Matches in SemOpenAlex for { <https://semopenalex.org/work/W4311839577> ?p ?o ?g. }
- W4311839577 abstract "High Energy Physics studies the fundamental forces and elementary particles of the Universe. With the unprecedented scale of experiments comes the challenge of accurate, ultra-low latency decision-making. Transformer Neural Networks (TNNs) have been proven to accomplish cutting-edge accuracy in classification for hadronic jet tagging. Nevertheless, software-centered solutions targeting CPUs and GPUs lack the inference speed required for real-time particle triggers, most notably those at the CERN Large Hadron Collider. This paper proposes a novel TNN-based architecture, efficiently mapped to Field-Programmable Gate Arrays, that outperforms GPU inference capabilities involving state-of-the-art neural network models by approximately 1000 times while preserving comparable classification accuracy. The design offers high customizability and aims to bridge the gap between hardware and software development by using High-Level Synthesis. Moreover, we propose a novel model-independent post-training quantization search algorithm that works in general hardware environments according to user-defined constraints. Experimental evaluation yields a 64% reduction in overall bit-widths with a 2% accuracy loss." @default.
- W4311839577 created "2022-12-31" @default.
- W4311839577 creator A5029466194 @default.
- W4311839577 creator A5038418461 @default.
- W4311839577 creator A5054475218 @default.
- W4311839577 creator A5057940557 @default.
- W4311839577 date "2022-12-05" @default.
- W4311839577 modified "2023-09-27" @default.
- W4311839577 title "Accelerating Transformer Neural Networks on FPGAs for High Energy Physics Experiments" @default.
- W4311839577 cites W2046076654 @default.
- W4311839577 cites W2219260334 @default.
- W4311839577 cites W2939653360 @default.
- W4311839577 cites W2962834855 @default.
- W4311839577 cites W3002842489 @default.
- W4311839577 cites W3002851826 @default.
- W4311839577 cites W3047848469 @default.
- W4311839577 cites W3098163999 @default.
- W4311839577 cites W3101191453 @default.
- W4311839577 cites W3101493857 @default.
- W4311839577 cites W3104395752 @default.
- W4311839577 cites W3104674222 @default.
- W4311839577 cites W3119613749 @default.
- W4311839577 cites W3121614399 @default.
- W4311839577 cites W3135348438 @default.
- W4311839577 cites W3162542754 @default.
- W4311839577 cites W3176081321 @default.
- W4311839577 cites W3176468986 @default.
- W4311839577 cites W3177366646 @default.
- W4311839577 cites W3199934250 @default.
- W4311839577 cites W4286001027 @default.
- W4311839577 cites W4289656095 @default.
- W4311839577 cites W4291653336 @default.
- W4311839577 cites W4293023328 @default.
- W4311839577 cites W4296473245 @default.
- W4311839577 cites W4308083513 @default.
- W4311839577 cites W4308479898 @default.
- W4311839577 cites W4312707063 @default.
- W4311839577 doi "https://doi.org/10.1109/icfpt56656.2022.9974463" @default.
- W4311839577 hasPublicationYear "2022" @default.
- W4311839577 type Work @default.
- W4311839577 citedByCount "0" @default.
- W4311839577 crossrefType "proceedings-article" @default.
- W4311839577 hasAuthorship W4311839577A5029466194 @default.
- W4311839577 hasAuthorship W4311839577A5038418461 @default.
- W4311839577 hasAuthorship W4311839577A5054475218 @default.
- W4311839577 hasAuthorship W4311839577A5057940557 @default.
- W4311839577 hasConcept C109214941 @default.
- W4311839577 hasConcept C113775141 @default.
- W4311839577 hasConcept C11413529 @default.
- W4311839577 hasConcept C118524514 @default.
- W4311839577 hasConcept C119599485 @default.
- W4311839577 hasConcept C121332964 @default.
- W4311839577 hasConcept C127413603 @default.
- W4311839577 hasConcept C154945302 @default.
- W4311839577 hasConcept C165801399 @default.
- W4311839577 hasConcept C199360897 @default.
- W4311839577 hasConcept C2776214188 @default.
- W4311839577 hasConcept C2777904410 @default.
- W4311839577 hasConcept C28855332 @default.
- W4311839577 hasConcept C41008148 @default.
- W4311839577 hasConcept C42935608 @default.
- W4311839577 hasConcept C50644808 @default.
- W4311839577 hasConcept C66322947 @default.
- W4311839577 hasConcept C87668248 @default.
- W4311839577 hasConcept C9390403 @default.
- W4311839577 hasConceptScore W4311839577C109214941 @default.
- W4311839577 hasConceptScore W4311839577C113775141 @default.
- W4311839577 hasConceptScore W4311839577C11413529 @default.
- W4311839577 hasConceptScore W4311839577C118524514 @default.
- W4311839577 hasConceptScore W4311839577C119599485 @default.
- W4311839577 hasConceptScore W4311839577C121332964 @default.
- W4311839577 hasConceptScore W4311839577C127413603 @default.
- W4311839577 hasConceptScore W4311839577C154945302 @default.
- W4311839577 hasConceptScore W4311839577C165801399 @default.
- W4311839577 hasConceptScore W4311839577C199360897 @default.
- W4311839577 hasConceptScore W4311839577C2776214188 @default.
- W4311839577 hasConceptScore W4311839577C2777904410 @default.
- W4311839577 hasConceptScore W4311839577C28855332 @default.
- W4311839577 hasConceptScore W4311839577C41008148 @default.
- W4311839577 hasConceptScore W4311839577C42935608 @default.
- W4311839577 hasConceptScore W4311839577C50644808 @default.
- W4311839577 hasConceptScore W4311839577C66322947 @default.
- W4311839577 hasConceptScore W4311839577C87668248 @default.
- W4311839577 hasConceptScore W4311839577C9390403 @default.
- W4311839577 hasFunder F4320334627 @default.
- W4311839577 hasLocation W43118395771 @default.
- W4311839577 hasOpenAccess W4311839577 @default.
- W4311839577 hasPrimaryLocation W43118395771 @default.
- W4311839577 hasRelatedWork W1995921138 @default.
- W4311839577 hasRelatedWork W2010122688 @default.
- W4311839577 hasRelatedWork W2063534976 @default.
- W4311839577 hasRelatedWork W2748952813 @default.
- W4311839577 hasRelatedWork W2777406049 @default.
- W4311839577 hasRelatedWork W2979314664 @default.
- W4311839577 hasRelatedWork W2995926156 @default.
- W4311839577 hasRelatedWork W3107443367 @default.
- W4311839577 hasRelatedWork W3149104646 @default.
- W4311839577 hasRelatedWork W3207688365 @default.
- W4311839577 isParatext "false" @default.
- W4311839577 isRetracted "false" @default.