Pith. sign in

REVIEW 3 major objections 5 minor 38 references

Scheduling Policies for Federated Learning in Wireless Networks

T0 review · 3 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 Federated learning's convergence rate in wireless networks is set by a single per-user quantity: the probability that an update is scheduled and successfully decoded.

desk verdict A useful framework with a real proof gap: the convergence-rate expressions are unsupported because η_t is equated to U_k^z when it should be G·U_k^z, but the core ideas and qualitative conclusions are salvageable. read the letter →

arxiv 1908.06287 v2 pith:EXF57ZTK submitted 2019-08-17 cs.IT eess.SPmath.IT

Howard H. Yang , Zuozhu Liu , Tony Q. S. Quek , H. Vincent Poor This is my paper · ORCID
classification cs.ITeess.SPmath.IT
keywords federatedlearningwirelessschedulingproportionalfairroundrobinrandomconvergenceratestochasticgeometrySINRthreshold
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

Federated learning in a wireless edge network cannot count on every user's update arriving: only a subset of user devices is scheduled in each round, and transmissions can be lost to interference. This paper develops an analytical model, based on random spatial layouts of access points and users, that expresses how fast federated learning converges in terms of a single per-user quantity: the probability that a user is both scheduled and successfully decoded under a given scheduling policy. Using that quantity, the paper derives closed-form expressions for the number of communication rounds needed to reach any target duality gap under random scheduling, round robin, and proportional fair scheduling. The paper's conclusion is that the best policy depends on the SINR decoding threshold: proportional fair converges fastest when the threshold is high, round robin when the threshold is low, and convergence slows sharply as the threshold rises. The practical upshot is that scheduling and signal-quality design directly control federated-learning training time, and that reducing the decoding threshold via compression or quantization can sometimes be worth more than improving local computation accuracy.

What carries the argument

The load-bearing object is the parameter update success probability $U_k^z = P(\gamma_{k,t} > \theta, S_{k,t}^z = 1)$: the joint probability that user $k$ is selected by the scheduler and that its packet decodes above the SINR threshold $\theta$. Algorithm 2 couples this to the optimization by setting the local learning rate $\eta_t$ to the time-averaged fraction of successful update indicators, so that users advance more aggressively when the channel is reliable and more cautiously when it is not. Lemma 2 shows that the expected increase in the dual objective in any round is no less than $U_k^z$ times the sum of local subproblem improvements, which converts the scheduling question into a calculation of $U_k^z$; Theorem 1 then turns that into round counts. The stochastic-geometry quantity $V(\theta,\alpha)$, an interference-plus-noise-to-signal measure, supplies the closed forms for $U_k^z$ under each policy.

What would settle it

Run Algorithm 2 on a convex task with known data size $n$ and record $\eta_t$ alongside the empirical fraction of successfully decoded scheduled updates; if $\eta_t$ converges to $G$ times that fraction rather than to the fraction itself, the premise of Lemma 2 fails and the predicted round counts $T_z$ would need rescaling.

Watch

Extended reading notes

Core claim

The central claim is Theorem 1: for any convergence target $\epsilon$, the wireless federated-learning procedure in Algorithm 2 — a distributed dual optimization method with an adaptive local learning rate — reaches an $\epsilon$ duality gap after at least $T_z \ge \log(\epsilon/n) / \log(1 - (1-\beta)U_k^z)$ communication rounds, where $\beta$ is the local-solver error level, $n$ is the total data size, and $U_k^z$ is the parameter update success probability of a user under scheduling policy $z$. The paper computes $U_k^z$ under three policies using stochastic geometry, obtaining closed forms for random scheduling, round robin, and proportional fair. From these expressions, the paper concludes that proportional fair scheduling outperforms the other two when the SINR threshold is high, because it selects users with the strongest channels; round robin is preferable when the SINR threshold is low, because almost every transmission succeeds and guaranteeing each user a turn matters more; and the convergence rate degrades rapidly as the SINR threshold increases. The analysis also identifies a trade-off between the number of scheduled users and subchannel bandwidth when the total spectrum is fixed.

Load-bearing premise

The load-bearing premise, used in the proof of Lemma 2 through eq. (26), is that the adaptive learning rate $\eta_t$ in Algorithm 2 converges exactly to each user's probability of a successful scheduled update, with no extra factor from the ratio of users to subchannels; the entire round-count bound of Theorem 1 rests on that alignment.

Editorial extensions

If this is right

  • If the central bound is correct, the communication-round count of wireless federated learning is determined by the scheduler through a single number $U_k^z$, so a new scheduling policy can be plugged into the same theorem by computing its update-success probability.
  • At high SINR thresholds, proportional fair scheduling can reduce required rounds by roughly a factor of $G$ relative to random or round robin, making channel-aware selection valuable exactly when transmissions are fragile.
  • At low SINR thresholds, round robin is preferable because per-transmission failures are rare and the binding constraint is guaranteeing each user a turn; fair time-sharing beats channel chasing.
  • Raising the decoding threshold degrades convergence quickly, so compressing or quantizing model updates to lower the required SINR is a direct lever on training time, sometimes outweighing gains from better local optimization — the paper's numerics show a 5 dB threshold reduction offsetting a six-fold larger error level.
  • A fixed spectrum budget forces a trade-off: scheduling more users per round increases the number of updates aggregated but splits the bandwidth and lowers success probability, and the analysis says the optimum lies at an interior number of subchannels.

Reading between the lines

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

  • Editorial inference: because the bound depends on $U_k^z$ only through the product $(1-\beta)U_k^z$, the framework suggests a general law of wireless federated learning — convergence speed is set by the probability that a user both computes and communicates successfully, so policies or protocols that raise either factor are interchangeable levers on training time.
  • Editorial inference: the SINR threshold $\theta$ can be reinterpreted as an abstraction of the entire physical layer — modulation, coding, quantization, and model size — so the model predicts that anything that moves $\theta$ down, such as stronger codes or fewer bits per parameter, buys convergence speed in a quantitatively predictable way.
  • Editorial inference: if the assumption that every access point sees statistically identical conditions is relaxed, the same per-cell success-probability machinery could be extended to study how uneven cell loads affect global convergence; the paper does not do this, but its single-cell analysis is a natural starting point.
  • Editorial inference: the three schedulers compared here are only a sample; the same analytic path would rate energy-aware, age-of-information-based, or delay-aware schedulers by computing their $U_k^z$, turning scheduler design for federated learning into a probabilistic optimization over one scalar.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The manuscript proposes a wireless federated learning (FL) algorithm (Algorithm 2) in which the local learning rate η_t is a running average of the number of successful UE parameter updates, adaptively coupling the optimization step to the wireless channel state. The authors model AP and UE locations as Poisson point processes, derive parameter update success probabilities under random scheduling (RS), round robin (RR), and proportional fair (PF) scheduling using stochastic geometry, and combine these quantities with a CoCoA-style dual coordinate ascent analysis to obtain explicit expressions for the number of communication rounds needed to reach an ε duality gap (Theorem 1). Based on these expressions, they conclude that PF outperforms RS and RR at high SINR thresholds, that RR is preferable at low SINR thresholds, and that the convergence rate degrades rapidly as the SINR threshold increases. The paper also includes SVM and CNN experiments that qualitatively support the scheduling-policy comparisons.

Significance. The paper addresses a timely and important problem: quantifying how wireless scheduling and inter-cell interference affect the convergence of FL. Its main contribution is a tractable analytical framework that connects physical-layer transmission success probabilities with an optimization-theoretic convergence rate. The success probabilities are derived from the channel model rather than fitted to the convergence goal, and the convergence analysis follows an established contraction framework, which is a genuine strength. If the proofs are repaired, the paper would provide a useful design tool for wireless FL systems and a clear basis for comparing scheduling policies. The experimental results are a further positive feature, although they do not by themselves compensate for the theoretical issue described below.

major comments (3)
  1. [Section IV-A, Eq. (26)] The assertion after Eq. (26) that U^z_k = E[η_t] is not correct as written. From Eq. (25), η_t = (1/(N t)) Σ_{l=0}^{t-1} Σ_{k=1}^K 1{S^z_{k,l}=1, γ_{k,l}>θ}. Since each indicator has expectation U^z_k by definition (16), the law of large numbers gives E[η_t] = (K/N) U^z_k = G U^z_k, not U^z_k. The equality U^z_k = E[η_t] holds only in the special case K = N (G = 1). This factor-of-G error is load-bearing because the proof of Lemma 2 in Appendix A uses E[η_t] as the convex-combination coefficient in Eq. (60). With the correct expectation, the contraction factor in Theorem 1, Eq. (30), should contain 1 − (1−β) G U^z_k rather than 1 − (1−β) U^z_k. Consequently, the round-count expression (30) and all corollary formulas that depend on it are not established as printed.
  2. [Corollary 2, Eq. (35), and Appendix D] There is an internal inconsistency in the treatment of RR. Definition (16) defines U^z_k as the joint probability P(γ_{k,t}>θ, S^z_{k,t}=1). For RR, a typical UE is scheduled in only one out of every G rounds, so the joint per-round success probability should be (1/G)(1/(1+V(θ,α))). However, Eq. (35) states U^RR_k = 1/(1+V(θ,α)) if scheduled and 0 otherwise, which is the success probability conditional on being scheduled, not the joint probability. Appendix D then multiplies the round count by G in Eq. (36) and in Eq. (70). This means the RR branch uses a different object than the RS and PF branches, so the comparison of RR with RS and PF in Theorem 1 is not well-defined on a common footing. The authors should either define U^z_k consistently as a per-round joint probability and remove the extra G factor, or explicitly re-derive the RR contraction over the G-round cycle.
  3. [Corollaries 1 and 3 and Section V] The numerical results in Section V are computed from the printed corollary formulas, which inherit the factor-of-G error from Theorem 1. For RS, the corrected contraction factor would be 1 − (1−β)/(1+V(θ,α)) rather than 1 − (1−β)/(G(1+V(θ,α))), assuming independent and identically distributed UEs. Similar corrections apply to the PF formula. The qualitative conclusions of the paper may survive this correction, since the G factors would partially cancel in the log-linear asymptotic comparisons of Remarks 6 and 7, but that is a repair that must be carried out explicitly. As it stands, Figures 3–5 do not validate the theorem as stated.
minor comments (5)
  1. [Eqs. (22), (25), (26), and throughout the appendix] The indicator function is rendered as "/BD" in several equations (e.g., Eq. (22), Eq. (25), Eq. (26), Eq. (57)–(59)). This appears to be a typesetting artifact and should be replaced with the standard indicator notation 1{·}.
  2. [Algorithm 2 and Eq. (25)] Algorithm 2 initializes η_0 = K/(2N), but the time-average expression in Eq. (25) does not include this initial value. For t ≥ 1, the recurrence (22) would yield η_t = (1/(N t)) Σ_{l=0}^{t-1} Σ_k 1{...} only if η_0 = 0. The role of the initialization should be clarified, or Eq. (25) should be stated as valid for t ≥ 1 with an explicit note that η_0 is not used.
  3. [Eq. (56)] The asymptotic expression after Eq. (56) writes T_NS ≳ log(n/ε)/log(β). Since β ∈ (0,1), both numerator and denominator would be negative, yielding a positive but confusing expression. This appears to be a sign typo and should read log(ε/n)/log(β), consistent with the paper's other round-count formulas.
  4. [Corollary 5] The proof of Corollary 5 is only a sketch. The statement that “both the parameter update success probability and the required communication rounds are increased by a factor of C” needs a more explicit derivation, since Theorem 1 as stated has no parameter C and the interaction between collecting more updates per global aggregation and the per-round contraction is not immediate.
  5. [Corollary 4] The approximation (1+θ)^{1/N} − 1 ≈ θ/N used in the proof of Corollary 4 is stated without conditions. Since the corollary is presented as a general result, the authors should specify the parameter regime in which this approximation is valid, or provide the exact expression.

Circularity Check

0 steps flagged · score 1.0 of 10

No material circularity: the convergence rate is assembled from independently derived SINR-based success probabilities plugged into a standard CoCoA-style contraction, and the self-citations used for stochastic-geometry tools are not load-bearing.

full rationale

The derivation chain is self-contained in the required sense: Lemma 1 obtains a dual reformulation by Fenchel-Rockafellar duality; Algorithm 2 performs distributed coordinate ascent; Lemma 2 bounds the expected dual increase in terms of U_z^k, the per-UE parameter update success probability defined in (16); Theorem 1 then converts the contraction factor 1 - (1 - beta) U_z^k into a round count; and Corollaries 1-3 compute U_z^k from the wireless model (PPP interference, Rayleigh fading, selection probabilities under RS, RR, and PF) using stochastic geometry. None of these steps fits a parameter to the convergence target, and none defines the predicted round count in terms of itself. The scheduling conclusions follow monotonically from the independently computed U_z^k values, so the comparison of PF versus RS versus RR is not forced by construction. The self-citations ([21]-[23], [28]) supply SINR-coverage and delay-analysis tools for the stochastic-geometry computations; the convergence contraction itself rests on the external CoCoA framework [37] and on convex-analysis results, so the self-citations are not load-bearing. The paper does contain a separate algebraic inconsistency: equation (25) defines eta_t as the time average of the number of successful UE updates divided by N, so E[eta_t] = (K/N) U_z^k = G U_z^k, not U_z^k as asserted in (26); this would affect the printed round-count formulas but is a correctness issue, not a circularity issue. Because no claimed prediction reduces by construction to its inputs, the circularity score is low; the only minor reason not to give a perfect zero is the presence of several author self-citations, none of which carries the central argument.

Assumptions & free parameters 0 free parameters · 7 assumptions · 0 invented entities

The central claim rests on the convexity and local-solver assumptions standard in distributed optimization, plus the stochastic-geometry network model. The paper introduces no fitted constants and no invented physical entities; the only hand-chosen constant is the 12/5 in the interference approximation, carried from prior work. The most serious load-bearing premise is the asserted equality between the algorithm's learning rate and the per-UE success probability (eq. 26), which is inconsistent with the algorithm's own update rule.

assumptions (7)
  • domain assumption Loss functions ℓ_i are 1/µ-smooth and the regularizer r is ζ-strongly convex (Assumptions 1 and 2, Section III).
    These convexity conditions justify strong duality, the Fenchel-Rockafellar dual (Lemma 1), and the smoothness/strong-convexity duality used in Lemma 2. They exclude non-convex models like the CNN tested in Section V-B, which the paper acknowledges.
  • domain assumption Each UE solves its local subproblem to error level β ∈ (0,1) per iteration (Assumption 3, Section IV-A).
    The contraction factor 1-(1-β)U^z_k in Theorem 1 depends linearly on β; if local solvers cannot meet this accuracy, the stated round count has no guarantee. This is a standard assumption in distributed coordinate descent [20].
  • ad hoc to paper The adaptive learning rate η_t converges to the per-UE update success probability U^z_k (eq. (26) in Section IV-A).
    This is the load-bearing coupling between the algorithm's local step sizes and the global aggregation success probability. From the algorithm's own definition in (22)/(25), E[η_t] equals G U^z_k, not U^z_k, so the premise is violated as written; Lemma 2 (Appendix A, eq. (60)) depends on it.
  • domain assumption Interfering UE locations are approximated by a non-homogeneous PPP with the void probability used in eq. (67), including the constant 12/5.
    The true uplink interference field is a Poisson-Voronoi perturbed lattice, for which exact Laplace transforms are unavailable; the approximation follows [38] and directly enters V(θ,α) and all U^z_k expressions. If the approximation is inaccurate, the closed-form convergence rates shift.
  • domain assumption AP locations form a homogeneous PPP with density λ, each AP has exactly K UEs uniformly distributed in its Voronoi cell, and all channels are i.i.d. Rayleigh block fading (Section II-A).
    This is the standard stochastic-geometry tractability model; it makes U^z_k computable and guarantees statistical equivalence of APs via Slivnyak's theorem. Deviating from it (e.g., correlated or non-uniform UE placement) breaks the derivations.
  • standard math Fenchel-Rockafellar duality and strong duality for problem (1) (Lemma 1 and Section IV-A).
    Used to transform the primal ERM into a dual objective whose duality gap measures solution quality. Requires convexity and a constraint qualification, which hold under Assumptions 1 and 2.
  • standard math Slivnyak's theorem and the stationarity of the PPP (Section III-B and Appendix C).
    Used to evaluate SINR and U^z_k at the typical AP; standard tool in stochastic geometry.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Scheduling Policies for Federated Learning in Wireless Networks." pith.science (2026). https://pith.science/paper/EXF57ZTK

@misc{pith2026190806287,
  author       = {Pith},
  title        = {Pith review of: Scheduling Policies for Federated Learning in Wireless Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EXF57ZTK}},
  note         = {Machine review of arXiv:1908.06287}
}
read the original abstract

Motivated by the increasing computational capacity of wireless user equipments (UEs), e.g., smart phones, tablets, or vehicles, as well as the increasing concerns about sharing private data, a new machine learning model has emerged, namely federated learning (FL), that allows a decoupling of data acquisition and computation at the central unit. Unlike centralized learning taking place in a data center, FL usually operates in a wireless edge network where the communication medium is resource-constrained and unreliable. Due to limited bandwidth, only a portion of UEs can be scheduled for updates at each iteration. Due to the shared nature of the wireless medium, transmissions are subjected to interference and are not guaranteed. The performance of FL system in such a setting is not well understood. In this paper, an analytical model is developed to characterize the performance of FL in wireless networks. Particularly, tractable expressions are derived for the convergence rate of FL in a wireless setting, accounting for effects from both scheduling schemes and inter-cell interference. Using the developed analysis, the effectiveness of three different scheduling policies, i.e., random scheduling (RS), round robin (RR), and proportional fair (PF), are compared in terms of FL convergence rate. It is shown that running FL with PF outperforms RS and RR if the network is operating under a high signal-to-interference-plus-noise ratio (SINR) threshold, while RR is more preferable when the SINR threshold is low. Moreover, the FL convergence rate decreases rapidly as the SINR threshold increases, thus confirming the importance of compression and quantization of the update parameters. The analysis also reveals a trade-off between the number of scheduled UEs and subchannel bandwidth under a fixed amount of available spectrum.

Figures

Figures reproduced from arXiv: 1908.06287 by the authors.

Figure 1
Figure 1. An illustration of the federated learning process: ( [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. A typical iteration round of the learning procedure: [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Normalized communication rounds vs UE number over su [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figures from the paper (4 more)
Figure 5
Figure 5. Figure 5: Normalized communication rounds vs number of subcha [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]
Figure 6
Figure 6. Figure 6: Comparison between Algorithm 1 and Algorithm 2: [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Test performance of the trained SVM with different sc [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]
Figure 8
Figure 8. Figure 8: Test performance of the trained CNNs with different s [PITH_FULL_IMAGE:figures/full_fig_p013_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 26 canonical work pages

  1. [1]

    A sur vey on mobile edge computing: The communication perspective,

    Y . Mao, C. Y ou, J. Zhang, K. Huang, and K. B. Letaief, “A sur vey on mobile edge computing: The communication perspective,” IEEE Commun. Surveys & Tutorials , vol. 19, no. 4, pp. 2322–2358, Aug. 2017

    work page 2017

  2. [2]

    H. H. Y ang and T. Q. S. Quek, Massive MIMO meets small cell: Backhaul and cooperation . SpringerBriefs in Computer Science, 2017

    work page 2017

  3. [3]

    Offloading i n mobile edge computing: Task allocation and computational frequen cy scaling,

    T. Q. Dinh, J. Tang, Q. D. La, and T. Q. S. Quek, “Offloading i n mobile edge computing: Task allocation and computational frequen cy scaling,” IEEE Trans. Commun. , vol. 65, no. 8, pp. 3571–3584, Aug. 2017

    2017

  4. [4]

    Deep learning f ramework for wireless systems: Applications to optical wireless com munications,

    H. Lee, S. H. Lee, T. Q. S. Quek, and I. Lee, “Deep learning f ramework for wireless systems: Applications to optical wireless com munications,” IEEE Commun. Mag. , vol. 57, no. 3, pp. 35–41, Mar. 2019

    work page 2019

  5. [5]

    Towar ds an intelligent edge: Wireless communication meets machine learning,

    G. Zhu, D. Liu, Y . Du, C. Y ou, J. Zhang, and K. Huang, “Towar ds an intelligent edge: Wireless communication meets machine learning,” Available as ArXiv:1809.00343 , 2018

    arXiv 2018

  6. [6]

    Wirele ss network intelligence at the edge,

    J. Park, S. Samarakoon, M. Bennis, and M. Debbah, “Wirele ss network intelligence at the edge,” Availabel as ArXiv:1812.02858 , 2018

    arXiv 2018

  7. [7]

    Federated opti miza- tion: Distributed optimization beyond the datacenter,

    J. Koneˇ cn` y, B. McMahan, and D. Ramage, “Federated opti miza- tion: Distributed optimization beyond the datacenter,” Available as ArXiv:1511.03575, 2015

    arXiv 2015

  8. [8]

    Federated learning: Strategies for improving co mmunication efficiency,

    J. Koneˇ cn` y, H. B. McMahan, F. X. Y u, P . Richt´ arik, A. T. Suresh, and D. Bacon, “Federated learning: Strategies for improving co mmunication efficiency,” arXiv preprint arXiv:1610.05492 , 2016

    arXiv 2016

Show all 38 references
  1. [9]

    Communication-efficient learning of deep networks from de centralized data,

    H. B. McMahan, E. Moore, D. Ramage, S. Hampson et al. , “Communication-efficient learning of deep networks from de centralized data,” Available as ArXiv:1602.05629 , 2016

    2016 arXiv

  2. [10]

    Client selection for federa ted learn- ing with heterogeneous resources in mobile edge,

    T. Nishio and R. Y onetani, “Client selection for federa ted learn- ing with heterogeneous resources in mobile edge,” arXiv preprint arXiv:1804.08333, 2018

    2018 arXiv

  3. [11]

    In -edge AI: Intelligentizing mobile edge computing, caching and co mmunication by federated learning,

    X. Wang, Y . Han, C. Wang, Q. Zhao, X. Chen, and M. Chen, “In -edge AI: Intelligentizing mobile edge computing, caching and co mmunication by federated learning,” arXiv preprint arXiv:1809.07857 , 2018

    2018 arXiv

  4. [12]

    LAG: Lazily aggregated gradient for communication-efficient distributed learnin g,

    T. Chen, G. B. Giannakis, T. Sun, and W. Yin, “LAG: Lazily aggregated gradient for communication-efficient distributed learnin g,” in Advances in Neural Information Processing Systems (NIPS) , Montreal, CANADA, Dec. 2018

    2018

  5. [13]

    Sparse communication for dist ributed gradient descent,

    A. F. Aji and K. Heafield, “Sparse communication for dist ributed gradient descent,” in Conference on Empirical Methods in Natural Language Processing (EMNLP) , Copenhagen, Denmark, Sep. 2017

    2017

  6. [14]

    Deep grad ient com- pression: Reducing the communication bandwidth for distri buted train- ing,

    Y . Lin, S. Han, H. Mao, Y . Wang, and W. J. Dally, “Deep grad ient com- pression: Reducing the communication bandwidth for distri buted train- ing,” in International Conference on Learning Representation (ICL R), V ancouver, Canada, May 2018

    2018

  7. [15]

    When edge meets learning: Adaptive control for res ource- constrained distributed machine learning,

    S. Wang, T. Tuor, T. Salonidis, K. K. Leung, C. Makaya, T. He, and K. Chan, “When edge meets learning: Adaptive control for res ource- constrained distributed machine learning,” in Proc. IEEE Conf. on Computer Commun. , Honolulu, HI, Apr. 2018, pp. 63–71

    2018

  8. [16]

    Adaptive federated learning in resource constrai ned edge com- puting systems,

    ——, “Adaptive federated learning in resource constrai ned edge com- puting systems,” IEEE J. Sel. Areas Commun. , vol. 37, no. 6, pp. 1205– 1221, Jun. 2019

    2019

  9. [17]

    Low-latency broadband an alog ag- gregation for federated edge learning,

    G. Zhu, Y . Wang, and K. Huang, “Low-latency broadband an alog ag- gregation for federated edge learning,” Available as ArXiv:1812.11494 , 2018

    2018 arXiv

  10. [18]

    Federated learni ng via over- the-air computation,

    K. Y ang, T. Jiang, Y . Shi, and Z. Ding, “Federated learni ng via over- the-air computation,” Available as ArXiv:1812.11750 , 2018

    2018 arXiv

  11. [19]

    Coded federate d comput- ing in wireless networks with straggling devices and imperf ect CSI,

    S. Ha, J. Zhang, O. Simeone, and J. Kang, “Coded federate d comput- ing in wireless networks with straggling devices and imperf ect CSI,” Available as ArXiv:1901.05239 , 2019

    1901 arXiv

  12. [20]

    Distributed optimization with arbitrary l ocal solvers,

    C. Ma, J. Koneˇ cn` y, M. Jaggi, V . Smith, M. I. Jordan, P . R icht´ arik, and M. Tak´ aˇ c, “Distributed optimization with arbitrary l ocal solvers,” Optimization Methods and Software , vol. 32, no. 4, pp. 813–848, 2017

    2017

  13. [21]

    The meta distribution of SIN R for small cell networks with temporal traffic,

    H. H. Y ang and T. Q. S. Quek, “The meta distribution of SIN R for small cell networks with temporal traffic,” in Proc. IEEE Int. Conf. Commun. , Shanghai, P . R. China, May 2019, pp. 1–6

    2019

  14. [22]

    Spatiotemporal analysis for SINR coverage in smal l cell net- works,

    ——, “Spatiotemporal analysis for SINR coverage in smal l cell net- works,” IEEE Trans. Commun. , vol. 67, no. 8, pp. 5520 – 5531, May 2019

    2019

  15. [23]

    Energy-efficien t design of MIMO heterogeneous networks with wireless backhaul,

    H. H. Y ang, G. Geraci, and T. Q. S. Quek, “Energy-efficien t design of MIMO heterogeneous networks with wireless backhaul,” IEEE Trans. Wireless Commun., vol. 15, no. 7, pp. 4914–4927, Jul. 2016

    2016

  16. [24]

    On stochastic geometry model ing of cellular uplink transmission with truncated channel inversion powe r control,

    H. ElSawy and E. Hossain, “On stochastic geometry model ing of cellular uplink transmission with truncated channel inversion powe r control,” IEEE Trans. Wireless Commun. , vol. 13, no. 8, pp. 4454–4469, Aug. 2014

    2014

  17. [25]

    Nearly unbiased variable selection unde r minimax con- cave penalty,

    C.-H. Zhang, “Nearly unbiased variable selection unde r minimax con- cave penalty,” The Annals of statistics , vol. 38, no. 2, pp. 894–942, 2010

    2010

  18. [26]

    Federated learn ing- enabled intelligent fog-radio access networks: Fundament al theory, key techniques, and future trends,

    Z. Zhao, C. Feng, H. H. Y ang, and X. Luo, “Federated learn ing- enabled intelligent fog-radio access networks: Fundament al theory, key techniques, and future trends,” IEEE Wireless Commun. Mag., submitted

  19. [27]

    Communication-efficient al gorithms for decentralized and stochastic optimization,

    G. Lan, S. Lee, and Y . Zhou, “Communication-efficient al gorithms for decentralized and stochastic optimization,” Math. Program. , pp. 1–48, Dec. 2018

    2018

  20. [28]

    Delay analysis of r andom scheduling and round robin in small cell networks,

    H. H. Y ang, Y . Wang, and T. Q. S. Quek, “Delay analysis of r andom scheduling and round robin in small cell networks,” IEEE Wireless Commun. Lett. , vol. 7, no. 6, pp. 978– 981, Dec. 2018

    2018

  21. [29]

    Cell-throughput analysis of th e proportional fair scheduler in the single-cell environment,

    J.-G. Choi and S. Bahk, “Cell-throughput analysis of th e proportional fair scheduler in the single-cell environment,” IEEE Trans. V ehicular Tech., vol. 56, no. 2, pp. 766–778, Mar. 2007

    2007

  22. [30]

    Convex optimization: Algorithms and compl exity,

    S. Bubeck, “Convex optimization: Algorithms and compl exity,” F ounda- tions and Trends R© in Machine Learning , vol. 8, no. 3-4, pp. 231–357, 2015

    2015

  23. [31]

    R. T. Rockafellar, Convex analysis . Princeton university press, 1970, vol. 28

    1970

  24. [32]

    Hiriart-Urruty and C

    J.-B. Hiriart-Urruty and C. Lemar´ echal, Fundamentals of convex anal- ysis. Springer Science & Business Media, 2012

    2012

  25. [33]

    Baccelli and B

    F. Baccelli and B. Blaszczyszyn, Stochastic Geometry and Wireless Networks. V olumn I: Theory . Now Publishers, 2009

    2009

  26. [34]

    Haenggi, Stochastic geometry for wireless networks

    M. Haenggi, Stochastic geometry for wireless networks . Cambridge University Press, 2012

    2012

  27. [35]

    Adding vs. averaging in distributed primal-dual optimiza tion,

    C. Ma, V . Smith, M. Jaggi, M. I. Jordan, P . Richt´ arik, an d M. Tak´ aˇ c, “Adding vs. averaging in distributed primal-dual optimiza tion,” in Int. Conf. Machine Learn. (ICML) , Lille, France, Jul. 2015

    2015

  28. [36]

    Conve rgence anal- ysis of quantized primal-dual algorithms in network utilit y maximization problems,

    E. Nekouei, T. Alpcan, G. N. Nair, and R. J. Evans, “Conve rgence anal- ysis of quantized primal-dual algorithms in network utilit y maximization problems,” IEEE Trans. Control of Network Systems , vol. 5, no. 1, pp. 284–297, Mar. 2018

    2018

  29. [37]

    CoCoA: A general framework for communication-efficient di stributed optimization,

    V . Smith, S. Forte, C. Ma, M. Tak´ aˇ c, M. I. Jordan, and M. Jaggi, “CoCoA: A general framework for communication-efficient di stributed optimization,” Journal of Machine Learning Research , vol. 18, no. 230, pp. 1–49, 2018

    2018

  30. [38]

    User point processes in cellular networks ,

    M. Haenggi, “User point processes in cellular networks ,” IEEE Wireless Commun. Lett. , vol. 6, no. 2, pp. 258–261, Apr. 2017

    2017

Pith tools

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