Pith. sign in

REVIEW 4 major objections 4 minor 59 references

Online Learning of Modular Bayesian Deep Receivers: Single-Step Adaptation with Streaming Data

T0 review · 4 major / 4 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read Deep wireless receivers can adapt to fading channels with a single Bayesian update per pilot block, replacing multi-epoch stochastic gradient training.

desk verdict A solid engineering contribution that reframes known Bayesian online-learning filters as modular, pipelined streaming updates for DeepSIC and shows real latency/BER gains on smooth time-varying channels; the main soft spot is sensitivity to the Gaussian Markov weight model, not the core mechanism. read the letter →

arxiv 2511.06045 v2 pith:Y74YWF5B submitted 2025-11-08 eess.SP cs.ITmath.IT

Yakov Gusakov , Osvaldo Simeone , Tirza Routtenberg , Nir Shlezinger This is my paper · ORCID
classification eess.SPcs.ITmath.IT
keywords onlinelearningBayesiandeepextendedKalmanfiltermodularreceiverarchitecturewirelessMIMOdetectionstreamingdataadaptationsingle-steptrainingnaturalgradient
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 aims to show that a deep neural network used as a wireless receiver can adapt to a rapidly changing channel with a single update per incoming pilot block, instead of the many passes over data that ordinary stochastic gradient training requires. The authors model the network weights themselves as a hidden state that drifts smoothly over time, so pilot symbols become noisy measurements of that state. A recursive Bayesian filter—essentially an extended Kalman filter—produces a new weight distribution in one step, and because the receiver is built from small modular subnetworks, each module can be filtered independently and in parallel. If correct, this makes on-device adaptation to fading, nonlinear channels practical under strict latency budgets.

What carries the argument

The central mechanism is a nonlinear state-space model in parameter space: the desired receiver weights θ_t follow p(θ_t|θ_{t-1}) = N(γθ_{t-1}, σ²I), and the pilot bits form a Bernoulli observation whose means are the network outputs. Recursive Bayesian filtering—specifically the conditional-moments extended Kalman filter (CM-EKF)—linearizes the network mapping at the predicted mean, producing a one-step Gaussian posterior update. For scalability, the covariance is restricted (VD-EKF for diagonal, Lo-Fi for low-rank), and the modular architecture DeepSIC (unfolded soft interference cancellation) allows each per-user module to run its own filter in parallel, cutting complexity from O(BQ²K³P²)

What would settle it

Take the same modular DeepSIC receiver and compare CM-EKF against GD-10 on a channel that switches abruptly between two very different fading states (e.g., independent per block, jumping every few pilots). If the Kalman filter's BER during tracking is no better than random re-initialization while SGD eventually recovers, the smooth-drift premise is refuted. Alternatively, sweep γ and σ² over orders of magnitude and see whether performance falls off sharply, indicating sensitivity to un-tuned hyperparameters.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that online training of a deep receiver can be recast as a state-space tracking problem—treating the desired DNN weights as a first-order Gaussian Markov process—so that each new pilot observation yields a closed-form Bayesian update of a Gaussian variational posterior (the conditional-moments extended Kalman filter, or CM-EKF). In the linearized-Gaussian limit this update coincides with a Bayesian online natural-gradient step, and the same update can be restricted to diagonal or low-rank covariances to reduce cost. Combined with the modular DeepSIC architecture, which unfolds soft interference cancellation into per-user subnetworks, each module

Load-bearing premise

The paper assumes the optimal receiver weights drift as a smoothly varying first-order Gaussian process with fixed constants γ and σ²; if real channel variations are abrupt or heavy-tailed, this model is misspecified and the filtering advantage over multi-epoch SGD may vanish.

Editorial extensions

If this is right

  • Single-step updates replace multi-epoch training, reducing adaptation latency to sub-millisecond per sample for modular receivers.
  • Module-wise filtering lets full-covariance Kalman updates run on compact subnetworks where they would be infeasible on a monolithic network.
  • Lo-Fi sits between diagonal and full-covariance filtering, offering a complexity-performance trade-off with linear-in-parameters cost.
  • The linearized-Gaussian BONG update reduces to CM-EKF, providing an ELBO-based justification for the same update rule.
  • The framework handles nonlinear channel distortions better than iterative SGD in the paper's tested synthetic and realistic scenarios.

Reading between the lines

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

  • Because the drift model is a fixed Gaussian Markov chain with user-chosen γ and σ², the method is likely to degrade on channels with abrupt, discontinuous changes (e.g., independent block fading); an adaptive or heavy-tailed drift prior would be a natural extension the paper does not explore.
  • The one-step advantage depends on pilots arriving at a rate at least comparable to the channel's coherence time; if pilots are too sparse, even a perfect filter cannot track, so the headline result implicitly assumes a pilot-density constraint.
  • The same state-space-plus-modular-filter recipe could transfer to other streaming model-based deep learning tasks with a natural modular decomposition, such as channel estimation or joint source-channel decoding.
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

4 major / 4 minor

Summary. The paper proposes an online learning framework for deep neural network receivers in time-varying wireless channels. It models the evolution of the desired receiver parameters as a first-order Gaussian Markov process and applies recursive Bayesian filtering—specifically the conditional-moments EKF and the Bayesian online natural gradient (BONG) with diagonal/low-rank covariance approximations (VD-EKF, Lo-Fi)—to perform a single parameter update per received pilot. The method is combined with the modular DeepSIC architecture, enabling module-wise updates, parallelization, and pipelining. Simulations on a synthetic rotation channel, Sionna, COST2100, and QuaDRiGa compare the proposed methods with SGD, GD, Bayes-by-backprop, NLMS, and MAML baselines, reporting BER and per-sample latency.

Significance. If validated, the contribution is significant: it offers a principled alternative to multi-epoch SGD for adapting deep receivers in dynamic channels, with a clear state-space formulation, open-source code, and comparisons against MAP/MMSE and several learning baselines. The central derivation is not circular; the EKF/BONG update equations follow from the assumed state-space model. The main deficiencies are empirical: the state-evolution hyperparameters are not reported, error bars are absent despite small trial counts, and the latency advantage is claimed more broadly than the data support.

major comments (4)
  1. [III-A, Eq. (13); IV-A] The state-evolution hyperparameters gamma and sigma^2 are load-bearing but not reported in the experimental section. With gamma<1 the prior pulls the weights toward zero and sigma^2 sets the effective step size of every filter update; the reported BER advantages of CM-EKF over SGD cannot be assessed without these values. Footnote 2 gives a code link, but a journal paper should be self-contained. Please report gamma, sigma^2, mu_0, and Sigma_0 for each experiment and include a sensitivity study over sigma^2 (and gamma) to show that the conclusions are not an artifact of a particular tuning.
  2. [IV-B, Table II] The claimed latency advantage is not supported for the full-covariance method. In Table II, CM-EKF on DeepSIC costs 1.712 ms per sample, versus 0.268 ms for GD-10; Lo-Fi costs 0.356 ms, also above GD-10. Only diagonal VD-EKF (0.103 ms) is faster. Hence the abstract's 'markedly reduced update latency' and Section IV-B's 'consistently achieves sub-millisecond adaptation' are contradicted by the reported table. If the latency claim is meant as time to reach a target BER over the stream, that metric should be reported separately.
  3. [IV-D/E, Figs. 7-10] Results are claimed to be averaged over 10 trials, but no error bars or confidence intervals are shown. In the nonlinear QuaDRiGa scenario, the text says CM-EKF 'decisively outperforms all other methods across all SNR values by several standard deviations,' yet the standard deviations are not displayed or tabulated. With only 10 trials and BER values spanning orders of magnitude, statistical significance is not established. Please add error bars or confidence intervals and report the number of trials for every curve.
  4. [II-A2 and V] The paper explicitly assumes the channel evolves smoothly over time, and all experiments use continuous variations (rotation, COST2100, QuaDRiGa). However, the Abstract and Conclusions claim 'increased robustness to channel dynamics' without this qualification. The Gaussian Markov model in Eq. (13) is likely misspecified under abrupt changes such as shadowing or handover. Either narrow the claim to smooth dynamics or add an experiment with a sudden channel change and show the filter's recovery.
minor comments (4)
  1. [IV-A2] The learning rates for SGD/GD and BBB are only described as 'manually optimized'; the actual values should be listed in a table or the appendix for reproducibility.
  2. [Table II] The table is hard to read because the rows for BBB and BONG/A1 are interleaved without clear grouping labels; reorganize the table so each algorithm family and covariance approximation is identified.
  3. [Fig. 6] Add axis labels and a colorbar; the grayscale decision regions are difficult to interpret in print.
  4. [III-B1, Eq. (21)] The dimensions of H_t should be stated explicitly in the text; currently the row/column convention is only implicit in the notation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the derivation is self-contained and validated against external baselines.

full rationale

The paper's central derivation—casting online learning as Bayesian tracking and deriving CM-EKF/BONG updates—is self-contained. The state-space model in Eq. (13) is an explicit modeling assumption, not defined in terms of the target BER or the reported results. The update equations (19)–(25) follow from standard Gaussian propagation and conditional-moments EKF, cited to external literature [50], [51]. The claimed single-step adaptation is then evaluated empirically against external references (MAP, MMSE, NLMS, and SGD/GD baselines) on synthetic and realistic channels; no fitted constant or training target is smuggled into the reported BER as a 'prediction.' The only self-citations are to prior architecture work (DeepSIC [23], modular Bayesian receivers [36]) and a prior conference version [1]; these provide context and architectural components rather than the load-bearing justification for the paper's central claim. The weakest assumption—the Gaussian Markov weight dynamics in Eq. (13)—is a limitation in model specification, not a circularity: the paper does not define its result in terms of that assumption, nor does it infer the assumption from the result it claims to demonstrate.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The mechanism rests on a hand-chosen state-space model for the weights, two scalar hyperparameters (γ, σ²) whose values are undisclosed, and the standard EKF linearization assumptions. No fundamentally new physical entities are introduced.

free parameters (3)
  • State evolution noise variance σ² = Not reported
    Controls the assumed rate of channel-driven parameter drift in Eq. (13). The paper gives no value or sensitivity analysis, yet the entire tracking behavior of CM-EKF and its variants depends on it.
  • State memory coefficient γ = Not reported
    Determines how much past information is retained in Eq. (13). Values in (0,1] are discussed qualitatively but no concrete value is given for the experiments.
  • Initial prior mean μ_0 and covariance Σ_0 = Not reported
    Initialization of the Gaussian prior p(θ_0)=N(μ_0, Σ_0) in Section III-A; likely set by standard initialization, but not specified.
assumptions (4)
  • ad hoc to paper First-order Gaussian Markov evolution of optimal DNN parameters, Eq. (13)
    The paper postulates p(θ_t | θ_{t-1}) = N(γ θ_{t-1}, σ² I) for the desired weights. This is the central modeling choice that makes Kalman filtering applicable; it is not derived from channel physics.
  • domain assumption Smooth temporal variation of the channel distribution
    Section II-A-2 assumes the conditional distribution P(r_t | s_t) evolves smoothly over time. Without this, the tracking formulation in parameter space is unjustified.
  • domain assumption First-order Taylor linearization of the DNN mapping is accurate
    Eq. (21) linearizes h_θ around the predictive mean. If the network is highly nonlinear over the parameter uncertainty scale, the EKF update is biased.
  • domain assumption Bernoulli observation model with DNN output as bit probability
    Eq. (14) assumes p(b_{t,i}=1 | x_t, θ_t) = h_{θ_t}(x_t)_i, i.e., the network sigmoid outputs are calibrated probabilities. This is an approximation for frequentist DNNs.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Online Learning of Modular Bayesian Deep Receivers: Single-Step Adaptation with Streaming Data." pith.science (2026). https://pith.science/paper/Y74YWF5B

@misc{pith2026251106045,
  author       = {Pith},
  title        = {Pith review of: Online Learning of Modular Bayesian Deep Receivers: Single-Step Adaptation with Streaming Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Y74YWF5B}},
  note         = {Machine review of arXiv:2511.06045}
}
read the original abstract

Deep neural network (DNN)-based receivers offer a powerful alternative to classical model-based designs for wireless communication, especially in complex and nonlinear propagation environments. However, their adoption is challenged by the rapid variability of wireless channels, which makes pre-trained static DNN-based receivers ineffective, and by the latency and computational burden of online stochastic gradient descent (SGD)-based learning. In this work, we propose an online learning framework that enables rapid low-complexity adaptation of DNN-based receivers. Our approach is based on two main tenets. First, we cast online learning as Bayesian tracking in parameter space, enabling a single-step adaptation, which deviates from multi-epoch SGD . Second, we focus on modular DNN architectures that enable parallel, online, and localized variational Bayesian updates. Simulations with practical communication channels demonstrate that our proposed online learning framework can maintain a low error rate with markedly reduced update latency and increased robustness to channel dynamics as compared to traditional gradient descent based method.

Figures

Figures reproduced from arXiv: 2511.06045 by the authors.

Figure 1
Figure 1. Illustration of proposed framework for single-step online adaptation [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Transmission scheme with synchronization and periodic pilot-data blocks. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. DeepSIC model architecture illustration. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Pipelined DeepSIC model architecture illustration. [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Symbol error rates, linear rotation channel. [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Decision zones after 500 blocks (45◦ counter-clockwise rotation) for four decoders - optimal decoding using the MAP rule, channel estimation using NLMS followed by maximum likelihood decoding, and two neural decoders trained using CM-EKF and SGD-8-4. 10 11 12 13 14 15 …
Figure 7
Figure 7. Figure 7: Online learning for modular vs. non-modular architectures. [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: BER vs. pilot index at SNR=4 dB. Synchronization from streaming data, with four different Sionna narrowband linear channel. 6 7 8 9 10 11 12 SNR [dB] 10 4 10 3 10 2 Average Tracking BER GD-10 BBB-DIAG-10 Lo-Fi BONG-EF CM-EKF 0 50 100 150 200 250 300 Channel Snapshot 10…
Figure 9
Figure 9. Figure 9: Learning from streaming data, COST2100 channel. Left: Average BER vs. SNR; Right: BER vs channel snapshot at SNR [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 10
Figure 10. Figure 10: Learning from streaming data, nonlinear QuaDRiGa channel. Left: Average BER vs. SNR; Right: BER vs channel snapshot at SNR [PITH_FULL_IMAGE:figures/full_fig_p011_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 4 linked inside Pith

  1. [18]

    On diagonal approximations to the extended Kalman filter for online training of Bayesian neural networks,

    P. G. Chang, K. P. Murphy, and M. Jones, “On diagonal approximations to the extended Kalman filter for online training of Bayesian neural networks,” inContinual Lifelong Learning Workshop at ACML 2022, 2022

    2022

  2. [19]

    Bayesian online natural gradient (BONG),

    M. Jones, P. Chang, and K. Murphy, “Bayesian online natural gradient (BONG),”Advances in Neural Information Processing Systems, vol. 37, pp. 131 104–131 153, 2024

    2024

  3. [54]

    Low-rank extended Kalman filtering for online learning of neural networks from streaming data,

    P. G. Chang, G. Dur´an-Mart´ın, A. Y . Shestopaloff, M. Jones, and K. Murphy, “Low-rank extended Kalman filtering for online learning of neural networks from streaming data,”arXiv preprint arXiv:2305.19535, 2023

    arXiv 2023

  4. [1]

    Rapid online Bayesian learning for deep receivers,

    Y . Gusakov, O. Simeone, T. Routtenberg, and N. Shlezinger, “Rapid online Bayesian learning for deep receivers,” inIEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2025

    2025

  5. [2]

    Holographic MIMO surfaces for 6G wireless networks: Opportunities, challenges, and trends,

    C. Huang, S. Hu, G. C. Alexandropoulos, A. Zappone, C. Yuen, R. Zhang, M. Di Renzo, and M. Debbah, “Holographic MIMO surfaces for 6G wireless networks: Opportunities, challenges, and trends,”IEEE Commun. Mag., vol. 27, no. 5, pp. 118–125, 2020

    2020

  6. [3]

    Reconfigurable intelligent surfaces: Principles and opportunities,

    Y . Liu, X. Liu, X. Mu, T. Hou, J. Xu, M. Di Renzo, and N. Al-Dhahir, “Reconfigurable intelligent surfaces: Principles and opportunities,”IEEE Commun. Surveys Tuts., vol. 23, no. 3, pp. 1546–1577, 2021

    2021

  7. [4]

    Dynamic metasurface antennas for 6G extreme massive MIMO com- munications,

    N. Shlezinger, G. C. Alexandropoulos, M. F. Imani, Y . C. Eldar, and D. R. Smith, “Dynamic metasurface antennas for 6G extreme massive MIMO com- munications,”IEEE Wireless Commun., vol. 28, no. 2, pp. 106–113, 2021

    2021

  8. [5]

    Wideband tHz multi-user downlink communications with leaky wave antennas,

    N. Lang, Y . Gabay, N. Shlezinger, T. Routtenberg, Y . Ghasempour, G. C. Alexandropoulos, and Y . C. Eldar, “Wideband tHz multi-user downlink communications with leaky wave antennas,”IEEE Trans. Wireless Commun., 2025, early access

    2025

Show all 59 references
  1. [6]

    Deep learning for wireless communications: An emerging interdisciplinary paradigm,

    L. Dai, R. Jiao, F. Adachi, H. V . Poor, and L. Hanzo, “Deep learning for wireless communications: An emerging interdisciplinary paradigm,” IEEE Wireless Commun., vol. 27, no. 4, pp. 133–139, 2020

    2020

  2. [7]

    An introduction to deep learning for the physical layer,

    T. O’Shea and J. Hoydis, “An introduction to deep learning for the physical layer,”IEEE Trans. on Cogn. Commun. Netw., vol. 3, no. 4, pp. 563–575, 2017

    2017

  3. [8]

    Artificial intelligence-empowered hybrid multiple-input/multiple-output beamforming: Learning to optimize for high-throughput scalable MIMO,

    N. Shlezinger, M. Ma, O. Lavi, N. T. Nguyen, Y . C. Eldar, and M. Juntti, “Artificial intelligence-empowered hybrid multiple-input/multiple-output beamforming: Learning to optimize for high-throughput scalable MIMO,” IEEE Veh. Technol. Mag., vol. 19, no. 3, pp. 58–67, 2024

    2024

  4. [9]

    Nine challenges in artificial intelligence and wireless communications for 6G,

    W. Tong and G. Y . Li, “Nine challenges in artificial intelligence and wireless communications for 6G,”IEEE Wireless Commun., vol. 29, no. 4, pp. 140–145, 2022

    2022

  5. [10]

    Adaptive and flexible model-based AI for deep receivers in dynamic channels,

    T. Raviv, S. Park, O. Simeone, Y . C. Eldar, and N. Shlezinger, “Adaptive and flexible model-based AI for deep receivers in dynamic channels,” IEEE Wireless Commun., vol. 31, no. 4, pp. 163–169, 2024

    2024

  6. [11]

    A note on implementation methodologies of deep learning-based signal detection for conventional MIMO transmitters,

    J. Xia, D. Deng, and D. Fan, “A note on implementation methodologies of deep learning-based signal detection for conventional MIMO transmitters,” IEEE Trans. Broadcast., vol. 66, no. 3, pp. 744–745, 2020

    2020

  7. [12]

    CRC-aided learned ensembles of belief-propagation polar decoders,

    T. Raviv, A. Goldman, O. Vayner, Y . Be’ery, and N. Shlezinger, “CRC-aided learned ensembles of belief-propagation polar decoders,” inIEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024, pp. 8856–8860

    2024

  8. [13]

    DeepRx: Fully convolutional deep learning receiver,

    M. Honkala, D. Korpi, and J. M. Huttunen, “DeepRx: Fully convolutional deep learning receiver,”IEEE Trans. Wireless Commun., vol. 20, no. 6, pp. 3925–3940, 2021

    2021

  9. [14]

    Machine learning for MU-MIMO receive processing in OFDM systems,

    M. Goutay, F. A. Aoudia, J. Hoydis, and J.-M. Gorce, “Machine learning for MU-MIMO receive processing in OFDM systems,”IEEE J. Sel. Areas Commun., vol. 39, no. 8, pp. 2318–2332, 2021

    2021

  10. [15]

    A hypernetwork based framework for non-stationary channel prediction,

    G. Liu, Z. Hu, L. Wang, H. Zhang, J. Xue, and M. Matthaiou, “A hypernetwork based framework for non-stationary channel prediction,” IEEE Trans. Veh. Technol., vol. 73, no. 6, pp. 8338–8351, 2024

    2024

  11. [16]

    Modular hypernetworks for scalable and adaptive deep MIMO receivers,

    T. Raviv and N. Shlezinger, “Modular hypernetworks for scalable and adaptive deep MIMO receivers,”arXiv preprint arXiv:2408.11920, 2024

    2024 arXiv

  12. [17]

    In-context learning for mimo equalization using transformer-based sequence models,

    M. Zecchin, K. Yu, and O. Simeone, “In-context learning for mimo equalization using transformer-based sequence models,” in2024 IEEE International Conference on Communications Workshops (ICC Workshops), 2024, pp. 1573–1578

    2024

  13. [20]

    A unifying framework for generalised Bayesian online learning in non-stationary environments,

    G. Duran-Martin, L. S ´anchez-Betancourt, A. Y . Shestopaloff, and K. Murphy, “A unifying framework for generalised Bayesian online learning in non-stationary environments,”arXiv preprint arXiv:2411.10153, 2024

    2024 arXiv

  14. [21]

    Durbin and S

    J. Durbin and S. J. Koopman,Time series analysis by state space methods. OUP Oxford, 2012, vol. 38

    2012

  15. [22]

    ViterbiNet: A deep learning based Viterbi algorithm for symbol detection,

    N. Shlezinger, N. Farsad, Y . C. Eldar, and A. J. Goldsmith, “ViterbiNet: A deep learning based Viterbi algorithm for symbol detection,”IEEE Trans. Wireless Commun., vol. 19, no. 5, pp. 3319–3331, 2020

    2020

  16. [23]

    DeepSIC: Deep soft interference cancellation for multiuser MIMO detection,

    N. Shlezinger, R. Fu, and Y . C. Eldar, “DeepSIC: Deep soft interference cancellation for multiuser MIMO detection,”IEEE Trans. Wireless Commun., vol. 20, no. 2, pp. 1349–1362, 2021

    2021

  17. [24]

    Model-based deep learning,

    N. Shlezinger and Y . C. Eldar, “Model-based deep learning,”Foundations and Trends® in Signal Processing, vol. 17, no. 4, pp. 291–416, 2023

    2023

  18. [25]

    Learning to demodulate from few pilots via offline and online meta-learning,

    S. Park, H. Jang, O. Simeone, and J. Kang, “Learning to demodulate from few pilots via offline and online meta-learning,”IEEE Trans. Signal Process., vol. 69, pp. 226 – 239, 2020

    2020

  19. [26]

    Online meta-learning for hybrid model-based deep receivers,

    T. Raviv, S. Park, O. Simeone, Y . C. Eldar, and N. Shlezinger, “Online meta-learning for hybrid model-based deep receivers,”IEEE Trans. Wireless Commun., vol. 22, no. 10, pp. 6415–6431, 2023

    2023

  20. [27]

    Asynchronous online adaptation via modular drift detection for deep receivers,

    N. Uzlaner, T. Raviv, N. Shlezinger, and K. Todros, “Asynchronous online adaptation via modular drift detection for deep receivers,”IEEE Trans. Wireless Commun., vol. 24, no. 5, pp. 4454–4468, 2025

    2025

  21. [28]

    End-to-end learning for OFDM: From neural receivers to pilotless communication,

    F. A. Aoudia and J. Hoydis, “End-to-end learning for OFDM: From neural receivers to pilotless communication,”IEEE Trans. Wireless Commun., vol. 21, no. 2, pp. 1049–1063, 2021

    2021

  22. [29]

    Adaptive neural network-based OFDM receivers,

    M. B. Fischer, S. D ¨orner, S. Cammerer, T. Shimizu, H. Lu, and S. Ten Brink, “Adaptive neural network-based OFDM receivers,” inIEEE Signal Processing Advances in Wireless Communications (SPAWC), 2022

    2022

  23. [30]

    Symbol-level online channel tracking for deep receivers,

    R. Finish, Y . Cohen, T. Raviv, and N. Shlezinger, “Symbol-level online channel tracking for deep receivers,” inIEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2022, pp. 8897–8901

    2022

  24. [31]

    Data augmentation for deep learning-based radio modulation classification,

    L. Huang, W. Pan, Y . Zhang, L. Qian, N. Gao, and Y . Wu, “Data augmentation for deep learning-based radio modulation classification,” IEEE Access, vol. 8, pp. 1498–1506, 2019

    2019

  25. [32]

    Data augmentation for deep receivers,

    T. Raviv and N. Shlezinger, “Data augmentation for deep receivers,”IEEE Trans. Wireless Commun., vol. 22, no. 11, pp. 8259–8274, 2023

    2023

  26. [33]

    Simeone,Machine Learning for Engineers

    O. Simeone,Machine Learning for Engineers. Cambridge University Press, 2022

    2022

  27. [34]

    Bayesian active meta-learning for reliable and efficient AI-based demodulation,

    K. M. Cohen, S. Park, O. Simeone, and S. Shamai, “Bayesian active meta-learning for reliable and efficient AI-based demodulation,”IEEE Trans. Signal Process., vol. 70, pp. 5366–5380, 2022

    2022

  28. [35]

    Robust Bayesian learning for reliable wireless AI: Framework and applications,

    M. Zecchin, S. Park, O. Simeone, M. Kountouris, and D. Gesbert, “Robust Bayesian learning for reliable wireless AI: Framework and applications,” IEEE Trans. on Cogn. Commun. Netw., vol. 9, no. 4, pp. 897–912, 2023

    2023

  29. [36]

    Uncertainty-aware and reliable neural MIMO receivers via modular Bayesian deep learning,

    T. Raviv, S. Park, O. Simeone, and N. Shlezinger, “Uncertainty-aware and reliable neural MIMO receivers via modular Bayesian deep learning,” IEEE Trans. Veh. Technol., 2025

    2025

  30. [37]

    A unifying Bayesian view of continual learning,

    S. Farquhar and Y . Gal, “A unifying Bayesian view of continual learning,” arXiv preprint arXiv:1902.06494, 2019

    1902 arXiv

  31. [38]

    Sionna: An open-source library for research on communication systems,

    J. Hoydis, S. Cammerer, F. Ait Aoudia, M. Nimier-David, L. Maggi, G. Marcus, A. Vem, and A. Keller, “Sionna: An open-source library for research on communication systems,” 2022, https://nvlabs.github.io/sionna/

    2022

  32. [39]

    The COST 2100 MIMO channel model,

    L. Liu, C. Oestges, J. Poutanen, K. Haneda, P. Vainikainen, F. Quitin, F. Tufvesson, and P. De Doncker, “The COST 2100 MIMO channel model,”IEEE Wireless Commun., vol. 19, no. 6, pp. 92–99, 2012

    2012

  33. [40]

    Quadriga: A 3-D multi-cell channel model with time evolution for enabling virtual field trials,

    S. Jaeckel, L. Raschkowski, K. B ¨orner, and L. Thiele, “Quadriga: A 3-D multi-cell channel model with time evolution for enabling virtual field trials,”IEEE Trans. Antennas Propag., vol. 62, no. 6, pp. 3242–3256, 2014

    2014

  34. [41]

    How much training is needed in multiple-antenna wireless links?

    B. Hassibi and B. M. Hochwald, “How much training is needed in multiple-antenna wireless links?”IEEE Trans. Inf. Theory, vol. 49, no. 4, pp. 951–963, 2003

    2003

  35. [42]

    Effects of channel aging in massive mimo systems,

    K. T. Truong and R. W. Heath, “Effects of channel aging in massive mimo systems,”Journal of Communications and Networks, vol. 15, no. 4, pp. 338–351, 2013

    2013

  36. [43]

    Bayesian deep learning and a probabilistic perspective of generalization,

    A. G. Wilson and P. Izmailov, “Bayesian deep learning and a probabilistic perspective of generalization,”Advances in Neural Information Processing Systems, vol. 33, pp. 4697–4708, 2020

    2020

  37. [44]

    Priors in Bayesian deep learning: A review,

    V . Fortuin, “Priors in Bayesian deep learning: A review,”International Statistical Review, vol. 90, no. 3, pp. 563–591, 2022

    2022

  38. [45]

    Concrete dropout,

    Y . Gal, J. Hron, and A. Kendall, “Concrete dropout,”Advances in Neural Information Processing Systems, vol. 30, 2017

    2017

  39. [46]

    Weight uncertainty in neural network,

    C. Blundell, J. Cornebise, K. Kavukcuoglu, and D. Wierstra, “Weight uncertainty in neural network,” inInternational conference on machine learning. PMLR, 2015, pp. 1613–1622

    2015

  40. [47]

    A survey of uncertainty in deep neural networks,

    J. Gawlikowski, C. R. N. Tassi, M. Ali, J. Lee, M. Humt, J. Feng, A. Kruspe, R. Triebel, P. Jung, R. Roscheret al., “A survey of uncertainty in deep neural networks,”Artificial Intelligence Review, vol. 56, no. Suppl 1, pp. 1513–1589, 2023. 13

    2023

  41. [48]

    Model-based deep learning: On the intersection of deep learning and optimization,

    N. Shlezinger, Y . C. Eldar, and S. P. Boyd, “Model-based deep learning: On the intersection of deep learning and optimization,”IEEE Access, vol. 10, pp. 115 384–115 398, 2022

    2022

  42. [49]

    Iterative soft interference cancellation for multiple antenna systems,

    W.-J. Choi, K.-W. Cheong, and J. M. Cioffi, “Iterative soft interference cancellation for multiple antenna systems,” inProc. IEEE WCNC, 2000

    2000

  43. [50]

    Iterative filtering and smoothing in nonlinear and non-Gaussian systems using conditional mo- ments,

    F. Tronarp, A. F. Garc ´ıa-Fern´andez, and S. S ¨arkk¨a, “Iterative filtering and smoothing in nonlinear and non-Gaussian systems using conditional mo- ments,”IEEE Signal Processing Letters, vol. 25, no. 3, pp. 408–412, 2018

    2018

  44. [51]

    Papoulis and S

    A. Papoulis and S. U. Pillai,Probability, Random Variables, and Stochastic Processes, 4th ed. New York: McGraw-Hill, 2002, see p. 127

    2002

  45. [52]

    New insights and perspectives on the natural gradient method,

    J. Martens, “New insights and perspectives on the natural gradient method,” J. Mach. Learn. Res., vol. 21, no. 1, Jan. 2020

    2020

  46. [53]

    SLANG: Fast structured covariance approximations for Bayesian deep learning with natural gradient,

    A. Mishkin, F. Kunstner, D. Nielsen, M. Schmidt, and M. E. Khan, “SLANG: Fast structured covariance approximations for Bayesian deep learning with natural gradient,” inAdvances in Neural Information Processing Systems, 2018

    2018

  47. [55]

    The unscented Kalman filter for nonlinear estimation,

    E. A. Wan and R. Van Der Merwe, “The unscented Kalman filter for nonlinear estimation,” inIEEE Adaptive Systems for Signal Processing, Communications, and Control Symposium, 2000, pp. 153–158

    2000

  48. [56]

    KOALA++: Efficient Kalman-based optimization of neural networks with gradient-covariance products,

    Z. Xia, A. Davtyan, and P. Favaro, “KOALA++: Efficient Kalman-based optimization of neural networks with gradient-covariance products,”arXiv preprint arXiv:2506.04432, 2025

    2025

  49. [57]

    Online convex programming and generalized infinitesimal gradient ascent,

    M. Zinkevich, “Online convex programming and generalized infinitesimal gradient ascent,” inInternational Conference on International Conference on Machine Learning, 2003, p. 928–935

    2003

  50. [58]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” inIEEE Conference on Computer Vision and Pattern Recognition, 2016, pp. 770–778

    2016

  51. [59]

    Model-agnostic meta-learning for fast adaptation of deep networks,

    C. Finn, P. Abbeel, and S. Levine, “Model-agnostic meta-learning for fast adaptation of deep networks,” inProceedings of the 34th International Conference on Machine Learning - Volume 70, ser. ICML’17. JMLR.org, 2017, p. 1126–1135

    2017

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.