Pith. sign in

REVIEW 5 major objections 6 minor 33 references

Fast State-Augmented Learning for Wireless Resource Allocation with Dual Variable Regression

T0 review · 5 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Dual-variable regression lets learned power control meet rate constraints twice as fast.

desk verdict A practical, incremental improvement to state-augmented learning that works empirically, but the theoretical justification for 'near-optimal initialization' rests on an unproven assumption and the excursion theorem is not correctly stated; worth a serious referee, not acceptance as-is. read the letter →

arxiv 2506.18748 v2 pith:WJGHFCW6 submitted 2025-06-23 eess.SP cs.LG

Yigit Berkay Uslu , Navid NaderiAlizadeh , Mark Eisen , Alejandro Ribeiro This is my paper · ORCID
classification eess.SPcs.LG
keywords wirelessresourceallocationpowercontrolgraphneuralnetworksstateaugmentationdualvariableregressionsubgradientdescentconstrainedlearningergodicconstraints
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 is trying to establish that a learned wireless resource allocation policy can be made to respect long-term per-user constraints much faster at inference time if the dual multipliers of the constrained optimization are initialized near their optimal values rather than at zero, and if those multipliers are learned by a second neural network instead of being tuned online. The authors propose a state-augmented graph neural network policy in which the dual variables are treated as graph signals, together with a dual-regression module that predicts a near-optimal starting multiplier vector from the network configuration and cheap features. They also show that training the policy on dual multipliers sampled from the actual dual descent trajectories of the trained policy, rather than from a fixed prior, substantially improves the learned Lagrangian maximization. In a 100-user power control case study, the proposed algorithm brings the 1st and 5th percentile user rates within a few percent of the minimum rate requirement in roughly half the time of the state-augmented baseline, and it transfers to larger networks. Theoretical results give a convergence rate for the expected dual iterates and an exponential bound on rare large excursions of the dual optimality gap, which supports the initialization strategy.

What carries the argument

The machinery is a state-augmented graph neural network pair operating on a graph whose nodes are transmitter-receiver pairs, whose edge weights are log-normalized channel gains, and whose node features are the current dual multipliers (primal-GNN) or regression features (dual-GNN). The primal-GNN maximizes the parametrized Lagrangian $L_\phi(\phi;\lambda,H)=f_0(r(H,p_\phi(H,\lambda;\phi)))+\lambda\cdot f(r(H,p_\phi(H,\lambda;\phi)))$ over $\phi$ for duals sampled from the dual descent buffer, and the dual-GNN solves a supervised regression problem whose targets are the time-averaged dual trajectories $\lambda^\dagger(H;\phi^*)$ and whose inputs include the fixed-power rate vector as an easy proxy for constraint difficulty. The identity that carries the argument is the near-equality $D_\lambda(H;\phi^*)\approx D_\lambda(H;\theta^\dagger)$ between the dual dynamics induced by the trained policy and those of the true Lagrangian maximizer; under this identity the regression targets are near-optimal, and the dual dynamics update formula (16) drives the system along a stochastic subgradient descent on the dual function. Supporting theory shows the dual function is convex in $\lambda$, its subgradients come from Danskin's theorem, the expected time-averaged dual iterate approaches the optimal set at rate $O(1/K)$, and excursions of the optimality gap beyond a neighborhood decay exponentially.

What would settle it

Compute the true optimal dual multipliers for a small network by solving the dual problem exactly, train the primal and dual GNNs, and compare the dual-GNN's predictions to the true optima; if predictions are accurate but the claimed half-time feasibility gain over zero initialization disappears when the primal policy has limited capacity, the central claim fails. A sharper version is to lower the capacity of the primal-GNN until the approximation error to $\theta^\dagger(\lambda)$ is large, then check whether the regression targets $\lambda^\dagger(H;\phi^*)$ stop being near-optimal and whether the runtime gain vanishes.

Watch

Extended reading notes

Core claim

The paper's central claim is that fast, feasible, near-optimal inference for ergodically constrained wireless resource allocation can be decomposed into two learned components: a state-augmented policy $p_\phi(H,\lambda;\phi)$ that approximates the Lagrangian maximizer for every dual input $\lambda$, and a dual-regression function $d_\psi(X(H),H;\psi)$ that predicts the near-optimal dual multiplier $\lambda^\dagger(H;\phi^*)$, defined as the expected time-average of the dual descent dynamics run under the trained policy. During online execution the policy receives the predicted multiplier as its initial dual state and then continues the stochastic dual descent updates $\lambda_{k+1}=[\lambda_k-\eta_\lambda f(\cdot)]_+$, so that feasibility is achieved by the ergodic behavior of the trajectory rather than by a single deterministic allocation. The authors argue that because the dual dynamics visit a neighborhood of the dual optimum regularly and rarely make large excursions, starting inside that neighborhood removes the slow initial phase in which zero-initialized multipliers generate suboptimal decisions. Empirically this is shown in a transmit power control problem where SA+DR reaches near-feasible rates for all users in about half the time of the SA baseline, while preserving or improving the mean rate.

Load-bearing premise

The load-bearing premise is that the trained state-augmented policy is close enough to a true Lagrangian maximizer that the dual descent trajectories used as regression targets are near-optimal; the paper states this as $D_\lambda(H;\phi^*)\approx D_\lambda(H;\theta^\dagger)$ and does not prove it, and it also relies on finite-time stochastic subgradients being nearly unbiased, which the authors acknowledge is violated for finite update windows.

Editorial extensions

If this is right

  • If the claim holds, learned policies can satisfy per-user ergodic constraints in finite time without solving an inner optimization online: one forward pass of the dual-GNN replaces the slow zero-initialized transient.
  • Training on dual descent trajectories converts the choice of dual sampling distribution from a tuned hyperparameter into a quantity that matches the actual dynamics, so the same recipe should transfer across step sizes and constraint severities without hand-tuning the dual prior.
  • Because feasibility comes from policy switching, alternating high-power and zero-power transmit decisions rather than a single fixed point, the approach predicts that any deployment must preserve the temporal randomization of the primal policy, not just its averaged output.
  • The convergence and excursion results imply the benefit of near-optimal initialization grows as the dual step size shrinks, since the $\|\lambda_0-\lambda^*\|^2/(2\eta_\lambda K)$ term in the finite-time bound becomes dominant; the paper's step-size ablations are consistent with this.
  • GNN parametrization means the primal and dual models are permutation equivariant and transferable: policies trained on 100-user networks keep near-feasible low-percentile rates on up to 400-user networks at fixed density.

Reading between the lines

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

  • An extension the author leaves implicit is that the same dual-regression initialization applies to any constrained learning problem with separable dual structure, not just power control; the regression features would need to be cheap proxies for constraint tightness, as the fixed-power rates are here.
  • The paper's assumption that the trained policy is close to the Lagrangian maximizer could be tested by comparing dual-GNN predictions against exactly computed optimal multipliers in small networks; the scatter plot in Fig. 3 is consistent with accurate prediction, but no exact comparison is reported.
  • In more interference-limited regimes where many constraints bind, the optimal multipliers are larger and less sparse, so the gains from learned initialization could be larger than in the reported setup; the authors note this direction explicitly.
  • Treating dual variables as graph signals suggests a continuous-time or asynchronous version of the algorithm, where different users' multipliers update at different rates; the convergence and excursion analysis would need to be re-derived for that setting.
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

5 major / 6 minor

Summary. The paper proposes a state-augmented graph neural network (GNN) approach to wireless resource allocation, augmented with a dual variable regression (DR) module that predicts near-optimal initial dual multipliers. During training, the primal GNN is trained on dual multipliers sampled from its own dual descent dynamics, and a secondary dual GNN is trained to predict time-averaged dual multipliers from network configurations and simple features. At inference, the dual multipliers are initialized with the dual-GNN predictions and then updated by stochastic dual descent while the primal GNN generates power allocations. The authors provide convergence and excursion results for stochastic dual gradient descent (DGD) and report experiments on a transmit power control problem showing that SA+DR reaches near-feasible rates roughly twice as fast as the SA baseline.

Significance. If the main claims hold, the paper offers a practically useful way to accelerate the transient phase of learned primal-dual resource allocation, and the GNN transferability results are a genuine strength. The empirical evaluation is extensive: it includes ablations against SA with zero initialization, an ablated uniform-prior SA, and non-learning baselines, plus a transfer experiment to larger networks. The paper also provides machine-checkable derivations of the convergence and excursion bounds, although those derivations contain errors that must be fixed. The central algorithmic contribution, however, rests on an unverified equivalence between the dual dynamics of the trained suboptimal policy and those of the exact Lagrangian maximizer, which is the main load-bearing assumption.

major comments (5)
  1. [Section III.B, Eqs. (19)-(22)] The claim that the dual regression targets are near-optimal multipliers is only supported by the assumption Dλ(H;φ*)≈Dλ(H;θ†) stated after Eq. (19). This assumption is not proven, and it is not validated by the experiments: Fig. 3 compares the dual-GNN predictions only to the self-generated targets λ†(H;φ*), not to an independent estimate of λ*(H). Since Proposition 1 applies to DGD iterates with exact Lagrangian maximizers θ†(λ), the theoretical grounding of the DR initialization collapses if this assumption fails. The authors should either prove the assumption under explicit conditions, validate it numerically against an independent solver, or reframe the DR initialization as a heuristic warm start and remove the term 'near-optimal initialization' from the theoretical claims.
  2. [Section IV.D, Eq. (38)] The excursion bound in Theorem 2 is malformed as printed. The denominator 1 + min{0, γ − Gmax − 1} is zero when γ = Gmax and negative for γ < Gmax + 1, making the right-hand side undefined or larger than 1 in some ranges. As a probability bound, P(G†0 ≥ γ | G0) must be a nonincreasing function of γ that is at most 1; the displayed formula does not satisfy this. Please restate the theorem with a corrected bound and verify it numerically for the stated parameter ranges.
  3. [Appendix E] The proof of Theorem 2 states that the stopped process H_{l∧L*} is 'a martingale by an appeal to Doob’s optional stopping theorem.' A stopped supermartingale is a supermartingale, not necessarily a martingale, and the optional stopping theorem for supermartingales gives E[H_{L*} | H0] ≤ H0, which is indeed what the proof later uses. The ''martingale'' claim is incorrect and should be removed; the proof should explicitly state the supermartingale optional stopping result and verify that its conditions (e.g., bounded or integrable stopping time) are satisfied.
  4. [Section IV.B, Proposition 1 and Appendix C] The transition from Eq. (58) to Eq. (59) is not justified. Inequality (58) bounds gθ(E[λ̄K]) − D*θ, but this does not imply the distance bound ||λ†K − λ*|| ≤ B(1 + ||λ0−λ*||²/(K η²S²)) without an additional strong convexity or moduli-of-continuity condition on gθ near the optimum. Corollary 2 provides an inclusion G_ϵ ⊆ Λ_{B_ϵ} only for a fixed ϵ, not for the K-dependent bound in (58). This is a load-bearing step in the motivation for the DR initialization, so the proof must be completed or the statement weakened.
  5. [Section IV.A, Assumption 1] The paper acknowledges that the unbiasedness assumption is violated for finite update windows T0, yet Proposition 1, Proposition 2, and Theorem 2 all rely on Assumption 1 without qualification. This is not fatal if the authors explicitly state that all theoretical guarantees are approximate for the actual algorithm and quantify the bias under a suitable T0-dependent condition. As written, the theorems claim more than the stated assumptions deliver.
minor comments (6)
  1. [Abstract] The abstract uses both 'viewing dual variables as graph signals' and later 'viewed as graph signals'; please harmonize the phrasing.
  2. [Eq. (1)] The notation Ht = H ⊙ Hst defines an elementwise product, but ⊙ is not defined in the text; please define it explicitly.
  3. [Fig. 3 caption] The caption says the green cone covers a 95% confidence interval, but no statistical procedure for constructing this interval is described in the text or supplementary material; please specify the method.
  4. [Section V.D] The statement that the time to reach near-feasible rates is 'roughly halved' is not quantified; please specify the exact threshold (e.g., within 5% of fmin) and report the measured time for SA+DR and SA.
  5. [Throughout] The abbreviation 'c.f.' is used inconsistently; please use 'cf.' and define the intended meaning, or replace with 'see' as appropriate.
  6. [References] Some references are to arXiv preprints or early-access papers; please update them to their final published versions where available.

Circularity Check

1 steps flagged · score 4.0 of 10

Dual-regression targets are generated by the trained primal policy itself; the paper's 'near-optimal' label rests on an explicit but unproven equality, so the theoretical grounding is partially self-referential.

  1. fitted input called prediction [Section III-B, Eq. (19); Section III-C, Eq. (20); Section V-D, Fig. 3]
    "where the regression targets are the near-optimal dual multipliers that are estimated by rolling out state-augmented dynamics that descend the state-augmented Lagrangian over the training dataset of network configurations ... Assuming Dλ(H;ϕ*) ≈ Dλ(H;θ†) for an optimal state-augmented policy ϕ*, (19) serves as a theoretically grounded method for near-optimal initialization of state-augmented dynamics."

    Eq. (19) defines λ†(H;ϕ*) as the expected time-average of the dual iterates generated by (16), i.e., by the trained state-augmented policy ϕ* itself. Eq. (20) then trains the dual-GNN dψ to regress onto these self-generated targets, and Eq. (22) uses dψ for initialization. The only bridge from these self-generated trajectory averages to true optimality is the asserted equality Dλ(H;ϕ*) ≈ Dλ(H;θ†), which is assumed rather than proved and is not tested against an independent estimate of λ*(H): Fig. 3 compares predictions only to the self-generated targets just described. Consequently, the 'predicted optimal dual multipliers' are, by construction, predictions of the algorithm's own dual trajectories; their near-optimality is an assumption, not a derived result.

full rationale

The core theoretical results (Proposition 1, Proposition 2, Theorem 2) are proven for the exact DGD iterates θ†(λ) and are not themselves circular: they are standard stochastic subgradient arguments developed in Appendices A-E. The potential circularity is confined to the dual-regression module. Eq. (19) defines the regression target λ†(H;ϕ*) as the expected time-average of the dual updates in (16), which are generated by the trained state-augmented policy ϕ*. The dual-GNN is then trained on these targets via (20)-(21) and used for initialization via (22). The paper's only bridge from these self-generated trajectory averages to the true optimal multipliers λ*(H) is the sentence 'Assuming Dλ(H;ϕ*) ≈ Dλ(H;θ†)'; this equality is asserted, not proven, and Fig. 3 validates the dual-GNN only against the same self-generated targets, never against an independent estimate of λ*(H). Under the stated assumption the method is theoretically grounded, so this is a conditional grounding rather than a fully circular derivation. The self-citations to [20] and [21] (Theorem 1, Definition 1, Lemma 3) are load-bearing for the base state-augmented guarantees, but they are prior published results with stated assumptions; they are not the mechanism that makes the dual-variable regression prediction equal to its own input. The empirical comparison in Fig. 1 (SA+DR vs SA and ablations) is an external check, which is why the paper is not scored higher. Score 4 reflects one central prediction target that is self-generated, with an explicit but unproven near-optimality assumption, while the main algorithm still has independent experimental content.

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

The central claim rests on stochastic approximation assumptions (unbiasedness and bounded moments), a Slater-type feasibility condition, a local strong convexity/smoothness condition introduced purely for the excursion bound, the unproven distributional match Dλ(H;φ*)≈Dλ(H;θ†), and the standard assumption of a small duality gap for the parametrized problem. No new physical entities are introduced.

free parameters (4)
  • Dual step size ηλ = 0.2
    Chosen as a hyperparameter for dual descent; affects the near-optimality gap O(ηλ) in (12) and the excursion bound. Not fitted to data.
  • Dual update window T0 = 5 time steps
    Chosen update period for dual multipliers in (16); determines the bias of stochastic subgradient estimates. Carried over from [21].
  • λmax scaling for dual-GNN = 50.0
    Scaling hyperparameter in the negated log-sigmoid output (97); chosen by hand.
  • Training epoch counts NSA, NDR = 100 each
    Training lengths chosen by hand; not fitted to achieve the theoretical results but affect the quality of φ* and ψ*.
assumptions (5)
  • domain assumption Assumption 1: unbiased stochastic subgradients with bounded second moment E[||ŝ(λ)||^2] ≤ S^2
    Used throughout the convergence analysis (Lemmas 2-3, Propositions 1-2, Theorem 2). Acknowledged by the authors to be violated in practice due to finite update window T0.
  • domain assumption Assumption 2: existence of a strictly feasible policy with gap ξ > 0
    Needed for Lemma 1's bound on dual multipliers and for the ergodic feasibility results (11), (17). Standard Slater-type condition.
  • ad hoc to paper Assumption 3: local L-smoothness and m-strong convexity of the dual function on an excursion subregion E
    Introduced solely for the exponential excursion bound in Theorem 2; not standard for this problem class and hard to verify.
  • ad hoc to paper Dλ(H;φ*) ≈ Dλ(H;θ†) for an optimal state-augmented policy φ*
    Stated in Section III.B to justify using the time-average of the learned policy's dual dynamics as near-optimal regression targets. Not proven; the actual regression targets are generated by φ*.
  • domain assumption Small duality gap of the parametrized problem (Remark 1)
    Relies on [24] to argue that working in the dual domain incurs small loss; not proven in this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Fast State-Augmented Learning for Wireless Resource Allocation with Dual Variable Regression." pith.science (2026). https://pith.science/paper/WJGHFCW6

@misc{pith2026250618748,
  author       = {Pith},
  title        = {Pith review of: Fast State-Augmented Learning for Wireless Resource Allocation with Dual Variable Regression},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WJGHFCW6}},
  note         = {Machine review of arXiv:2506.18748}
}
read the original abstract

We consider resource allocation problems in multi-user wireless networks, where the goal is to optimize a network-wide utility function subject to constraints on the ergodic average performance of users. We demonstrate how a state-augmented graph neural network (GNN) parametrization for the resource allocation policy circumvents the drawbacks of the ubiquitous dual subgradient methods by representing the network configurations (or states) as graphs and viewing dual variables as dynamic inputs to the model, treated as graph signals supported over the graphs. Lagrangian maximizing state-augmented policies are learned during the offline training phase, and the dual variables evolve through gradient updates while executing the learned state-augmented policies during the inference phase. Our main contributions are to illustrate how near-optimal initialization of dual multipliers for faster inference can be accomplished with dual variable regression, leveraging a secondary GNN parametrization, and how maximization of the Lagrangian over the multipliers sampled from the dual descent dynamics substantially improves the training of state-augmented models. We demonstrate the superior performance of the proposed algorithm with extensive numerical experiments in a case study of transmit power control. Finally, we prove a convergence result and an exponential probability bound on the excursions of the dual function (iterate) optimality gaps.

Figures

Figures reproduced from arXiv: 2506.18748 by the authors.

Figure 1
Figure 1. Comparison of time evolution of mean, 1st, and 5th percentiles of all receiver ergodic rates for all the algorithms. [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗
Figure 2
Figure 2. Comparison of dual multipliers, constraints and power allocations over time steps with (SA+DR) and without (SA) [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Scatter plot of target (optimal) dual multipliers against [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: (a) Example policy switching behavior of two neighboring (mutually-interfering) users. Dashed lines indicate the time [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: Transferability of the policies learned on networks with [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Progress of the SA+DR training algorithm. Leftmost plot shows the state-augmented Lagrangian and the objective [PITH_FULL_IMAGE:figures/full_fig_p015_6.png]
Figure 8
Figure 8. Figure 8: Scatter plot of DR features and against the DR [PITH_FULL_IMAGE:figures/full_fig_p015_8.png]
Figure 9
Figure 9. Figure 9: (a) Plot of mean and 5th percentile rates for varying dual step size choices. We see that loss of objective utility [PITH_FULL_IMAGE:figures/full_fig_p017_9.png]
Figure 10
Figure 10. Figure 10: Additional examples of SA+DR and SA algorithm roll-outs for several test network users. [PITH_FULL_IMAGE:figures/full_fig_p017_10.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 29 canonical work pages

  1. [1]

    ITLinQ: A new approach for spectrum sharing in device-to-device communication systems,

    Navid Naderializadeh and Amir Salman Avestimehr, “ITLinQ: A new approach for spectrum sharing in device-to-device communication systems,” IEEE Journal on Selected Areas in Communications , vol. 32, no. 6, pp. 1139–1151, 2014

    work page 2014

  2. [2]

    ITLinQ+: An improved spectrum sharing mechanism for device-to-device communications,

    Xinping Yi and Giuseppe Caire, “ITLinQ+: An improved spectrum sharing mechanism for device-to-device communications,” in 2015 49th Asilomar Conference on Signals, Systems and Computers . IEEE, 2015, pp. 1310–1314

    work page 2015

  3. [3]

    FPLinQ: A cooperative spectrum sharing strategy for d2d communications,

    Kaiming Shen and Wei Yu, “FPLinQ: A cooperative spectrum sharing strategy for d2d communications,” in 2017 IEEE International Sympo- sium on Information Theory (ISIT) . IEEE, 2017, pp. 2323–2327

    work page 2017

  4. [4]

    Optimal wireless resource alloca- tion with random edge graph neural networks,

    Mark Eisen and Alejandro Ribeiro, “Optimal wireless resource alloca- tion with random edge graph neural networks,” IEEE Transactions on Signal Processing, vol. 68, pp. 2977–2991, 2020

    work page 2020

  5. [5]

    Resource management in wireless networks via multi- agent deep reinforcement learning,

    Navid Naderializadeh, Jaroslaw J Sydir, Meryem Simsek, and Hosein Nikopour, “Resource management in wireless networks via multi- agent deep reinforcement learning,” IEEE Transactions on Wireless Communications, vol. 20, no. 6, pp. 3507–3523, 2021

    work page 2021

  6. [6]

    Intelligent O-RAN for Beyond 5G and 6G Wireless Networks

    Solmaz Niknam, Abhishek Roy, Harpreet S Dhillon, Sukhdeep Singh, Rahul Banerji, Jeffery H Reed, Navrati Saxena, and Seungil Yoon, “Intelligent o-ran for beyond 5g and 6g wireless networks,” arXiv preprint arXiv:2005.08374, 2020

    work page Pith review arXiv 2005

  7. [7]

    Unfolding wmmse using graph neural networks for efficient power allocation,

    Arindam Chowdhury, Gunjan Verma, Chirag Rao, Ananthram Swami, and Santiago Segarra, “Unfolding wmmse using graph neural networks for efficient power allocation,” IEEE Transactions on Wireless Commu- nications, vol. 20, no. 9, pp. 6004–6017, 2021

    work page 2021

  8. [8]

    Unsupervised learning for asynchronous resource allocation in ad-hoc wireless net- works,

    Zhiyang Wang, Mark Eisen, and Alejandro Ribeiro, “Unsupervised learning for asynchronous resource allocation in ad-hoc wireless net- works,” in 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2021, pp. 8143–8147

    work page 2021

Show all 33 references
  1. [9]

    Edge artificial intelligence for 6g: Vision, enabling technologies, and appli- cations,

    Khaled B Letaief, Yuanming Shi, Jianmin Lu, and Jianhua Lu, “Edge artificial intelligence for 6g: Vision, enabling technologies, and appli- cations,” IEEE Journal on Selected Areas in Communications , vol. 40, no. 1, pp. 5–36, 2021

    2021

  2. [10]

    Link scheduling using graph neural networks,

    Zhongyuan Zhao, Gunjan Verma, Chirag Rao, Ananthram Swami, and Santiago Segarra, “Link scheduling using graph neural networks,” IEEE Transactions on Wireless Communications , vol. 22, no. 6, pp. 3997– 4012, 2023

    2023

  3. [11]

    Modular meta-learning for power control via random edge graph neural networks,

    Ivana Nikoloska and Osvaldo Simeone, “Modular meta-learning for power control via random edge graph neural networks,” arXiv preprint arXiv:2108.13178, 2021

    2021 arXiv

  4. [12]

    Power allocation for wireless federated learning using graph neural networks,

    Boning Li, Ananthram Swami, and Santiago Segarra, “Power allocation for wireless federated learning using graph neural networks,” in ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) . IEEE, 2022, pp. 5243–5247

    2022

  5. [13]

    Regularization strategy aided robust unsupervised learning for wireless resource allocation,

    Hao Huang, Yun Lin, Guan Gui, Haris Gacanin, Hikmet Sari, and Fumiyuki Adachi, “Regularization strategy aided robust unsupervised learning for wireless resource allocation,” IEEE Transactions on Vehicular Technology, vol. 72, no. 7, pp. 9647–9652, 2023

    2023

  6. [14]

    Learning to slice wi-fi networks: A state-augmented primal-dual approach,

    Yi ˘git Berkay Uslu, Roya Doostnejad, Alejandro Ribeiro, and Navid NaderiAlizadeh, “Learning to slice wi-fi networks: A state-augmented primal-dual approach,” in 2024 IEEE Global Communications Confer- ence (GLOBECOM), 2024, pp. 4521–4527

    2024

  7. [15]

    Opportunistic routing in wireless communications via learn- able state-augmented policies,

    Sourajit Das, Navid NaderiAlizadeh, Rahul Mangharam, and Alejandro Ribeiro, “Opportunistic routing in wireless communications via learn- able state-augmented policies,” arXiv preprint arXiv:2405.05748, 2025

    2025

  8. [16]

    Diffusion model based resource allocation strategy in ultra-reliable wireless networked control systems,

    Amirhassan Babazadeh Darabi and Sinem Coleri, “Diffusion model based resource allocation strategy in ultra-reliable wireless networked control systems,” IEEE Communications Letters , vol. 29, no. 1, pp. 85–89, 2025

    2025

  9. [17]

    Learning optimal resource allocations in wireless systems,

    Mark Eisen, Clark Zhang, Luiz FO Chamon, Daniel D Lee, and Alejandro Ribeiro, “Learning optimal resource allocations in wireless systems,” IEEE Transactions on Signal Processing , vol. 67, no. 10, pp. 2775–2790, 2019

    2019

  10. [18]

    Optimal resource allocation in wireless communi- cation and networking,

    Alejandro Ribeiro, “Optimal resource allocation in wireless communi- cation and networking,” EURASIP Journal on Wireless Communications and Networking, vol. 2012, no. 1, pp. 1–19, 2012

    2012

  11. [19]

    Learning resilient radio resource management policies with graph neural net- works,

    Navid NaderiAlizadeh, Mark Eisen, and Alejandro Ribeiro, “Learning resilient radio resource management policies with graph neural net- works,” IEEE Transactions on Signal Processing, vol. 71, pp. 995–1009, 2023

    2023

  12. [20]

    State augmented constrained reinforcement learning: Overcoming the limitations of learning with rewards,

    Miguel Calvo-Fullana, Santiago Paternain, Luiz FO Chamon, and Ale- jandro Ribeiro, “State augmented constrained reinforcement learning: Overcoming the limitations of learning with rewards,” arXiv preprint arXiv:2102.11941, 2021

    2021 arXiv

  13. [21]

    State- augmented learnable algorithms for resource management in wireless networks,

    Navid NaderiAlizadeh, Mark Eisen, and Alejandro Ribeiro, “State- augmented learnable algorithms for resource management in wireless networks,” IEEE Transactions on Signal Processing , 2022

    2022

  14. [22]

    Boyd and Lieven Vandenberghe, Convex optimization , Cambridge University Press, 2004

    Stephen P. Boyd and Lieven Vandenberghe, Convex optimization , Cambridge University Press, 2004

    2004

  15. [23]

    5, Springer Science & Business Media, 2012

    John M Danskin, The theory of max-min and its application to weapons allocation problems, vol. 5, Springer Science & Business Media, 2012

    2012

  16. [24]

    Near-optimal solutions of constrained learning problems,

    Juan Elenter, Luiz F. O. Chamon, and Alejandro Ribeiro, “Near-optimal solutions of constrained learning problems,” in The Twelfth International Conference on Learning Representations , 2024

    2024

  17. [25]

    Ergodic stochastic optimization algorithms for wireless communication and networking,

    Alejandro Ribeiro, “Ergodic stochastic optimization algorithms for wireless communication and networking,” IEEE Transactions on Signal Processing, vol. 58, no. 12, pp. 6369–6386, 2010

    2010

  18. [26]

    Inexact stochastic mirror descent for two-stage nonlinear stochastic programs,

    Vincent Guigues, “Inexact stochastic mirror descent for two-stage nonlinear stochastic programs,” 2020

    2020

  19. [27]

    Graph embedding- based wireless link scheduling with few training samples,

    Mengyuan Lee, Guanding Yu, and Geoffrey Ye Li, “Graph embedding- based wireless link scheduling with few training samples,” IEEE Transactions on Wireless Communications , vol. 20, no. 4, pp. 2282– 2294, 2020

    2020

  20. [28]

    A graph neural network approach for scalable wireless power control,

    Yifei Shen, Yuanming Shi, Jun Zhang, and Khaled B Letaief, “A graph neural network approach for scalable wireless power control,” arXiv preprint arXiv:1907.08487, 2019

    1907 arXiv

  21. [29]

    Ultra-dense networks in 5G: Interference management via non- orthogonal multiple access and treating interference as noise,

    Navid Naderializadeh, Oner Orhan, Hosein Nikopour, and Shilpa Tal- war, “Ultra-dense networks in 5G: Interference management via non- orthogonal multiple access and treating interference as noise,” in 2017 IEEE 86th Vehicular Technology Conference (VTC-Fall) . IEEE, 2017, pp. 1–6

    2017

  22. [30]

    A convergence theorem for non negative almost supermartingales and some applications,

    H. Robbins and D. Siegmund, “A convergence theorem for non negative almost supermartingales and some applications,” Herbert Robbins Selected Papers, p. 111–135, 1985

    1985

  23. [31]

    Shor, n. z., minimization methods for non-differentiable functions.,

    W. Schirotzek, “Shor, n. z., minimization methods for non-differentiable functions.,” ZAMM - Journal of Applied Mathematics and Mechanics / Zeitschrift f ¨ur Angewandte Mathematik und Mechanik , vol. 66, no. 11, pp. 575–575, 1986

    1986

  24. [32]

    Distributed network optimization with heuristic rational agents,

    Ceyhun Eksin and Alejandro Ribeiro, “Distributed network optimization with heuristic rational agents,” IEEE Transactions on Signal Processing, vol. 60, no. 10, pp. 5396–5411, 2012

    2012

  25. [33]

    Doob, Stochastic Processes, John Wiley & Sons, New York, 1953

    Joseph L. Doob, Stochastic Processes, John Wiley & Sons, New York, 1953. 13 SUPPLEMENTARY MATERIALS A. Training State-Augmented (SA) Policies In practice, we can neither optimize the expectations in (14) and (21) over the statistical distribution DH nor compute the regression ...

    1953

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.