FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2941817504> ?p ?o ?g. }

• W2941817504 abstract "Important applications such as mobile computing require reducing the computational costs of neural network inference. Ideally, applications would specify their preferred tradeoff between accuracy and speed, and the network would optimize this end-to-end, using classification error to remove parts of the network cite{lecun1990optimal,mozer1989skeletonization,BMVC2016_104}. Increasing speed can be done either during training -- e.g., pruning filters cite{li2016pruning} -- or during inference -- e.g., conditionally executing a subset of the layers cite{aig}. We propose a single end-to-end framework that can improve inference efficiency in both settings. We introduce a batch activation loss and use Gumbel reparameterization to learn network structure cite{aig,jang2016categorical}. We train end-to-end against batch activation loss combined with classification loss, and the same technique supports pruning as well as conditional computation. We obtain promising experimental results for ImageNet classification with ResNet cite{he2016resnet} (45-52% less computation) and MobileNetV2 cite{sandler2018mobilenetv2} (19-37% less computation)." @default.
• W2941817504 created "2019-05-03" @default.
• W2941817504 creator A5042850991 @default.
• W2941817504 creator A5060577899 @default.
• W2941817504 creator A5070253419 @default.
• W2941817504 date "2018-12-11" @default.
• W2941817504 modified "2023-09-27" @default.
• W2941817504 title "An end-to-end approach for speeding up neural network inference" @default.
• W2941817504 cites W1613249581 @default.
• W2941817504 cites W1686810756 @default.
• W2941817504 cites W1934410531 @default.
• W2941817504 cites W2095705004 @default.
• W2941817504 cites W2108598243 @default.
• W2941817504 cites W2114766824 @default.
• W2941817504 cites W2125389748 @default.
• W2941817504 cites W2134273960 @default.
• W2941817504 cites W2194775991 @default.
• W2941817504 cites W2242818861 @default.
• W2941817504 cites W2325237720 @default.
• W2941817504 cites W2331143823 @default.
• W2941817504 cites W2474389331 @default.
• W2941817504 cites W2495425901 @default.
• W2941817504 cites W2547875792 @default.
• W2941817504 cites W2551895583 @default.
• W2941817504 cites W2562731582 @default.
• W2941817504 cites W2598097916 @default.
• W2941817504 cites W2784822746 @default.
• W2941817504 cites W2797569913 @default.
• W2941817504 cites W2884751099 @default.
• W2941817504 cites W2886851211 @default.
• W2941817504 cites W2899771611 @default.
• W2941817504 cites W2919115771 @default.
• W2941817504 cites W2951569836 @default.
• W2941817504 cites W2951595529 @default.
• W2941817504 cites W2952339051 @default.
• W2941817504 cites W2962677625 @default.
• W2941817504 cites W2963055657 @default.
• W2941817504 cites W2963114950 @default.
• W2941817504 cites W2963163009 @default.
• W2941817504 cites W2963363373 @default.
• W2941817504 cites W2963382930 @default.
• W2941817504 cites W2963446712 @default.
• W2941817504 cites W2963674932 @default.
• W2941817504 cites W2963828549 @default.
• W2941817504 cites W2964081807 @default.
• W2941817504 cites W2964233199 @default.
• W2941817504 cites W2964259004 @default.
• W2941817504 cites W2964299589 @default.
• W2941817504 cites W3037590790 @default.
• W2941817504 cites W3097096317 @default.
• W2941817504 hasPublicationYear "2018" @default.
• W2941817504 type Work @default.
• W2941817504 sameAs 2941817504 @default.
• W2941817504 citedByCount "1" @default.
• W2941817504 countsByYear W29418175042020 @default.
• W2941817504 crossrefType "posted-content" @default.
• W2941817504 hasAuthorship W2941817504A5042850991 @default.
• W2941817504 hasAuthorship W2941817504A5060577899 @default.
• W2941817504 hasAuthorship W2941817504A5070253419 @default.
• W2941817504 hasConcept C108010975 @default.
• W2941817504 hasConcept C11413529 @default.
• W2941817504 hasConcept C119857082 @default.
• W2941817504 hasConcept C154945302 @default.
• W2941817504 hasConcept C2776214188 @default.
• W2941817504 hasConcept C41008148 @default.
• W2941817504 hasConcept C45374587 @default.
• W2941817504 hasConcept C50644808 @default.
• W2941817504 hasConcept C6557445 @default.
• W2941817504 hasConcept C74296488 @default.
• W2941817504 hasConcept C86803240 @default.
• W2941817504 hasConceptScore W2941817504C108010975 @default.
• W2941817504 hasConceptScore W2941817504C11413529 @default.
• W2941817504 hasConceptScore W2941817504C119857082 @default.
• W2941817504 hasConceptScore W2941817504C154945302 @default.
• W2941817504 hasConceptScore W2941817504C2776214188 @default.
• W2941817504 hasConceptScore W2941817504C41008148 @default.
• W2941817504 hasConceptScore W2941817504C45374587 @default.
• W2941817504 hasConceptScore W2941817504C50644808 @default.
• W2941817504 hasConceptScore W2941817504C6557445 @default.
• W2941817504 hasConceptScore W2941817504C74296488 @default.
• W2941817504 hasConceptScore W2941817504C86803240 @default.
• W2941817504 hasLocation W29418175041 @default.
• W2941817504 hasOpenAccess W2941817504 @default.
• W2941817504 hasPrimaryLocation W29418175041 @default.
• W2941817504 hasRelatedWork W2771111398 @default.
• W2941817504 hasRelatedWork W2785668315 @default.
• W2941817504 hasRelatedWork W2807434070 @default.
• W2941817504 hasRelatedWork W2893585013 @default.
• W2941817504 hasRelatedWork W2901523331 @default.
• W2941817504 hasRelatedWork W2924888702 @default.
• W2941817504 hasRelatedWork W2944859147 @default.
• W2941817504 hasRelatedWork W2963287528 @default.
• W2941817504 hasRelatedWork W2990873889 @default.
• W2941817504 hasRelatedWork W2995428361 @default.
• W2941817504 hasRelatedWork W3001642847 @default.
• W2941817504 hasRelatedWork W3005908409 @default.
• W2941817504 hasRelatedWork W3048959830 @default.
• W2941817504 hasRelatedWork W3102230092 @default.
• W2941817504 hasRelatedWork W3102982004 @default.
• W2941817504 hasRelatedWork W3110217300 @default.

• next