Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Wireless Link Scheduling with State-Augmented Graph Neural Networks

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read A graph neural network that receives Lagrangian dual variables as time-varying inputs can schedule wireless links to meet per-link minimum transmission requirements while staying close to the maximum-independent-set performance bound.

desk verdict A competent incremental application of state-augmented GNNs to fair link scheduling, but the constraint-satisfaction claim is not established by the current evidence. read the letter →

arxiv 2505.07598 v1 pith:NYVN7AHH submitted 2025-05-12 eess.SP

classification eess.SP
keywords linkschedulinggraphneuralnetworksstateaugmentationLagrangiandualityconstrainedoptimizationwirelessadhocfairnessmaximumindependentset
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

The paper tries to establish that a learned link-scheduling policy can maximize long-run successful transmissions in a wireless ad hoc network while guaranteeing each link at least a preset minimum average transmission rate. It does this by making the scheduling policy a graph neural network whose input includes a running penalty signal, the Lagrangian dual variable, so the policy can react to which links are falling behind. If true, this would give a practical, scalable scheduler that approaches the theoretical maximum-independent-set bound while enforcing fairness, something greedy or probabilistic heuristics do not achieve. The evidence is numerical: on unseen conflict graphs with roughly 500 links, the learned policy stays close to the MIS upper bound and keeps most per-link violations under 10% of the requirement.

What carries the argument

The central object is the state-augmented GNN policy $\Phi(A,H;\lambda)$: a graph neural network whose input graph signal is the current Lagrangian dual vector $\lambda_t$, so the same weights produce different schedules depending on which links are short of their minimum requirement. The GNN is built from graph convolutional filters $y_l = \sum_{k=0}^{K-1} h_{lk} A^k x_{l-1}$ followed by pointwise nonlinearities and a final sigmoid thresholded at 0.5, with the conflict graph adjacency matrix $A$ as the graph shift operator. The execution-time dual update $\lambda_{t+1} = [\lambda_t - \eta_\lambda(\Phi(A,H;\lambda_t)^\top \odot ([1_K - A\Phi(A,H;\lambda_t)]_+ - \Delta))]_+$ supplies the dynamic input, so scheduling decisions react to which constraints are being violated. Training maximizes the expected augmented Lagrangian over dual vectors sampled from $\mathcal{U}[0,2]^K$.

What would settle it

Run the execution phase on a conflict graph where one link's minimum requirement $\Delta$ exceeds what any conflict-free schedule can give it, and record that link's dual variable; if the dual grows far beyond the training interval $[0,2]$ and scheduling decisions for neighboring links degrade, the unverified extrapolation assumption is the failure point.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that state augmentation with Lagrangian dual variables turns a static graph-neural-network scheduler into a time-adaptive policy for a constrained time-averaged problem. The scheduling policy $\Phi(A,H;\lambda)$ takes the conflict graph adjacency matrix $A$ and the current dual vector $\lambda_t$ as inputs and outputs a binary schedule; $\lambda_t$ evolves by projected gradient descent on the Lagrangian, increasing when a link falls below its minimum transmission requirement $\Delta$ and decreasing when the requirement is met. The training objective is the expectation over randomly sampled dual vectors of the augmented Lagrangian, so the network learns a family of schedules indexed by constraint-violation state rather than one fixed schedule. The numerical claim is that this policy keeps the average number of successful transmissions close to the maximum-independent-set upper bound, about 25% of links active, while satisfying the per-link minimum transmission constraints for the feasible majority, with residual violations concentrated below 10% of $\Delta$ and attributed to infeasible links.

Load-bearing premise

The load-bearing assumption is that the neural network, trained only on penalty signals sampled from the range zero to two, still schedules well when those signals are updated online and drift outside that range; the paper gives no experiment or proof for that extrapolation.

Editorial extensions

If this is right

  • A single trained SAGNN can schedule unseen conflict graphs from the same distribution without computing maximum independent sets online.
  • Per-link fairness constraints can be enforced with a cheap online dual update, so the method scales to hundreds of links at each time step.
  • In the tested regime the gap to the MIS upper bound is small enough that fairness does not force a large loss in total successful transmissions.
  • Residual constraint violations concentrate below 10% of the requirement, suggesting the method degrades gracefully even when a few links are infeasible.

Reading between the lines

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

  • One implication the paper leaves implicit is that the dual value itself could serve as an infeasibility alarm: a link whose dual keeps growing under the online update is likely one whose minimum requirement no conflict-free schedule can meet.
  • Because training samples dual vectors only from $[0,2]^K$, a natural testable extension is to clip the online dual update or widen the training range and see whether performance on hard links improves.
  • The same dual-augmentation idea should transfer to weighted or SINR-based interference models, since the binary conflict assumption $1_K - As(t)$ is a simplification of the interference structure.
  • The paper evaluates graphs of roughly 500 links; GNN transferability suggests the policy may extend to larger networks, but that size extrapolation is untested here.
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

4 major / 5 minor

Summary. The paper considers long-term average link scheduling in wireless ad hoc networks under per-link minimum transmission requirements. It formulates a constrained binary optimization problem, introduces a Lagrangian dual formulation, and proposes a State-Augmented Graph Neural Network (SAGNN) that takes the dual variables as time-varying inputs to the scheduling policy. The policy is trained by maximizing the expected augmented Lagrangian over randomly sampled dual variables, and executed with a stochastic dual subgradient update. Experiments on grid-derived conflict graphs compare SAGNN with p-persistent and random-MIS baselines, reporting near-MIS-bound average successful transmissions and small residual constraint violations.

Significance. The state-augmentation idea is sensible and practically motivated: feeding constraint-violation indicators as inputs is a plausible way to make a learned scheduler adaptive to fairness constraints, and the GNN parameterization provides permutation equivariance and transferability. The paper provides a code link, uses test graphs not seen in training, and gives a clear problem formulation. If the empirical claims held, SAGNN would be a useful baseline for GNN-based constrained scheduling. However, the current evidence does not establish the central feasibility claim, the theoretical support is imported rather than proved for the learned policy, and the comparison set is too narrow. The contribution at present is a promising proposal rather than a fully validated method.

major comments (4)
  1. [Section IV.B and IV.C, Fig. 5, Eq. (18)] The claim in Section IV.B that 'the minimum transmission satisfied for each link' is not established by the reported evidence. Section IV.C and Fig. 5 show that a small percentage of links violate the constraint, and the paper attributes these residual violations to 'inherently infeasible' cases without performing any feasibility check. This matters because infeasible links are exactly those whose constraint residual (s_i(t) - Delta_i) remains negative, so the dual update (18) will drive lambda_i(t) upward and out of the uniform [0,2] training range. For a conflict graph with average degree 6, there can exist links whose maximum achievable schedule fraction is below the tested Delta = 0.15, so some of the reported violations are likely to correspond to out-of-distribution lambda inputs. The paper should report the empirical distribution of lambda_t during execution, provide a feasibility analysis (e.g., an upper bound on each link's achievable long-term transmission fraction), and either remove the unqualified feasibility claim or explicitly qualify it as holding only for feasible links.
  2. [Section III, Eqs. (12)-(13) and Section IV.A] The only theoretical support for the algorithm, Eqs. (12)-(13), is imported from references [13] and [18] and applies to the exact primal-dual algorithm in the limit T -> infinity, not to the learned GNN policy with a fixed architecture and finite training. The paper explicitly motivates SAGNN by arguing that these infinite-time guarantees fail in finite time, but it provides no finite-time guarantee for the learned policy, no bound on the dual iterates, and no analysis of how the policy extrapolates beyond the lambda range used at training time. Since execution (Algorithm 2) feeds lambda_t generated by Eq. (18) into a policy trained on lambda ~ U[0,2]^K, the central claims of near-optimality and feasibility rest on an unverified generalization assumption. At minimum, the authors should empirically evaluate the trained policy on lambda inputs drawn from a much wider range (e.g., U[0,10]^K) and demonstrate that performance and constraint satisfaction do not degrade outside the training distribution.
  3. [Section IV, experimental comparison] The baseline set (p-persistent and random MIS, with and without collision avoidance) is too weak to support the contribution. There is no comparison to the GNN-based scheduler of reference [10], which is the most directly relevant learned baseline, nor to the original state-augmented learning algorithm of reference [13] on which the method builds. The claim that SAGNN 'outperforms heuristic baselines' is true but does not demonstrate an advantage over existing learning-based approaches. Please add these comparisons or justify why they are inapplicable. In addition, the MIS-based upper bound of about 25% of links is computed without the per-link minimum-transmission constraints and is therefore not a valid upper bound for the constrained problem; the comparison against this loose bound should be reconsidered or replaced by a constrained upper bound.
  4. [Section IV.A and Figures 2-4] The statistical evidence is thin. Figures 2 and 3 are based on only three independent runs, and Figure 4 reports averages over 100 graphs without error bars, confidence intervals, or significance tests. The paper also does not report the variance of the constraint violation across links or runs, nor the number of links whose dual variables leave the training range during execution. To support the central feasibility and performance claims, please report standard errors or per-run curves, the number of seeds, and a quantitative statement of how many links remain infeasible and how their violation levels compare with the infeasibility threshold.
minor comments (5)
  1. [Eq. (16) and Eq. (6)] The expression lambda^T (Phi^T ⊙ ([1 - A Phi]_+ - Delta)) mixes row and column vectors inconsistently; the element-wise product should be performed between same-size column vectors, and the notation should be cleaned up throughout the Lagrangian definitions.
  2. [Algorithm 2, line 5] The notation '∇λ Lλ(H*)' is ambiguous because Lλ depends on λ both through the dual term and through the policy input; please specify that the update uses the partial derivative with respect to λ treating the scheduled decisions as fixed, as written explicitly in Eq. (18).
  3. [Section IV.A and Fig. 4] The text says the algorithm is evaluated on 50 unseen graphs in Section IV, but Fig. 4 states 100 graphs; please reconcile these numbers and specify how the test set is split.
  4. [References] References [10] and [17] are the same paper (Zhao et al.), and references [13] and [19] are also duplicates (NaderiAlizadeh et al.); please consolidate them into single entries.
  5. [Section IV.A] The statement that 'the minimum transmission requirement is not considered during training' is confusing; clarify that Delta is a constant in the maximization over H and therefore does not affect the gradient, while the dual variable lambda encodes the constraint information.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: SAGNN's central claims are validated on held-out test graphs, and the self-citations are background motivation, not load-bearing inputs to the derivation.

full rationale

The paper's central claim is that the state-augmented GNN policy performs close to the MIS upper bound while approximately satisfying per-link minimum transmission requirements, and that it outperforms heuristic baselines. This claim is supported by evaluation on graphs not used in training: the paper generates 10 training graphs and 50 test graphs (also 100 graphs in one experiment), and reports averages over unseen instances. The GNN is trained by maximizing the expected augmented Lagrangian over randomly sampled dual variables, but the reported successes are measured by simulating the trained policy on held-out conflict graphs, so the prediction is not forced by construction. The dual variables are inputs to the policy, not fitted outputs renamed as predictions; there is no equation in which the claimed result equals the fitted objective by definition. The convergence statements in Eqs. (12)-(13) are explicitly imported from prior work [13,18] and are just as explicitly said to hold only as time goes to infinity, with the paper stating 'these results hold only as the number of iterations, i.e. the operation time goes to infinity.' They motivate the state-augmentation design but are not used to prove the finite-time behavior of the learned SAGNN, so they are not load-bearing for the empirical verdict. The self-citations to [13]-[15] and [19] for state augmentation are background attribution of a known technique, not an appeal to an unverified theorem that forces the conclusion. The paper does contain an evidentiary gap: Section IV.B claims 'the minimum transmission satisfied for each link,' while Section IV.C and Fig. 5 show a small percentage of links falling short, and the attribution of these violations to 'inherently infeasible' cases is not checked. That is a correctness or support issue, not circularity. Overall, no load-bearing step reduces the derivation to its own inputs.

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

The central claim rests on the Lagrangian duality framework inherited from prior work, on a binary conflict graph model of interference, and on the untested assumption that a GNN trained over sampled λ values will generalize to the dynamic λ range during execution. The resilience factor and dual learning rate are free parameters tuned for the reported results. No new physical or mathematical entities are introduced.

free parameters (4)
  • Resilience factor = 0.05 for Δ=0.1, 0.1 for Δ=0.125 and 0.15
    A hyperparameter from resilient constrained learning [33], set differently for each transmission requirement to mitigate infeasibility. It directly affects constraint violation outcomes and is tuned per setting, so it is a free parameter for the reported results.
  • Dual learning rate ηλ = 2
    Chosen by hand; controls the speed of dual variable adaptation. The paper does not justify the value or report sensitivity to it.
  • λ sampling distribution = U[0,2]^K, with 30% zeros or 25% twos in some runs
    The distribution used to train the GNN over dual variables is selected by the authors; the paper notes ad-hoc modifications in some runs. The policy's behavior during execution depends on this training distribution.
  • Scheduling threshold = 0.5
    The sigmoid output of the GNN is thresholded at 0.5 to obtain binary schedules; this is a design choice not derived from the problem.
assumptions (4)
  • domain assumption The Lagrangian dual framework yields near-optimal solutions for the integer scheduling problem with fairness constraints.
    The paper cites [13,18] for almost-sure feasibility and near-optimality of the primal-dual policy. These guarantees apply to the exact algorithm, and the paper assumes they carry over to the GNN parameterization, which is not proven.
  • domain assumption Successful transmission of a scheduled link depends only on the binary conflict graph: a link succeeds if and only if no conflicting link is scheduled simultaneously.
    The objective in Eq. (1) uses [1K - A s(t)]+, which abstracts physical layer interference into pairwise conflicts. This ignores cumulative interference, SINR thresholds, and channel effects, and is a modeling assumption for the entire paper.
  • ad hoc to paper A GNN of fixed architecture can represent the Lagrangian-maximizing scheduling policy for arbitrary graphs and dual variables.
    The paper assumes the GNN parameterization is expressive enough to approximate the optimal policy, but provides no approximation or universality analysis for this specific class of policies.
  • ad hoc to paper Training on dual variables sampled from U[0,2]^K generalizes to the dual variable trajectory generated during execution, which may fall outside this range.
    Execution updates λ via Eq. (18) with no bound; for infeasible links λ may grow beyond 2. The paper does not analyze out-of-distribution behavior of the GNN.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Wireless Link Scheduling with State-Augmented Graph Neural Networks." pith.science (2026). https://pith.science/paper/NYVN7AHH

@misc{pith2026250507598,
  author       = {Pith},
  title        = {Pith review of: Wireless Link Scheduling with State-Augmented Graph Neural Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NYVN7AHH}},
  note         = {Machine review of arXiv:2505.07598}
}
read the original abstract

We consider the problem of optimal link scheduling in large-scale wireless ad hoc networks. We specifically aim for the maximum long-term average performance, subject to a minimum transmission requirement for each link to ensure fairness. With a graph structure utilized to represent the conflicts of links, we formulate a constrained optimization problem to learn the scheduling policy, which is parameterized with a graph neural network (GNN). To address the challenge of long-term performance, we use the state-augmentation technique. In particular, by augmenting the Lagrangian dual variables as dynamic inputs to the scheduling policy, the GNN can be trained to gradually adapt the scheduling decisions to achieve the minimum transmission requirements. We verify the efficacy of our proposed policy through numerical simulations and compare its performance with several baselines in various network settings.

Figures

Figures reproduced from arXiv: 2505.07598 by the authors.

Figure 1
Figure 1. An example wireless network. Left: Connectivity graph [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Evolution of constraint violation during training for [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Evolution of the average successful transmissions in [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Number of total and successful transmissions for [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Box plot of the distribution of values for links violating [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Long-Horizon Wireless Link Scheduling with State-Augmented Graph Neural Networks

    eess.SP 2026-07 conditional novelty 5.0 of 10

    A state-augmented GNN that imitates dual subgradient descent produces near-optimal, constraint-satisfying long-horizon wireless link schedules.

Reference graph

Works this paper leans on

32 extracted references · 23 canonical work pages · cited by 1 Pith paper

  1. [18]

    Ergodic stochastic optimization algorithms for wireless communication and networking,

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

  2. [1]

    FlashLinQ: A synchronous distributed scheduler for peer-to- peer ad hoc networks,

    X. Wu, S. Tavildar, S. Shakkottai, T. Richardson, J. Li, R. Laroia, and A. Jovicic, “FlashLinQ: A synchronous distributed scheduler for peer-to- peer ad hoc networks,” IEEE/ACM Transactions on Networking , vol. 21, no. 4, pp. 1215–1228, 2013

  3. [2]

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

    N. Naderializadeh and A. S. 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

  4. [3]

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

    X. Yi and G. 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

  5. [4]

    FPLinQ: A cooperative spectrum sharing strategy for device-to-device communications,

    K. Shen and W. Yu, “FPLinQ: A cooperative spectrum sharing strategy for device-to-device communications,” in 2017 IEEE International Symposium on Information Theory (ISIT) , 2017, pp. 2323–2327

  6. [5]

    Spatial deep learning for wireless scheduling,

    W. Cui, K. Shen, and W. Yu, “Spatial deep learning for wireless scheduling,” IEEE Journal on Selected Areas in Communications , vol. 37, no. 6, pp. 1248–1261, 2019

  7. [6]

    Grlinq: A distributed link scheduling mechanism with graph reinforcement learning,

    Z. Shan, X. Yi, L. Liang, C.-S. Liao, and S. Jin, “Grlinq: A distributed link scheduling mechanism with graph reinforcement learning,” in 2024 IEEE International Symposium on Information Theory (ISIT) , 2024, pp. 2436–2441

  8. [7]

    Topological interference management through index coding,

    S. A. Jafar, “Topological interference management through index coding,” IEEE Transactions on Information Theory , vol. 60, no. 1, pp. 529–568, 2013

Show all 32 references
  1. [8]

    Interference networks with no csit: Impact of topology,

    N. Naderializadeh and A. S. Avestimehr, “Interference networks with no csit: Impact of topology,” IEEE Transactions on Information Theory , vol. 61, no. 2, pp. 917–938, 2014

  2. [9]

    Gblinks: Gnn-based beam selection and link activation for ultra-dense d2d mmwave networks,

    S. He, S. Xiong, W. Zhang, Y . Yang, J. Ren, and Y . Huang, “Gblinks: Gnn-based beam selection and link activation for ultra-dense d2d mmwave networks,” IEEE Transactions on Communications , vol. 70, no. 5, pp. 3451–3466, 2022

  3. [11]

    Geometric machine learning over riemannian manifolds for wireless link scheduling,

    R. Shelim and A. S. Ibrahim, “Geometric machine learning over riemannian manifolds for wireless link scheduling,” IEEE Access, vol. 10, pp. 22 854– 22 864, 2022

  4. [12]

    Graph neural network-based scheduling for multi-uav-enabled communications in d2d networks,

    P. Li, L. Wang, W. Wu, F. Zhou, B. Wang, and Q. Wu, “Graph neural network-based scheduling for multi-uav-enabled communications in d2d networks,” Digital Communications and Networks , vol. 10, no. 1, pp. 45–52, 2024. [Online]. Available: https://www.sciencedirect.com/science/ ...

  5. [14]

    Learning state-augmented policies for information routing in communication networks,

    S. Das, N. NaderiAlizadeh, and A. Ribeiro, “Learning state-augmented policies for information routing in communication networks,” IEEE Transactions on Signal Processing , 2024

  6. [15]

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

    Y . B. Uslu, R. Doostnejad, A. Ribeiro, and N. NaderiAlizadeh, “Learning to slice wi-fi networks: A state-augmented primal-dual approach,” arXiv preprint arXiv:2405.05748, 2024

  7. [16]

    An optimal maximal independent set algorithm for bounded-independence graphs,

    J. Schneider and R. Wattenhofer, “An optimal maximal independent set algorithm for bounded-independence graphs,” Distributed Computing , vol. 22, pp. 349–361, 2010

  8. [17]

    Link scheduling using graph neural networks,

    Z. Zhao, G. Verma, C. Rao, A. Swami, and S. Segarra, “Link scheduling using graph neural networks,” IEEE Transactions on Wireless Communi- cations, vol. 22, no. 6, pp. 3997–4012, 2023

  9. [19]

    State-augmented learn- able algorithms for resource management in wireless networks,

    N. NaderiAlizadeh, M. Eisen, and A. Ribeiro, “State-augmented learn- able algorithms for resource management in wireless networks,” IEEE Transactions on Signal Processing , vol. 70, pp. 5898–5912, 2022

  10. [20]

    Convolutional neural network architectures for signals supported on graphs,

    F. Gama, A. G. Marques, G. Leus, and A. Ribeiro, “Convolutional neural network architectures for signals supported on graphs,” IEEE Transactions on Signal Processing , vol. 67, no. 4, pp. 1034–1049, 2019

  11. [21]

    The graph neural network model,

    F. Scarselli, M. Gori, A. C. Tsoi, M. Hagenbuchner, and G. Monfardini, “The graph neural network model,” IEEE transactions on neural networks , vol. 20, no. 1, pp. 61–80, 2008

  12. [22]

    Graph signal processing: Overview, challenges, and applications,

    A. Ortega, P. Frossard, J. Kova ˇcevi´c, J. M. Moura, and P. Vandergheynst, “Graph signal processing: Overview, challenges, and applications,” Pro- ceedings of the IEEE , vol. 106, no. 5, pp. 808–828, 2018

  13. [23]

    Learning decentralized wireless resource allocations with graph neural networks,

    Z. Wang, M. Eisen, and A. Ribeiro, “Learning decentralized wireless resource allocations with graph neural networks,” IEEE Transactions on Signal Processing, vol. 70, pp. 1850–1863, 2022

  14. [24]

    Optimal wireless resource allocation with ran- dom edge graph neural networks,

    M. Eisen and A. Ribeiro, “Optimal wireless resource allocation with ran- dom edge graph neural networks,” ieee transactions on signal processing , vol. 68, pp. 2977–2991, June 2020

  15. [25]

    Graph neural networks for scalable radio resource management: Architecture design and theoretical analysis,

    Y . Shen, Y . Shi, J. Zhang, and K. B. Letaief, “Graph neural networks for scalable radio resource management: Architecture design and theoretical analysis,” IEEE Journal on Selected Areas in Communications , vol. 39, no. 1, pp. 101–115, 2020

  16. [26]

    Graph neural networks: Architectures, stability, and transferability,

    L. Ruiz, F. Gama, and A. Ribeiro, “Graph neural networks: Architectures, stability, and transferability,” Proceedings of the IEEE , vol. 109, no. 5, pp. 660–682, 2021

  17. [27]

    Stability of graph scattering transforms,

    F. Gama, A. Ribeiro, and J. Bruna, “Stability of graph scattering transforms,” Advances in Neural Information Processing Systems , vol. 32, pp. 8038–8048, 2019

  18. [28]

    Transferability Properties of Graph Neural Networks,

    L. Ruiz, L. F. Chamon, and A. Ribeiro, “Transferability Properties of Graph Neural Networks,” arXiv preprint arXiv:2112.04629 , 2021

  19. [29]

    Geometric graph filters and neural net- works: Limit properties and discriminability trade-offs,

    Z. Wang, L. Ruiz, and A. Ribeiro, “Geometric graph filters and neural net- works: Limit properties and discriminability trade-offs,” IEEE Transactions on Signal Processing , 2024

  20. [32]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,”

  21. [33]

    Resilient constrained learning,

    I. Hounie, A. Ribeiro, and L. F. O. Chamon, “Resilient constrained learning,” in Advances in Neural Information Processing Systems , A. Oh, T. Naumann, A. Globerson, K. Saenko, M. Hardt, and S. Levine, Eds., vol. 36. Curran Associates, Inc., 2023, pp. 71 767–71 798. [Online]. ...

  22. [35]

    Available: https://arxiv.org/abs/1412.6980

    [Online]. Available: https://arxiv.org/abs/1412.6980

  23. [2015]

    Available: http://arxiv.org/abs/1502.03167

    [Online]. Available: http://arxiv.org/abs/1502.03167

  24. [2017]

    Available: http://arxiv.org/abs/1710.10370

    [Online]. Available: http://arxiv.org/abs/1710.10370

Pith tools

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