Matches in SemOpenAlex for { <https://semopenalex.org/work/W4379115874> ?p ?o ?g. }
Showing items 1 to 96 of
96
with 100 items per page.
- W4379115874 abstract "With ever increasing depth and width in deep neural networks to achieve state-of-the-art performance, deep learning computation has significantly grown, and dot-products remain dominant in overall computation time. Most prior works are built on conventional dot-product where weighted input summation is used to represent the neuron operation. However, another implementation of dot-product based on the notion of angles and magnitudes in the Euclidean space has attracted limited attention. This paper proposes DeepCAM, an inference accelerator built on two critical innovations to alleviate the computation time bottleneck of convolutional neural networks. The first innovation is an approximate dot-product built on computations in the Euclidean space that can replace addition and multiplication with simple bit-wise operations. The second innovation is a dynamic size content addressable memory-based (CAM-based) accelerator to perform bit-wise operations and accelerate the CNNs with a lower computation time. Our experiments on benchmark image recognition datasets demonstrate that DeepCAM is up to <tex xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>$523times$</tex> and <tex xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>$3498times$</tex> faster than Eyeriss and traditional CPUs like Intel Skylake, respectively. Furthermore, the energy consumed by our DeepCAM approach is <tex xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>$2.16times$</tex> to <tex xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>$109times$</tex> less compared to Eyeriss." @default.
- W4379115874 created "2023-06-03" @default.
- W4379115874 creator A5004629816 @default.
- W4379115874 creator A5011045826 @default.
- W4379115874 creator A5050310538 @default.
- W4379115874 creator A5050796355 @default.
- W4379115874 date "2023-04-01" @default.
- W4379115874 modified "2023-09-27" @default.
- W4379115874 title "DeepCAM: A Fully CAM-based Inference Accelerator with Variable Hash Lengths for Energy-efficient Deep Neural Networks" @default.
- W4379115874 cites W1536792390 @default.
- W4379115874 cites W1985123706 @default.
- W4379115874 cites W2099328942 @default.
- W4379115874 cites W2612094043 @default.
- W4379115874 cites W2793950911 @default.
- W4379115874 cites W2919115771 @default.
- W4379115874 cites W2920326572 @default.
- W4379115874 cites W2979473749 @default.
- W4379115874 cites W2988640543 @default.
- W4379115874 cites W3005619596 @default.
- W4379115874 cites W3024621361 @default.
- W4379115874 cites W3033076639 @default.
- W4379115874 cites W3082020764 @default.
- W4379115874 cites W3095256020 @default.
- W4379115874 cites W3098780461 @default.
- W4379115874 cites W3140854437 @default.
- W4379115874 cites W3187235458 @default.
- W4379115874 cites W3190062760 @default.
- W4379115874 cites W4293023010 @default.
- W4379115874 doi "https://doi.org/10.23919/date56975.2023.10137068" @default.
- W4379115874 hasPublicationYear "2023" @default.
- W4379115874 type Work @default.
- W4379115874 citedByCount "0" @default.
- W4379115874 crossrefType "proceedings-article" @default.
- W4379115874 hasAuthorship W4379115874A5004629816 @default.
- W4379115874 hasAuthorship W4379115874A5011045826 @default.
- W4379115874 hasAuthorship W4379115874A5050310538 @default.
- W4379115874 hasAuthorship W4379115874A5050796355 @default.
- W4379115874 hasBestOaLocation W43791158742 @default.
- W4379115874 hasConcept C108583219 @default.
- W4379115874 hasConcept C11413529 @default.
- W4379115874 hasConcept C13280743 @default.
- W4379115874 hasConcept C149635348 @default.
- W4379115874 hasConcept C154945302 @default.
- W4379115874 hasConcept C185798385 @default.
- W4379115874 hasConcept C186450821 @default.
- W4379115874 hasConcept C202444582 @default.
- W4379115874 hasConcept C205649164 @default.
- W4379115874 hasConcept C2524010 @default.
- W4379115874 hasConcept C2776214188 @default.
- W4379115874 hasConcept C2780513914 @default.
- W4379115874 hasConcept C32900221 @default.
- W4379115874 hasConcept C33923547 @default.
- W4379115874 hasConcept C41008148 @default.
- W4379115874 hasConcept C45374587 @default.
- W4379115874 hasConcept C50644808 @default.
- W4379115874 hasConcept C80444323 @default.
- W4379115874 hasConcept C81363708 @default.
- W4379115874 hasConcept C90673727 @default.
- W4379115874 hasConceptScore W4379115874C108583219 @default.
- W4379115874 hasConceptScore W4379115874C11413529 @default.
- W4379115874 hasConceptScore W4379115874C13280743 @default.
- W4379115874 hasConceptScore W4379115874C149635348 @default.
- W4379115874 hasConceptScore W4379115874C154945302 @default.
- W4379115874 hasConceptScore W4379115874C185798385 @default.
- W4379115874 hasConceptScore W4379115874C186450821 @default.
- W4379115874 hasConceptScore W4379115874C202444582 @default.
- W4379115874 hasConceptScore W4379115874C205649164 @default.
- W4379115874 hasConceptScore W4379115874C2524010 @default.
- W4379115874 hasConceptScore W4379115874C2776214188 @default.
- W4379115874 hasConceptScore W4379115874C2780513914 @default.
- W4379115874 hasConceptScore W4379115874C32900221 @default.
- W4379115874 hasConceptScore W4379115874C33923547 @default.
- W4379115874 hasConceptScore W4379115874C41008148 @default.
- W4379115874 hasConceptScore W4379115874C45374587 @default.
- W4379115874 hasConceptScore W4379115874C50644808 @default.
- W4379115874 hasConceptScore W4379115874C80444323 @default.
- W4379115874 hasConceptScore W4379115874C81363708 @default.
- W4379115874 hasConceptScore W4379115874C90673727 @default.
- W4379115874 hasFunder F4320306076 @default.
- W4379115874 hasLocation W43791158741 @default.
- W4379115874 hasLocation W43791158742 @default.
- W4379115874 hasOpenAccess W4379115874 @default.
- W4379115874 hasPrimaryLocation W43791158741 @default.
- W4379115874 hasRelatedWork W2731899572 @default.
- W4379115874 hasRelatedWork W2999805992 @default.
- W4379115874 hasRelatedWork W3011074480 @default.
- W4379115874 hasRelatedWork W3116150086 @default.
- W4379115874 hasRelatedWork W3133861977 @default.
- W4379115874 hasRelatedWork W3192840557 @default.
- W4379115874 hasRelatedWork W4200173597 @default.
- W4379115874 hasRelatedWork W4291897433 @default.
- W4379115874 hasRelatedWork W4312417841 @default.
- W4379115874 hasRelatedWork W4321369474 @default.
- W4379115874 isParatext "false" @default.
- W4379115874 isRetracted "false" @default.
- W4379115874 workType "article" @default.