FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2900768265> ?p ?o ?g. }

• W2900768265 abstract "In this paper, we propose to train convolutional neural networks (CNNs) with both binarized weights and activations, leading to quantized models specifically} for mobile devices with limited power capacity and computation resources. Previous works on quantizing CNNs seek to approximate the floating-point information using a set of discrete values, which we call value approximation, but typically assume the same architecture as the full-precision networks. In this paper, however, we take a novel 'structure approximation' view for quantization---it is very likely that a different architecture may be better for best performance. In particular, we propose a `network decomposition' strategy, named textbf{Group-Net}, in which we divide the network into groups. In this way, each full-precision group can be effectively reconstructed by aggregating a set of homogeneous binary branches. In addition, we learn effective connections among groups to improve the representational capability. Moreover, the proposed Group-Net shows strong generalization to other tasks. For instance, we extend Group-Net for highly accurate semantic segmentation by embedding rich context into the binary structure. Experiments on both classification and semantic segmentation tasks demonstrate the superior performance of the proposed methods over various popular architectures. In particular, we outperform the previous best binary neural networks in terms of accuracy and major computation savings." @default.
• W2900768265 created "2018-11-29" @default.
• W2900768265 creator A5006294869 @default.
• W2900768265 creator A5032352025 @default.
• W2900768265 creator A5070976480 @default.
• W2900768265 creator A5076928390 @default.
• W2900768265 creator A5089444805 @default.
• W2900768265 date "2018-11-22" @default.
• W2900768265 modified "2023-09-23" @default.
• W2900768265 title "Structured Binary Neural Networks for Accurate Image Classification and Semantic Segmentation" @default.
• W2900768265 cites W1522301498 @default.
• W2900768265 cites W1902934009 @default.
• W2900768265 cites W2031489346 @default.
• W2900768265 cites W2117539524 @default.
• W2900768265 cites W2144794286 @default.
• W2900768265 cites W2186615578 @default.
• W2900768265 cites W2242818861 @default.
• W2900768265 cites W2267635276 @default.
• W2900768265 cites W2274287116 @default.
• W2900768265 cites W2279098554 @default.
• W2900768265 cites W2469490737 @default.
• W2900768265 cites W2524428287 @default.
• W2900768265 cites W2553303224 @default.
• W2900768265 cites W2554592357 @default.
• W2900768265 cites W2560017826 @default.
• W2900768265 cites W2593245696 @default.
• W2900768265 cites W2604700561 @default.
• W2900768265 cites W2612445135 @default.
• W2900768265 cites W2613718673 @default.
• W2900768265 cites W2618939455 @default.
• W2900768265 cites W2622213512 @default.
• W2900768265 cites W2630837129 @default.
• W2900768265 cites W2736941579 @default.
• W2900768265 cites W2737121650 @default.
• W2900768265 cites W2752077245 @default.
• W2900768265 cites W2753301142 @default.
• W2900768265 cites W2765761462 @default.
• W2900768265 cites W2769312802 @default.
• W2900768265 cites W2771727678 @default.
• W2900768265 cites W2786548044 @default.
• W2900768265 cites W2786672677 @default.
• W2900768265 cites W2787752464 @default.
• W2900768265 cites W2796438033 @default.
• W2900768265 cites W2893165708 @default.
• W2900768265 cites W2899771611 @default.
• W2900768265 cites W2949605076 @default.
• W2900768265 cites W2949650786 @default.
• W2900768265 cites W2950179405 @default.
• W2900768265 cites W2950458216 @default.
• W2900768265 cites W2951104886 @default.
• W2900768265 cites W2951192860 @default.
• W2900768265 cites W2951283768 @default.
• W2900768265 cites W2951583185 @default.
• W2900768265 cites W2952632681 @default.
• W2900768265 cites W2953328958 @default.
• W2900768265 cites W2962746461 @default.
• W2900768265 cites W2963122961 @default.
• W2900768265 cites W2963125010 @default.
• W2900768265 cites W2963674932 @default.
• W2900768265 cites W2963778169 @default.
• W2900768265 cites W2964133305 @default.
• W2900768265 cites W2964309882 @default.
• W2900768265 doi "https://doi.org/10.48550/arxiv.1811.10413" @default.
• W2900768265 hasPublicationYear "2018" @default.
• W2900768265 type Work @default.
• W2900768265 sameAs 2900768265 @default.
• W2900768265 citedByCount "3" @default.
• W2900768265 countsByYear W29007682652019 @default.
• W2900768265 countsByYear W29007682652020 @default.
• W2900768265 crossrefType "posted-content" @default.
• W2900768265 hasAuthorship W2900768265A5006294869 @default.
• W2900768265 hasAuthorship W2900768265A5032352025 @default.
• W2900768265 hasAuthorship W2900768265A5070976480 @default.
• W2900768265 hasAuthorship W2900768265A5076928390 @default.
• W2900768265 hasAuthorship W2900768265A5089444805 @default.
• W2900768265 hasBestOaLocation W29007682651 @default.
• W2900768265 hasConcept C11413529 @default.
• W2900768265 hasConcept C115961682 @default.
• W2900768265 hasConcept C153180895 @default.
• W2900768265 hasConcept C154945302 @default.
• W2900768265 hasConcept C28855332 @default.
• W2900768265 hasConcept C33923547 @default.
• W2900768265 hasConcept C41008148 @default.
• W2900768265 hasConcept C41608201 @default.
• W2900768265 hasConcept C45374587 @default.
• W2900768265 hasConcept C48372109 @default.
• W2900768265 hasConcept C50644808 @default.
• W2900768265 hasConcept C75294576 @default.
• W2900768265 hasConcept C80444323 @default.
• W2900768265 hasConcept C81363708 @default.
• W2900768265 hasConcept C89600930 @default.
• W2900768265 hasConcept C94375191 @default.
• W2900768265 hasConceptScore W2900768265C11413529 @default.
• W2900768265 hasConceptScore W2900768265C115961682 @default.
• W2900768265 hasConceptScore W2900768265C153180895 @default.
• W2900768265 hasConceptScore W2900768265C154945302 @default.
• W2900768265 hasConceptScore W2900768265C28855332 @default.
• W2900768265 hasConceptScore W2900768265C33923547 @default.
• W2900768265 hasConceptScore W2900768265C41008148 @default.
• W2900768265 hasConceptScore W2900768265C41608201 @default.

• next