A 0.086-mm 12.7-pJ/SOP 64k-Synapse 256-Neuron Online-Learning Digital Spiking Neuromorphic Processor in 28-nm CMOS

C Frenkel, M Lefebvre, JD Legat… - IEEE transactions on …, 2018 - ieeexplore.ieee.org
Shifting computing architectures from von Neumann to event-based spiking neural networks
(SNNs) uncovers new opportunities for low-power processing of sensory data in …

Spiking neural networks hardware implementations and challenges: A survey

M Bouvier, A Valentian, T Mesquida… - ACM Journal on …, 2019 - dl.acm.org
Neuromorphic computing is henceforth a major research field for both academic and
industrial actors. As opposed to Von Neumann machines, brain-inspired processors aim at …

Bottom-up and top-down approaches for the design of neuromorphic processing systems: tradeoffs and synergies between natural and artificial intelligence

C Frenkel, D Bol, G Indiveri - Proceedings of the IEEE, 2023 - ieeexplore.ieee.org
While Moore's law has driven exponential computing power expectations, its nearing end
calls for new avenues for improving the overall system performance. One of these avenues …

SpiNNaker 2: A 10 million core processor system for brain simulation and machine learning-Keynote presentation

C Mayr, S Hoeppner, S Furber - … Process Architectures 2017 & …, 2019 - ebooks.iospress.nl
SpiNNaker is an ARM-based processor platform optimized for the simulation of spiking
neural networks. This brief describes the roadmap in going from the current SPINNaker1 …

Bottom-up and top-down neural processing systems design: Neuromorphic intelligence as the convergence of natural and artificial intelligence

CP Frenkel, D Bol, G Indiveri - ArXiv. org, 2021 - zora.uzh.ch
While Moore's law has driven exponential computing power expectations, its nearing end
calls for new avenues for improving the overall system performance. One of these avenues …

[HTML][HTML] GPUs outperform current HPC and neuromorphic solutions in terms of speed and energy when simulating a highly-connected cortical model

JC Knight, T Nowotny - Frontiers in neuroscience, 2018 - frontiersin.org
While neuromorphic systems may be the ultimate platform for deploying spiking neural
networks (SNNs), their distributed nature and optimization for specific types of models …

The SpiNNaker 2 processing element architecture for hybrid digital neuromorphic computing

S Höppner, Y Yan, A Dixius, S Scholze… - arXiv preprint arXiv …, 2021 - arxiv.org
This paper introduces the processing element architecture of the second generation
SpiNNaker chip, implemented in 22nm FDSOI. On circuit level, the chip features adaptive …

Comparing Loihi with a SpiNNaker 2 prototype on low-latency keyword spotting and adaptive robotic control

Y Yan, TC Stewart, X Choo, B Vogginger… - Neuromorphic …, 2021 - iopscience.iop.org
We implemented two neural network based benchmark tasks on a prototype chip of the
second-generation SpiNNaker (SpiNNaker 2) neuromorphic system: keyword spotting and …

Memory-efficient deep learning on a SpiNNaker 2 prototype

C Liu, G Bellec, B Vogginger, D Kappel… - Frontiers in …, 2018 - frontiersin.org
The memory requirement of deep learning algorithms is considered incompatible with the
memory restriction of energy-efficient hardware. A low memory footprint can be achieved by …

Efficient recurrent architectures through activity sparsity and sparse back-propagation through time

A Subramoney, KK Nazeer, M Schöne, C Mayr… - arXiv preprint arXiv …, 2022 - arxiv.org
Recurrent neural networks (RNNs) are well suited for solving sequence tasks in resource-
constrained systems due to their expressivity and low computational requirements …