- W4308479822 hasConceptScore W4308479822C76155785 @default.
- W4308479822 hasConceptScore W4308479822C82876162 @default.
- W4308479822 hasIssue "11" @default.
- W4308479822 hasLocation W43084798221 @default.
- W4308479822 hasOpenAccess W4308479822 @default.
- W4308479822 hasPrimaryLocation W43084798221 @default.
- W4308479822 hasRelatedWork W2981519481 @default.
- W4308479822 hasRelatedWork W3093101573 @default.
- W4308479822 hasRelatedWork W3131663603 @default.
- W4308479822 hasRelatedWork W3162668736 @default.
- W4308479822 hasRelatedWork W3168991979 @default.
- W4308479822 hasRelatedWork W3208397797 @default.
- W4308479822 hasRelatedWork W4213074318 @default.
- W4308479822 hasRelatedWork W4286233754 @default.
- W4308479822 hasRelatedWork W4318348488 @default.
- W4308479822 hasRelatedWork W4319995060 @default.
- W4308479822 hasVolume "41" @default.
- W4308479822 isParatext "false" @default.
- W4308479822 isRetracted "false" @default.
- W4308479822 workType "article" @default.