On Analog Gradient Descent Learning over Multiple Access Fading Channels

Kobi Cohen; Tomer Sery

REVIEW 2 major objections 5 minor 47 references

On Analog Gradient Descent Learning over Multiple Access Fading Channels

T0 review · 2 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read This paper develops GBMA, a gradient-descent protocol in which nodes transmit analog local gradients over a fading multiple access channel, and proves that its convergence rate approaches the centralized gradient-descent rate as the…

desk verdict A solid convergence analysis for a distortion-tolerant analog gradient aggregation scheme over fading MAC, with one load-bearing assumption in Theorem 1 that needs explicit statement or repair. read the letter →

arxiv 1908.07463 v1 pith:ULOCMBVX submitted 2019-08-20 cs.LG cs.ITmath.ITstat.ML

Tomer Sery , Kobi Cohen This is my paper · ORCID

classification cs.LGcs.ITmath.ITstat.ML MSC 68Q3290C2594A05

keywords distributedlearninggradientdescentmultipleaccesschannelfederatedfadingchannelsover-the-aircomputationconvergenceanalysisenergyscaling

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

Gradient-Based Multiple Access (GBMA) is a distributed learning protocol in which every node sends an analog function of its local gradient over the same channel, and the network edge updates the parameter estimate directly from the noisy, faded superposition. The paper's central claim is that despite using no power control or beamforming to cancel fading, GBMA can approach the convergence rate of centralized gradient descent as the number of nodes grows: linear $O(c^k)$ for strongly convex losses and $O(1/k)$ for convex losses with Lipschitz gradients. It also gives energy scaling laws: setting per-node energy $E_N = \Omega(N^{\epsilon-2})$ makes the additive noise term vanish, and under a bounded-gradient-magnitude condition the fading distortion term vanishes as well. A sympathetic reader would care because this is a concrete argument that the communication bottleneck of federated learning need not grow with the number of participants when gradients are aggregated over the air.

What carries the argument

The central object is the noisy distorted gradient estimate $v_k = \frac{1}{N}\sum_{n=1}^N h_{n,k} g_n(\theta_k) + w_k$, obtained after matched filtering the received superposition of analog gradient signals. The proof machinery combines the standard smoothness and strong-convexity inequalities of Lemmas 3 and 4 with the channel index of dispersion $D = \sigma_h^2/\mu_h$; the constant stepsize $\beta$ is chosen as a function of $\mu_h$, $L$, $\mu$, $N$, and $D$ so that the contraction coefficient $c$ stays below 1. A second lemma bounds each local gradient norm through the compact parameter set's diameter $\delta$, converting per-node distortion into the decaying term $\sigma_h^2 L^2 \delta(2+\delta)/N$.

What would settle it

Fix a strongly convex smooth objective $F(\theta) = \frac{1}{N}\sum_n f_n(\theta)$ whose individual minimizers $\theta_n^*$ lie outside the compact set $\Theta$, run GBMA with large $N$ and $E_N = N^{\epsilon-2}$, and measure the error $\mathbb{E}[F(\theta_k)] - F(\theta^*)$ after many iterations; if the error floor does not shrink toward the centralized gradient-descent bound as $N$ increases, the theorem's asymptotic claim fails. A direct check would compute the bound in equation (39) with $\|\theta^* - \theta_n^*\| > \delta$ to see whether the distortion term retains an $N$-independent constant.

Watch

Extended reading notes

Core claim

The discovery is a finite-sample error bound for GBMA: for strongly convex objectives, $$\mathbb{E}[F(\theta_k)] - F(\$\theta$^*) \le c^k $r_0^{2}$ \frac{L}{2} + \frac{L\$beta^{2}$}{2(1-c)}\left(\frac{\$sigma_h^{2}$ \delta $L^{2}$(2+\delta)}{N} + \frac{d\$sigma_w^{2}$}{E_N $N^{2}$}\right),$$ and analogous $O(1/k)$ bounds for convex objectives. The proof tracks three additive effects: initial distance, gradient distortion from random channel gains, and additive channel noise. Because the distortion term scales as $\sigma_h^2/N$ and the noise term as $\sigma_w^2/(E_N N^2)$, both disappear as $N \to \infty$ when $E_N = \Omega(N^{\epsilon-2})$, leaving exactly the centralized gradient-descent rate.

Load-bearing premise

The load-bearing premise is that every local loss's own minimizer lies inside the compact parameter set $\Theta$, so the distance from the global optimum to each local minimizer can be bounded by $\Theta$'s diameter $\delta$; the proof uses this to keep the fading distortion term proportional to $1/N$, and if a local minimizer sits far outside $\Theta$ that term keeps a constant offset that does not vanish as $N$ grows.

Editorial extensions

If this is right

Adding nodes does not increase bandwidth requirements, because all nodes share the same set of $d$ orthogonal waveforms and the aggregated channel noise is independent of $N$.
Under strong convexity, GBMA with a constant stepsize converges linearly despite using noisy, faded gradients, in contrast to SGD-based methods that typically achieve only $O(1/k)$ and require decreasing stepsizes.
Per-node transmission energy can be driven to zero as the network grows, with total network energy made arbitrarily small, while the error still approaches the centralized gradient-descent error.
For convex losses, the expected squared per-node gradient magnitude may grow like $B(N) = O(N^{1-\epsilon})$ without ruining the asymptotic convergence rate, unlike classical SGD analyses that require bounded gradients.

Reading between the lines

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

The same proof structure suggests that other first-order methods that tolerate multiplicative noise, such as normalized or accelerated gradient variants, could inherit the same energy-scaling behavior, but the paper does not analyze them.
A testable extension would allow fading to be correlated across iterations; under i.i.d. fading the distortion averages out, but temporally correlated channels would likely leave a bias that the current $N \to \infty$ argument cannot remove.
The results imply a design rule for federated systems: when many devices participate, the network can spend less energy per device and still match a centralized oracle, though privacy considerations are not addressed in this paper.

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.

Desk Editor's Note

read the letter

Colleague,

Bottom line: this paper gives the first convergence analysis I know of for an over-the-air gradient aggregation scheme that does not use power control or beamforming—just phase correction. The GBMA scheme is simple and the central claim, that you can approach centralized GD rates as N grows while letting per-node energy shrink, is credible under the stated conditions. The proof machinery is mostly standard optimization inequalities, cleanly applied.

What is genuinely new: showing that fading distortion can be tolerated without explicit channel inversion, and the accompanying energy scaling laws. The analysis is not fitted to experiments; the bounds come from the model. The experiments are supportive but not reproducible (no code or data release), and the comparison to FDM-GD is a bit cherry-picked at -50 dB, but that is secondary.

Soft spots, in order of importance:

1. Theorem 1's proof, Eq. (39), uses ||θ* − θ_n*|| ≤ δ where θ_n* is the unconstrained minimizer of f_n. That requires every local minimizer to lie inside the compact set Θ, and the paper never states it. Without it, the variance term can be Θ(N) instead of O(1/N), so the additive term in (16) need not vanish. This is a real gap in the proof of the main theorem. It can likely be fixed by adding a boundedness condition like sup_n ||θ_n* − θ*||, or by assuming (1/N^2) Σ ||∇f_n(θ*)||^2 → 0. Read the appendix carefully; the issue is exactly as the stress-test note describes.

2. Condition (17) in Theorem 2 is acknowledged as technical, but the argument that it is satisfied for large N is only heuristic. It is testable, so not fatal, but it deserves a proper statement.

3. The claim that total transmission energy can be made arbitrarily small ignores downlink broadcast costs and the channel estimation overhead. That is an overstatement, though the per-node scaling law itself is fine.

These do not sink the paper. The core idea—that fading distortion averages out without power control—is correct in the i.i.d. channel model if the local-minimizer condition is added. The authors should be asked to fix that assumption before publication. The citation pattern is fair; the related work on CA-DSGD and FEEL is adequately covered.

Who is this for: researchers in wireless edge learning, especially those working on over-the-air aggregation. It deserves a serious referee. I would not desk-reject it.

Recommendation: send to peer review with a request for a revised Theorem 1 that either states the local-minimizer condition up front or bounds the cross term differently. The central argument holds up once that patch is in.

Referee Report

2 major / 5 minor

Summary. The paper studies distributed empirical risk minimization over a fading multiple-access channel, where each node transmits an analog phase-corrected function of its local gradient and the network edge performs a gradient-descent update using the noisy, fading-distorted superposition. The authors propose the Gradient-Based Multiple Access (GBMA) algorithm and provide finite-sample convergence bounds for strongly convex and convex losses with Lipschitz gradients, together with energy-scaling laws that aim to let GBMA approach the centralized GD convergence rate as the number of nodes N grows. The analytical results are supplemented by experiments on the Million Song Dataset and a source-localization problem.

Significance. If the theoretical claims are correct, the paper offers a valuable and timely result for wireless federated learning: it shows that over-the-air gradient aggregation without power control or beamforming can asymptotically match centralized GD, with per-node transmission energy scaling down as N^{ε-2}. The finite-sample bounds are clean, the energy-scaling framework is explicit, and the experiments support the qualitative behavior. However, the proof of the main strongly convex result (Theorem 1) contains a gap in the treatment of the local gradient distortion, and this gap directly affects the central asymptotic claim.

major comments (2)

[Appendix VIII-A, Eq. (39)] The proof of Theorem 1 bounds the fading-distortion term by introducing θ_n^*, the minimizer of the local loss f_n, and then uses ||θ* - θ_n^*|| ≤ δ, where δ is the diameter of Θ. This step implicitly requires θ_n^* ∈ Θ, which is not assumed in Section II or in the statement of Theorem 1. This gap is load-bearing: if local minimizers lie outside Θ, the term σ_h^2/N^2 Σ_n E||∇f_n(θ_k)||^2 need not vanish as N→∞. A concrete counterexample is d=1 with f_n(θ) = 1/2(θ - a_n)^2 and a_n = n - (N+1)/2, for which F(θ) = 1/2 θ^2 + const and θ* = 0 lies inside any Θ centered at 0, while Σ_n a_n^2 = Θ(N^3), so the distortion term is Θ(N^2) and the additive term in Eq. (16) does not converge to zero. The theorem should be amended with an explicit condition, e.g., sup_n ||θ_n^* - θ*|| bounded or (1/N^2) Σ_n E||∇f_n(θ*)||^2 → 0, or the proof must bound the distortion term without relying on θ_n^*.
[Section II and Theorem 1] The proof of Theorem 1 also assumes that each local function f_n has an unconstrained minimizer θ_n^* and that ∇f_n(θ_n^*) = 0. The stated assumptions (convexity and Lipschitz gradient on Θ, or even on R^d) do not guarantee the existence of such a minimizer; for example, f_n(θ) = e^{-θ} is convex and L-smooth on certain domains but has no minimizer. The theorem should either assume the existence of θ_n^* explicitly or restate the argument in terms of a bound on the spread of local gradients at θ*, which is the quantity that actually determines whether the fading distortion vanishes.

minor comments (5)

[Section V-C] The energy scaling range N^{ε-2} ≲ EN ≲ N^{-ε-1} is nonempty only for 0 < ε ≤ 1/2; this constraint should be stated when the scaling laws are presented.
[Section V-C, Theorem 2 discussion] The comment that condition (17) can be satisfied for any finite k by choosing N large enough is correct, but it should be made explicit that the asymptotic O(1/k) statement is for a fixed iteration horizon k while N→∞; for a uniform-in-k statement an additional argument is needed because E||∇F(θ_i)||^2 tends to 0 as i grows.
[Section VI-A] There are minor typographical issues, such as 'we start by examining' at the beginning of Section VI-A, and the legend in Figs. 2(b) and 3(b) appears to have a missing symbol for ǫ.
[References] The reference to Nesterov's lecture notes is incomplete; the author and a formal citation should be provided.
[Section II] The symbol Θ is defined as a compact convex set, but the update in Eq. (9) does not include a projection; the paper should clarify whether the analysis is for the unconstrained problem or whether the iterates are assumed to remain in Θ.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the convergence bounds are derived from convex-optimization inequalities and channel statistics, with no fitted input renamed as a prediction.

full rationale

I find no significant circularity. The central claims of The paper—Theorems 1 and 2 and the energy scaling laws—are derived from standard convex-optimization inequalities (Lemmas 3 and 4), the stated channel model, and explicit assumptions on strong convexity, Lipschitz gradients, and i.i.d. fading statistics. No parameter is fitted to a subset of the data and then reported as a prediction; the experimental 'upper bounds' shown in Figures 2 and 3 are the closed-form bounds from the theorems evaluated at the simulation parameters, not fits to the empirical curves. Self-citations such as [7], [8], and [44] are contextual references to prior work on model-dependent MAC inference or to SGD analysis and do not carry the load of the convergence proof. The proof of Theorem 1 does contain a genuine gap: equation (39) introduces the local minimizer θ*_n and bounds ||θ* − θ*_n|| ≤ δ, which requires that each θ*_n lies in the compact set Θ; this condition is never stated, and without it the distortion term need not vanish as N→∞. However, this is a missing or incorrect assumption—a correctness risk—not circularity, because the claimed convergence is not defined in terms of the local minimizers and the proof could be repaired by adding an explicit boundedness condition such as sup_n ||θ*_n − θ*|| < ∞. Accordingly, the paper's derivation is self-contained against the centralized GD benchmark, and the circularity score is 0.

Assumptions & free parameters 2 free parameters · 6 assumptions · 0 invented entities

The derivation is self-contained given the channel and data assumptions; the main burden is the i.i.d. fading and nonzero-mean channel gains, plus the technical conditions B(N) and local minimizer location.

free parameters (2)

stepsize β = chosen within bound (15) for strongly convex case, (18) and (20) for convex case
The constant stepsize is a design parameter. The bounds require knowledge of the Lipschitz constant L, strong convexity μ, and channel statistics μ_h, σ_h^2, which are typically unknown, making the parameter selection rule hard to implement in practice.
transmission energy E_N = Ω(N^{ε-2}) for some ε>0
The energy scaling law is a design choice; the theorem shows this makes the noise term vanish. The exponent ε is not specified, so the result is asymptotic.

assumptions (6)

domain assumption The fading channel gains h_{n,k} are i.i.d. across nodes and time with mean μ_h and variance σ_h^2, and are independent of the data and parameter.
Stated in Section II. Used in equations (31)-(34) to compute the expectation and second moment of the received signal; without this, the distorted gradient would be biased.
domain assumption Each node knows its channel phase (or corrects it with error less than π/4) so the effective channel gain has nonzero mean.
Stated in Section III after equation (5). If the phase correction is poor, μ_h could be near zero, eliminating the gradient signal.
standard math The local loss functions f_n are convex with L_n-Lipschitz gradients, and in Theorem 1 also μ_n-strongly convex.
Section II and Definitions 2 and 3. These are standard optimization assumptions used in the proof lemmas.
domain assumption For Theorem 2(b), E[||∇f_n(θ)||^2] ≤ B(N) for all θ and n, with B(N)=O(N^{1-ε}).
Section V-B, after equation (20). This scaling is needed to make the fading distortion term vanish as N→∞; it is a nonstandard assumption on how local gradient norms grow with network size.
ad hoc to paper The unconstrained minimizer θ_n* of each local loss lies in the compact parameter set Θ.
Used implicitly in equation (39) of Appendix VIII-A to bound ||θ* - θ_n*|| by the diameter δ. This is not stated in the paper.
ad hoc to paper Condition (17) of Theorem 2: E[||∇F(θ_i)||^2] > dσ_w^2/(E_N N^2) for all i.
Section V-B, Theorem 2. This technical condition ensures the error sequence is nonincreasing, and is not checked in practice.

how reviews work

0 comments

Cite this review

Pith. "Pith review of On Analog Gradient Descent Learning over Multiple Access Fading Channels." pith.science (2026). https://pith.science/paper/ULOCMBVX

@misc{pith2026190807463,
  author       = {Pith},
  title        = {Pith review of: On Analog Gradient Descent Learning over Multiple Access Fading Channels},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ULOCMBVX}},
  note         = {Machine review of arXiv:1908.07463}
}

read the original abstract

We consider a distributed learning problem over multiple access channel (MAC) using a large wireless network. The computation is made by the network edge and is based on received data from a large number of distributed nodes which transmit over a noisy fading MAC. The objective function is a sum of the nodes' local loss functions. This problem has attracted a growing interest in distributed sensing systems, and more recently in federated learning. We develop a novel Gradient-Based Multiple Access (GBMA) algorithm to solve the distributed learning problem over MAC. Specifically, the nodes transmit an analog function of the local gradient using common shaping waveforms and the network edge receives a superposition of the analog transmitted signals used for updating the estimate. GBMA does not require power control or beamforming to cancel the fading effect as in other algorithms, and operates directly with noisy distorted gradients. We analyze the performance of GBMA theoretically, and prove that it can approach the convergence rate of the centralized gradient descent (GD) algorithm in large networks. Specifically, we establish a finite-sample bound of the error for both convex and strongly convex loss functions with Lipschitz gradient. Furthermore, we provide energy scaling laws for approaching the centralized convergence rate as the number of nodes increases. Finally, experimental results support the theoretical findings, and demonstrate strong performance of GBMA using synthetic and real data.

Figures

Figures reproduced from arXiv: 1908.07463 by the authors.

**Figure 2.** Simulation results for prediction of a release year o [PITH_FULL_IMAGE:figures/full_fig_p019_2.png] view at source ↗

**Figure 3.** Simulation results for prediction of a release year o [PITH_FULL_IMAGE:figures/full_fig_p020_3.png] view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: Simulation results for prediction of a release year o [PITH_FULL_IMAGE:figures/full_fig_p020_4.png]

**Figure 5.** Figure 5: Simulation results for source localization. A compa [PITH_FULL_IMAGE:figures/full_fig_p021_5.png]

**Figure 6.** Figure 6: Simulation results for source localization. The tot [PITH_FULL_IMAGE:figures/full_fig_p022_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 41 canonical work pages

[1]

A convergent increme ntal gradient method with a constant step size,

D. Blatt, A. Hero, and H. Gauchman, “A convergent increme ntal gradient method with a constant step size,” SIAM Journal on Optimization , vol. 18, pp. 29–51, 01 2007

work page 2007
[2]

Incremental adaptive strate gies over distributed networks,

C. G. Lopes and A. H. Sayed, “Incremental adaptive strate gies over distributed networks,” IEEE Transactions on Signal Processing, vol. 55, no. 8, pp. 4064–4077, Aug 2007

work page 2007
[3]

Stochastic inc remental gradient descent for estimation in sensor network s,

S. S. Ram, A. Nedic, and V . V . V eeravalli, “Stochastic inc remental gradient descent for estimation in sensor network s,” in 2007 Conference Record of the F orty-First Asilomar Confere nce on Signals, Systems and Computers , Nov 2007, pp. 582–586

work page 2007
[4]

Incremental s tochastic subgradient algorithms for convex optimization ,

S. S. Ram, A. Nedi´ c, and V . V . V eeravalli, “Incremental s tochastic subgradient algorithms for convex optimization ,” SIAM Journal on Optimization , vol. 20, no. 2, pp. 691–717, 2009

work page 2009
[5]

Lag: Lazily agg regated gradient for communication-efﬁcient distributed learning,

T. Chen, G. Giannakis, T. Sun, and W. Yin, “Lag: Lazily agg regated gradient for communication-efﬁcient distributed learning,” in Advances in Neural Information Processing Systems , 2018, pp. 5050–5060

work page 2018
[6]

Federated optimization: Distributed machine learning f or on-device intelligence,

J. Koneˇ cn` y, H. B. McMahan, D. Ramage, and P . Richt´ arik , “Federated optimization: Distributed machine learning f or on-device intelligence,” arXiv preprint arXiv:1610.02527 , 2016

arXiv 2016
[7]

Performance analysis of likelih ood-based multiple access for detection over fading channe ls,

K. Cohen and A. Leshem, “Performance analysis of likelih ood-based multiple access for detection over fading channe ls,” IEEE Transactions on Information Theory , vol. 59, no. 4, pp. 2471–2481, 2013

work page 2013
[8]

Spectrum and energy efﬁcient multiple access for de tection in wireless sensor networks,

——, “Spectrum and energy efﬁcient multiple access for de tection in wireless sensor networks,” IEEE Transactions on Signal Processing, vol. 66, no. 22, pp. 5988–6001, 2018

work page 2018

Show all 47 references

[9]

Machine learning at the wirele ss edge: Distributed stochastic gradient descent over-the -air,

M. M. Amiri and D. Gunduz, “Machine learning at the wirele ss edge: Distributed stochastic gradient descent over-the -air,” arXiv preprint arXiv:1901.00844 , 2019

1901 arXiv
[10]

Over-the-air machine lear ning at the wireless edge,

M. M. Amiri and D. G¨ und¨ uz, “Over-the-air machine lear ning at the wireless edge,” in Proc. IEEE Int. W orkshop on Signal Process. Advances in Wireless Commun.(SPAWC) , 2019. August 21, 2019 DRAFT 32

2019
[11]

Federated learning over wire less fading channels,

M. M. Amiri and D. Gunduz, “Federated learning over wire less fading channels,” arXiv preprint arXiv:1907.09769 , 2019

1907 arXiv
[12]

Collaborative m achine learning at the wireless edge with blind transmitter s,

M. M. Amiri, T. M. Duman, and D. Gunduz, “Collaborative m achine learning at the wireless edge with blind transmitter s,” arXiv preprint arXiv:1907.03909 , 2019

1907 arXiv
[14]

Broadband analog aggregation for low-latency fed erated edge learning (extended version),

——, “Broadband analog aggregation for low-latency fed erated edge learning (extended version),” arXiv preprint arXiv:1812.11494v3, 2018

2018 arXiv
[15]

Energy-efﬁcie nt radio resource allocation for federated edge learning,

Q. Zeng, Y . Du, K. K. Leung, and K. Huang, “Energy-efﬁcie nt radio resource allocation for federated edge learning,” arXiv preprint arXiv:1907.06040 , 2019

1907 arXiv
[16]

Federated learni ng via over-the-air computation,

K. Yang, T. Jiang, Y . Shi, and Z. Ding, “Federated learni ng via over-the-air computation,” arXiv preprint arXiv:1812.11750, 2018

2018 arXiv
[17]

Dece ntralized detection with censoring sensors,

S. Appadwedula, V . V . V eeravalli, and D. L. Jones, “Dece ntralized detection with censoring sensors,” IEEE Trans. on Signal Process. , vol. 56, no. 4, pp. 1362–1373, Apr. 2008

2008
[18]

Hierarchical c ensoring sensors for change detection,

N. Patwari, A. O. Hero, and B. M. Sadler, “Hierarchical c ensoring sensors for change detection,” Statistical Signal Process., 2003 IEEE W orkshop on , pp. 21–24, Sep. 2003

2003
[19]

Energy efﬁcient signal dete ction in sensor networks using ordered transmissions,

R. S. Blum and B. M. Sadler, “Energy efﬁcient signal dete ction in sensor networks using ordered transmissions,” IEEE Trans. on Signal Process. , vol. 56, no. 7, pp. 3229–3235, Jul. 2008

2008
[20]

Ordering for estimation and optimization i n energy efﬁcient sensor networks,

R. S. Blum, “Ordering for estimation and optimization i n energy efﬁcient sensor networks,” IEEE Transactions on Signal Processing, vol. 59, no. 6, pp. 2847–2856, 2011

2011
[21]

Ordering f or reduced transmission energy detection in sensor network s testing a shift in the mean of a gaussian graphical model,

J. Zhang, Z. Chen, R. S. Blum, X. Lu, and W. Xu, “Ordering f or reduced transmission energy detection in sensor network s testing a shift in the mean of a gaussian graphical model,” IEEE Transactions on Signal Processing , vol. 65, no. 8, pp. 2178–2189, 2017

2017
[22]

Energy-efﬁcient decision fusion for distri buted detection in wireless sensor networks,

N. Sriranga, K. G. Nagananda, R. S. Blum, A. Saucan, and P . K. V arshney, “Energy-efﬁcient decision fusion for distri buted detection in wireless sensor networks,” in 21st International Conference on Information Fusion (FUSI ON), 2018, pp. 1541– 1547

2018
[23]

Energy-efﬁcient detection in w ireless sensor networks using likelihood ratio and channel state information,

K. Cohen and A. Leshem, “Energy-efﬁcient detection in w ireless sensor networks using likelihood ratio and channel state information,” IEEE Journal on Selected Areas in Comm., , vol. 29, no. 8, pp. 1671–1683, Sep. 2011

2011
[24]

Asymptotically cons istent one-bit detection in large sensor networks,

P . Braca, S. Marano, and V . Matta, “Asymptotically cons istent one-bit detection in large sensor networks,” in Proc. European Signal Process. Conf. , Aug. 2011, pp. 1035–1039

2011
[25]

Single-transmission distributed detection via o rder statistics,

——, “Single-transmission distributed detection via o rder statistics,” IEEE Transactions on Signal Processing , vol. 60, no. 4, pp. 2042–2048, Apr. 2012

2012
[26]

Type-based decentralized detect ion in wireless sensor networks,

K. Liu and A. Sayeed, “Type-based decentralized detect ion in wireless sensor networks,” IEEE Trans. on Signal Process. , vol. 55, no. 5, pp. 1899 –1910, May 2007

1910
[27]

A likeliho od-based multiple access for estimation in sensor networks ,

S. Marano, V . Matta, T. Lang, and P . Willett, “A likeliho od-based multiple access for estimation in sensor networks ,” IEEE Trans. on Signal Process. , vol. 55, no. 11, pp. 5155–5166, Nov. 2007

2007
[28]

Type based estimation over multi access channels,

G. Mergen and L. Tong, “Type based estimation over multi access channels,” IEEE Transactions on Signal Processing , vol. 54, no. 2, pp. 613–626, 2006

2006
[29]

Asymptotic detectio n performance of type-based multiple access over multiacce ss fading channels,

G. Mergen, V . Naware, and L. Tong, “Asymptotic detectio n performance of type-based multiple access over multiacce ss fading channels,” IEEE Trans. on Signal Process. , vol. 55, no. 3, pp. 1081 –1092, Mar. 2007. August 21, 2019 DRAFT 33

2007
[30]

Distributed detection in sensor networks over fading channels with mult iple antennas at the fusion centre,

I. Nevat, G. W. Peters, and I. B. Collings, “Distributed detection in sensor networks over fading channels with mult iple antennas at the fusion centre,” IEEE transactions on signal processing , vol. 62, no. 3, pp. 671–683, 2014

2014
[31]

Event d etection in sensor networks with non-linear ampliﬁers via m ixture series expansion,

P . Zhang, I. Nevat, G. W. Peters, and L. Clavier, “Event d etection in sensor networks with non-linear ampliﬁers via m ixture series expansion,” IEEE Sensors Journal , vol. 16, no. 18, pp. 6939–6946, 2016

2016
[32]

Type-based random access fo r distributed detection over multiaccess fading channels,

A. Anandkumar and L. Tong, “Type-based random access fo r distributed detection over multiaccess fading channels, ” IEEE Transactions on Signal Processing , vol. 55, no. 10, pp. 5032–5043, 2007

2007
[33]

Decision fusion over nonc oherent fading multiaccess channels,

F. Li, J. S. Evans, and S. Dey, “Decision fusion over nonc oherent fading multiaccess channels,” IEEE Transactions on Signal Processing, vol. 59, no. 9, p. 4367, 2011

2011
[34]

Optimal resour ce allocation for detection of a gaussian process using a mac in wsns,

J. A. Maya, L. R. V ega, and C. G. Galarza, “Optimal resour ce allocation for detection of a gaussian process using a mac in wsns,” IEEE Transactions on Signal Processing , vol. 63, no. 8, pp. 2057–2069, 2015

2015
[35]

Exploiting spa tial correlation in energy constrained distributed detect ion,

J. A. Maya, C. G. Galarza, and L. R. V ega, “Exploiting spa tial correlation in energy constrained distributed detect ion,” arXiv preprint arXiv:1509.04119 , 2015

2015 arXiv
[36]

Adaptive federated learning in resource constrained edge computing systems,

S. Wang, T. Tuor, T. Salonidis, K. K. Leung, C. Makaya, T. He, and K. Chan, “Adaptive federated learning in resource constrained edge computing systems,” IEEE Journal on Selected Areas in Communications , vol. 37, no. 6, pp. 1205–1221, 2019

2019
[37]

Optimizing pipelined computation and communication for latency-constrained ed ge learning,

N. Skatchkovsky and O. Simeone, “Optimizing pipelined computation and communication for latency-constrained ed ge learning,” arXiv preprint arXiv:1906.04488 , 2019

1906 arXiv
[38]

An incremental cl ustered gradient method for wireless sensor networks,

A. Mahmud, M. A. Adnan, and M. Shopon, “An incremental cl ustered gradient method for wireless sensor networks,” in 2018 21st Saudi Computer Society National Computer Confere nce (NCC) , April 2018, pp. 1–6

2018
[39]

Wireless compressi ve sensing over fading channels with distributed sparse ran dom projections,

T. Wimalajeewa and P . K. V arshney, “Wireless compressi ve sensing over fading channels with distributed sparse ran dom projections,” IEEE Transactions on Signal and Information Processing ove r Networks , vol. 1, no. 1, pp. 33–44, 2015

2015
[40]

Introductory lectures on convex programming volume i : Basic course,

“Introductory lectures on convex programming volume i : Basic course,” Lecture notes series / , vol. 3, no. 4, 1998

1998
[41]

Pegasos: Primal estimated sub-gradient solver for svm,

S. Shalev-Shwartz, Y . Singer, N. Srebro, and A. Cotter, “Pegasos: Primal estimated sub-gradient solver for svm,” Mathematical programming, vol. 127, no. 1, pp. 3–30, 2011

2011
[42]

Beyond the regret minimization ba rrier: optimal algorithms for stochastic strongly-convex optimization,

E. Hazan and S. Kale, “Beyond the regret minimization ba rrier: optimal algorithms for stochastic strongly-convex optimization,” The Journal of Machine Learning Research , vol. 15, no. 1, pp. 2489–2512, 2014

2014
[43]

Rob ust stochastic approximation approach to stochastic progr am- ming,

A. Nemirovski, A. Juditsky, G. Lan, and A. Shapiro, “Rob ust stochastic approximation approach to stochastic progr am- ming,” SIAM Journal on optimization , vol. 19, no. 4, pp. 1574–1609, 2009

2009
[44]

On projected stoch astic gradient descent algorithm with weighted averaging f or least squares regression,

K. Cohen, A. Nedi´ c, and R. Srikant, “On projected stoch astic gradient descent algorithm with weighted averaging f or least squares regression,” IEEE Transactions on Automatic Control , vol. 62, no. 11, pp. 5974–5981, 2017

2017
[45]

UCI machine learning repository,

M. Lichman, “UCI machine learning repository,” in University of California, Irvine, School of Information an d Computer Sciences, http://archive.ics.uci.edu/ml, 2013

2013
[46]

The million song dataset,

T. Bertin-Mahieux, D. P . Ellis, B. Whitman, and P . Lamer e, “The million song dataset,” in Proceedings of the 12th International Conference on Music Information Retrieval ( ISMIR 2011) , 2011

2011
[47]

Decentralized source loca lization and tracking [wireless sensor networks],

M. G. Rabbat and R. D. Nowak, “Decentralized source loca lization and tracking [wireless sensor networks],” in 2004 IEEE International Conference on Acoustics, Speech, and Si gnal Processing, vol. 3, May 2004, pp. iii–921

2004
[48]

Distributed optimization in se nsor networks,

M. Rabbat and R. Nowak, “Distributed optimization in se nsor networks,” in Third International Symposium on Information Processing in Sensor Networks, 2004. IPSN 2004 , April 2004, pp. 20–27. August 21, 2019 DRAFT

2004

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

A convergent increme ntal gradient method with a constant step size,

D. Blatt, A. Hero, and H. Gauchman, “A convergent increme ntal gradient method with a constant step size,” SIAM Journal on Optimization , vol. 18, pp. 29–51, 01 2007

work page 2007

[2] [2]

Incremental adaptive strate gies over distributed networks,

C. G. Lopes and A. H. Sayed, “Incremental adaptive strate gies over distributed networks,” IEEE Transactions on Signal Processing, vol. 55, no. 8, pp. 4064–4077, Aug 2007

work page 2007

[3] [3]

Stochastic inc remental gradient descent for estimation in sensor network s,

S. S. Ram, A. Nedic, and V . V . V eeravalli, “Stochastic inc remental gradient descent for estimation in sensor network s,” in 2007 Conference Record of the F orty-First Asilomar Confere nce on Signals, Systems and Computers , Nov 2007, pp. 582–586

work page 2007

[4] [4]

Incremental s tochastic subgradient algorithms for convex optimization ,

S. S. Ram, A. Nedi´ c, and V . V . V eeravalli, “Incremental s tochastic subgradient algorithms for convex optimization ,” SIAM Journal on Optimization , vol. 20, no. 2, pp. 691–717, 2009

work page 2009

[5] [5]

Lag: Lazily agg regated gradient for communication-efﬁcient distributed learning,

T. Chen, G. Giannakis, T. Sun, and W. Yin, “Lag: Lazily agg regated gradient for communication-efﬁcient distributed learning,” in Advances in Neural Information Processing Systems , 2018, pp. 5050–5060

work page 2018

[6] [6]

Federated optimization: Distributed machine learning f or on-device intelligence,

J. Koneˇ cn` y, H. B. McMahan, D. Ramage, and P . Richt´ arik , “Federated optimization: Distributed machine learning f or on-device intelligence,” arXiv preprint arXiv:1610.02527 , 2016

arXiv 2016

[7] [7]

Performance analysis of likelih ood-based multiple access for detection over fading channe ls,

K. Cohen and A. Leshem, “Performance analysis of likelih ood-based multiple access for detection over fading channe ls,” IEEE Transactions on Information Theory , vol. 59, no. 4, pp. 2471–2481, 2013

work page 2013

[8] [8]

Spectrum and energy efﬁcient multiple access for de tection in wireless sensor networks,

——, “Spectrum and energy efﬁcient multiple access for de tection in wireless sensor networks,” IEEE Transactions on Signal Processing, vol. 66, no. 22, pp. 5988–6001, 2018

work page 2018