FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3035325130> ?p ?o ?g. }

• W3035325130 endingPage "2048" @default.
• W3035325130 startingPage "2035" @default.
• W3035325130 abstract "Model quantization is a widely used technique to compress and accelerate deep neural network (DNN) inference. Emergent DNN hardware accelerators begin to support mixed precision (1-8 bits) to further improve the computation efficiency, which raises a great challenge to find the optimal bitwidth for each layer: it requires domain experts to explore the vast design space trading off among accuracy, latency, energy, and model size, which is both time-consuming and sub-optimal. Conventional quantization algorithm ignores the different hardware architectures and quantizes all the layers in a uniform way. In this paper, we introduce the Hardware-Aware Automated Quantization (HAQ) framework which leverages the reinforcement learning to automatically determine the quantization policy, and we take the hardware accelerator's feedback in the design loop. Rather than relying on proxy signals such as FLOPs and model size, we employ a hardware simulator to generate direct feedback signals (latency and energy) to the RL agent. Compared with conventional methods, our framework is fully automated and can specialize the quantization policy for different neural network architectures and hardware architectures. Our framework effectively reduced the latency by 1.4-1.95x and the energy consumption by 1.9x with negligible loss of accuracy compared with the fixed bitwidth (8 bits) quantization. Our framework reveals that the optimal policies on different hardware architectures (i.e., edge and cloud architectures) under different resource constraints (i.e., latency, energy, and model size) are drastically different. We interpreted the implication of different quantization policies, which offer insights for both neural network architecture design and hardware architecture design." @default.
• W3035325130 created "2020-06-19" @default.
• W3035325130 creator A5021398655 @default.
• W3035325130 creator A5052596733 @default.
• W3035325130 creator A5055965638 @default.
• W3035325130 creator A5065001783 @default.
• W3035325130 creator A5070926896 @default.
• W3035325130 date "2020-06-11" @default.
• W3035325130 modified "2023-09-30" @default.
• W3035325130 title "Hardware-Centric AutoML for Mixed-Precision Quantization" @default.
• W3035325130 cites W2002555321 @default.
• W3035325130 cites W2108598243 @default.
• W3035325130 cites W2194775991 @default.
• W3035325130 cites W2300242332 @default.
• W3035325130 cites W2531409750 @default.
• W3035325130 cites W2554302513 @default.
• W3035325130 cites W2799238091 @default.
• W3035325130 cites W2810704618 @default.
• W3035325130 cites W2886851211 @default.
• W3035325130 cites W2962851801 @default.
• W3035325130 cites W2962861284 @default.
• W3035325130 cites W2963122961 @default.
• W3035325130 cites W2963163009 @default.
• W3035325130 cites W2963363373 @default.
• W3035325130 cites W2963367920 @default.
• W3035325130 cites W2963821229 @default.
• W3035325130 doi "https://doi.org/10.1007/s11263-020-01339-6" @default.
• W3035325130 hasPublicationYear "2020" @default.
• W3035325130 type Work @default.
• W3035325130 sameAs 3035325130 @default.
• W3035325130 citedByCount "10" @default.
• W3035325130 countsByYear W30353251302020 @default.
• W3035325130 countsByYear W30353251302021 @default.
• W3035325130 countsByYear W30353251302022 @default.
• W3035325130 countsByYear W30353251302023 @default.
• W3035325130 crossrefType "journal-article" @default.
• W3035325130 hasAuthorship W3035325130A5021398655 @default.
• W3035325130 hasAuthorship W3035325130A5052596733 @default.
• W3035325130 hasAuthorship W3035325130A5055965638 @default.
• W3035325130 hasAuthorship W3035325130A5065001783 @default.
• W3035325130 hasAuthorship W3035325130A5070926896 @default.
• W3035325130 hasBestOaLocation W30353251302 @default.
• W3035325130 hasConcept C111919701 @default.
• W3035325130 hasConcept C113775141 @default.
• W3035325130 hasConcept C11413529 @default.
• W3035325130 hasConcept C119599485 @default.
• W3035325130 hasConcept C127413603 @default.
• W3035325130 hasConcept C13164978 @default.
• W3035325130 hasConcept C138236772 @default.
• W3035325130 hasConcept C149635348 @default.
• W3035325130 hasConcept C154945302 @default.
• W3035325130 hasConcept C173608175 @default.
• W3035325130 hasConcept C199360897 @default.
• W3035325130 hasConcept C2742236 @default.
• W3035325130 hasConcept C2776214188 @default.
• W3035325130 hasConcept C2776221188 @default.
• W3035325130 hasConcept C2777904410 @default.
• W3035325130 hasConcept C28855332 @default.
• W3035325130 hasConcept C41008148 @default.
• W3035325130 hasConcept C42935608 @default.
• W3035325130 hasConcept C45374587 @default.
• W3035325130 hasConcept C50644808 @default.
• W3035325130 hasConcept C65232700 @default.
• W3035325130 hasConcept C76155785 @default.
• W3035325130 hasConcept C79974875 @default.
• W3035325130 hasConcept C82876162 @default.
• W3035325130 hasConcept C9390403 @default.
• W3035325130 hasConcept C97541855 @default.
• W3035325130 hasConceptScore W3035325130C111919701 @default.
• W3035325130 hasConceptScore W3035325130C113775141 @default.
• W3035325130 hasConceptScore W3035325130C11413529 @default.
• W3035325130 hasConceptScore W3035325130C119599485 @default.
• W3035325130 hasConceptScore W3035325130C127413603 @default.
• W3035325130 hasConceptScore W3035325130C13164978 @default.
• W3035325130 hasConceptScore W3035325130C138236772 @default.
• W3035325130 hasConceptScore W3035325130C149635348 @default.
• W3035325130 hasConceptScore W3035325130C154945302 @default.
• W3035325130 hasConceptScore W3035325130C173608175 @default.
• W3035325130 hasConceptScore W3035325130C199360897 @default.
• W3035325130 hasConceptScore W3035325130C2742236 @default.
• W3035325130 hasConceptScore W3035325130C2776214188 @default.
• W3035325130 hasConceptScore W3035325130C2776221188 @default.
• W3035325130 hasConceptScore W3035325130C2777904410 @default.
• W3035325130 hasConceptScore W3035325130C28855332 @default.
• W3035325130 hasConceptScore W3035325130C41008148 @default.
• W3035325130 hasConceptScore W3035325130C42935608 @default.
• W3035325130 hasConceptScore W3035325130C45374587 @default.
• W3035325130 hasConceptScore W3035325130C50644808 @default.
• W3035325130 hasConceptScore W3035325130C65232700 @default.
• W3035325130 hasConceptScore W3035325130C76155785 @default.
• W3035325130 hasConceptScore W3035325130C79974875 @default.
• W3035325130 hasConceptScore W3035325130C82876162 @default.
• W3035325130 hasConceptScore W3035325130C9390403 @default.
• W3035325130 hasConceptScore W3035325130C97541855 @default.
• W3035325130 hasIssue "8-9" @default.
• W3035325130 hasLocation W30353251301 @default.
• W3035325130 hasLocation W30353251302 @default.
• W3035325130 hasLocation W30353251303 @default.

• next