IEEE Transactions on Emerging Topics in Computing · 2023 · 14 citations · 25 references
ComputingComputer EngineeringComputer ScienceEdge ComputingHardware-software Co-designMachine LearningEdge ArchitectureEdge DevicesHardware AccelerationNetwork-on-chip DesignPower-aware DesignSpiking Neural NetworksEmbedded Hardware Co-designEnergy-efficient ArchitectureEnergy-efficient ComputationHardware-software Co-optimizationInference EnginesOptical Flow EstimationResource-constrained DevicesTraining
Spiking Neural Networks (SNNs) are bio-plausible models that hold great potential for realizing energy-efficient implementations of sequential tasks on resource-constrained edge devices. However, commercial edge platforms based on standard GPUs are not optimized to deploy SNNs, resulting in high energy and latency. While analog In-Memory Computing (IMC) platforms can serve as energy-efficient inference engines, they are accursed by the immense energy, latency, and area requirements of high-precision ADCs (HP-ADC), overshadowing the benefits of in-memory computations. We propose a hardware/software co-design methodology to deploy SNNs into an ADC-Less IMC architecture using sense-amplifiers as 1-bit ADCs replacing conventional HP-ADCs and alleviating the above issues. Our proposed framework incurs minimal accuracy degradation by performing hardware-aware training and is able to scale beyond simple image classification tasks to more complex sequential regression tasks. Experiments on complex tasks of optical flow estimation and gesture recognition show that progressively increasing the hardware awareness during SNN training allows the model to adapt and learn the errors due to the non-idealities associated with ADC-Less IMC. Also, the proposed ADC-Less IMC offers significant energy and latency improvements, <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$2-7\times$</tex-math></inline-formula> and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$8.9-24.6\times$</tex-math></inline-formula> , respectively, depending on the SNN model and the workload, compared to HP-ADC IMC.
25
Loihi: A Neuromorphic Manycore Processor with On-Chip Learning
Mike Davies, Narayan Srinivasa, Tsung-Han Lin et al. · IEEE Micro · 2018 · 3.7K citations
Engineering, Computer Engineering, Computational Neuroscience +16
Emre Neftci, Hesham Mostafa, Friedemann Zenke · IEEE Signal Processing Magazine · 2019 · 1.4K citations
Life Sciences, Computational Engineering, Computational Neuroscience +12