Fast and energy-efficient neuromorphic deep learning with first-spike times

Göltz, Julian; Kriener, Laura; Baumbach, Andreas; Billaudelle, Sebastian; Breitwieser, Oliver; Cramer, Benjamin; Dold, Dominik; Kungl, Ákos F.; Senn, Walter; Schemmel, Johannes; Meier, K.; Petrovici, Mihai A.

doi:10.48550/arxiv.1912.11443

Cited by 13 publications

(16 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Despite the efforts to improve the efficiency of TTFS coding in deep SNNs [17,36], their improvements have been restricted by the conversionbased training algorithms. In several studies, SNNs were directly trained, but their methods were not validated as applicable to deep SNNs [12,37,38]. A recent study suggested direct training methods…”

Section: Training Methods Of Deep Snnsmentioning

confidence: 99%

T2FSNN: Deep Spiking Neural Networks with Time-to-first-spike Coding

Park

Kim

et al. 2020

2020 57th ACM/IEEE Design Automation Conference (DAC)

View full text Add to dashboard Cite

The tremendous energy consumption of deep neural networks (DNNs) has become a serious problem in deep learning. Spiking neural networks (SNNs), which mimic the operations in the human brain, have been studied as prominent energy-efficient neural networks. Due to their event-driven and spatiotemporally sparse operations, SNNs show possibilities for energy-efficient processing. To unlock their potential, deep SNNs have adopted temporal coding such as time-to-first-spike (TTFS) coding, which represents the information between neurons by the first spike time. With TTFS coding, each neuron generates one spike at most, which leads to a significant improvement in energy efficiency. Several studies have successfully introduced TTFS coding in deep SNNs, but they showed restricted efficiency improvement owing to the lack of consideration for efficiency during training. To address the aforementioned issue, this paper presents training methods for energyefficient deep SNNs with TTFS coding. We introduce a surrogate DNN model to train the deep SNN in a feasible time and analyze the effect of the temporal kernel on training performance and efficiency. Based on the investigation, we propose stochastically relaxed activation and initial value-based regularization for the temporal kernel parameters. In addition, to reduce the number of spikes even further, we present temporal kernel-aware batch normalization. With the proposed methods, we could achieve comparable training results with significantly reduced spikes, which could lead to energy-efficient deep SNNs.

show abstract

Section: Training Methods Of Deep Snnsmentioning

confidence: 99%

T2FSNN: Deep Spiking Neural Networks with Time-to-first-spike Coding

Park

Kim

et al. 2020

2020 57th ACM/IEEE Design Automation Conference (DAC)

View full text Add to dashboard Cite

show abstract

“…We have generalized this method to include an exact, closed-form expression for finite membrane time constants [10], [11] and applied it to a 3-layer network emulated on BrainScaleS-2 (Fig. 2).…”

Section: I I E X P E R I M E N T Smentioning

confidence: 99%

“…This duration scales proportionally to the chosen synaptic and membrane time constants, which in our case were set to 5 µs. Taking into consideration relaxation times between patterns, our setup is able to handle a pattern throughput of at least 20 kHz, independently of emulated network size [11].…”

Section: I I E X P E R I M E N T Smentioning

confidence: 99%

Versatile Emulation of Spiking Neural Networks on an Accelerated Neuromorphic Substrate

Billaudelle

Stradmann

Schreiber

et al. 2020

2020 IEEE International Symposium on Circuits and Systems (ISCAS)

View full text Add to dashboard Cite

We present first experimental results on the novel BrainScaleS-2 neuromorphic architecture based on an analog neuro-synaptic core and augmented by embedded microprocessors for complex plasticity and experiment control. The high acceleration factor of 1000 compared to biological dynamics enables the execution of computationally expensive tasks, by allowing the fast emulation of long-duration experiments or rapid iteration over many consecutive trials. The flexibility of our architecture is demonstrated in a suite of five distinct experiments, which emphasize different aspects of the BrainScaleS-2 system.

show abstract

“…Although the HICANN architecture has been successfully used to implement deep multi-layer networks using rate-based spiking models [34] and back-propagation based training, is looses some of its power-efficiency by emulating a Perceptron model. Encoding the activation in the time between spikes can enhance the efficiency significantly [35]. In all spiking solutions the network operates in continuous time and therefore the size of the network is limited to the number of neurons and synapses available on the chip.…”

Section: Analog Inference: Rate-based Extension Of Hicann-xmentioning

confidence: 99%

Accelerated Analog Neuromorphic Computing

Schemmel¹,

Billaudelle²,

Phillip³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

This paper presents the concepts behind the BrainScales (BSS) accelerated analog neuromorphic computing architecture. It describes the second-generation BrainScales-2 (BSS-2) version and its most recent in-silico realization, the HICANN-X Application Specific Integrated Circuit (ASIC), as it has been developed as part of the neuromorphic computing activities within the European Human Brain Project (HBP). While the first generation is implemented in an 180 nm process, the second generation uses 65 nm technology. This allows the integration of a digital plasticity processing unit, a highly-parallel micro processor specially built for the computational needs of learning in an accelerated analog neuromorphic systems.The presented architecture is based upon a continuous-time, analog, physical model implementation of neurons and synapses, resembling an analog neuromorphic accelerator attached to build-in digital compute cores. While the analog part emulates the spike-based dynamics of the neural network in continuous-time, the latter simulates biological processes happening on a slower time-scale, like structural and parameter changes. Compared to biological time-scales, the emulation is highly accelerated, i.e. all time-constants are several orders of magnitude smaller than in biology. Programmable ion channel emulation and inter-compartmental conductances allow the modeling of nonlinear dendrites, back-propagating action-potentials as well as NMDA and Calcium plateau potentials. To extend the usability of the analog accelerator, it also supports vector-matrix multiplication. Thereby, BSS-2 supports inference of deep convolutional networks as well as local-learning with complex ensembles of spiking neurons within the same substrate. A prerequisite to successful training is the calibratability of the underlying analog circuits across the full range of process variations. For this purpose a custom software toolbox has been developed, that facilitates complex calibrated Monte-Carlo simulations.

show abstract

Fast and energy-efficient neuromorphic deep learning with first-spike times

Cited by 13 publications

References 29 publications

T2FSNN: Deep Spiking Neural Networks with Time-to-first-spike Coding

T2FSNN: Deep Spiking Neural Networks with Time-to-first-spike Coding

Versatile Emulation of Spiking Neural Networks on an Accelerated Neuromorphic Substrate

Accelerated Analog Neuromorphic Computing

Contact Info

Product

Resources

About