Design Considerations for Energy-efficient Inference on Edge Devices
Walid A. Hanafy, Tergel Molom-Ochir, Rohan Shenoy
Abstract
The emergence of low-power accelerators has enabled deep learning models to be executed on mobile or embedded edge devices without relying on cloud resources. The energy-constrained nature of these devices requires a judicious choice of a deep learning model and system configuration parameter to meet application needs while optimizing energy used during deep learning inference.
Topics & Concepts
InferenceComputer scienceDeep learningEnhanced Data Rates for GSM EvolutionEdge deviceCloud computingMobile deviceEdge computingEnergy (signal processing)Artificial intelligenceEfficient energy usePower (physics)Distributed computingMachine learningComputer engineeringOperating systemElectrical engineeringEngineeringStatisticsQuantum mechanicsPhysicsMathematicsAdvanced Neural Network ApplicationsAdvanced Memory and Neural ComputingIoT and Edge/Fog Computing