Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Communication-Efficient Personalized Federal Graph Learning via Low-Rank Decomposition

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read CEFGL claims that decomposing each client's model into a low-rank shared component and a sparse personalized component simultaneously improves classification accuracy under non-IID graph data and cuts communication cost by an order of…

desk verdict CEFGL's idea is plausible and its experiments are broad, but the algorithm as written is dimensionally invalid and inconsistent with itself; the headline numbers cannot be reproduced from the paper, so it needs major revision, not desk rejection. read the letter →

arxiv 2412.13442 v1 pith:FNY6KM6G submitted 2024-12-18 cs.LG cs.AIcs.DC

Ruyue Liu , Rong Yin , Xiangzhen Bo , Xiaoshuai Hao , Xingrui Zhou , Yong Liu , Can Ma , Weiping Wang This is my paper · ORCID
classification cs.LGcs.AIcs.DC
keywords FederatedGraphLearningLow-RankDecompositionPersonalizedSparseModelsCommunicationEfficiencyNon-IIDDataQuantizationCompression
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 proposes CEFGL, a federated graph learning algorithm aimed at two problems at once: clients' graph data are non-IID, so one shared model fits none of them well, and federated training is communication-bound. CEFGL decomposes each client model into a low-rank component that captures knowledge shared across clients and a sparse component that captures client-specific knowledge, trained through a dual-channel encoder. It adds a correction term to reduce client drift, allows many local updates between communication rounds, and compresses parameters by quantization. The paper reports that across sixteen datasets, under both IID and non-IID splits and with three GNN backbones, CEFGL achieves the best average classification accuracy compared with seven baselines. In the cross-dataset CHEM setting, it claims a 5.64% accuracy gain over FedStar while cutting communication bits by a factor of 18.58 and communication time by a factor of 1.65.

What carries the argument

The central mechanism is parameter decomposition in the spirit of robust PCA: each client model is written as a low-rank component $W$ (global, shared) plus a sparse component $S$ (local, personalized), with convex relaxations of rank to trace norm and of $\ell^0$ to $\ell^1$. A dual-channel encoder trains $W$ and $S$ separately; a Scaffnew-style correction term $h_i^t$ counters client drift; local updates run many steps before communication; and a probabilistic skip in Algorithm 1 lets clients skip rounds. Quantization $Q_r$ compresses uplink and downlink messages to as few as 4 bits per parameter. The server performs truncated SVD aggregation, so the shared model stays low-rank.

What would settle it

Re-run the CHEM cross-dataset experiment with GIN using only the published Algorithm 1 updates (Eqs. 6-9), the stated hyperparameters ($\nu=0.5$, $\alpha=0.6$, $\mu=0.0001$, $\lambda=0.001$, $p=0.5$), and 4-bit quantization, measuring accuracy and total transmitted bits over 200 rounds; if the accuracy gain over FedStar and the 18.58x bit reduction do not appear, the central claim fails. A simpler check is to verify that the aggregated global model after truncated SVD has rank near the claimed low-rank ratios in Table 3.

Watch

Extended reading notes

Core claim

The central claim is that a federated graph classifier can be made both more accurate under data heterogeneity and dramatically cheaper to communicate by decomposing model parameters into a low-rank global part and a sparse personalized part, rather than transmitting dense full models. The paper argues that the low-rank component encodes common knowledge across clients, the sparse component encodes local knowledge, and the two together give each client a hybrid model that adapts locally without losing shared structure. It further claims the correction-term update and probabilistic communication skipping make the method robust to client dropout, while 4-bit quantization barely hurts accuracy. The headline numbers are stated against FedStar: +5.64% accuracy on the cross-dataset CHEM benchmark with GIN, 18.58x fewer communication bits, and 1.65x less communication time per round.

Load-bearing premise

The paper's experimental numbers depend on the update equations in Section 4.1 and Algorithm 1 being the exact optimization steps that were run; if those equations do not match the implementation, the reported accuracy and communication gains cannot be reproduced.

Editorial extensions

If this is right

  • If CEFGL's claims hold, federated graph classification on heterogeneous cross-domain data can be made both more accurate and cheaper than state-of-the-art baselines, removing a major practical barrier to deployment.
  • The 18.58x reduction in communication bits means large GNN models can be trained federated on bandwidth-constrained or mobile clients without sacrificing accuracy.
  • The robustness to client dropout, with accuracy maintained even when the dropout rate is sampled from a Beta(10,1) distribution, implies the method suits real networks with unstable participation.
  • The finding that 10% of sparse parameters suffice for personalization suggests local personalization does not need a large private model, reducing on-device storage too.

Reading between the lines

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

  • The same low-rank-global/sparse-local split is not graph-specific; it could be applied to federated image or text models, where the sparse component would play the role of a lightweight adapter.
  • The communication probability $p$ is effectively a tunable privacy-efficiency knob: lower $p$ means the server sees fewer updates and cannot reconstruct exact gradients, so the method could pair naturally with differential privacy.
  • A testable extension is to replace truncated SVD with a learned or adaptive rank selection per round, which might improve accuracy on datasets where the optimal rank varies across clients.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. This paper proposes CEFGL, a personalized federated graph learning algorithm that decomposes client model parameters into a low-rank shared component and a sparse private component, combines local training with correction terms, uses probabilistic communication skipping and quantization to reduce communication cost, and evaluates the method on sixteen graph-classification datasets with GCN, GIN, and GraphSage backbones. The central empirical claim is that CEFGL outperforms FedStar and other federated graph learning baselines in accuracy while substantially reducing communication bits and time; the headline example is a 5.64 percentage point accuracy improvement on the CHEM cross-dataset setting with GIN, together with an 18.58x reduction in communication bits and a 1.65x reduction in communication time.

Significance. If the claims hold, the paper would make a useful empirical contribution: a single framework that handles non-IID graph data by combining sparse personalization with low-rank global sharing, while reducing communication through quantization and less frequent synchronization. The experimental comparison is extensive, covering three GNN architectures, multiple FGL baselines, sensitivity analyses, dropout robustness, and an ablation of the two model components; these are genuine strengths. However, the load-bearing algorithmic description is not internally consistent as written: the update equations contain dimensionally invalid terms, the skip-communication branch of the pseudocode is incoherent, and the main results are not matched by the ablation table. The empirical claims are therefore not yet tied to a well-defined, reproducible algorithm.

major comments (4)
  1. [§4.1, Eq. (6) and Eq. (8)] Equation (6) is dimensionally invalid: the term (α/2)||Θ^t − W_i^t|| is a scalar norm added to the parameter matrix W_i^t, so the right-hand side is not a matrix update of the same shape as W_i^t. Equation (8) has the same defect, since +||S_i^t||_1 adds a scalar to the matrix S_i^t. If the intended updates are proximal-style corrections, the equations must be rewritten (for example, with (α/2)(Θ^t − W_i^t) or with the norm placed inside the objective and its gradient taken), and Algorithm 1 lines 6 and 8 must be made consistent. As written, the update rules in Algorithm 1 cannot be executed, which blocks reproduction of the reported accuracy and communication results.
  2. [Algorithm 1, line 14] The skip-communication branch sets Θ^{t+1} = W_i^{t+1} when j = 0, but W_i^{t+1} is a client-local variable that is not transmitted to the server on a skipped round. The branch is also outside the client loop, so the index i is undefined at that point. Choosing a single client's model as the global model would contradict the idea of skipping communication and would likely destroy shared knowledge. The pseudocode needs to specify what the server and clients actually do on skipped rounds (for example, Θ^{t+1} = Θ^t while clients continue local optimization), and the client-selection logic in lines 3–4 needs to be reconciled with the communication-skip decision.
  3. [Table 5 vs. Table 1 and Table 4] The full-model ablation row (W and S, GIN) reports CHEM = 83.03, whereas Table 1 reports GIN 'Ours' CHEM = 85.43 ± 1.14 for the same setting and backbone, and Table 4 reports CHEM = 85.03 for the stated default sparsity regime. Since the ablation's full configuration should coincide with the main result, this discrepancy suggests either different hyperparameters, different communication probability, a different number of rounds, or an error. The authors need to explain this mismatch before the headline 5.64% improvement over FedStar can be evaluated.
  4. [§4.1, Figure 1, and Remark 1] The role of W_i^t and S_i^t in the forward model is ambiguous. Equation (6) updates W_i using a loss f_i(W_i^t; D_i), while Eq. (8) updates S_i using a loss f_i(Θ^t + S_i^t; D_i), and the Figure 1 caption says that during local training the model parameters W_i^t are frozen while the sparsification components are fine-tuned. Remark 1 further states that W_i is not actually low-rank during local training. The paper needs to state the exact forward model and the exact order of operations (simultaneous or sequential training of W_i and S_i, and which parameters are frozen at each stage) so that Algorithm 1 matches the implementation used in the experiments.
minor comments (5)
  1. [Title and Abstract] The title uses 'Federal Graph Learning' while the body uses 'Federated Graph Learning'; the title should be corrected to 'Federated'.
  2. [§5.2] The text mentions a baseline 'FedPerGCN' that is not listed in the baseline enumeration in §5.1; this is likely a typo for FedGCN or FedPer and should be corrected.
  3. [§4.3, Eqs. (10)–(11)] The quantizer Q_r is described as involving independent random variables, but Eq. (11) gives a deterministic rounding formula. Please either provide the actual probability distribution (as in QSGD) or state explicitly that deterministic rounding is used, since the current definition is ambiguous.
  4. [§5.1 and Table 2] The text lists the IID single-dataset setups as DD, NCI1, and IMDB-MULTI, but Table 2's third column is IMDB-BINARY; please reconcile the dataset name.
  5. [§4.2, Eq. (9)] Equation (9) uses Θ^{t+1}_i on the left-hand side for what should be a single global model, and it does not show the quantization applied in Algorithm 1 line 12; please correct the notation and specify whether quantization is applied after the low-rank projection.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: CEFGL's accuracy and communication gains are empirical benchmark results against external baselines, and no equation reduces to a fitted input or self-citation chain.

full rationale

The central claims are experimental: CEFGL is compared with FedAvg, FedProx, FedPer, GCFL, SpreadGNN, FedGCN, and FedStar on sixteen public datasets (Section 5.1), and the reported 5.64% accuracy gain and 18.58x communication reduction are presented as measured outcomes over 200 rounds (Section 5.9, Figure 6). None of the update equations (6)-(9) defines the reported accuracy as a function of a fitted parameter, so no prediction reduces to its input by construction. The low-rank and sparse decomposition is motivated by external RPCA [42], and the correction term is inspired by external Scaffnew [33]; these are design inspirations, not self-citations carrying the result. The only self-citations ([2], [8]) are introductory examples of GNN use in the literature and are not load-bearing. Hyperparameters (alpha, mu, lambda, p, r) are tuned and then reported, which is standard empirical practice and falls under soundness rather than circularity. The paper does contain reproducibility concerns: Eq. (6) and Eq. (8) add scalar norms to matrices, the skip-communication branch in Algorithm 1 is ambiguous, and Remark 1 admits that W_i^t is not actually low-rank during local training; however, these are correctness issues, not circularity. The benchmark comparisons are external and falsifiable, so the derivation chain is self-contained with respect to circularity.

Assumptions & free parameters 7 free parameters · 4 assumptions · 0 invented entities

The method introduces no new physical entities; the low-rank global model, sparse private model, and correction term are learned parameters, not independent inventions. The free parameters are hyperparameters tuned on validation data; their values are load-bearing for the reported accuracy and communication trade-offs.

free parameters (7)
  • mu (low-rank hyperparameter) = 0.0001 (default); Table 3 suggests 0.001 performs better
    Controls the low-rank regularizer/TSVD threshold; strongly affects accuracy and communication cost.
  • lambda (sparse hyperparameter) = 0.001 (default); Table 4 varies it
    Controls sparsity of the personalized component; affects accuracy and model density.
  • nu (RPCA trade-off parameter, Eq. 4) = 0.5
    Balances rank and sparsity in the motivating objective; not validated.
  • alpha (balance parameter, Eq. 6) = 0.6
    Weights the global model prior in local training; not validated.
  • p (communication probability) = 0.5
    Determines how often clients skip communication; varied in Figure 3.
  • r (quantization bits) = 4 bits
    Compression level; varied in Figure 4.
  • beta (Top-k sparsity ratio) = around 0.1
    Used in Figure 2 for Top-k sparsification; chosen after sensitivity analysis.
assumptions (4)
  • standard math Convex relaxation of rank and L0 via nuclear norm and L1 (Eq. 5) is a valid surrogate for the RPCA objective.
    Standard result from Candes et al. [42]; not proved here but accepted.
  • domain assumption A single low-rank global model can represent shared knowledge across heterogeneous graph datasets and a sparse per-client model can capture local knowledge.
    This is the core design assumption of CEFGL, stated in Section 4.1; it is not proved and is only tested empirically.
  • domain assumption The correction-term update and server aggregation in Eqs. (7)-(9) converge to a useful global model.
    Remark 2 gives an informal argument assuming local convergence; no formal convergence analysis is provided.
  • ad hoc to paper The skip-communication branch (Algorithm 1 line 14), which sets the global model to a single client's model, does not harm training.
    This step is unusual and only heuristically motivated.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Communication-Efficient Personalized Federal Graph Learning via Low-Rank Decomposition." pith.science (2026). https://pith.science/paper/FNY6KM6G

@misc{pith2026241213442,
  author       = {Pith},
  title        = {Pith review of: Communication-Efficient Personalized Federal Graph Learning via Low-Rank Decomposition},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FNY6KM6G}},
  note         = {Machine review of arXiv:2412.13442}
}
read the original abstract

Federated graph learning (FGL) has gained significant attention for enabling heterogeneous clients to process their private graph data locally while interacting with a centralized server, thus maintaining privacy. However, graph data on clients are typically non-IID, posing a challenge for a single model to perform well across all clients. Another major bottleneck of FGL is the high cost of communication. To address these challenges, we propose a communication-efficient personalized federated graph learning algorithm, CEFGL. Our method decomposes the model parameters into low-rank generic and sparse private models. We employ a dual-channel encoder to learn sparse local knowledge in a personalized manner and low-rank global knowledge in a shared manner. Additionally, we perform multiple local stochastic gradient descent iterations between communication phases and integrate efficient compression techniques into the algorithm. The advantage of CEFGL lies in its ability to capture common and individual knowledge more precisely. By utilizing low-rank and sparse parameters along with compression techniques, CEFGL significantly reduces communication complexity. Extensive experiments demonstrate that our method achieves optimal classification accuracy in a variety of heterogeneous environments across sixteen datasets. Specifically, compared to the state-of-the-art method FedStar, the proposed method (with GIN as the base model) improves accuracy by 5.64\% on cross-datasets setting CHEM, reduces communication bits by a factor of 18.58, and reduces the communication time by a factor of 1.65.

Figures

Figures reproduced from arXiv: 2412.13442 by the authors.

Figure 1
Figure 1. Illustration of the CEFGL framework. The process begins with the client transforming the input feature data into vectors of consistent dimensions using a mul￾tilayer perceptron (MLP). The client then performs local training to obtain the initial model. During this phase, the client trains the model with its local data while keeping the model parameters Wt i frozen. Simultaneously, the sparsification components, whic… view at source ↗
Figure 2
Figure 2. Variation of test accuracy and loss with the number of communication rounds [PITH_FULL_IMAGE:figures/full_fig_p017_2.png] view at source ↗
Figure 3
Figure 3. Variation of test accuracy and loss with the number of communication rounds [PITH_FULL_IMAGE:figures/full_fig_p019_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Test accuracy and loss of CEFGL after employing quantization compression [PITH_FULL_IMAGE:figures/full_fig_p020_4.png]
Figure 5
Figure 5. Figure 5: (a) Accuracy on CHEM when clients accidentally drop out with different prob [PITH_FULL_IMAGE:figures/full_fig_p021_5.png]
Figure 6
Figure 6. Figure 6: (a) Variation of test accuracy and loss with the number of communication [PITH_FULL_IMAGE:figures/full_fig_p022_6.png]
Figure 7
Figure 7. Figure 7: t-SNE visualization of local graph representations derived from the last hidden [PITH_FULL_IMAGE:figures/full_fig_p023_7.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. Federated Learning with Graph-Based Aggregation for Traffic Forecasting

    cs.LG 2025-07 conditional novelty 4.0 of 10

    Two graph-based averaging rules for federated traffic forecasting achieve slightly lower RMSE than recent graph-based federated baselines on METR-LA and PEMS-BAY.

Reference graph

Works this paper leans on

47 extracted references · 38 canonical work pages · cited by 1 Pith paper

  1. [1]

    T. N. Kipf, M. Welling, Semi-supervised classification with graph con- volutional networks, CoRR abs/1609.02907 (2016). 24

    arXiv 2016

  2. [2]

    R. Liu, R. Yin, Y. Liu, W. Wang, Aswt-sgnn: Adaptive spectral wavelet transform-based self-supervised graph neural network, in: Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 38, 2024, pp. 13990–13998

    work page 2024

  3. [3]

    X. Su, Z. You, D. Huang, L. Wang, L. Wong, B. Ji, B. Zhao, Biomedical knowledge graph embedding with capsule network for multi-label drug- drug interaction prediction, IEEE Transactions on Knowledge and Data Engineering 35 (6) (2023) 5640–5651

    work page 2023

  4. [4]

    J. Liu, F. Hu, Q. Zou, P. Tiwari, H. Wu, Y. Ding, Drug reposi- tioning by multi-aspect heterogeneous graph contrastive learning and positive-fusion negative sampling strategy, Information Fusion 112 (2024) 102563

    work page 2024

  5. [5]

    Y. Wu, Y. Chen, Z. Yin, W. Ding, I. King, A survey on graph embedding techniques for biomedical data: Methods and applications, Information Fusion 100 (2023) 101909

    2023

  6. [6]

    Y. Ding, N. Robinson, C. Tong, Q. Zeng, C. Guan, Lggnet: Learning from local-global-graph representations for brain–computer interface, IEEE Transactions on Neural Networks and Learning Systems 35 (7) (2024) 9773–9786

    work page 2024

  7. [7]

    G´ orriz, I

    J. G´ orriz, I. ´Alvarez Ill´ an, A. ´Alvarez Marquina, J. Arco, M. Atz- mueller, F. Ballarini, E. Barakova, G. Bologna, P. Bonomini, G. Castellanos-Dominguez, D. Castillo-Barnes, S. Cho, R. Contr- eras, J. Cuadra, E. Dom ´ ınguez, F. Dom ´ ınguez-Mateos, R. Duro, D. Elizondo, A. Fern´ andez-Caballero, E. Fernandez-Jover, M. For- moso, N. Gallego-Molina, ...

    work page 2023

  8. [8]

    R. Liu, R. Yin, Y. Liu, W. Wang, Unbiased and augmentation-free self-supervised graph representation learning, Pattern Recognition 149 (2024) 110274

    2024

Show all 47 references
  1. [9]

    Y. Li, J. Huan, J. Shen, L. Chen, J. Cao, Y. Cheng, Social network large-scale group decision-making considering dynamic trust relation- ships and historical preferences of decision makers in opinion evolution, Information Fusion (2024) 102837

    2024

  2. [10]

    H. Peng, W. Zeng, J. Tang, M. Wang, H. Huang, X. Zhao, Open knowl- edge graph completion with negative-aware representation learning and multi-source reliability inference, Information Fusion 115 (2025) 102729

    2025

  3. [11]

    Zhang, L

    Z. Zhang, L. Bai, L. Zhu, Ts-align: A temporal similarity-aware entity alignment model for temporal knowledge graphs, Information Fusion 112 (2024) 102581

    2024

  4. [12]

    W. Fan, X. Zhao, Q. Li, T. Derr, Y. Ma, H. Liu, J. Wang, J. Tang, Ad- versarial attacks for black-box recommender systems via copying trans- ferable cross-domain user profiles, IEEE Transactions on Knowledge and Data Engineering 35 (12) (2023) 12415–12429

    2023

  5. [13]

    Paz-Ruza, A

    J. Paz-Ruza, A. Alonso-Betanzos, B. Guijarro-Berdi˜ nas, B. Cancela, C. Eiras-Franco, Sustainable transparency on recommender systems: Bayesian ranking of images for explainability, Information Fusion 111 (2024) 102497

    2024

  6. [14]

    Himeur, A

    Y. Himeur, A. Alsalemi, A. Al-Kababji, F. Bensaali, A. Amira, C. Sar- dianos, G. Dimitrakopoulos, I. Varlamis, A survey of recommender systems for energy efficiency in buildings: Principles, challenges and prospects, Information Fusion 72 (2021) 1–21

    2021

  7. [15]

    W. Zhu, X. Zhou, S. Lan, W. Wang, Z. Hou, Y. Ren, T. Pan, A dual branch graph neural network based spatial interpolation method for traf- fic data inference in unobserved locations, Information Fusion 114 (2025) 102703. 26

    2025

  8. [16]

    D. Xu, H. Peng, Y. Tang, H. Guo, Hierarchical spatio-temporal graph convolutional neural networks for traffic data imputation, Information Fusion 106 (2024) 102292

    2024

  9. [17]

    J. W. de Kok, M. ´A. A. de la Hoz, Y. de Jong, V. Brokke, P. W. Elbers, P. Thoral, A. Castillejo, T. Trenor, J. M. Castellano, A. E. Bronchalo, et al., A guide to sharing open healthcare data under the general data protection regulation, Scientific data 10 (1) (2023) 404

    2023

  10. [18]

    K. Wu, Towards a universal cognitive tool: designing accessible visual- ization for people with intellectual and developmental disabilities, ACM SIGACCESS Accessibility and Computing (131) (2021) 1–6

    2021

  11. [19]

    McMahan, E

    B. McMahan, E. Moore, D. Ramage, S. Hampson, B. A. y Arcas, Communication-efficient learning of deep networks from decentralized data, in: Artificial intelligence and statistics, PMLR, 2017, pp. 1273– 1282

    2017

  12. [20]

    Sabah, Y

    F. Sabah, Y. Chen, Z. Yang, A. Raheem, M. Azam, N. Ahmad, R. Sar- war, Communication optimization techniques in personalized federated learning: Applications, challenges and future directions, Information Fusion (2024) 102834

    2024

  13. [21]

    Huang, D

    W. Huang, D. Wang, X. Ouyang, J. Wan, J. Liu, T. Li, Multimodal fed- erated learning: Concept, methods, applications and future directions, Information Fusion 112 (2024) 102576

    2024

  14. [22]

    C. Wang, B. Chen, G. Li, H. Wang, Automated graph neural net- work search under federated learning framework, IEEE Transactions on Knowledge and Data Engineering 35 (10) (2023) 9959–9972

    2023

  15. [23]

    xuan Hu, C

    H. xuan Hu, C. Cao, Q. Hu, Y. Zhang, Federated learning enabled graph convolutional autoencoder and factorization machine for potential friendship prediction in social networks, Information Fusion 102 (2024) 102042

    2024

  16. [24]

    H. Xie, J. Ma, L. Xiong, C. Yang, Federated graph classification over non-iid graphs, Advances in neural information processing systems 34 (2021) 18839–18852. 27

    2021

  17. [25]

    Z. Ye, X. Zhang, X. Chen, H. Xiong, D. Yu, Adaptive clustering based personalized federated learning framework for next poi recommendation with location noise, IEEE Transactions on Knowledge and Data Engi- neering 36 (5) (2024) 1843–1856

    2024

  18. [26]

    S. P. Karimireddy, S. Kale, M. Mohri, S. J. Reddi, S. U. Stich, A. T. Suresh, Scaffold: Stochastic controlled averaging for on-device federated learning, CoRR abs/1910.06378 (2019)

    2019 arXiv

  19. [27]

    Collins, H

    L. Collins, H. Hassani, A. Mokhtari, S. Shakkottai, Exploiting shared representations for personalized federated learning, in: International conference on machine learning, PMLR, 2021, pp. 2089–2099

    2021

  20. [28]

    T Dinh, N

    C. T Dinh, N. Tran, J. Nguyen, Personalized federated learning with moreau envelopes, Advances in Neural Information Processing Systems 33 (2020) 21394–21405

    2020

  21. [29]

    T. Li, A. K. Sahu, M. Zaheer, M. Sanjabi, A. Talwalkar, V. Smith, Fed- erated optimization in heterogeneous networks, Proceedings of Machine learning and systems 2 (2020) 429–450

    2020

  22. [30]

    J. Luo, S. Wu, Adapt to adaptation: Learning personalization for cross- silo federated learning, in: The 31st International Joint Conference on Artificial Intelligence., Vol. 2022, 2022, p. 2166

    2022

  23. [31]

    Huang, L

    T. Huang, L. Shen, Y. Sun, W. Lin, D. Tao, Fusion of global and local knowledge for personalized federated learning, CoRR abs/2302.11051 (2023)

    2023 arXiv

  24. [32]

    A. Z. Tan, H. Yu, L. Cui, Q. Yang, Towards personalized federated learning, IEEE Transactions on Neural Networks and Learning Systems 34 (12) (2023) 9587–9603

    2023

  25. [33]

    Mishchenko, G

    K. Mishchenko, G. Malinovsky, S. Stich, P. Richt´ arik, Proxskip: Yes! local gradient steps provably lead to communication acceleration! fi- nally!, in: International Conference on Machine Learning, PMLR, 2022, pp. 15750–15769

    2022

  26. [34]

    X. Yang, H. Yu, X. Gao, H. Wang, J. Zhang, T. Li, Federated con- tinual learning via knowledge fusion: A survey, IEEE Transactions on Knowledge and Data Engineering (2024) 1–20. 28

    2024

  27. [35]

    Huang, L

    Y. Huang, L. Chu, Z. Zhou, L. Wang, J. Liu, J. Pei, Y. Zhang, Personal- ized cross-silo federated learning on non-iid data, in: Proceedings of the AAAI conference on artificial intelligence, Vol. 35, 2021, pp. 7865–7873

    2021

  28. [36]

    C. He, E. Ceyani, K. Balasubramanian, M. Annavaram, S. Avestimehr, Spreadgnn: Decentralized multi-task federated learning for graph neural networks on molecular data, in: Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 36, 2022, pp. 6865–6873

    2022

  29. [37]

    F. Chen, P. Li, T. Miyazaki, C. Wu, Fedgraph: Federated graph learn- ing with intelligent sampling, IEEE Transactions on Parallel and Dis- tributed Systems 33 (8) (2022) 1775–1786

    2022

  30. [38]

    Ghosh, J

    A. Ghosh, J. Chung, D. Yin, K. Ramchandran, An efficient framework for clustered federated learning, Advances in Neural Information Pro- cessing Systems 33 (2020) 19586–19597

    2020

  31. [39]

    Zhang, S

    C. Zhang, S. Zhang, S. Yu, J. James, Graph-based traffic forecasting via communication-efficient federated learning, in: 2022 IEEE Wireless Communications and Networking Conference (WCNC), IEEE, 2022, pp. 2041–2046

    2022

  32. [40]

    Y. Tan, Y. Liu, G. Long, J. Jiang, Q. Lu, C. Zhang, Federated learning on non-iid graphs via structural knowledge sharing, in: Proceedings of the AAAI conference on artificial intelligence, Vol. 37, 2023, pp. 9953– 9961

    2023

  33. [41]

    Y. Yao, W. Jin, S. Ravi, C. Joe-Wong, Fedgcn: Convergence- communication tradeoffs in federated training of graph convolutional networks, Advances in neural information processing systems 36 (2024)

    2024

  34. [42]

    E. J. Cand` es, X. Li, Y. Ma, J. Wright, Robust principal component analysis?, Journal of the ACM (JACM) 58 (3) (2011) 1–37

    2011

  35. [43]

    M. G. Arivazhagan, V. Aggarwal, A. K. Singh, S. Choudhary, Federated learning with personalization layers, CoRR abs/1912.00818 (2019)

    2019 arXiv

  36. [44]

    K. Xu, W. Hu, J. Leskovec, S. Jegelka, How powerful are graph neural networks?, in: 7th International Conference on Learning Representa- tions, ICLR 2019, New Orleans, LA, USA, May 6-9, 2019, OpenRe- view.net, 2019. 29

    2019

  37. [45]

    W. L. Hamilton, Z. Ying, J. Leskovec, Inductive representation learning on large graphs, in: I. Guyon, U. von Luxburg, S. Bengio, H. M. Wallach, R. Fergus, S. V. N. Vishwanathan, R. Garnett (Eds.), Advances in Neu- ral Information Processing Systems 30: Annual Conference on Ne...

    2017

  38. [46]

    Alistarh, D

    D. Alistarh, D. Grubic, J. Li, R. Tomioka, M. Vojnovic, Qsgd: Communication-efficient sgd via gradient quantization and encoding, Advances in neural information processing systems 30 (2017)

    2017

  39. [47]

    Van der Maaten, G

    L. Van der Maaten, G. Hinton, Visualizing data using t-sne., Journal of machine learning research 9 (11) (2008). 30

    2008

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.