Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Spiking Neural Network Accelerator Architecture for Differential-Time Representation using Learned Encoding

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A feedforward spiking neural network using only LIF neurons and a learned patch-based encoding surpasses 99% on MNIST and runs on a mid-sized FPGA in under 295 microseconds.

desk verdict Solid incremental hardware architecture; load-bearing gap in learned-encoding-to-spike conversion undermines the integrated claim. read the letter →

arxiv 2501.07952 v1 pith:7SA7XQGG submitted 2025-01-14 cs.NE eess.SP

Daniel Windhager , Lothar Ratschbacher , Bernhard A. Moser , Michael Lunglmayr This is my paper · ORCID
classification cs.NEeess.SP
keywords spikingneuralnetworksFPGAacceleratordifferential-timeencodinglearnedspikeLIFneuronsMNISTquantization-awaretrainingmultiplier-freeinference
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper sets out to show that a spiking neural network does not need a fixed, hand-designed spike encoding and does not need multipliers to reach high accuracy on MNIST. Its two moves are to fold the encoding into learning—the first LIF layer receives raw pixel values from image patches and learns which pixels matter, with weights restricted to $\{-1,0,1\}$ or powers of two—and to process the resulting spikes in hardware using a differential-time representation that stores only the time between spikes. On MNIST, the trained network reaches 99.03% accuracy with an 800-512-256-10 LIF-only architecture, and the synthesized accelerator fits on a ZCU102 with 56,520 LUTs, 91 block RAMs, and no DSP blocks while sustaining 3,400 images/s. The paper's significance is that it connects two open problems in SNN deployment—how to encode data and how to process spikes cheaply—in a single multiplier-free design.

What carries the argument

The argument rests on two coupled mechanisms. The first is learned patch encoding: a first layer of LIF neurons receives raw pixel values directly from image patches, with weights constrained during training to $\{-1,0,1\}$ or powers of two, so the network learns a spike encoding while keeping inference multiplier-free. The second is the differential-time hardware datapath: spikes are carried as pairs of a synapse index and a delta time (the time since the previous spike), which decouples spike time from processing time. The Spike Sorter serializes incoming parallel spikes by propagating the earliest spike through a cascade of nearest-neighbor comparisons and reconstructs the synapse index from the comparison bits; the Neuron Core is a single register, comparator, multiplexer, and adder, applying decay by bit-shifting the potential; the Leading-One Position Detector reserializes output spikes that remain sorted; and the Layer Controller broadcasts the common decay and delays the delta time so all spikes triggered by the same input share the same output time.

What would settle it

Take the trained 9x9 ternary encoder and implement the described ZCU102 design, then feed raw MNIST pixel patches straight into the Spike Sorter without any intermediate rate or latency encoding. If the accuracy on the test set drops below 99.03%, or if the design cannot pass timing at 300 MHz because multi-bit pixel values must be converted into event times, then the central claim is falsified.

Watch

Extended reading notes

Core claim

In the authors' terms, the discovery is that encoding, normally a preprocessing step fixed before training, can be treated as part of the learned network. The first layer consists of LIF neurons that receive sequences of raw pixel values from 9x9 image patches, so the network itself decides which pixels and which relative timings carry information. Training with quantization-aware constraints leaves weights in $\{-1,0,1\}$ or $\{2^i : i\in\mathbb{Z}\}$, so inference requires only additions, negations, and bit shifts; with the decay constant $\beta=0.5$ and threshold $\theta=1$, the neuron update itself reduces to shifts and comparisons. On MNIST this reaches 99.03% accuracy with a 400-800-512-256-10 feedforward LIF-only network, and the companion accelerator—built on the differential-time representation in which each spike carries only its time difference from the previous spike—processes 3,400 images/s on a ZCU102 at 300 MHz using 56,520 LUTs, 91 BRAMs, and no DSP blocks. The authors conclude that a purely LIF, multiplier-free feedforward SNN can deliver accuracy comparable to state-of-the-art SNN accelerators on MNIST while using fewer resources.

Load-bearing premise

The load-bearing premise is that the learned encoding trained on raw pixel values can actually be fed into the accelerator's datapath, which only carries a synapse index and a time difference, but the paper never specifies how a multi-bit pixel intensity becomes a spike time without an extra amplitude channel.

Editorial extensions

If this is right

  • If the central claim is right, an SNN accelerator can run a >99% MNIST classifier with no DSP slices, since all weights are ternary or powers of two and decay is a bit shift.
  • Learned encoding removes the usual fixed rate- or latency-encoding front end, so the input layer can be a small patch-based LIF layer rather than one neuron per pixel; with 9x9 patches this reduces the first layer from 784 to 400 neurons.
  • The differential-time Spike Sorter plus global Layer Controller means all neurons in a layer share one decay signal, so each neuron core costs only a register, a comparator, a multiplexer, and an adder.
  • At 3,400 images/s on a ZCU102, the design reaches sub-295-microsecond inference, placing it in the range of real-time edge applications despite being fully feedforward and LIF-only.
  • Since restricting weights to $\{-1,0,1\}$ loses almost nothing in accuracy compared with arbitrary weights, the network is relying more on which pixels fire than on their exact intensities, which justifies the multiplier-free constraint.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A testable extension the paper leaves implicit: the same learned-encoding idea could be applied to larger image datasets like CIFAR-10, where a patch-based LIF front end would need to show whether the accuracy gap to convolutional SNNs remains small without multipliers.
  • The paper reports the encoder was trained and evaluated offline; an inferred consequence is that loading the trained weights into the accelerator should reproduce 99.03% bit-exactly, which is directly checkable by co-simulation or on-board measurement.
  • The nearest-neighbor comparison cascade in the Spike Sorter is described for four inputs, but the same tournament structure should scale to any power-of-two synapse count with a short critical path; that scaling is a natural follow-up experiment.
  • Because the learned encoding resembles an attention-like selection of important pixels, one could probe what the first-layer weights have learned (e.g., stroke versus background regions) and use that as a regularizer for training on new datasets.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes an FPGA accelerator for feedforward SNNs using a differential-time spike representation, together with a learned patch-based input encoding in which first-layer weights are quantized to powers of two or to {-1, 0, 1}. The central claims are that the combined system uses only LIF neurons, requires no multiplications, surpasses 99% accuracy on MNIST, and fits on a ZCU102 FPGA with 56,520 LUTs, 91 BRAMs, 0 DSPs, and 3400 images/s. The architecture consists of a Spike Sorter, Neuron Cores with shift-based decay, a Leading-One Position Detector, and a Layer Controller. The evaluation reports averaged accuracy over 96 training runs for several patch sizes and a synthesis comparison with prior accelerators.

Significance. If the open interface question is resolved, this is a useful contribution: a multiplier-free, LIF-only SNN accelerator with a concrete resource/performance table is a valuable data point for edge neuromorphic inference, and letting the network learn which pixels matter during encoding is an attractive idea. The paper gives specific, checkable synthesis numbers and compares against prior FPGA designs with the same board, which is a strength. Its main limitation is not the empirical method itself but the incompletely specified bridge between the learned encoding layer and the differential-time hardware; until that bridge is described, the integrated claim is not fully demonstrated.

major comments (3)
  1. [Section III and Section IV, Figs. 2-3] The accelerator data path carries only a synapse index and a delta time: the Neuron Core adds the retrieved weight w_i for each incoming event, and there is no per-event amplitude or value field. The paper states in Section III that pixels are handed directly to the neuron without any preprocessing or encoding, and in Section IV that the encoding was done offline, but it never specifies how multi-bit MNIST pixel intensities are converted into the differential-time spike events that Fig. 2 and Fig. 3 can process. If pixel intensities are represented as spike times, the 'without any preprocessing' statement is false; if they are thresholded to binary events, intensity information is discarded and the reported 99.03% accuracy applies to a different input representation. This is the load-bearing gap between the learned encoding and the hardware implementation.
  2. [Section III, Fig. 5] The training protocol is under-specified: the paper does not give the number of simulation timesteps, the surrogate-gradient method, the loss function, the optimizer, the learning-rate schedule, the batch size, the number of epochs, or the standard deviation of accuracy over the 96 runs shown in Fig. 5. Without these details, the central accuracy claim of over 99% cannot be reproduced or critically assessed; please report the full protocol and the dispersion of the results.
  3. [Section I-A, Eq. (1)] The statement that setting theta = 1 and beta = 0.5 is 'without loss of generality' is only justified by a citation to [13]. This assumption is load-bearing because the shift-based decay in Section II-B and the multiplier-free hardware claim depend on beta being exactly 0.5. The paper should either reproduce the scaling argument that makes this restriction lossless or state explicitly any normalization constraints, rather than deferring to the prior work.
minor comments (5)
  1. [Table I] The column header 'Energy [W]' uses a power unit; rename it to 'Power [W]' or report true energy per inference.
  2. [Fig. 5] The figure shows averaged results over 96 runs but no error bars or confidence intervals; please add dispersion information, especially because some patch-size differences appear small.
  3. [Footnote 2] The accuracy comparison to [13] is described as equal by construction, so the table's accuracy column for [13] should be marked as a consistency check rather than an independent result; the resource and throughput comparison remains valid.
  4. [Section II-C] The phrase 'The spike inputs of a layer, consisting of a synapse index and delta time respectively, are generated in parallel' is unclear; clarify how the LOPD serialization interacts with the parallel spike generation.
  5. [Acknowledgment] There is a typo in the acknowledgment: 'in in the frame' should be 'in the frame'; please copyedit the text.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the MNIST accuracy and FPGA resource figures are empirical measurements or synthesis results, not derivations from fitted inputs.

full rationale

The paper's central claims are (i) a trained feedforward LIF SNN reaches 99.03% MNIST accuracy, and (ii) a differential-time FPGA accelerator implements it with reported resource and latency figures. Claim (i) is an empirical measurement: Section III describes training with quantization-aware weight restrictions over five patch sizes, repeated over 96 runs and averaged, with results shown in Fig. 5. It is not derived from a fitted parameter or from the hardware. Claim (ii) is a Vivado v2023.2 synthesis result (Section IV, Table I), benchmarked against external designs [24], [25] and the authors' prior architecture [13]. The accuracy column for [13] in Table I is explicitly stated in footnote 2 to equal the current design 'by definition' because the same network is run on the prior accelerator; this is a baseline comparison, not a predicted quantity, so it is not a circular derivation. The only self-citation carrying justificatory weight is [13] for the 'without loss of generality' setting of θ=1 and β=0.5; this is a mathematical normalization argument cited from the authors' prior work, and the paper's accuracy and synthesis results do not reduce to that citation. No step in the derivation chain equates a prediction with an input by construction. A genuine gap exists at the interface between the learned encoding and the hardware: Section III says pixels are 'handed directly to the neuron, without any preprocessing or encoding', while Section IV says 'encoding was done offline', and the hardware data path (Fig. 2/3) carries only a synapse index and a delta time, not pixel amplitudes. That missing pixel-to-spike mapping is an implementation and verification weakness, but it is not circular reasoning. Therefore the circularity score is 0.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central results rest on standard LIF dynamics, the authors' prior differential-time representation, and the unstated compatibility of the learned encoding with the spike-based hardware. No code or bitstream is provided, so the synthesis numbers are taken on faith.

free parameters (4)
  • Membrane decay rate beta = 0.5
    Set by hand; authors assert without loss of generality citing [13], but no derivation appears in this paper. The hardware implements decay as a bit shift, which relies on beta = 0.5.
  • Input patch size p = 9
    Chosen from a sweep over p in [5,9] shown in Fig. 5; the final network uses 9x9 patches, so this is a model selection parameter.
  • Weight quantization set = {-1,0,1}
    Of the two tested quantizations (powers of two and ternary), the ternary set was used for the synthesized design; the choice affects accuracy.
  • Network structure = Y-800-512-256-10 with Y = (28-p+1)^2
    Fixed by hand; no architecture search is reported.
assumptions (5)
  • domain assumption Leaky-Integrate-and-Fire dynamics with reset by subtraction, Eq. (1)
    The entire network and hardware are built on this neuron model; its validity for the claimed accuracy is assumed.
  • domain assumption Differential-time spike representation from [13] decouples spike time from processing time
    The architecture inherits this representation from the authors' prior work, which is cited but not fully re-derived.
  • ad hoc to paper Setting threshold theta = 1 and decay beta = 0.5 is without loss of generality
    Asserted in Section I-A with a reference to [13]; no proof is given in this paper, and the shift-based decay depends on beta = 0.5.
  • ad hoc to paper The offline learned encoding produces standard binary spike trains compatible with the accelerator
    Section IV says 'encoding was done offline', but the paper does not describe how continuous pixel values become differential-time spike events.
  • domain assumption Xilinx Vivado v2023.2 synthesis reports reflect physical hardware behavior
    Resource and timing numbers in Table I are based on synthesis, not on measurements from a fabricated or board-tested implementation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Spiking Neural Network Accelerator Architecture for Differential-Time Representation using Learned Encoding." pith.science (2026). https://pith.science/paper/7SA7XQGG

@misc{pith2026250107952,
  author       = {Pith},
  title        = {Pith review of: Spiking Neural Network Accelerator Architecture for Differential-Time Representation using Learned Encoding},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7SA7XQGG}},
  note         = {Machine review of arXiv:2501.07952}
}
read the original abstract

Spiking Neural Networks (SNNs) have garnered attention over recent years due to their increased energy efficiency and advantages in terms of operational complexity compared to traditional Artificial Neural Networks (ANNs). Two important questions when implementing SNNs are how to best encode existing data into spike trains and how to efficiently process these spike trains in hardware. This paper addresses both of these problems by incorporating the encoding into the learning process, thus allowing the network to learn the spike encoding alongside the weights. Furthermore, this paper proposes a hardware architecture based on a recently introduced differential-time representation for spike trains allowing decoupling of spike time and processing time. Together these contributions lead to a feedforward SNN using only Leaky-Integrate and Fire (LIF) neurons that surpasses 99% accuracy on the MNIST dataset while still being implementable on medium-sized FPGAs with inference times of less than 295us.

Figures

Figures reproduced from arXiv: 2501.07952 by the authors.

Figure 1
Figure 1. Hardware architecture of network with multiple neuron layers. Data [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. Neuron Core requiring only one register, one comparator, a multiplexer [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 4
Figure 4. Patch-wise selection of the pixels with subsequent neuron activation [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: Results of network evaluation on MNIST dataset with patch sizes [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Lightweight LIF-only SNN accelerator using differential time encoding

    cs.NE 2025-05 conditional novelty 4.0 of 10

    A LIF-only SNN accelerator using differential time encoding reports 99.03% MNIST accuracy on FPGA and ASIC with no multiplication operations.

Reference graph

Works this paper leans on

25 extracted references · 20 canonical work pages · cited by 1 Pith paper

  1. [13]

    SNN Architecture for Differential Time Encoding Using Decoupled Processing Time

    D. Windhager, B. A. Moser, and M. Lunglmayr, “Snn architecture for differential time encoding using decoupled processing time,” 2023. [Online]. Available: https://arxiv.org/abs/2311.14447

    work page Pith review arXiv 2023

  2. [1]

    MNIST Handwritten Digit Database,

    Y . LeCun, C. Cortes, and C. Burges, “MNIST Handwritten Digit Database,” 2010. [Online]. Available: http://yann.lecun.com/exdb/mnist

    work page 2010

  3. [2]

    The Heidelberg spiking datasets for the systematic evaluation of spiking neural networks

    B. Cramer, Y . Stradmann, J. Schemmel, and F. Zenke, “The heidelberg spiking datasets for the systematic evaluation of spiking neural networks,” CoRR, vol. abs/1910.07407, 2019. [Online]. Available: http://arxiv.org/abs/1910.07407

    work page Pith review arXiv 1910

  4. [3]

    Speech Commands: A Dataset for Limited-V ocabulary Speech Recognition,

    P. Warden, “Speech Commands: A Dataset for Limited-V ocabulary Speech Recognition,” ArXiv e-prints , Apr. 2018. [Online]. Available: https://arxiv.org/abs/1804.03209

    arXiv 2018

  5. [4]

    Darwin: A neuromorphic hardware co-processor based on spiking neural networks,

    D. Ma, J. Shen, Z. Gu, M. Zhang, X. Zhu, X. Xu, Q. Xu, Y . Shen, and G. Pan, “Darwin: A neuromorphic hardware co-processor based on spiking neural networks,” Journal of Systems Architecture , vol. 77, pp. 43–51, 2017

    work page 2017

  6. [5]

    Energy efficient parallel neuromorphic architectures with approximate arithmetic on fpga,

    Q. Wang, Y . Li, B. Shao, S. Dey, and P. Li, “Energy efficient parallel neuromorphic architectures with approximate arithmetic on fpga,” Neurocomputing, vol. 221, pp. 146–158, 2017. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S0925231216311213

    work page 2017

  7. [6]

    Advancing neuromorphic com- puting with loihi: A survey of results and outlook,

    M. Davies, A. Wild, G. Orchard, Y . Sandamirskaya, G. A. F. Guerra, P. Joshi, P. Plank, and S. R. Risbud, “Advancing neuromorphic com- puting with loihi: A survey of results and outlook,” Proceedings of the IEEE, vol. 109, no. 5, pp. 911–934, 2021

    2021

  8. [7]

    Scalable noc-based neuro- morphic hardware learning and inference,

    H. Fang, A. Shrestha, D. Ma, and Q. Qiu, “Scalable noc-based neuro- morphic hardware learning and inference,” in IJCNN, 2018, pp. 1–8

    work page 2018

Show all 25 references
  1. [8]

    Snava—a real-time multi-fpga multi-model spiking neural network simulation architecture,

    A. Sripad, G. Sanchez, M. Zapata, V . Pirrone, T. Dorta, S. Cambria, A. Marti, K. Krishnamourthy, and J. Madrenas, “Snava—a real-time multi-fpga multi-model spiking neural network simulation architecture,” Neural Networks , vol. 97, pp. 28–45, 2018

    2018

  2. [9]

    A Fast and Energy-Efficient SNN Processor With Adaptive Clock/Event-Driven Computation Scheme and Online Learning,

    S. Li, Z. Zhang, R. Mao, J. Xiao, L. Chang, and J. Zhou, “A Fast and Energy-Efficient SNN Processor With Adaptive Clock/Event-Driven Computation Scheme and Online Learning,” IEEE Trans. on Circuits and Systems I: Regular Papers , vol. 68, no. 4, pp. 1543–1552, 2021

    2021

  3. [10]

    Truenorth: Design and tool flow of a 65 mw 1 million neuron programmable neurosynaptic chip,

    F. Akopyan, J. Sawada, A. Cassidy, R. Alvarez-Icaza, J. Arthur, P. Merolla, N. Imam, Y . Nakamura, P. Datta, G.-J. Nam, B. Taba, M. Beakes, B. Brezzo, J. B. Kuang, R. Manohar, W. P. Risk, B. Jackson, and D. S. Modha, “Truenorth: Design and tool flow of a 65 mw 1 million neuron...

    2015

  4. [11]

    A low power and low latency fpga-based spiking neural network accelerator,

    H. Liu, Y . Chen, Z. Zeng, M. Zhang, and H. Qu, “A low power and low latency fpga-based spiking neural network accelerator,” in 2023 International Joint Conference on Neural Networks (IJCNN) , 2023, pp. 1–8

    2023

  5. [12]

    The spinnaker project,

    S. B. Furber, F. Galluppi, S. Temple, and L. A. Plana, “The spinnaker project,” Proceedings of the IEEE , vol. 102, no. 5, pp. 652–665, 2014

    2014

  6. [14]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. Kaiser, and I. Polosukhin, “Attention is all you need,” CoRR, vol. abs/1706.03762, 2017. [Online]. Available: http://arxiv.org/abs/1706.03762

    2017 arXiv

  7. [15]

    An optimized multi-layer spiking neural network implementation in fpga without multipliers,

    A. Mehrabi, Y . Bethi, A. van Schaik, and S. Afshar, “An optimized multi-layer spiking neural network implementation in fpga without multipliers,” Procedia Computer Science , vol. 222, pp. 407–414, 2023, international Neural Network Society Workshop on Deep Learning Innovation...

    2023

  8. [17]

    Spiker+: a framework for the generation of efficient spiking neural networks fpga accelerators for inference at the edge,

    ——, “Spiker+: a framework for the generation of efficient spiking neural networks fpga accelerators for inference at the edge,” 2024. [Online]. Available: https://arxiv.org/abs/2401.01141

    2024 arXiv

  9. [18]

    Minitaur, an event-driven fpga-based spiking network accelerator,

    D. Neil and S.-C. Liu, “Minitaur, an event-driven fpga-based spiking network accelerator,” IEEE Transactions on V ery Large Scale Integration (VLSI) Systems , vol. 22, no. 12, pp. 2621–2628, 2014

    2014

  10. [19]

    Fast and low-power leading-one detectors for energy-efficient logarithmic computing,

    M. S. Ansari, S. Gandhi, B. F. Cockburn, and J. Han, “Fast and low-power leading-one detectors for energy-efficient logarithmic computing,” IET Computers & Digital Techniques , vol. 15, no. 4, pp. 241–250, 2021. [Online]. Available: https://ietresearch.onlinelibrary. wiley.com...

    2021 doi

  11. [20]

    Leading one detectors and leading one position detectors - an evolutionary design methodology,

    K. Kunaraj and R. Seshasayanan, “Leading one detectors and leading one position detectors - an evolutionary design methodology,” Canadian Journal of Electrical and Computer Engineering , vol. 36, no. 3, pp. 103– 110, 2013

    2013

  12. [21]

    Vlsi implementations of low-power leading-one detector circuits,

    K. Abed and R. Siferd, “Vlsi implementations of low-power leading-one detector circuits,” in Proceedings of the IEEE SoutheastCon 2006 , 2006, pp. 279–284

    2006

  13. [22]

    Approximate leading one detector design for a hardware-efficient mitchell multiplier,

    S. Gandhi, M. S. Ansari, B. F. Cockburn, and J. Han, “Approximate leading one detector design for a hardware-efficient mitchell multiplier,” in 2019 IEEE Canadian Conference of Electrical and Computer Engi- neering (CCECE) , 2019, pp. 1–4

    2019

  14. [23]

    Training spiking neural networks using lessons from deep learning,

    J. K. Eshraghian, M. Ward, E. Neftci, X. Wang, G. Lenz, G. Dwivedi, M. Bennamoun, D. S. Jeong, and W. D. Lu, “Training spiking neural networks using lessons from deep learning,” Proceedings of the IEEE , vol. 111, no. 9, pp. 1016–1054, 2023

    2023

  15. [24]

    Syncnn: Evaluating and accel- erating spiking neural networks on fpgas,

    S. Panchapakesan, Z. Fang, and J. Li, “Syncnn: Evaluating and accel- erating spiking neural networks on fpgas,” in 2021 31st International Conference on Field-Programmable Logic and Applications (FPL) , 2021, pp. 286–293

    2021

  16. [25]

    Encoding, model, and architecture: Systematic optimization for spiking neural network in fpgas,

    H. Fang, Z. Mei, A. Shrestha, Z. Zhao, Y . Li, and Q. Qiu, “Encoding, model, and architecture: Systematic optimization for spiking neural network in fpgas,” in 2020 IEEE/ACM International Conference On Computer Aided Design (ICCAD) , 2020, pp. 1–9

    2020

  17. [2022]

    Available: https://arxiv.org/abs/2201.06993

    [Online]. Available: https://arxiv.org/abs/2201.06993

    arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.