FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3177366646> ?p ?o ?g. }

• W3177366646 endingPage "686" @default.
• W3177366646 startingPage "675" @default.
• W3177366646 abstract "Although the quest for more accurate solutions is pushing deep learning research towards larger and more complex algorithms, edge devices demand efficient inference and therefore reduction in model size, latency and energy consumption. One technique to limit model size is quantization, which implies using fewer bits to represent weights and biases. Such an approach usually results in a decline in performance. Here, we introduce a method for designing optimally heterogeneously quantized versions of deep neural network models for minimum-energy, high-accuracy, nanosecond inference and fully automated deployment on chip. With a per-layer, per-parameter type automatic quantization procedure, sampling from a wide range of quantizers, model energy consumption and size are minimized while high accuracy is maintained. This is crucial for the event selection procedure in proton–proton collisions at the CERN Large Hadron Collider, where resources are strictly limited and a latency of $${mathcal{O}}(1),upmu{rm{s}}$$ is required. Nanosecond inference and a resource consumption reduced by a factor of 50 when implemented on field-programmable gate array hardware are achieved. With edge computing on custom hardware, real-time inference with deep neural networks can reach the nanosecond timescale. An important application in this regime is event processing at particle collision detectors like those at the Large Hadron Collider (LHC). To ensure high performance as well as reduced resource consumption, a method is developed, and made available as an extension of the Keras library, to automatically design optimal quantization of the different layers in a deep neural network." @default.
• W3177366646 created "2021-07-05" @default.
• W3177366646 creator A5000947451 @default.
• W3177366646 creator A5004366207 @default.
• W3177366646 creator A5007162189 @default.
• W3177366646 creator A5013207401 @default.
• W3177366646 creator A5032688100 @default.
• W3177366646 creator A5039731897 @default.
• W3177366646 creator A5048866796 @default.
• W3177366646 creator A5061275953 @default.
• W3177366646 creator A5061839726 @default.
• W3177366646 creator A5084994817 @default.
• W3177366646 date "2021-06-21" @default.
• W3177366646 modified "2023-10-17" @default.
• W3177366646 title "Automatic heterogeneous quantization of deep neural networks for low-latency inference on the edge for particle detectors" @default.
• W3177366646 cites W1999085092 @default.
• W3177366646 cites W2013305145 @default.
• W3177366646 cites W2102596689 @default.
• W3177366646 cites W2194775991 @default.
• W3177366646 cites W2300242332 @default.
• W3177366646 cites W2513568085 @default.
• W3177366646 cites W2525740295 @default.
• W3177366646 cites W2584551591 @default.
• W3177366646 cites W2727238169 @default.
• W3177366646 cites W2759398875 @default.
• W3177366646 cites W2883780447 @default.
• W3177366646 cites W2884150179 @default.
• W3177366646 cites W2889499839 @default.
• W3177366646 cites W2891946740 @default.
• W3177366646 cites W2895432151 @default.
• W3177366646 cites W2907463061 @default.
• W3177366646 cites W2963029056 @default.
• W3177366646 cites W2963163009 @default.
• W3177366646 cites W2963363373 @default.
• W3177366646 cites W2963521187 @default.
• W3177366646 cites W2963568120 @default.
• W3177366646 cites W2981751377 @default.
• W3177366646 cites W2982041622 @default.
• W3177366646 cites W2982083293 @default.
• W3177366646 cites W2982479999 @default.
• W3177366646 cites W2999066514 @default.
• W3177366646 cites W3002851826 @default.
• W3177366646 cites W3093982621 @default.
• W3177366646 cites W3101493857 @default.
• W3177366646 cites W3102169921 @default.
• W3177366646 cites W3104395752 @default.
• W3177366646 cites W3119613749 @default.
• W3177366646 cites W3175799302 @default.
• W3177366646 cites W3210266748 @default.
• W3177366646 cites W4242577057 @default.
• W3177366646 cites W4245844998 @default.
• W3177366646 doi "https://doi.org/10.1038/s42256-021-00356-5" @default.
• W3177366646 hasPublicationYear "2021" @default.
• W3177366646 type Work @default.
• W3177366646 sameAs 3177366646 @default.
• W3177366646 citedByCount "67" @default.
• W3177366646 countsByYear W31773666462020 @default.
• W3177366646 countsByYear W31773666462021 @default.
• W3177366646 countsByYear W31773666462022 @default.
• W3177366646 countsByYear W31773666462023 @default.
• W3177366646 crossrefType "journal-article" @default.
• W3177366646 hasAuthorship W3177366646A5000947451 @default.
• W3177366646 hasAuthorship W3177366646A5004366207 @default.
• W3177366646 hasAuthorship W3177366646A5007162189 @default.
• W3177366646 hasAuthorship W3177366646A5013207401 @default.
• W3177366646 hasAuthorship W3177366646A5032688100 @default.
• W3177366646 hasAuthorship W3177366646A5039731897 @default.
• W3177366646 hasAuthorship W3177366646A5048866796 @default.
• W3177366646 hasAuthorship W3177366646A5061275953 @default.
• W3177366646 hasAuthorship W3177366646A5061839726 @default.
• W3177366646 hasAuthorship W3177366646A5084994817 @default.
• W3177366646 hasBestOaLocation W31773666462 @default.
• W3177366646 hasConcept C109214941 @default.
• W3177366646 hasConcept C111919701 @default.
• W3177366646 hasConcept C113775141 @default.
• W3177366646 hasConcept C11413529 @default.
• W3177366646 hasConcept C119599485 @default.
• W3177366646 hasConcept C121332964 @default.
• W3177366646 hasConcept C127413603 @default.
• W3177366646 hasConcept C138236772 @default.
• W3177366646 hasConcept C154945302 @default.
• W3177366646 hasConcept C2776214188 @default.
• W3177366646 hasConcept C2780165032 @default.
• W3177366646 hasConcept C28855332 @default.
• W3177366646 hasConcept C41008148 @default.
• W3177366646 hasConcept C50644808 @default.
• W3177366646 hasConcept C76155785 @default.
• W3177366646 hasConcept C79974875 @default.
• W3177366646 hasConcept C82876162 @default.
• W3177366646 hasConcept C87668248 @default.
• W3177366646 hasConcept C94915269 @default.
• W3177366646 hasConceptScore W3177366646C109214941 @default.
• W3177366646 hasConceptScore W3177366646C111919701 @default.
• W3177366646 hasConceptScore W3177366646C113775141 @default.
• W3177366646 hasConceptScore W3177366646C11413529 @default.
• W3177366646 hasConceptScore W3177366646C119599485 @default.
• W3177366646 hasConceptScore W3177366646C121332964 @default.
• W3177366646 hasConceptScore W3177366646C127413603 @default.

• next