Matches in SemOpenAlex for { <https://semopenalex.org/work/W3036249069> ?p ?o ?g. }
Showing items 1 to 91 of
91
with 100 items per page.
- W3036249069 abstract "Optical neural networks (ONNs) demonstrate orders-of-magnitude higher speed in deep learning acceleration than their electronic counterparts. However, limited control precision and device variations induce accuracy degradation in practical ONN implementations. To tackle this issue, we propose a quantization scheme that adapts a full-precision ONN to low-resolution voltage controls. Moreover, we propose a protective regularization technique that dynamically penalizes quantized weights based on their estimated noise-robustness, leading to an improvement in noise robustness. Experimental results show that the proposed scheme effectively adapts ONNs to limited-precision controls and device variations. The resultant four-layer ONN demonstrates higher inference accuracy with lower variances than baseline methods under various control precisions and device noises." @default.
- W3036249069 created "2020-06-25" @default.
- W3036249069 creator A5002117195 @default.
- W3036249069 creator A5011883763 @default.
- W3036249069 creator A5038900138 @default.
- W3036249069 creator A5066920215 @default.
- W3036249069 creator A5073993140 @default.
- W3036249069 creator A5090386129 @default.
- W3036249069 date "2020-03-01" @default.
- W3036249069 modified "2023-10-16" @default.
- W3036249069 title "ROQ: A Noise-Aware Quantization Scheme Towards Robust Optical Neural Networks with Low-bit Controls" @default.
- W3036249069 cites W2076476236 @default.
- W3036249069 cites W2146747949 @default.
- W3036249069 cites W2399958287 @default.
- W3036249069 cites W2550362495 @default.
- W3036249069 cites W2613989746 @default.
- W3036249069 cites W2752849906 @default.
- W3036249069 cites W2898741639 @default.
- W3036249069 cites W2904694042 @default.
- W3036249069 cites W2909619795 @default.
- W3036249069 cites W2945445229 @default.
- W3036249069 cites W2946522000 @default.
- W3036249069 cites W2952285760 @default.
- W3036249069 cites W3013837193 @default.
- W3036249069 cites W3098290163 @default.
- W3036249069 cites W4247624284 @default.
- W3036249069 doi "https://doi.org/10.23919/date48585.2020.9116521" @default.
- W3036249069 hasPublicationYear "2020" @default.
- W3036249069 type Work @default.
- W3036249069 sameAs 3036249069 @default.
- W3036249069 citedByCount "11" @default.
- W3036249069 countsByYear W30362490692020 @default.
- W3036249069 countsByYear W30362490692021 @default.
- W3036249069 countsByYear W30362490692022 @default.
- W3036249069 countsByYear W30362490692023 @default.
- W3036249069 crossrefType "proceedings-article" @default.
- W3036249069 hasAuthorship W3036249069A5002117195 @default.
- W3036249069 hasAuthorship W3036249069A5011883763 @default.
- W3036249069 hasAuthorship W3036249069A5038900138 @default.
- W3036249069 hasAuthorship W3036249069A5066920215 @default.
- W3036249069 hasAuthorship W3036249069A5073993140 @default.
- W3036249069 hasAuthorship W3036249069A5090386129 @default.
- W3036249069 hasConcept C104317684 @default.
- W3036249069 hasConcept C11413529 @default.
- W3036249069 hasConcept C127313418 @default.
- W3036249069 hasConcept C127413603 @default.
- W3036249069 hasConcept C154945302 @default.
- W3036249069 hasConcept C185592680 @default.
- W3036249069 hasConcept C24326235 @default.
- W3036249069 hasConcept C2776214188 @default.
- W3036249069 hasConcept C28855332 @default.
- W3036249069 hasConcept C3019883945 @default.
- W3036249069 hasConcept C3020199158 @default.
- W3036249069 hasConcept C41008148 @default.
- W3036249069 hasConcept C50644808 @default.
- W3036249069 hasConcept C55493867 @default.
- W3036249069 hasConcept C62649853 @default.
- W3036249069 hasConcept C63479239 @default.
- W3036249069 hasConceptScore W3036249069C104317684 @default.
- W3036249069 hasConceptScore W3036249069C11413529 @default.
- W3036249069 hasConceptScore W3036249069C127313418 @default.
- W3036249069 hasConceptScore W3036249069C127413603 @default.
- W3036249069 hasConceptScore W3036249069C154945302 @default.
- W3036249069 hasConceptScore W3036249069C185592680 @default.
- W3036249069 hasConceptScore W3036249069C24326235 @default.
- W3036249069 hasConceptScore W3036249069C2776214188 @default.
- W3036249069 hasConceptScore W3036249069C28855332 @default.
- W3036249069 hasConceptScore W3036249069C3019883945 @default.
- W3036249069 hasConceptScore W3036249069C3020199158 @default.
- W3036249069 hasConceptScore W3036249069C41008148 @default.
- W3036249069 hasConceptScore W3036249069C50644808 @default.
- W3036249069 hasConceptScore W3036249069C55493867 @default.
- W3036249069 hasConceptScore W3036249069C62649853 @default.
- W3036249069 hasConceptScore W3036249069C63479239 @default.
- W3036249069 hasLocation W30362490691 @default.
- W3036249069 hasOpenAccess W3036249069 @default.
- W3036249069 hasPrimaryLocation W30362490691 @default.
- W3036249069 hasRelatedWork W1968403749 @default.
- W3036249069 hasRelatedWork W2071213595 @default.
- W3036249069 hasRelatedWork W2098514804 @default.
- W3036249069 hasRelatedWork W2220182755 @default.
- W3036249069 hasRelatedWork W2351279671 @default.
- W3036249069 hasRelatedWork W2373250155 @default.
- W3036249069 hasRelatedWork W2386387936 @default.
- W3036249069 hasRelatedWork W2584698272 @default.
- W3036249069 hasRelatedWork W3018461255 @default.
- W3036249069 hasRelatedWork W3131251371 @default.
- W3036249069 isParatext "false" @default.
- W3036249069 isRetracted "false" @default.
- W3036249069 magId "3036249069" @default.
- W3036249069 workType "article" @default.