Pith. sign in

REVIEW 3 major objections 6 minor 37 references

Reservoir computing with all-optical non-fading memory in a self-pulsing microresonator network

T0 review · 3 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read A passive silicon photonic network stores spike timing for at least 75 microseconds.

desk verdict Genuinely new experimental demonstration of non-fading memory in a 64-ring photonic reservoir, but the reported scores are likely inflated by test-set selection across 1500+ configurations. read the letter →

arxiv 2411.17272 v1 pith:3XRO4MTW submitted 2024-11-26 physics.optics

Alessio Lugnan , Stefano Biasi , Alessandro Foradori , Peter Bienstman , Lorenzo Pavesi This is my paper · ORCID
classification physics.optics
keywords reservoircomputingphotonicneuromorphicmicroresonatornetworkself-pulsingnon-fadingmemorysiliconphotonicsmultistabilityopticalsignalprocessing
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

This paper reports an experimental photonic reservoir computer built from 64 coupled silicon microring resonators that stores information about its optical input for tens of microseconds, far longer than the native carrier and thermal lifetimes of silicon. The central claim is that the network's self-pulsing dynamics and multistability give it non-fading memory, so a linear readout can infer the timing of a single input pulse, or the rate of a pulse train, from the network state long after the input has ended. This matters because many optical sensors produce signals on millisecond or slower timescales, while photonic processors typically operate in microseconds; a reservoir that retains input information bridges that gap without optical-to-electrical conversion. The authors demonstrate the memory at two timescales differing by a factor of about five, and report that spike timing information was still readable at least 75 microseconds after the perturbation.

What carries the argument

A microring resonator (MRR) is a ring waveguide coupled to bus waveguides whose resonance wavelength shifts when light generates free carriers and heat. In a network of 64 coupled MRRs, two-photon absorption, free-carrier absorption and dispersion, and the thermo-optic effect act with different lifetimes (carriers about 1-45 ns, thermal about 60-280 ns), producing self-pulsing oscillations and multistable states. A single strong input drives the whole network into this nonlinear regime; the perturbation switches the network into a neighboring stable dynamical state that persists for the rest of the nonlinear stage. The readout is a linear regressor or classifier applied to the final samples of the output waveform, after choosing among four downsampling ratios; the reservoir's role is to expand the input into a high-dimensional, temporally persistent state that the linear readout can map to the target.

What would settle it

Measure the output after the pump power is dropped and check whether it returns to the pre-stage linear-regime waveform before the next nonlinear stage begins; if the residual differs, or if classification scores drop sharply when the low-power reset interval is lengthened or when the sample order is not randomized, the non-fading memory claim would be weakened.

Watch

Extended reading notes

Core claim

The discovery is the first experimental demonstration of physical reservoir computing with all-optical non-fading memory. A constant pump drives the 64-ring network into a self-pulsing state; a short perturbation (a single pulse or a spike train) durably alters that dynamical state, and the altered state can be read out at the end of a nonlinear stage, well after the perturbation ends. Linear regression and logistic-regression readouts on the output waveforms recover the perturbation's timing or rate. In the timing task the information was retained for at least 75 microseconds, with classification accuracy close to 100 percent at both short (20 microsecond) and long (100 microsecond) nonlinear-stage timescales; in the rate task with randomized perturbation start, the reservoir state carried rate information for at least 40 microseconds. The baseline without the reservoir was near zero for the timing task, attributing the performance to the reservoir rather than to direct signal transmission.

Load-bearing premise

The network fully resets to the linear regime between nonlinear stages, so each training sample starts from the same state and the reported scores come from memory within a single excitation rather than from correlations between neighboring samples.

Editorial extensions

If this is right

  • Spike timing information is present in the network state at least 75 microseconds after the pulse ends, so a slow photodetector with sub-MHz bandwidth can read the processed result.
  • The same hardware solves regression and classification for both pulse timing and pulse-train rate, at two timescales about a factor of five apart, by tuning input wavelength and power.
  • Because the baseline score is near zero for the timing task, the memory lives in the reservoir dynamics rather than in the input waveform itself.
  • The long-timescale rate task requires randomizing the perturbation start; otherwise the readout can exploit when the pulse train ends rather than its spike rate.
  • The approach enables all-optical preprocessing for fiber-optic sensors, where pulse timing maps to sensor position and the readout delay can be tens of microseconds.

Reading between the lines

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

  • A direct test of the reset assumption would be to vary the duration of the low-power interval between nonlinear stages; if scores degrade when the interval is shortened, part of the reported memory may come from correlations between neighboring samples rather than from within-excitation storage.
  • The demonstrated 75-microsecond retention suggests that distributed fiber sensing, where pulse timing encodes location, could work with a sensing range set by the readout delay rather than by the native silicon carrier lifetime.
  • Multi-wavelength excitation could let several independent reservoirs share one chip, multiplying the readout feature space without increasing footprint.
  • If the memory is truly non-fading within the nonlinear stage, lengthening the pump stage should extend the readout delay well beyond 100 microseconds, with the practical limit set by thermal drift and the reset requirement.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The manuscript reports an experimental photonic reservoir computer built from a fully passive 8×8 network of 64 coupled silicon microring resonators, driven into self-pulsing and multistable nonlinear dynamics by a continuous-wave input. The authors demonstrate that a linear readout can infer the timing of a single input pulse and the rate of a pulse train from the network output sampled well after the perturbation has ended, at two timescales differing by about a factor of five. The central claim is that the reservoir exhibits non-fading, all-optical memory for tens of microseconds, enabling physical reservoir computing on signals whose timescales are much slower than the intrinsic photonic response. The paper includes baseline measurements with a non-resonant input, a random-start variant to remove a shortcut for the rate task, and cross-validated readout training.

Significance. If the quantitative claims survive scrutiny, this is an important experimental step: it demonstrates that a compact, CMOS-compatible, fully passive photonic network can store input information for timescales far beyond the intrinsic carrier and thermal lifetimes of silicon, and it shows how to use that memory for practical signal-processing tasks such as timing and rate inference. The authors provide useful controls, including a non-resonant baseline and a randomized perturbation-start condition, and they are transparent about the preprocessing and training pipeline. The main weakness is that the reported scores are selected as the best over a large grid of physical configurations, with no independent validation set, which inflates the apparent performance and weakens the quantitative memory-duration claim. The reproducibility of the study is also limited by the unavailability of code and data at the time of review.

major comments (3)
  1. [Section 4.2 (last paragraph) and Fig. 5] The test scores in Fig. 5 are maxima over 1520 configurations (4 downsampling ratios × 19 wavelengths × 20 powers), selected on the test set itself. This procedure is explicitly described in Section 4.2: 'The test scores presented in Fig. 5 represent the best results selected from this parameter space.' Selecting the maximum of 1520 noisy test estimates is substantially optimistically biased; for a null effect, the expected maximum can be many standard errors above the true value. Since the baseline is a single non-resonant configuration, the 'best reservoir vs one baseline' comparison does not establish that the reservoir outperforms the baseline fairly. The manuscript should either use a nested train/validation/test split for configuration selection, or report the full distribution of scores over the 1520 configurations (e.g., percentiles, or the score for a pre-specified or random configuration), together with the non-resonant baseline evaluated under the same selection procedure.
  2. [Section 2.3, Fig. 5b] The quantitative claims that spike timing information was 'stored in the photonic network state for at least 75 µs' and that classification accuracies are 'close to 100%' are based on the best selected configuration from the 1520-configuration grid. Because of the test-set selection, these numerical values are not reliable as unbiased estimates of the memory duration or achievable accuracy. The authors should provide unbiased estimates (for example, by evaluating a fixed configuration chosen on a validation set, or by aggregating scores across configurations) and give confidence intervals that account for the selection procedure. Alternatively, they should explicitly reframe the reported numbers as upper bounds or selected maxima rather than typical or unbiased performance.
  3. [Section 2.1, Step 4] Step 4 states that lowering the input power 'leaves the network in the linear regime for a long enough time to reset its memory,' but no measurement verifies the reset duration or that each NL stage starts from the same network state. The paper does not report the length of the linear-regime interval between NL stages, nor any control showing that the pre-NL-stage state is independent of the previous perturbation label. Although the randomization of sample order makes cross-sample leakage a noise source rather than a systematic shortcut, the interpretation that memory is non-fading within a single excitation but resets between excitations requires explicit support. Please report the linear-regime duration and, ideally, a control such as the correlation between the output at the start of an NL stage and the previous perturbation label.
minor comments (6)
  1. [Fig. 5 caption] The word 'respectivley' should be 'respectively'.
  2. [Section 2.3] The claim that memory persists 'for at least 75 µs' would be clearer if the calculation were spelled out: for the long timescale, the NL stage is 100 µs, the latest perturbation ends at 19.8 µs, and the readout interval is the last 5.12 µs, so the gap between the end of the perturbation and the start of the readout is approximately 75 µs. Please state this explicitly.
  3. [Fig. 3] The markers indicating optimum configurations are not keyed to the specific ML tasks in the main text or the figure caption; a table or a more descriptive legend would help the reader connect the marked operating points to Fig. 5.
  4. [Section 2.2] In the description of the '1 feature per port' variation, the phrase 'averaging the readout over time' should specify that the average is taken over the readout interval only, not over the entire NL stage.
  5. [References] Reference [25] is a preprint; if a peer-reviewed version exists, the authors should cite it, or at least note that the non-fading memory effect relies on this unpublished work.
  6. [Code availability] The code availability statement says the code will be uploaded to Zenodo after publication; providing the code and data at revision time would strengthen reproducibility and make the selection-bias analysis easier for reviewers to verify.

Circularity Check

0 steps flagged · score 1.0 of 10

Experimental demonstration is self-contained; test-set selection is a statistical concern, not circularity.

full rationale

This paper is an experimental demonstration, not a formal derivation, so the circularity tests that apply to fitted parameters or self-citation chains have limited purchase. The central claim—non-fading memory enabling spike-timing and spike-rate inference after tens of microseconds—is supported by direct measurements of the 64-MRR network output, by an out-of-resonance baseline, and by random-perturbation-start controls. The authors' prior work [25] is cited as motivation and as supporting evidence for the underlying multistability/spike-rate sensitivity, but the present experiments independently re-measure the effect in a larger network, so the self-citation is not load-bearing in the sense of forcing the conclusion. The only notable methodological weakness is the explicit selection of the best test scores over the full wavelength/power/downsampling parameter space without a separate validation set; this can inflate the reported quantitative scores and weakens the strength of the 'at least 75 µs' quantitative claim. This is a statistical model-selection bias, not a circular reduction: the reported score is the maximum of many test estimates rather than a prediction forced by construction, and the qualitative separation from baselines and random-start controls remains meaningful. No equation in the paper defines a target in terms of an input, and no fitted parameter is renamed as a prediction. Therefore no significant circularity is present.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The paper introduces no new physical entities. Its central claim rests on known MRR nonlinear dynamics, an assumed reset between samples, and the sufficiency of six output ports. The free parameters are experimental operating points and ML hyperparameters selected to maximize performance.

free parameters (3)
  • Reservoir operating point (input frequency and power) = 19 frequencies from 192.68 to 192.86 THz, 20 powers from about 5.5 to 16.4 mW; best per task selected from these 380…
    The best input frequency and power for each ML task were chosen based on test performance, making this a free parameter search that can inflate reported scores.
  • Readout downsampling ratio = Chosen from {400, 40, 20, 10} for short timescales and {1600, 160, 80, 20} for long timescales
    The best downsampling ratio per measurement was selected based on validation performance; this is a standard ML hyperparameter but contributes to the reported best scores.
  • L2 regularization strength = One of 8 values from 1e-8 to 1e-1
    Regularization strength was selected by inner cross-validation; this is a standard hyperparameter and not central to the physical claim.
assumptions (3)
  • domain assumption Silicon MRR nonlinear dynamics (TPA, free-carrier effects, thermo-optic effect) follow the standard coupled-mode models from refs. 7 to 12.
    The paper explains self-pulsing and multistability using known silicon photonics nonlinearities and does not derive a new model.
  • domain assumption Each nonlinear stage starts from the same reset linear state after the input power is lowered.
    Dataset validity requires independent samples; the reset is stated in Section 2.1, Step 4, but not directly verified.
  • domain assumption The six output ports provide a sufficient projection of the reservoir state for the linear readout.
    The ML results depend on this projection; no state-space analysis is provided to justify completeness.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Reservoir computing with all-optical non-fading memory in a self-pulsing microresonator network." pith.science (2026). https://pith.science/paper/3XRO4MTW

@misc{pith2026241117272,
  author       = {Pith},
  title        = {Pith review of: Reservoir computing with all-optical non-fading memory in a self-pulsing microresonator network},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3XRO4MTW}},
  note         = {Machine review of arXiv:2411.17272}
}
read the original abstract

Photonic neuromorphic computing may offer promising applications for a broad range of photonic sensors, including optical fiber sensors, to enhance their functionality while avoiding loss of information, energy consumption, and latency due to optical-electrical conversion. However, time-dependent sensor signals usually exhibit much slower timescales than photonic processors, which also generally lack energy-efficient long-term memory. To address this, we experimentally demonstrate a first implementation of physical reservoir computing with non-fading memory for multi-timescale signal processing. This is based on a fully passive network of 64 coupled silicon microring resonators. Our compact photonic reservoir is capable of hosting energy-efficient nonlinear dynamics and multistability. It can process and retain input signal information for an extended duration, at least tens of microseconds. Our reservoir computing system can learn to infer the timing of a single input pulse and the spike rate of an input spike train, even after a relatively long period following the end of the input excitation. We demonstrate this operation at two different timescales, with approximately a factor of 5 difference. This work presents a novel approach to extending the memory of photonic reservoir computing and its timescale of application.

Figures

Figures reproduced from arXiv: 2411.17272 by the authors.

Figure 1
Figure 1. Relevant previous works. Here we summarize previous works that inspired this research, with a short description and a sketch about the most relevant content. We divide these works into three groups, corresponding to three ways of considering nonlinear MRRs for neuromorphic computing. for the measurements and the considered ML tasks ( [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Working principles of photonic RC with non-fading memory. a Our reservoir (schematics at the center) consists of a network of coupled silicon MRRs, whose resonance is centered at different wavelengths due to fabrication errors (see plots on the top, showing MRR stored energy v.s. wavelength for few MRRs). Moreover, if the input laser power is high enough (milliwatts), the resonance wavelength of the excited MRRs is … view at source ↗
Figure 3
Figure 3. Self-pulsing frequency as a function of laser frequency and power. a Color-maps of the SP frequency (i.e. inverse of the fundamental period of the SP waveform) at the different output ports. Ten different markers on the maps (see the legends below the maps) indicate the input parameter combinations that optimize the performance of a specific ML task (described in Sections 2.2 and 2.3). In the legend, ‘t.s.’ and ‘r.b… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Input waveform types, corresponding to different ML tasks. a Example input waveform segment where the NL stages are perturbed by a single pulse at different timings w.r.t. the NL stage start. (The unperturbed NL stage in the middle was employed as reference and can be …
Figure 5
Figure 5. Figure 5: ML performances. a Examples of the linear regression test prediction (light-blue dots) versus the actual ML target (orange lines). These examples correspond to the score values displayed in the bar plots of b or c, as indicated by gray arrows. b, c, d, e Bar plots of t…
Figure 6
Figure 6. Figure 6: Experimental setup. Laser light (wavelength around 1550 nm) is modulated into the desired waveform and injected into the on-chip photonic network. The input wavelength is selected by using a tunable continuous wave laser, while the input power is set via a variable opt…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 33 canonical work pages

  1. [1]

    Photonics for artificial intelligence and neuromorphic computing

    Bhavin J Shastri, Alexander N Tait, Thomas Ferreira de Lima, Wolfram HP Pernice, Harish Bhaskaran, C David Wright, and Paul R Prucnal. Photonics for artificial intelligence and neuromorphic computing. Nature Photonics, 15(2):102–114, 2021

    2021

  2. [2]

    Special session: Neuromorphic hardware design and reliability from traditional cmos to emerging technologies

    Fabio Pavanello, Elena Ioana Vatajelu, Alberto Bosio, Thomas Van Vaerenbergh, Peter Bienstman, Benoit Charbonnier, Alessio Carpegna, Stefano Di Carlo, and Alessandro Savino. Special session: Neuromorphic hardware design and reliability from traditional cmos to emerging technologies. In 2023 IEEE 41st VLSI Test Symposium (VTS), pages 1–10. IEEE, 2023

    work page 2023

  3. [3]

    Reconfigurable neuromorphic computing: Materials, devices and integration

    Minyi Xu, Xinrui Chen, Yehao Guo, Yang Wang, Dong Qiu, Xinchuan Du, Yi Cui, Xianfu Wang, and Jie Xiong. Reconfigurable neuromorphic computing: Materials, devices and integration. Advanced Materials, page 2301063, 2023

    work page 2023

  4. [4]

    2022 roadmap on neuromorphic computing and engineering

    Dennis V Christensen, Regina Dittmann, Bernabe Linares-Barranco, Abu Sebastian, Manuel Le Gallo, An- drea Redaelli, Stefan Slesazeck, Thomas Mikolajick, Sabina Spiga, Stephan Menzel, et al. 2022 roadmap on neuromorphic computing and engineering. Neuromorphic Computing and Engineering, 2(2):022501, 2022

    work page 2022

  5. [5]

    Adaptive extreme edge computing for wearable devices

    Erika Covi, Elisa Donati, Xiangpeng Liang, David Kappel, Hadi Heidari, Melika Payvand, and Wei Wang. Adaptive extreme edge computing for wearable devices. Frontiers in Neuroscience, 15:611300, 2021

    work page 2021

  6. [6]

    Silicon microring resonators

    Wim Bogaerts, Peter De Heyn, Thomas Van Vaerenbergh, Katrien De V os, Shankar Kumar Selvaraja, Tom Claes, Pieter Dumon, Peter Bienstman, Dries Van Thourhout, and Roel Baets. Silicon microring resonators. Laser & Photonics Reviews, 6(1):47–73, 2012

    work page 2012

  7. [7]

    Nonlinear silicon photonics

    Juerg Leuthold, Christian Koos, and Wolfgang Freude. Nonlinear silicon photonics. Nature photonics, 4(8):535– 544, 2010

    work page 2010

  8. [8]

    Nonlinear silicon photonics

    Massimo Borghi, Claudio Castellan, Stefano Signorini, Alessandro Trenti, and Lorenzo Pavesi. Nonlinear silicon photonics. Journal of Optics, 19(9):093002, 2017

    work page 2017

Show all 37 references
  1. [9]

    Johnson, Matthew Borselli, and Oskar Painter

    Thomas J. Johnson, Matthew Borselli, and Oskar Painter. Self-induced optical modulation of the transmission through a high-q silicon microdisk resonator. Opt. Express, 14(2):817–831, Jan 2006

    2006

  2. [10]

    On the effect of the thermal cross-talk in a photonic feed-forward neural network based on silicon microresonators

    Stefano Biasi, Riccardo Franchi, Davide Bazzanella, and Lorenzo Pavesi. On the effect of the thermal cross-talk in a photonic feed-forward neural network based on silicon microresonators. Frontiers in Physics, 10:1093191, 2022

    2022

  3. [11]

    Simplified description of self- pulsation and excitability by thermal and free-carrier effects in semiconductor microcavities

    Thomas Van Vaerenbergh, Martin Fiers, Joni Dambre, and Peter Bienstman. Simplified description of self- pulsation and excitability by thermal and free-carrier effects in semiconductor microcavities. Phys. Rev. A, 86:063808, Dec 2012

    2012

  4. [12]

    On the modeling of thermal and free carrier nonlinearities in silicon-on-insulator microring resonators

    Massimo Borghi, Davide Bazzanella, Mattia Mancinelli, and Lorenzo Pavesi. On the modeling of thermal and free carrier nonlinearities in silicon-on-insulator microring resonators. Optics Express, 29(3):4363–4377, 2021

    2021

  5. [13]

    Optical bistability and pulsating behaviour in silicon-on-insulator ring resonator structures

    Gino Priem, Pieter Dumon, Walter Bogaerts, Dries Van Thourhout, Geert Morthier, and Roel Baets. Optical bistability and pulsating behaviour in silicon-on-insulator ring resonator structures. Optics express, 13(23):9623– 9628, 2005. 12 RESEARCH ARTICLE

    2005

  6. [14]

    Thirty years in silicon photonics: a personal view

    Lorenzo Pavesi. Thirty years in silicon photonics: a personal view. Frontiers in Physics, page 709, 2021

    2021

  7. [15]

    Chaotic dynamics in coupled resonator sequences

    Mattia Mancinelli, Massimo Borghi, Fernando Ramiro-Manzano, JM Fedeli, and Lorenzo Pavesi. Chaotic dynamics in coupled resonator sequences. Optics express, 22(12):14505–14516, 2014

    2014

  8. [16]

    Cascadable excitability in microrings

    Thomas Van Vaerenbergh, Martin Fiers, Pauline Mechet, Thijs Spuesens, Rajesh Kumar, Geert Morthier, Ben- jamin Schrauwen, Joni Dambre, and Peter Bienstman. Cascadable excitability in microrings. Optics express, 20(18):20292–20308, 2012

    2012

  9. [17]

    All-optical spiking neuron based on passive microresonator

    Jinlong Xiang, Axel Torchy, Xuhan Guo, and Yikai Su. All-optical spiking neuron based on passive microresonator. Journal of Lightwave Technology, 38(15):4019–4029, 2020

    2020

  10. [18]

    All-optical silicon microring spiking neuron

    Jinlong Xiang, Yujia Zhang, Yaotian Zhao, Xuhan Guo, and Yikai Su. All-optical silicon microring spiking neuron. Photonics Research, 10(4):939–946, 2022

    2022

  11. [19]

    Rigorous dynamic model of a silicon ring resonator with phase change material for a neuromorphic node

    Alessio Lugnan, Santiago García-Cuevas Carrillo, C David Wright, and Peter Bienstman. Rigorous dynamic model of a silicon ring resonator with phase change material for a neuromorphic node. Optics Express, 30(14):25177– 25194, 2022

    2022

  12. [20]

    On-chip spiking neural networks based on add-drop ring microresonators and electrically reconfigurable phase-change material photonic switches

    Qiang Zhang, Ning Jiang, Yiqun Zhang, Anran Li, Huanhuan Xiong, Gang Hu, Yongsheng Cao, and Kun Qiu. On-chip spiking neural networks based on add-drop ring microresonators and electrically reconfigurable phase-change material photonic switches. Photonics Research, 12(4):755–766, 2024

    2024

  13. [21]

    The taiji microresonator as an unidirectional spiking neuron

    Stefano Biasi, Alessandro Foradori, Riccardo Franchi, Alessio Lugnan, Peter Bienstman, and Lorenzo Pavesi. The taiji microresonator as an unidirectional spiking neuron. arXiv preprint arXiv:2410.03257, 2024

    2024 arXiv

  14. [22]

    Photonic neural networks based on integrated silicon microresonators

    Stefano Biasi, Giovanni Donati, Alessio Lugnan, Mattia Mancinelli, Emiliano Staffoli, and Lorenzo Pavesi. Photonic neural networks based on integrated silicon microresonators. Intelligent Computing, 3:0067, 2024

    2024

  15. [23]

    An array of microresonators as a photonic extreme learning machine

    Stefano Biasi, Riccardo Franchi, Lorenzo Cerini, and Lorenzo Pavesi. An array of microresonators as a photonic extreme learning machine. APL Photonics, 8(9), 2023

    2023

  16. [24]

    Emergent self-adaptation in an integrated photonic neural network for backpropagation-free learning

    Alessio Lugnan, Samarth Aggarwal, Frank Brückerhoff-Plückelmann, C David Wright, Wolfram HP Pernice, Harish Bhaskaran, and Peter Bienstman. Emergent self-adaptation in an integrated photonic neural network for backpropagation-free learning. arXiv preprint arXiv:2312.03802, 2023

    2023 arXiv

  17. [25]

    Exploring the potential of self-pulsing optical microresonators for spiking neural networks and event detection

    Stefano Biasi, Alessio Lugnan, Davide Micheli, and Lorenzo Pavesi. Exploring the potential of self-pulsing optical microresonators for spiking neural networks and event detection. PREPRINT available at Research Square, 2024

    2024

  18. [26]

    echo state

    Herbert Jaeger. The “echo state” approach to analysing and training recurrent neural networks-with an erratum note. Bonn, Germany: German National Research Center for Information Technology GMD Technical Report, 148(34):13, 2001

    2001

  19. [27]

    Hands-on reservoir computing: a tutorial for practical implementation

    Matteo Cucchi, Steven Abreu, Giuseppe Ciccone, Daniel Brunner, and Hans Kleemann. Hands-on reservoir computing: a tutorial for practical implementation. Neuromorphic Computing and Engineering, 2(3):032002, 2022

    2022

  20. [28]

    Pattern recognition in a bucket

    Chrisantha Fernando and Sampsa Sojakka. Pattern recognition in a bucket. In European conference on artificial life, pages 588–597. Springer, 2003

    2003

  21. [29]

    Recent advances in physical reservoir computing: A review

    Gouhei Tanaka, Toshiyuki Yamane, Jean Benoit Héroux, Ryosho Nakane, Naoki Kanazawa, Seiji Takeda, Hidetoshi Numata, Daiju Nakano, and Akira Hirose. Recent advances in physical reservoir computing: A review. Neural Networks, 115:100–123, 2019

    2019

  22. [30]

    Photonic neuromorphic information processing and reservoir computing

    Alessio Lugnan, Andrew Katumba, Floris Laporte, Matthias Freiberger, Stijn Sackesyn, Chonghuai Ma, Emmanuel Gooskens, Joni Dambre, and Peter Bienstman. Photonic neuromorphic information processing and reservoir computing. APL Photonics, 5(2), 2020

    2020

  23. [31]

    Dimensions of timescales in neuromorphic computing systems

    Herbert Jaeger, Dirk Doorakkers, Celestine Lawrence, and Giacomo Indiveri. Dimensions of timescales in neuromorphic computing systems. arXiv preprint arXiv:2102.10648, 2021

    2021 arXiv

  24. [32]

    Emerging opportunities and challenges for the future of reservoir computing

    Min Yan, Can Huang, Peter Bienstman, Peter Tino, Wei Lin, and Jie Sun. Emerging opportunities and challenges for the future of reservoir computing. Nature Communications, 15(1):2056, 2024

    2024

  25. [33]

    Distributed optical fiber sensing: Review and perspective

    Ping Lu, Nageswara Lalam, Mudabbir Badar, Bo Liu, Benjamin T Chorpening, Michael P Buric, and Paul R Ohodnicki. Distributed optical fiber sensing: Review and perspective. Applied Physics Reviews, 6(4), 2019

    2019

  26. [34]

    Large-scale neural network in passive silicon photonics for biologically plausible learning

    Alessio Lugnan, Alessandro Foradori, Stefano Biasi, Peter Bienstman, and Lorenzo Pavesi. Large-scale neural network in passive silicon photonics for biologically plausible learning. In Machine Learning in Photonics , volume 13017, pages 127–131. SPIE, 2024

    2024

  27. [35]

    Micro ring resonators as building blocks for an all-optical high-speed reservoir-computing bit-pattern-recognition system

    Charis Mesaritakis, Vassilis Papataxiarhis, and Dimitris Syvridis. Micro ring resonators as building blocks for an all-optical high-speed reservoir-computing bit-pattern-recognition system. JOSA B, 30(11):3048–3055, 2013. 13 RESEARCH ARTICLE

    2013

  28. [36]

    Scikit-learn: Machine learning in python

    Fabian Pedregosa, Gaël Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, et al. Scikit-learn: Machine learning in python. the Journal of machine Learning research, 12:2825–2830, 2011

    2011

  29. [37]

    Shortcut learning in deep neural networks

    Robert Geirhos, Jörn-Henrik Jacobsen, Claudio Michaelis, Richard Zemel, Wieland Brendel, Matthias Bethge, and Felix A Wichmann. Shortcut learning in deep neural networks. Nature Machine Intelligence, 2(11):665–673, 2020. 7 Acknowledgments This project has received funding from...

    2020

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.