Matches in SemOpenAlex for { <https://semopenalex.org/work/W3176724974> ?p ?o ?g. }
- W3176724974 abstract "Large machine learning models achieve unprecedented performance on various tasks and have evolved as the go-to technique. However, deploying these compute and memory hungry models on resource constraint environments poses new challenges. In this work, we propose mathematically provable Representer Sketch, a concise set of count arrays that can approximate the inference procedure with simple hashing computations and aggregations. Representer Sketch builds upon the popular Representer Theorem from kernel literature, hence the name, providing a generic fundamental alternative to the problem of efficient inference that goes beyond the popular approach such as quantization, iterative pruning and knowledge distillation. A neural network function is transformed to its weighted kernel density representation, which can be very efficiently estimated with our sketching algorithm. Empirically, we show that Representer Sketch achieves up to 114x reduction in storage requirement and 59x reduction in computation complexity without any drop in accuracy." @default.
- W3176724974 created "2021-07-05" @default.
- W3176724974 creator A5024993683 @default.
- W3176724974 creator A5086798890 @default.
- W3176724974 creator A5091829195 @default.
- W3176724974 date "2021-06-21" @default.
- W3176724974 modified "2023-09-25" @default.
- W3176724974 title "Efficient Inference via Universal LSH Kernel." @default.
- W3176724974 cites W1499864241 @default.
- W3176724974 cites W1540155273 @default.
- W3176724974 cites W1821462560 @default.
- W3176724974 cites W1999085092 @default.
- W3176724974 cites W2037757210 @default.
- W3176724974 cites W2080745194 @default.
- W3176724974 cites W2114766824 @default.
- W3176724974 cites W2147717514 @default.
- W3176724974 cites W2153635508 @default.
- W3176724974 cites W2161591461 @default.
- W3176724974 cites W2162006472 @default.
- W3176724974 cites W2163605009 @default.
- W3176724974 cites W2194775991 @default.
- W3176724974 cites W2293634267 @default.
- W3176724974 cites W2319920447 @default.
- W3176724974 cites W2524428287 @default.
- W3176724974 cites W2604086375 @default.
- W3176724974 cites W2604653570 @default.
- W3176724974 cites W2626778328 @default.
- W3176724974 cites W2776585113 @default.
- W3176724974 cites W2782812883 @default.
- W3176724974 cites W2788362551 @default.
- W3176724974 cites W2805454539 @default.
- W3176724974 cites W2883070812 @default.
- W3176724974 cites W2905810301 @default.
- W3176724974 cites W2913243980 @default.
- W3176724974 cites W2958537367 @default.
- W3176724974 cites W2963674932 @default.
- W3176724974 cites W2964299589 @default.
- W3176724974 cites W2978017171 @default.
- W3176724974 cites W3016019667 @default.
- W3176724974 cites W3036957384 @default.
- W3176724974 cites W3102335301 @default.
- W3176724974 cites W3115029918 @default.
- W3176724974 cites W3205138306 @default.
- W3176724974 cites W587794757 @default.
- W3176724974 hasPublicationYear "2021" @default.
- W3176724974 type Work @default.
- W3176724974 sameAs 3176724974 @default.
- W3176724974 citedByCount "0" @default.
- W3176724974 crossrefType "posted-content" @default.
- W3176724974 hasAuthorship W3176724974A5024993683 @default.
- W3176724974 hasAuthorship W3176724974A5086798890 @default.
- W3176724974 hasAuthorship W3176724974A5091829195 @default.
- W3176724974 hasConcept C108010975 @default.
- W3176724974 hasConcept C11413529 @default.
- W3176724974 hasConcept C114614502 @default.
- W3176724974 hasConcept C119857082 @default.
- W3176724974 hasConcept C154945302 @default.
- W3176724974 hasConcept C2524010 @default.
- W3176724974 hasConcept C2776036281 @default.
- W3176724974 hasConcept C2776214188 @default.
- W3176724974 hasConcept C2779231336 @default.
- W3176724974 hasConcept C33923547 @default.
- W3176724974 hasConcept C41008148 @default.
- W3176724974 hasConcept C45374587 @default.
- W3176724974 hasConcept C6557445 @default.
- W3176724974 hasConcept C74193536 @default.
- W3176724974 hasConcept C80444323 @default.
- W3176724974 hasConcept C86803240 @default.
- W3176724974 hasConceptScore W3176724974C108010975 @default.
- W3176724974 hasConceptScore W3176724974C11413529 @default.
- W3176724974 hasConceptScore W3176724974C114614502 @default.
- W3176724974 hasConceptScore W3176724974C119857082 @default.
- W3176724974 hasConceptScore W3176724974C154945302 @default.
- W3176724974 hasConceptScore W3176724974C2524010 @default.
- W3176724974 hasConceptScore W3176724974C2776036281 @default.
- W3176724974 hasConceptScore W3176724974C2776214188 @default.
- W3176724974 hasConceptScore W3176724974C2779231336 @default.
- W3176724974 hasConceptScore W3176724974C33923547 @default.
- W3176724974 hasConceptScore W3176724974C41008148 @default.
- W3176724974 hasConceptScore W3176724974C45374587 @default.
- W3176724974 hasConceptScore W3176724974C6557445 @default.
- W3176724974 hasConceptScore W3176724974C74193536 @default.
- W3176724974 hasConceptScore W3176724974C80444323 @default.
- W3176724974 hasConceptScore W3176724974C86803240 @default.
- W3176724974 hasLocation W31767249741 @default.
- W3176724974 hasOpenAccess W3176724974 @default.
- W3176724974 hasPrimaryLocation W31767249741 @default.
- W3176724974 hasRelatedWork W15705661 @default.
- W3176724974 hasRelatedWork W1621301074 @default.
- W3176724974 hasRelatedWork W174961330 @default.
- W3176724974 hasRelatedWork W2433350887 @default.
- W3176724974 hasRelatedWork W2468625129 @default.
- W3176724974 hasRelatedWork W2563338646 @default.
- W3176724974 hasRelatedWork W2572411918 @default.
- W3176724974 hasRelatedWork W2573350151 @default.
- W3176724974 hasRelatedWork W2765391880 @default.
- W3176724974 hasRelatedWork W2784713518 @default.
- W3176724974 hasRelatedWork W2794404674 @default.
- W3176724974 hasRelatedWork W2805880533 @default.
- W3176724974 hasRelatedWork W2900042804 @default.