Matches in SemOpenAlex for { <https://semopenalex.org/work/W3100148727> ?p ?o ?g. }
Showing items 1 to 100 of
100
with 100 items per page.
- W3100148727 endingPage "257" @default.
- W3100148727 startingPage "241" @default.
- W3100148727 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. Increasing speed can be done either during training – e.g., pruning filters – or during inference – e.g., conditionally executing a subset of the layers. We propose a single end-to-end framework that can improve inference efficiency in both settings. We use a combination of batch activation loss and classification loss, and Gumbel reparameterization to learn network structure. We train end-to-end, and the same technique supports pruning as well as conditional computation. We obtain promising experimental results for ImageNet classification with ResNet (45–52% less computation)." @default.
- W3100148727 created "2020-11-23" @default.
- W3100148727 creator A5042850991 @default.
- W3100148727 creator A5060577899 @default.
- W3100148727 creator A5070253419 @default.
- W3100148727 date "2020-01-01" @default.
- W3100148727 modified "2023-10-12" @default.
- W3100148727 title "Channel Selection Using Gumbel Softmax" @default.
- W3100148727 cites W1934410531 @default.
- W3100148727 cites W2108598243 @default.
- W3100148727 cites W2194775991 @default.
- W3100148727 cites W2331143823 @default.
- W3100148727 cites W2474389331 @default.
- W3100148727 cites W2551895583 @default.
- W3100148727 cites W2562731582 @default.
- W3100148727 cites W2884751099 @default.
- W3100148727 cites W2886851211 @default.
- W3100148727 cites W2919115771 @default.
- W3100148727 cites W2928560789 @default.
- W3100148727 cites W2962677625 @default.
- W3100148727 cites W2963055657 @default.
- W3100148727 cites W2963126366 @default.
- W3100148727 cites W2963163009 @default.
- W3100148727 cites W2963363373 @default.
- W3100148727 cites W2963382930 @default.
- W3100148727 cites W2964081807 @default.
- W3100148727 cites W3097096317 @default.
- W3100148727 doi "https://doi.org/10.1007/978-3-030-58583-9_15" @default.
- W3100148727 hasPublicationYear "2020" @default.
- W3100148727 type Work @default.
- W3100148727 sameAs 3100148727 @default.
- W3100148727 citedByCount "10" @default.
- W3100148727 countsByYear W31001487272021 @default.
- W3100148727 countsByYear W31001487272022 @default.
- W3100148727 countsByYear W31001487272023 @default.
- W3100148727 crossrefType "book-chapter" @default.
- W3100148727 hasAuthorship W3100148727A5042850991 @default.
- W3100148727 hasAuthorship W3100148727A5060577899 @default.
- W3100148727 hasAuthorship W3100148727A5070253419 @default.
- W3100148727 hasBestOaLocation W31001487272 @default.
- W3100148727 hasConcept C105795698 @default.
- W3100148727 hasConcept C108010975 @default.
- W3100148727 hasConcept C11413529 @default.
- W3100148727 hasConcept C119857082 @default.
- W3100148727 hasConcept C127162648 @default.
- W3100148727 hasConcept C137610916 @default.
- W3100148727 hasConcept C147581598 @default.
- W3100148727 hasConcept C154945302 @default.
- W3100148727 hasConcept C173608175 @default.
- W3100148727 hasConcept C188441871 @default.
- W3100148727 hasConcept C2776214188 @default.
- W3100148727 hasConcept C31258907 @default.
- W3100148727 hasConcept C33923547 @default.
- W3100148727 hasConcept C41008148 @default.
- W3100148727 hasConcept C45374587 @default.
- W3100148727 hasConcept C50644808 @default.
- W3100148727 hasConcept C6557445 @default.
- W3100148727 hasConcept C68339613 @default.
- W3100148727 hasConcept C81917197 @default.
- W3100148727 hasConcept C86803240 @default.
- W3100148727 hasConceptScore W3100148727C105795698 @default.
- W3100148727 hasConceptScore W3100148727C108010975 @default.
- W3100148727 hasConceptScore W3100148727C11413529 @default.
- W3100148727 hasConceptScore W3100148727C119857082 @default.
- W3100148727 hasConceptScore W3100148727C127162648 @default.
- W3100148727 hasConceptScore W3100148727C137610916 @default.
- W3100148727 hasConceptScore W3100148727C147581598 @default.
- W3100148727 hasConceptScore W3100148727C154945302 @default.
- W3100148727 hasConceptScore W3100148727C173608175 @default.
- W3100148727 hasConceptScore W3100148727C188441871 @default.
- W3100148727 hasConceptScore W3100148727C2776214188 @default.
- W3100148727 hasConceptScore W3100148727C31258907 @default.
- W3100148727 hasConceptScore W3100148727C33923547 @default.
- W3100148727 hasConceptScore W3100148727C41008148 @default.
- W3100148727 hasConceptScore W3100148727C45374587 @default.
- W3100148727 hasConceptScore W3100148727C50644808 @default.
- W3100148727 hasConceptScore W3100148727C6557445 @default.
- W3100148727 hasConceptScore W3100148727C68339613 @default.
- W3100148727 hasConceptScore W3100148727C81917197 @default.
- W3100148727 hasConceptScore W3100148727C86803240 @default.
- W3100148727 hasLocation W31001487271 @default.
- W3100148727 hasLocation W31001487272 @default.
- W3100148727 hasOpenAccess W3100148727 @default.
- W3100148727 hasPrimaryLocation W31001487271 @default.
- W3100148727 hasRelatedWork W2804941026 @default.
- W3100148727 hasRelatedWork W2899244816 @default.
- W3100148727 hasRelatedWork W2901381533 @default.
- W3100148727 hasRelatedWork W2908692794 @default.
- W3100148727 hasRelatedWork W2941817504 @default.
- W3100148727 hasRelatedWork W3098411449 @default.
- W3100148727 hasRelatedWork W3110217300 @default.
- W3100148727 hasRelatedWork W3128179121 @default.
- W3100148727 hasRelatedWork W3207072508 @default.
- W3100148727 hasRelatedWork W4289123518 @default.
- W3100148727 isParatext "false" @default.
- W3100148727 isRetracted "false" @default.
- W3100148727 magId "3100148727" @default.
- W3100148727 workType "book-chapter" @default.