Matches in SemOpenAlex for { <https://semopenalex.org/work/W2912890062> ?p ?o ?g. }
Showing items 1 to 86 of
86
with 100 items per page.
- W2912890062 abstract "As optimizing neural networks to improve processing speed has become a necessary process in industrial deployment, emerging challenges due to limited storage and computation when adopting neural networks on embedded system need to be addressed. There are some previous studies focusing on training neural networks with low precision weights and activations, like Binarized neural network and DoReFa-Net. Meanwhile, an alternative approach is deployed to reduce number of operations and model size, like MobileNet and SqueezeNet. In this work, by adopting SQ(Scale Quantization) method to quantify activations of each neuron, we have gained a more satisfying result especially in compact network models. For SqueezeDet on KITTI dataset, SQ achieves 22.86% more accuracy on means correct detections, while means location and classification error decrease by 41.84% and 68.57%, compared with DoReFa-Net with 2-bits activations. Additionally, we also propose the SQT(Scale Quantization with Threshold) method to further improve the quantization flexibility. It is worth mentioned that our approach is primarily designed for activations quantization, but it can also be complemented with other orthogonal methods to further achieving efficient neural network processing at the edge." @default.
- W2912890062 created "2019-02-21" @default.
- W2912890062 creator A5008970005 @default.
- W2912890062 creator A5012839164 @default.
- W2912890062 creator A5029924540 @default.
- W2912890062 creator A5039756596 @default.
- W2912890062 creator A5044243518 @default.
- W2912890062 creator A5081016267 @default.
- W2912890062 date "2018-11-01" @default.
- W2912890062 modified "2023-09-27" @default.
- W2912890062 title "Activations Quantization for Compact Neural Networks" @default.
- W2912890062 cites W2094756095 @default.
- W2912890062 cites W2279098554 @default.
- W2912890062 cites W2289252105 @default.
- W2912890062 cites W2337344472 @default.
- W2912890062 cites W2405920868 @default.
- W2912890062 cites W2442974303 @default.
- W2912890062 cites W2469490737 @default.
- W2912890062 cites W2518281301 @default.
- W2912890062 cites W2554302513 @default.
- W2912890062 cites W2583037806 @default.
- W2912890062 cites W2612445135 @default.
- W2912890062 cites W2724243553 @default.
- W2912890062 cites W2724359148 @default.
- W2912890062 cites W2962988160 @default.
- W2912890062 cites W2963087201 @default.
- W2912890062 cites W2963114950 @default.
- W2912890062 cites W2964299589 @default.
- W2912890062 doi "https://doi.org/10.1109/icdsp.2018.8631650" @default.
- W2912890062 hasPublicationYear "2018" @default.
- W2912890062 type Work @default.
- W2912890062 sameAs 2912890062 @default.
- W2912890062 citedByCount "1" @default.
- W2912890062 countsByYear W29128900622022 @default.
- W2912890062 crossrefType "proceedings-article" @default.
- W2912890062 hasAuthorship W2912890062A5008970005 @default.
- W2912890062 hasAuthorship W2912890062A5012839164 @default.
- W2912890062 hasAuthorship W2912890062A5029924540 @default.
- W2912890062 hasAuthorship W2912890062A5039756596 @default.
- W2912890062 hasAuthorship W2912890062A5044243518 @default.
- W2912890062 hasAuthorship W2912890062A5081016267 @default.
- W2912890062 hasConcept C105339364 @default.
- W2912890062 hasConcept C105795698 @default.
- W2912890062 hasConcept C111919701 @default.
- W2912890062 hasConcept C113775141 @default.
- W2912890062 hasConcept C11413529 @default.
- W2912890062 hasConcept C138236772 @default.
- W2912890062 hasConcept C154945302 @default.
- W2912890062 hasConcept C2780598303 @default.
- W2912890062 hasConcept C28855332 @default.
- W2912890062 hasConcept C33923547 @default.
- W2912890062 hasConcept C41008148 @default.
- W2912890062 hasConcept C45374587 @default.
- W2912890062 hasConcept C50644808 @default.
- W2912890062 hasConcept C79974875 @default.
- W2912890062 hasConceptScore W2912890062C105339364 @default.
- W2912890062 hasConceptScore W2912890062C105795698 @default.
- W2912890062 hasConceptScore W2912890062C111919701 @default.
- W2912890062 hasConceptScore W2912890062C113775141 @default.
- W2912890062 hasConceptScore W2912890062C11413529 @default.
- W2912890062 hasConceptScore W2912890062C138236772 @default.
- W2912890062 hasConceptScore W2912890062C154945302 @default.
- W2912890062 hasConceptScore W2912890062C2780598303 @default.
- W2912890062 hasConceptScore W2912890062C28855332 @default.
- W2912890062 hasConceptScore W2912890062C33923547 @default.
- W2912890062 hasConceptScore W2912890062C41008148 @default.
- W2912890062 hasConceptScore W2912890062C45374587 @default.
- W2912890062 hasConceptScore W2912890062C50644808 @default.
- W2912890062 hasConceptScore W2912890062C79974875 @default.
- W2912890062 hasLocation W29128900621 @default.
- W2912890062 hasOpenAccess W2912890062 @default.
- W2912890062 hasPrimaryLocation W29128900621 @default.
- W2912890062 hasRelatedWork W2040210784 @default.
- W2912890062 hasRelatedWork W2105100451 @default.
- W2912890062 hasRelatedWork W2347472543 @default.
- W2912890062 hasRelatedWork W2386852690 @default.
- W2912890062 hasRelatedWork W2912890062 @default.
- W2912890062 hasRelatedWork W2996131497 @default.
- W2912890062 hasRelatedWork W3151873305 @default.
- W2912890062 hasRelatedWork W3216454443 @default.
- W2912890062 hasRelatedWork W3217204427 @default.
- W2912890062 hasRelatedWork W4210396250 @default.
- W2912890062 isParatext "false" @default.
- W2912890062 isRetracted "false" @default.
- W2912890062 magId "2912890062" @default.
- W2912890062 workType "article" @default.