FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2905102070> ?p ?o ?g. }

• W2905102070 endingPage "3336" @default.
• W2905102070 startingPage "3329" @default.
• W2905102070 abstract "The advancement of deep models poses great challenges to real-world deployment because of the limited computational ability and storage space on edge devices. To solve this problem, existing works have made progress to prune or quantize deep models. However, most existing methods rely heavily on a supervised training process to achieve satisfactory performance, acquiring large amount of labeled training data, which may not be practical for real deployment. In this paper, we propose a novel layer-wise quantization method for deep neural networks, which only requires limited training data (1% of original dataset). Specifically, we formulate parameters quantization for each layer as a discrete optimization problem, and solve it using Alternative Direction Method of Multipliers (ADMM), which gives an efficient closed-form solution. We prove that the final performance drop after quantization is bounded by a linear combination of the reconstructed errors caused at each layer. Based on the proved theorem, we propose an algorithm to quantize a deep neural network layer by layer with an additional weights update step to minimize the final error. Extensive experiments on benchmark deep models are conducted to demonstrate the effectiveness of our proposed method using 1% of CIFAR10 and ImageNet datasets. Codes are available in: https://github.com/csyhhu/L-DNQ" @default.
• W2905102070 created "2018-12-22" @default.
• W2905102070 creator A5028904070 @default.
• W2905102070 creator A5078565848 @default.
• W2905102070 creator A5082984558 @default.
• W2905102070 date "2019-07-17" @default.
• W2905102070 modified "2023-10-16" @default.
• W2905102070 title "Deep Neural Network Quantization via Layer-Wise Optimization Using Limited Training Data" @default.
• W2905102070 cites W1686810756 @default.
• W2905102070 cites W1724438581 @default.
• W2905102070 cites W2114766824 @default.
• W2905102070 cites W2125389748 @default.
• W2905102070 cites W2164278908 @default.
• W2905102070 cites W2194775991 @default.
• W2905102070 cites W2233116163 @default.
• W2905102070 cites W2267635276 @default.
• W2905102070 cites W2286365479 @default.
• W2905102070 cites W2593245696 @default.
• W2905102070 cites W2606482332 @default.
• W2905102070 cites W2739789140 @default.
• W2905102070 cites W2787752464 @default.
• W2905102070 cites W2962963202 @default.
• W2905102070 cites W2963114950 @default.
• W2905102070 cites W2963206030 @default.
• W2905102070 cites W2963452728 @default.
• W2905102070 cites W2963516298 @default.
• W2905102070 cites W2963547822 @default.
• W2905102070 cites W2963674932 @default.
• W2905102070 cites W2964008506 @default.
• W2905102070 cites W587794757 @default.
• W2905102070 doi "https://doi.org/10.1609/aaai.v33i01.33013329" @default.
• W2905102070 hasPublicationYear "2019" @default.
• W2905102070 type Work @default.
• W2905102070 sameAs 2905102070 @default.
• W2905102070 citedByCount "20" @default.
• W2905102070 countsByYear W29051020702019 @default.
• W2905102070 countsByYear W29051020702020 @default.
• W2905102070 countsByYear W29051020702021 @default.
• W2905102070 countsByYear W29051020702022 @default.
• W2905102070 countsByYear W29051020702023 @default.
• W2905102070 crossrefType "journal-article" @default.
• W2905102070 hasAuthorship W2905102070A5028904070 @default.
• W2905102070 hasAuthorship W2905102070A5078565848 @default.
• W2905102070 hasAuthorship W2905102070A5082984558 @default.
• W2905102070 hasBestOaLocation W29051020701 @default.
• W2905102070 hasConcept C105339364 @default.
• W2905102070 hasConcept C108583219 @default.
• W2905102070 hasConcept C111919701 @default.
• W2905102070 hasConcept C11413529 @default.
• W2905102070 hasConcept C126255220 @default.
• W2905102070 hasConcept C13280743 @default.
• W2905102070 hasConcept C137836250 @default.
• W2905102070 hasConcept C138236772 @default.
• W2905102070 hasConcept C154945302 @default.
• W2905102070 hasConcept C185798385 @default.
• W2905102070 hasConcept C205649164 @default.
• W2905102070 hasConcept C28855332 @default.
• W2905102070 hasConcept C2984842247 @default.
• W2905102070 hasConcept C33923547 @default.
• W2905102070 hasConcept C41008148 @default.
• W2905102070 hasConcept C50644808 @default.
• W2905102070 hasConcept C79974875 @default.
• W2905102070 hasConcept C98045186 @default.
• W2905102070 hasConceptScore W2905102070C105339364 @default.
• W2905102070 hasConceptScore W2905102070C108583219 @default.
• W2905102070 hasConceptScore W2905102070C111919701 @default.
• W2905102070 hasConceptScore W2905102070C11413529 @default.
• W2905102070 hasConceptScore W2905102070C126255220 @default.
• W2905102070 hasConceptScore W2905102070C13280743 @default.
• W2905102070 hasConceptScore W2905102070C137836250 @default.
• W2905102070 hasConceptScore W2905102070C138236772 @default.
• W2905102070 hasConceptScore W2905102070C154945302 @default.
• W2905102070 hasConceptScore W2905102070C185798385 @default.
• W2905102070 hasConceptScore W2905102070C205649164 @default.
• W2905102070 hasConceptScore W2905102070C28855332 @default.
• W2905102070 hasConceptScore W2905102070C2984842247 @default.
• W2905102070 hasConceptScore W2905102070C33923547 @default.
• W2905102070 hasConceptScore W2905102070C41008148 @default.
• W2905102070 hasConceptScore W2905102070C50644808 @default.
• W2905102070 hasConceptScore W2905102070C79974875 @default.
• W2905102070 hasConceptScore W2905102070C98045186 @default.
• W2905102070 hasIssue "01" @default.
• W2905102070 hasLocation W29051020701 @default.
• W2905102070 hasOpenAccess W2905102070 @default.
• W2905102070 hasPrimaryLocation W29051020701 @default.
• W2905102070 hasRelatedWork W2279398222 @default.
• W2905102070 hasRelatedWork W2791691546 @default.
• W2905102070 hasRelatedWork W2915754718 @default.
• W2905102070 hasRelatedWork W2941846814 @default.
• W2905102070 hasRelatedWork W3082895349 @default.
• W2905102070 hasRelatedWork W3157803868 @default.
• W2905102070 hasRelatedWork W3176779650 @default.
• W2905102070 hasRelatedWork W3193983877 @default.
• W2905102070 hasRelatedWork W4288040045 @default.
• W2905102070 hasRelatedWork W4299822940 @default.
• W2905102070 hasVolume "33" @default.
• W2905102070 isParatext "false" @default.
• W2905102070 isRetracted "false" @default.

• next